Gene Therapy & Molecular Biology Volume 4 Issue A by Niko Boulikas

GENE THERAPY & MOLECULAR BIOLOGY FROM BASIC MECHANISMS TO CLINICAL APPLICATIONS Volume 4 December 1999 Published by Gene Therapy Press

Gene Therapy and Molecular Biology Vol 4

Table of contents Gene Ther & Mol Biol Vol 4, December 1999

Pages

Article title

Type of Article

Authors (corresponding author is in boldface)

1-10

The peculiar organization of telomeres in Drosophila melanogaster

Review Article

Laura Fanti and Sergio Pimpinelli

11-22

Somatic cell nuclear transfer as a tool for investigating ageing processes in mammals

Review Article

Paul G. Shiels

23-31

Semliki Forest virus vectors for in vitro Review Article and in vivo applications

Kenneth Lundstrom, Christophe Schweitzer, J. Grayson Richards, Markus U. Ehrengruber, Francois Jenck, and Cornel Mülhardt

33-43

Ovine adenovirus vectors promote efficient gene delivery in vivo

Research Article

Peter Löser, Günter Cichon, Gary S. Jennings, Gerald W. Both, and Christian Hofmann

45-58

Efficient expression of ribozyme and reduction of stromelysin mRNA in cultured cells and tissue from rabbit knee via Adeno-associated Virus (AAV)

Research Article

Elisabeth Roberts, Piruz Nahreini, Kristi Jensen, Ira von Carlowitz, Karyn Bouhana, Stephen Hunt III, Thale Jarvis, Larry Couture, and Dennis Macejak

59-74

Review Article Glial cell line-derived neurotrophic factor (GDNF) gene therapy in an aged rat model of Parkinson's disease

Bronwen Connor and Martha C. Bohn

75-82

Structural insights into NF!B/I!B signaling

Review Article

Gourisankar Ghosh, De-Bin Huang, and Tom Huxford

83-98

Mammalian c-Jun N-terminal kinase

Review Article

Yi-Rong Chen and Tse-Hua

Table of Contents page 1

Gene Therapy and Molecular Biology Vol 4

pathway and STE20-related kinases

Tan

99-107

Nucleocytoplasmic trafficking and glucocorticoid receptor function

Review Article

Robert J.G. HachĂŠ, Joanne G.A. Savory and Yvonne A. Lefebvre

109-118

Ribozyme-dependent inactivation of lacZ mRNA in E. coli: a feasibility study to set up a rapid in vivo system for screening HIV-1 RNA-specific ribozymes

Research Article

Maria Fe C. Medina and Sadhna Joshi

119-132

Direct redox modulation of p53 protein: potential sources of redox control and potential outcomes

Review Article

Hsiao-Huei Wu, Mark Sherman, Yate-Ching Yuan, and Jamil Momand

133-141

HIV-1 DNA integration: advancing anti-HIV-1 gene therapy approaches by blocking and modulating the process

Review Article

Aaron Geist, Mohamad BouHamdan, Giuseppe Nunnari, Roger J. Pomerantz, and Joseph Kulkosky

143-158

Functional organization of the nuclear lamina

Research Article

Michal Goldberg, Einav Nili, Gady Cojocaru, Yonatan B. Tzur, Raanan Berger, Michael Brandies, Gideon Rechavi, Yosef Gruenbaum and Amos J. Simon

159-170

Models of cationic liposome mediated transfection

Research Article

Aaron Ahearn and Robert Malone

171-176

Essentials of radionanotargeting using oligodeoxynucleotides

Review Article

Kalevi J. A. Kairemo, Antti P. Jekunen, and Mikko Tenhunen

177-182

ets-1 mRNA as target for antisense radio-oligonucleotide therapy in melanoma cells

Research Article

Kalevi J. A. Kairemo, Ketil Thorstensen, Merete Mack, Mikko Tenhunen, and Antti P. Jekunen

183-191

Electronic microarray for DNA analysis

Review Article

Lana Feng and Michael Nerenberg

193-202

Rapid generation of recombinant herpes simplex virus vectors expressing the bacterial lacZ gene under the control of neuronal promoters

Research Article

Alistair McGregor, Alexis Roberts, R. Wayne Davies, J. Barklie Clements and Alasdair R. MacLean

203-208

Recent progress in gene therapy for

Review Article

Yasushi Ikuno, Andrius

Table of Contents page 2

Gene Therapy and Molecular Biology Vol 4

eye diseases

Kazlauskas

209-219

The role of HSV amplicon vectors in cancer gene therapy

221-232

Research Quantitative detection of CFTR Article mRNA in gene transfer studies in human, murine and simian respiratory tissues in vitro and in vivo

Elena Nicolis, Paola Melotti, Anna Tamanini, Monika Lusky, Majid Methali, Andrea Pavirani, and Giulio Cabrini

233-248

Gene therapy for prostate cancer

Review Article

Mitchell S Steiner and Jeffrey R Gingrich

249-260

The HSV-TK/GCV gene therapy for brain tumors

Review Article

Naoto Adachi, Dilek L KĂśnĂź, Karl Frei, Peter Roth, and Yasuhiro Yonekawa

261-274

Establishment of tumor cell lines by transient expression of immortalizing genes

Review Article

Liangping Li

275-284

Nuclear matrix and nucleotide excision

Research Article

Piotr Widlak and Joanna Rzeszowska-Wolny

repair:damage-recognition proteins are

Review Article

Kutubuddin Mahmood, Khaled Tolba, Howard J. Federoff, and Joseph D. Rosenblatt

not constitutive components of the nuclear matrix 26

285-290

Functional improvement in ligament scar tissue following antisense gene therapy: A model system for in vivo engineering of connective tissues

Review Article

David A. Hart, N. Nakamura, R. Boorman, L. Marchuk, H. Hiraoka, Y. Kaneda, N.G. Shrive and C.B. Frank

291-296

3-aminobenzamide: a novel drug to induce in vivo DNA hypermethylation

Review Article

Giuseppe Zardo, Anna Reale, Mariagrazia Perilli, Adriana de Capoa and Paola Caiafa

297-301

Mechanically stretching single chromatin fibers

Review Article

Sanford H. Leuba, Mikhail A. Karymov, Yanzhang Liu, Stuart M. Lindsay and Jordanka Zlatanova

303-312

Gene potentiation: Forming longrange open chromatin structures

Review Article

Susan M. Wykes and Stephen A. Krawetz

313-322

Glycine clock: Eubacteria first, Archaea next, Protoctista, Fungi, Planta and Animalia at last

Research Article

Edward N. Trifonov

Table of Contents page 3

Gene Therapy and Molecular Biology Vol 4

323-338

Nuclear prostaglandin receptors

Review Article

Mousumi Bhattacharya, Daya R. Varma, and Sylvain Chemtob

339-348

Target of rapamycin (TOR) signaling coordinates tRNA and 5S rRNA gene transcription with growth rate in yeast

Review Article

Michael C. Schultz

349-362

DNA structural and sequence determinants for nucleosome positioning

Review Article

Daniel J. Fitzgerald and John N. Anderson

363-368

Regulation of transcription by bent DNA through chromatin structure

Review Article

Ryoiti Kiyama, Yoshiaki Onishi, Chanane Wanapirak, and Yuko Wada-Kiyama

369-385

Crosstalk between intra- and extracellular factors in the development of prolactinomas in the anterior pituitary

Review Article

HĂŠlĂ¨ne Courvoisier

387-395

Cks1 mediates the effects of mutant p53 proteins on the mitotic spindle cell cycle checkpoint

Review Article

M. L. Hixon, and Antonio Gualberto

397-404

BRCA1 function in transcription

Review Article

Neelima Mondal and Jeffrey D. Parvin

405-410

Coding end resolution in scid recombination-inducible cell lines

Review Article

Matthew L. Brown, Sandra Lew and Yung Chang

411-416

Pak protein kinases as mediators of Ras signaling and cell transformation. A place for Pak on the Ras MAP: More than just another JNK bond

Review Article

Yi Tang, Albert Chen, Ya Zhuo, Qi Wang, Albert Pahk and Jeffrey Field

417-424

Negative regulation of cytoplasmic protein tyrosine kinase activity by adaptor proteins

Review Article

Gael Manes, Paul Bello and Serge Roche

Table of Contents page 4

Gene Therapy and Molecular Biology Vol 4, page 1 Gene Ther Mol Biol Vol 4, 1-10. December 1999.

The peculiar organization of telomeres in Drosophila melanogaster Review Article

Laura Fanti1, 2 and Sergio Pimpinelli2 1

Istituto di Genetica, Università di Bari, 70126 Bari, Italy; Istituto Pasteur, Fondazione Cenci Bolognetti and Dipartimento di Genetica e Biologia molecolare, Università “La Sapienza”, 00185 Roma, Italy _________________________________________________________________________________________________________________ Correspondence: Sergio Pimpinelli, Dipartimento di Genetica e Biologia molecolare, Università di Roma “La Sapienza” Piazzale Aldo Moro, 500185 Roma, Italy; Tel. 39-06-49912876; Fax. 39-06-4456866; E-mail: pimpinelli@axcasp.caspur.it Abbreviations: HP1, Heterochromatin Protein 1; TPE, telomere position effects; hTRT, human telomerase reverse transcriptase subunit; TAS, Telomere Associated Sequences; ORF, open reading frame; LBR, lamin B receptor; LTR, long terminal repeat Key Words: Telomeres, Drosophila melanogaster, telomerase, HeT-A, TART, heterochromatin protein, 1 (HP1), telomere capping Received: 21 August 1999; accepted: 4 September 1999

Summary Telomeres are specialized DNA-protein complexes at the ends of eukaryotic chromosomes. They protect the chromosome extremities by preventing potential chromosome damages such as loss of terminal sequences during chromosome replication and by preventing chromosome fusions and degradation. The telomeres of most organisms contain short terminal GC-rich repeats due to the activity of telomerase, a ribonucleoprotein DNA polymerase. Telomeres have gained an exceptional widespread interest in human biology because of their involvement in aging and carcinogenic processes. Despite the fact that the telomere concept was elaborated in Drosophila melanogaster, this organism lacks telomerase and its telomeres are dramatically different from those in other organisms. Telomere loss in Drosophila is prevented by two specific non-LTR transposons, called HeT-A and TART, that appear to be dispensable for chromosome stability. Recent studies have permitted to isolate proteins involved in telomere stability in Drosophila. In particular it has been shown that the heterochromatin protein 1 (HP1) plays an essential role in telomere capping. HP1 is of a special interest because it is a highly conserved protein; HP1 homologues have been identified in many different organisms. Most important, three HP1-like proteins have been found in humans. Future studies will tell us if some of the human HP1 proteins have conserved a functional telomeric localization as in Drosophila.

Telomeres protect the extremities of chromosomes by preventing loss of terminal sequences during DNA replication thus preventing chromosome fusions and degradation. The dynamic spatial order of chromosomes during mitotic and meiotic cycles is also determined by telomeres, by their interaction with both the nuclear envelope and nuclear matrix (reviewed in Dernburg et al. 1996). Moreover, telomeres show a peculiar genetic effect on gene expression called telomere position effects (TPE) (Hazelrigg et al., 1984; Levis et al., 1985; see also Sandell and Zakian, 1992 and Shore, 1996 for reviews). The telomeres of eukaryotes are usually composed of conserved, short, tandemly-repeated, GC-rich sequences (see Henderson, 1995 for review) (see Table 1).

I. Introduction Telomeres are specialized terminal structures in linear chromosomes that are essential for the stability of eukaryotic genomes (for review see Biessmann and Mason, 1992; Zakian, 1995, 1996). The telomere concept was elaborated for the first time by Muller (l938, 1940) to explain his failure to recover terminally deleted chromosomes after X-irradiation in Drosophila. He observed that the recovered broken chromosomes were always capped by other chromosome fragments. Experiments performed by Barbara McClintock provided a strong support to the telomere concept; she showed that, in maize, chromosomes that lack a telomere fuse, generate a dicentric bridge during mitosis and initiate a chromosome breakage-fusion-bridge cycle (McClintock, 1941).

Fanti and Pimpinelli: Organization of telomeres in Drosophila Table 1. Telomeric Repeat Sequences in Eukaryotes

Group

Organism

Telomeric sequence

Vertebrates Insects Nematodes Higher plants

Human, mouse, Xenopus Bombyx mori Ascaris lumbricoides, C. elegans Parascaris univalens Arabidopsis Chlamydomonas Tetrahymena, Glaucoma Paramecium Oxytricha, Stylonychia, Euplotes Physarum, Didymium Dictyostelium Trypanosoma, Crithidia Plasmodium Neurospora Schizosaccharomyces pombe Saccharomyces cerevisiae Candida glabrata Candida albicans Candida tropicalis Candida maltosa Candida guillermondii Candida pseudotropicalis Kluyveromyces lactis

TTAGGG TTAGG TTAGGC TTGCA TTTAGGG TTTTAGGG TTGGGG TTGGG(T/G) TTTTGGGG TTAGGG AG(1-8) TTAGGG TTAGGG(T/C) TTAGGG TTAC(A)(C)G(1-8) G(2-3)(TG)(1-6)T (consensus) GGGGTCTGGGTGCTG GGTGTACGGATGTCTAACTTCTT GGTGTA[C/A]GGATGTCACGATCATT GGTGTACGGATGCAGACTCGCTT GGTGTAC GGTGTACGGATTTGATTAGTTATGT GGTGTACGGATTTGATTAGGTATGT

Algae Ciliates

Slime molds Flagellates Sporozoan Filamentous fungi Fission yeasts Budding yeasts

capping and, thus, in chromosome stability. Moreover, there are also proteins involved in chromosome orientation, nuclear architecture and in telomere gene silencing (see Fang and Cech for review). However, only recently, it has been reported a clear example of chromosome fusion induced by a dominant negative allele of the human telomeric DNA-binding TRF2 protein demonstrating that this protein is required for the protection of telomeric ends (van Steensel et al., 1998). Recently, telomeres have gained an exceptionally widespread interest in human biology and biomedicine. Based on different sets of data, not always concordant, there is an intense debate about the involvement of telomeres in cell aging and carcinogenic processes in humans (see Harley, 1995; de Lange, 1995, 1998 for reviews). The telomere hypothesis of aging and immortalization (see Figure 2 for a schematic model) assigns a significant role of telomere shortening as a cause of cellular senescence. Specifically, the maintenance of the length of telomere sequences could be involved in the process of cell immortalization and, hence, in carcinogenesis.

This sequence conservation is due to a common mechanism for telomere synthesis that involves the telomerase, a ribonucleoprotein DNA polymerase. This enzyme prevents the shortening of terminal sequences at every replication round by adding short, tandem, GC-rich sequences onto the chromosome end (for review see Greider, 1995) (Figure 1). The telomeric tandem repeats seem to be also essential for chromosome stability. In yeast, for example, chromosomes lacking telomeric DNA are lost (Sandell and Zakian, 1993). Thus, besides the telomere replication function, the DNA telomeric sequences also play an essential role in capping functions in organisms with telomerase activity. Along with telomeric DNA sequences and telomerase, several other proteins involved in telomere metabolism have been isolated in different organisms. Different classes of proteins have been characterized in terms of their interaction with telomeric DNA or with other proteins. From a functional point of view, such proteins could be differentially involved in telomere

Gene Therapy and Molecular Biology Vol 4, page 3

Figure 1. Mechanism of telomere elongation by telomerase which resolves the end-replication problem. After removal of RNA primers and ligation of the Okazaki fragments, an unreplicated region on the lagging-strand is led. De novo elongation of the leading-strand by telomerase permits the terminal gap on the lagging-strand to be filled by the conventional replication mechanism.

divisions, continue to divide, showing normal karyotypes and youthful phenotypes. The implications of these findings in carcinogenesis are also significant. They suggest, in fact, that telomerase-dependent immortalization is probably required for tumour progression, and that telomerase silencing could be an essential component of a tumour suppressor system.

Several lines of compelling evidence support this view. It has been shown that telomerase activity is present in germ cells, although it is not detectable in various somatic cell types. As a consequence, the somatic chromosomes progressively lose telomeric DNA sequences at any replicative round. This telomere loss eventually leads to a permanent cell cycle arrest, probably due to a checkpoint mechanism that recognizes the damaged DNA at the chromosome extremities. In contrast, in immortalized or cancerous cells the telomerase results are usually reactivated and the telomeres are maintained (see Harley, 1995; Autexier and Greider, 1996; Shay and Bacchetti, 1997 for reviews). Most recently, a decisive support for the biological role of telomere loss in human cell aging has been provided by Bodnar et al. (1998). Since all the telomerase subunits in human somatic cells are present except for that of the reverse transcriptase subunit (hTRT), telomerase activity was induced in different somatic cells by transfecting the hTRT subunit and examined the cellsâ&#x20AC;&#x2122; proliferative potential. The results showed that the activation of telomerase-induced telomere elongation coincides with a strong increase in proliferative potential. These cells, instead of senescing after a defined number of

II. The peculiarity of Drosophila telomeres Despite the fact that telomeres were discovered in Drosophila, Drosophila telomeres are dramatically different from those in other organisms regarding at least the replication functions. Several studies have shown that Drosophila lacks telomerase and that peculiar transposable elements seem to act as buffers against telomere loss during DNA replication. In addition, these elements are not involved in chromosome stability thus suggesting that replicating functions and telomere capping should be separate. The recent discovery that HP1 functions as a telomeric "cap" protein give support to the view that telomere stability in Drosophila mainly, if not exclusively, depends on specific proteins.

Fanti and Pimpinelli: Organization of telomeres in Drosophila

Figure 2. The model of telomere function in cell aging and immortalization. Telomerase activity maintains the telomeres in germ line cells. In somatic cells telomerase is inactive, and, consequently, the telomeres of some chromosome become shorter, leading to a cell cycle arrest (Hayflick limit) (Hayflick, 1965). Somatic cells can, however, bypass the cell cycle arrest by transformation events. In this case, the cells eventually have the majority of their chromosomes with critically short telomeres and enter crisis. Some clones reactivate telomerase and become immortal, acquiring the capacity to grow indefinitely.

Figure 3. The molecular structure of the RNA transposition intermediate of Drosophila HeT-A and TART telomeric retrotransposons. The 3' and 5' ends are shown in blue. The ORFs correspond to the differently colored boxes, and their orientation is indicated by arrows. (Adapted from Pardue et al., 1997).

TART (Rubin, 1978; Young et al., 1983; Levis et al., 1993; Danilevskaya et al., 1994; for reviews see also Mason and Biessmann, 1995; Pardue, 1995). The Drosophila telomeres (see Figure 3 for a description of their structure) seem to be extended by the addition of copies of these retrotransposons.

A. The HeT-A and TART transposons are involved in telomere replication The Drosophila telomeres contain arrays of two nonLTR retrotransposon-like elements called HeT-A and 4

Gene Therapy and Molecular Biology Vol 4, page 5 The HeT-A and TART copies added at the ends of chromosomes are generated from RNA templates by the activity of a reverse transcriptase (see Figure 4 for a model). This mechanism of transposition has suggested that these retrotransposons could be evolutionarily related to the telomerase (Pardue et al., 1997). The Drosophila telomeres contain also other sequences called TAS (Telomere Associated Sequences) (Karpen and Spradling, 1992; Walter et al., 1995). However, all these elements are dispensable for chromosome stability during meiosis and mitosis. Chromosomes carrying terminal deletions in Drosophila have been recovered (Mason et al., 1984; Levis, 1989). Their molecular analysis has shown that, in many cases, their ends completely lack all the normal telomere elements and that these chromosomes continue to lose terminal DNA sequences. Nevertheless, these broken chromosomes are stably transmitted through many generations as normal capped chromosomes (Biessmann et al., 1990; Levis, 1989). Occasionally, the HeT-A and TART elements transpose to the receding ends of the broken chromosomes (Biessmann et al., 1992; Biessmann et al., 1994; Sheen and Levis, 1994). These observations suggest that, in Drosophila, these elements are essential for telomere elongation (Figure 4) but dispensable for chromosome stability. Thus, unlike other organisms with telomerase-dependent telomeres, it appears that in Drosophila, while the replication function of the telomere seem to depend on specific DNA sequences, the capping function is probably attributable to one or more proteins (Biessmann et al., 1990).

B. The Heterochromatin Protein 1 (HP1) is involved in telomere capping Recent studies have provided support for a role of HP1 in telomere capping. Indeed, it was recently shown that HP1 protein is a structural component of all Drosophila telomeres, and that, when mutated, it induced telomeric fusions, thus suggesting its functional role in telomere capping (Fanti et al., 1998). HP1 is a chromosomal protein which is mainly located in the heterochromatin of both polytene (James and Elgin, 1986; James et al., 1989) and mitotic (Kellum et al., 1995; Fanti et al., 1998) chromosomes of Drosophila melanogaster . Several features of this protein are already known. HP1 is a 206 amino acid protein encoded by the modifier of position effect variegation, the Su(var)205 locus (Eissenberg et al., 1990 ) (Figure 5). A conserved amino acidic motif, called "chromo domain" has been identified in the amino-terminal region of HP1 (Paro and Hogness, 1991) and another additional domain, called "chromo shadow domain" was also identified in the carboxy-terminal region (Aasland and Stewart, 1995) (Figure 5). Both domains are likely involved in protein-protein interactions (Paro and Hogness, 1991; Aasland and Stewart, 1995). Moreover, it has been shown that the nuclear targeting activity of the protein depends on a portion of the carboxyterminal domain while the amino-and carboxy-terminal halves have an independent capacity to bind heterochromatin (Powers and Eissenberg, 1993; Platero et al., 1995). A detailed cytogenetic analysis of HP1 has shown that HP1 is a stable component of all the telomeres in Drosophila, including the ends of stable terminal deletions lacking the telomeric transposons (see Figure 6 and 7 for examples) (Fanti et al., 1998).

Figure 4. Diagram showing the current view of Drosophila telomeres. Violet bars represent the HeT-A elements and orange and green bars represent the TART elements. The blue pointed bars represent the poly(A) ends by which both elements are attached to the chromosome. The transcripts of the two elements are used both as mRNAs and as templates for telomere elongation. (Adapted from Pardue et al., 1997).

Fanti and Pimpinelli: Organization of telomeres in Drosophila

Figure 5. Molecular structure of HP1 protein and HP1 encoding locus. (A) Genomic map of Su(var)2-5 locus. Yellow lines represent introns while boxes represent exons. Orange boxes represent the open reading frame. (B) Corresponding protein map of wild type HP1 with the indicated domains: the green and red boxes correspond to the chromo and chromo shadow domains, respectively. The HP1 nuclear targeting activity is restricted to about the last 58 amino acids of the carboxyterminus while both the N- and C-terminal halves of the protein display independent heterochromatin binding activity. So far, only the carboxy-terminal half of HP1 has been shown to possess a telomere binding activity.

Figure 6. HP1 immuno-pattern on Drosophila wild type polytene chromosomes detected using the C1A9 HP1 monoclonal antibody. This protein is strongly concentrated at the chromocenter (big arrow), is present in many euchromatic regions (arrowheads) and is stably localized at all telomeres (arrows).

Gene Therapy and Molecular Biology Vol 4, page 7

Figure 7. HP1 immunopattern on mitotic chromosomes of Drosophila larval neuroblasts as revealed with the C1A9 HP1 monoclonal antibody. The protein is abundantly accumulated in the pericentric heterochromatin. Immunosignals are also present at the telomeres (arrows) and euchromatic arms. The numbers indicate the autosomes; X = X chromosome; Y = Y chromosome.

The effects of mutations in HP1 on telomere behavior were also analysed (Fanti et al., 1998). The results showed that the absence of HP1 in mutant cells causes multiple telomere-telomere fusions, resulting in a striking spectrum of abnormal chromosome configurations (Figure 8). Analysis of different metaphase and anaphase configurations has shown that the telomeric fusions induce the formation of chromosome bridges during anaphase causing extensive chromosome breakages. These fusions probably, although not exclusively, involve DNA end fusions rather than proteinaceous bridges. These observations suggested that HP1 is, most likely, a telomeric "cap" protein which is essential for telomere stability independent of the type of sequences at the chromosome termini (Fanti et al., 1998).

Another interesting gene causing telomeric fusions has been described. It has been shown that mutations of UbcD1, encoding a class I ubiquitin-conjugating (E2) enzyme, also cause frequent telomere-telomere associations during both mitosis and male meiosis in Drosophila (Cenci et al., 1997). The telomeric associations present in UbcD1 mutants are, however, resolved during mitotic anaphase and do not cause chromosome breakages, thus suggesting that in these mutants the telomeres remain associated by proteinaceous bridges rather than by DNA fusions. The most plausible explanation of these results is that the telomeres are normally associated during interphase by UbcD1 target proteins. In UbcD1 mutants the failure to degrade these proteins maintains the telomeric associations after interphase (Cenci et al., 1997).

Fanti and Pimpinelli: Organization of telomeres in Drosophila

Figure 8. Effect of an HP1 null-mutation on telomere stability. (A) DAPI stained neuroblastic metaphase of a wild type female larva. (B) Ring configuration involving all the chromosomes in a neuroblast cell of HP1 mutant female larva. (C) The same configuration as in (B) showing the HeT-A signals after in situ hybridization with the corresponding probe.

(Shaffer et al. cited in Elgin, 1996). The same LBR protein also interacts with a human chromo domain protein homologous to Drosophila HP1 (Ye and Worman, 1996). It is not unreasonable to suppose that UbcD1 is involved in the degradation of HP1 interacting proteins, like the lamin B receptor, that mediate the ordered interaction among telomeres and/or interaction of telomeres with other structures like the inner nuclear membrane.

The observation that mutations to the UbcD1 locus also cause telomeric fusions raises the question of a possible functional interaction between the HP1 and UbcD1 proteins. A comparison of Su(var)2-5 and UbcD1 mutant effects seems to exclude that HP1 is a direct target of the UbcD1 enzyme. While the telomere fusions are caused by the absence of HP1, those observed in UbcD1 mutants are believed to be caused by the presence of an undegraded protein. In support of this view, we observed that HP1 is present at the telomeres of UbcD1 mutant cells (unpublished observations). Moreover, in the two mutants the telomeric fusions are differentially resolved in anaphase in that in UbcD1 mutants, the anaphasic chromosome bridges do not cause chromosome breakage (Cenci et al., 1997). Other possible explanations are suggested by a set of recent data suggesting that HP1 may also mediate the association of both the heterochromatin and telomeres with the inner nuclear membrane. It has recently been shown that the lamin B receptor (LBR), an integral protein of the inner nuclear membrane, interacts with the Drosophila HP1 in a yeast two-hybrid assay

C. Is the telomere function of HP1 conserved? Several studies have revealed that HP1 is a highly conserved chromosomal protein. HP1 homologues have been identified in several insect species, in plants, and in mammals, including mouse and human (Singh et al., 1991; Saunders et al., 1992; Lorentz et al., 1994; Wreggett et al., 1994; Nicol and Jeppesen, 1994; Furuta et al., 1997). Intriguingly, in mammals, the HP1 homologous proteins are similar to the Drosophila HP1 in being enriched in pericentric heterochromatin (Wreggett et al., 1994) or localized in

Gene Therapy and Molecular Biology Vol 4, page 9 Elgin SCR (1996) Heterochromatin and gene regulation in Drosophila. Curr Opin Genet Dev 6, 193-202. Eissenberg JC, James TC, Foster-Hartnett DM, Hartnett T, Ngan V, and Elgin SCR (1990) Mutation in a heterochromatin-specific chromosomal protein is associated with suppression of positioneffect variegation in Drosophila melanogaster. Proc Natl Acad Sci USA 87, 9923-9927. Fang G, and Cech TR (1995) Telomere proteins. In telomeres. Cold Spring Harbor Lab Press 69-105. Fanti L, Giovinazzo G, Berloco M, and Pimpinelli S (1998) The heterochromatin protein 1 prevents telomere fusions in Drosophila. Molecular Cell 2, 527-538. Furuta K, Chan EKL, Kiyosawa K, Reimer G, Luderschimdt C, Tan EM (1997) Heterochromatin protein HP1Hs! (p25b) and its localization with centromeres in mitosis. Chromosoma 106, 1119. Greider CW (1995) Telomerase biochemistry and regulation. In Telomeres. Cold Spring Harbor Lab Press 35-68. Harley CB (1995) Telomeres and aging. In Telomeres. Cold Spring Harbor Lab Press 247-263. Hazelrigg T, Levis R, and Rubin GM (1984) Transformation of white locus DNA in Drosophila: dosage compensation, zeste interaction, and position effects. Cell 36, 469-481. Hayflick L (1965) The limited in vitro lifetime of human diploid cell strain. Exp Cell Res 37, 614-636. Henderson E (1995) Telomere DNA structure. In Telomeres. Cold Spring Harbor Lab Press 11-34. Horsley D, Hutchings A, Butcher GW, and Singh PB (1996) M32, a murine homologue of Drosophila heterochromatin protein 1 (HP1), localises to euchromatin within interphase nuclei and is largely excluded from constitutive heterochromatin. Cytogenet. Cell Genet 73, 308-311. James TC, and Elgin SCR (1986) Identification of nonhistone chromosomal protein associated with heterochromatin in Drosophila and its gene. Mol Cell Biol 6, 3862-3872. James TC, Eissenberg JC, Craig C, Dietrich V, Hobson A, and Elgin SCR. (1989) Distribution patterns of HP1, a heterochromatinassociated nonhistone chromosomal protein of Drosophila. Eur J Cell Biol 50, 170-180. Karpen GH, and Spradling AC (1992) Analysis of subtelomeric heterochromatin in the Drosophila minichromosome Dp 1187 by single-P element insertional mutagenesis. Genetics 132, 737-753. Kellum R, Raff JW, and Alberts B (1995) Heterochromatin protein 1 distribution during development and during cell cycle in Drosophila embryos. J Cell Sci 108, 1407-1418. Levis R (1989) Viable deletions of a telomere from a Drosophila chromosome. Cell 58, 791-801. Levis R, Hazelrigg T, and Rubin GM (1985) Effects of genomic position on the expression of transduced copies of the white gene of Drosophila. Science 229, 558-561. Levis RW, Ganesan R, Houtchens K, Tolar LA, Sheen FM (1993) Transposons in place of telomeric repeats at a Drosophila telomere. Cell 75, 1083-1093. Lorentz A, Osterman K, Fleck O, and Schmidt H (1994) Switching gene swi6, involved in the repression of silent mating-type loci in fission yeast, encodes a homologue of chromatin-associated proteins from Drosophila and mammals. Gene 143, 139-143. Mason JM, and Biessmann H (1995) The unusual telomeres of Drosophila. Trends Genet 11, 58-62.

euchromatin (Horsley et al., 1996). In particular, three HP1 like proteins were identified in human. It has been shown that one of them, the HP1Hs!, is localized to the centromeres of metaphase chromosomes (Furuta et al., 1997). From all these observations we think not unreasonable to propose that some HP1 like proteins have conserved a functionally important telomeric localization in other species including human. In conclusion, we think that HP1 could become a powerful starting tool to dissect the Drosophila telomeres by identification and characterization of HP1-interacting telomeric proteins. As a consequence, we believe that these approaches will permit to identify, by homology comparisons, other telomeric proteins in human cells and, thus, to make a significant contribution to the analysis of human telomere organization.

References Aasland R, and Stewart AF (1995) The chromo shadow domain, a second chromo domain in heterochromatin-binding protein 1, HP1. Nucl Acid Res 23, 3168-3173. Autexier C, and Greider CV (1996) Telomerase and cancer: revisiting the telomere hypothesis. Trends Biochem Sci 21, 387-391. Biessmann H, Carter SB, and Mason JM (1990) Chromosome ends in Drosophila without telomeric DNA sequences. Proc Natl Acad Sci USA 87, 1758-1761. Biessmann H, and Mason JM (1992) Genetics and molecular biology of telomeres. Adv in Genet 30, 185-249. Biessmann H, Champion LE, O’Hair M, Ikenaga K, Kasravi B, and Mason JM (1992) Frequent transposition of Drosophila melanogaster HeT-A transposable elements to receding chromosome ends. The EMBO J 11, 4459-4469. Biessmann H, Mason JM, Ferry C, d’Hulst M, Valgeirsdottir K, Traverse KL, and Pardue ML (1994) Addition of telomereassociated HeT DNA sequences “heals” broken chromosome ends in Drosophila. Cell 61, 663-673. Bodnar AG, Oullette M, Frolkis M, Holt SE, Chiu CP, Morin GB, Harley CB, Shay JW, Lichtsteiner S, Wright WE (1998) Extension of life-span by introduction of telomerase into normal human cells. Science 279, 349-352. Cenci G, Rawson RB, Belloni G, Castrillon DH, Tudor M, Petrucci R, Goldberg ML, Wasserman SA, and Gatti M (1997) UbcD1, a Drosophila ubiquitin-conjugating enzyme required for proper telomere behavior. Genes and Dev 11, 863-875. Danilevskaya O, Slot F, Pavlova M, and Pardue MM (1994) Structure of the Drosophila HeT-A transposon: a retrotransposon-like element forming telomeres. Chromosoma 103, 215-224. de Lange T, (1995) Telomere Dynamics and genome instability in Human cancer. In Telomeres. Cold Spring Harbor Lab Press 265-293. de Lange T, (1998) Telomeres and senescence: ending the debate. Science 279, 334-335. Dernburg AF, Sedat J, Zacheus Cande W, Bass HW (1996) Cytology of telomeres. In Telomeres. Cold Spring Harbor Lab Press 295-338.

Fanti and Pimpinelli: Organization of telomeres in Drosophila Mason JM, Strobel E, and Green MM (1984) mu-2: mutator gene in Drosophila that potentiates the induction of terminal deficiencies. Proc Natl Acad Sci USA 81, 6090-6094. McClintock B (1941) The stability of broken ends of chromosomes in Zea mays. Genetics 26, 234-282 Muller HJ (1938) The remaking of chromosomes. Collect Net 8, 182-195. Muller HJ (1940) An analysis of the process of structural change in chromosomes of Drosophila. J Genet 40, 1-66. Nicol L, and Jeppesen P (1994) Human autoimmune sera recognize a conserved 26 kD protein associated with mammalian heterochromatin that is homologous to heterochromatin protein 1 of Drosophila. Chromosome Res 2, 245-253. Pardue ML (1995) Drosophila telomeres: another way to end it all. In Telomeres. Cold Spring Harbor Lab Press 339-370. Pardue ML, Danilevskaya ON, Traverse KL, and Lowenhaupt K (1997) Evolutionary links between telomeres and transposable elements. Genetica 100, 73-84. Paro R, and Hogness DS (1991) The polycomb protein shares a homologous domain with a heterochromatin-associated protein of Drosophila. Proc Natl Acad Sci USA 88, 263267. Platero JS, Hartnett T, and Eissenberg JC (1995) Functional analysis of the chromo domain of HP1. The EMBO J 14, 3977-3986. Powers JA, and Eissenberg JC (1993) Overlapping domains of the heterochromatin-associated protein HP1 mediate nuclear localization and heterochromatin binding. J Cell Biol 120, 291-299. Rubin GM (1978) Isolation of a telomeric DNA sequence from Drosophila melanogaster. Cold Spring Harbor Symp Quant Biol 42, 1041-1046. Sandell LL, and Zakian, VA (1992) Telomeric position effect in yeast. Trends Cell Biol 2, 10-14. Sandell LL, and Zakian VA (1993) Loss of a yeast telomere: arrest, recovery and chromosome loss. Cell 75, 729-739. Saunders WS, Chue C, Goebl M, Craig C, Clark RF, Powers JA, Eissenberg JC, Elgin SC R, Rothfield NF, and Earnshaw WC

(1992) Molecular cloning of a human homologue of Drosophila heterochromatin protein HP1 using anti-centromere autoantibodies with anti-chromo specificity. J Cell Sci 104, 573582. Shay JW, and Bacchetti S (1997) A survey of telomerase activity in human cancer. Eur J Cancer 33, 787-791. Sheen FM, and Levis RW (1994) Transposition of the LINE-like retrotransposon TART to Drosophila chromosome termini. Proc Natl Acad Sci USA 91, 12510-12514. Shore D (1996) Telomere Position Effects and transcriptional silencing in the yeast Saccharomyces cerevisiae. In Telomeres. Cold Spring Harbor Lab Press 139-191. Singh PB, Miller JR, Pearce J, Kothary R, Burton RD, Paro R, James TC, and Gaunt SJ (1991) A sequence motif found in a Drosophila heterochromatin protein is conserved in animal and plants. Nucl Acids Res 19, 789-794. Walter MF, Jang C, Kasravi B, Donath J, Mechler BM, Mason JM, Biessmann H (1995) DNA organization and polymorphism of wild-type Drosophila telomere region. Chromosoma 104, 229241. van Steensel B, Smogorzewska A, and de Lange T (1998) TRF2 protects human telomeres from end-to-end fusions. Cell 92, 401413. Wreggett KA, Hill F, James PS, Hutchings A, Butcher GW, and Singh PB (1994) A mammalian homologue of Drosophila heterochromatin protein 1 (HP1) is a component of constitutive heterochromatin. Cytogenet. Cell Genet 66, 99-103. Ye Q, and Worman HJ (1996) Interaction between an integral protein of the nuclear envelope inner membrane and human chromo domain protein homologous to Drosophila HP1. J Biol Chem 271, 14653-14656. Young BS, Pession A, Traverse KL, French C, and Pardue ML (1983) Telomere regions in Drosophila share complex DNA sequences with pericentric heterochromatin. Cell 34, 85-94 Zakian VA (1995) Telomeres: beginning to understand the end. Science 270, 1601-1607. Zakian VA (1996) Telomere functions: lessons from yeast. Trends Cell Biol 6, 29-33.

Sergio Pimpinelli

Laura Fanti 10

Gene Therapy and Molecular Biology Vol 4, page 11 Gene Ther Mol Biol Vol 4, 11-22. December 1999.

Somatic cell nuclear transfer as a tool for investigating ageing processes in mammals Review Article

Paul G. Shiels PPL Therapeutics, Roslin, Edinburgh EH25 9PP, Scotland ______________________________________________________________________________________________________ Correspondence: Tel: +44(0)131 440-4777; Fax: +44(0)131 440-4888; E-mail: pshiels@ppl-therapeutics.com Key words: Mammalian cloning, nuclear transfer, ageing, aging, oxidative damage, telomere, rDNA, mitochondria. n u c l e o l u s Received: 1 September 1999; accepted: 10 September 1999

Summary The development of nuclear transfer technology has opened up a new frontier in the investigation of t h e p r o c e s s e s w h i c h c o n t r i b u t e t o a g e i n g i n m a m m a l s . T h i s r e v i e w s e e k s t o a s s e s s the individual hypotheses that have been proposed to account for the development of the ageing phenotype and to ask how they correlate with observations made on cloned mammals. In sheep derived by nuclear transfer there appears t o be prematurely shortened telomeres, indicative o f increased age. The animals, however, are physiologically normal, consistent with a redox model of ageing where mitochondrial damage is the key contributory factor. The application of nuclear transfer technology to the study of ageing phenomena and i t s use i n experimentally redressing aspects o f the ageing phenotype i s discussed.

oxidative damage and fragmentation of mitochondrial DNA (reviewed by Ozawa 1997; Osiewacz, 1997), or upon age dependent demethylation of DNA. These are not necessarily mutually exclusive, but all require critical testing in vivo.

I. Introduction Ageing can be defined as an increase in molecular chaos over time. This is generally manifest as a change in phenotype and an associated exponential increase in the likelihood of mortality. The latter part of that definition was first described by Gompertz (1825), but it has only been in more recent times that the molecular and cellular events giving rise to the age related phenotypic changes have begun to be unraveled. The seminal work of Hayflick and Moorehead (1961) in describing replicative senescence for human fibroblasts in vitro, gave a new impetus to studies in gerontology and has paved the way for a description of the molecular basis of ageing. Contemporary investigations in the nematode Caenorhabditis elegans and in lower eukaryotes, principally Saccharomyces cerevisiae (reviewed by Guarente, 1997), have given direct insight into which genes and molecular processes underlie the basis of ageing at the cellular level and provide a basic paradigm for ageing in higher organisms. The most celebrated model of ageing is still, however, based upon in vitro senescence data and centres around telomere shortening as a molecular clock. Other molecular models have been proposed, based on

Nuclear transplantation provides a powerful tool for examining the relationship between age related changes at the molecular level and ageing in the whole animal. The advent of Dolly, derived by nuclear transplantation of an adult nucleus from a mammary cell of a six year old animal (Wilmut et al, 1997), provides for the first time, the ability to directly test these models in an in vivo context. The facility to recreate a higher organism while circumventing the germline and potentially, the resetting of any molecular clocks, provides a new approach to the investigation of the relationship between factors determining physiological age and cellular senescence.

II. Models of ageing A. Telomere based clocks The telomere hypothesis of cellular ageing (Olovnikov, 1973; Cooke and Smith, 1986; Harley, 1991; Bodnar et al. 1998) espouses that the loss of telomeric DNA through incomplete replication of chromosome ends and lack of telomerase to repair damage, provides a mitotic clock that 11

Shiels: Somatic cell nuclear transfer and mammalian ageing eventually signals cell death. Once a critical loss of telomeric sequences has occurred from the chromosome ends, replicative capacity is compromised and the cell dies. Telomere shortening has been causally implicated in human cellular senescence (Harley et al 1990; Allsopp et al, 1992), disease (Oexle and Zwirner, 1997) and by general implication, the physiological ageing process in higher animals. Telomeres are specialised structures found at the end of eukaryotic chromosomes, consisting of simple repetitive DNA; in mammals telomeres comprise (TTAGGG)n (Moyzis et al, 1988). They have at least three roles in the maintenance of chromosome structure and integrity, (i) a capping function that is to protect DNA ends from fusion, recombination and degradation, (ii) attachment of the chromosome to the nuclear envelope and (iii) facilitation of the complete replication of chromosome ends (Olovnikov 1973). In man, the latter is achieved via the mediacy of a unique ribonucleoprotein complex, termed telomerase, which has the capacity to synthesise telomeric DNA de novo onto the 3â&#x20AC;&#x2122; end of the parental G rich strand, using the telomerase RNA component as template. This then allows DNA polymerase to conventionally complete the synthesis of the daughter strand. Telomerase principally functions in the human germ line, stem cells and haemopoeitic cells, but not in somatic tissues where telomere damage accumulates over time. Consequently, telomeres in man shorten during replicative ageing (Cooke and Smith, 1986; de Lange, 1992; Frenck et al, 1998).

The telomere capping function appears to be mediated by a combination of a unique tertiary structure and specific telomere binding proteins. Conventionally, this was thought to be achieved either by a specific DNA binding protein attaching to the linear chromosome end (Gottschling and Zakian, 1986) or by a distinct structure composed of a quartet of G residues at the single stranded terminus of the telomere (Williamson et al, 1989). As yet, there has been no direct in vivo evidence for either of these features in higher eukaryotes. Electron microscopy has shown that the mammalian telomere takes the form of a loop, termed the t-loop, created by the telomere DNA folding back on itself to form a lariat whose leading end is the telomeric 3â&#x20AC;&#x2122; G strand overhang. This is envisaged as invading adjacent duplex telomeric repeats, thus creating a displacement loop (D loop). Duplex DNA binding proteins are proposed to bind along the telomeric repeats of the t-loop, while a specialised DNA binding protein stabilises the D-loop lariat junction (Griffith et al, 1999). This is illustrated in Figure 1. Reconciliation of a telomere based model of ageing and cellular senescence with a telomere structure based on a t-loop, poses some intriguing questions. Does, for example, encasement within the duplex serve to distinguish the terminal sequence from damage accrued DNA breaks? Physical isolation of the telomere from normal DNA damage responses would have a number of implications for models seeking to explain the loss and gain of telomeric DNA.

Figure 1. Hypothesised structure for mammalian t e l o m e r e s . The telomere is proposed to loop back upon itself with the 3â&#x20AC;&#x2122; overhang invading the adjacent duplex creating a displacement loop (D-loop) which is stabilised by specific telomere binding proteins ( yellow ovoids). Other duplex DNA binding protein complexes (red ovoids) are proposed to engage along the tloop stabilising the whole structure.

Gene Therapy and Molecular Biology Vol 4, page 13 intriguing, especially in view of the correlation between telomere shortening and ageing in mammals. In S. cerevisiae, however, telomeres do not shorten with age, and an inverse correlation between telomere length and lifespan has been observed (Austriaco and Guarente1997). Conversely, telomere length in yeast correlates positively with telomere silencing. In the absence of telomere erosion with age, longer telomeres could be envisaged as better competitors for silencing complexes, with resultant exacerbation of rDNA circle generation. These observations are not irreconcilable with observations in mammals where telomere shortening occurs with increasing age. In this instance the release of Sir complex analogues for recruitment to nucleolar or other sites would result as a consequence of telomere erosion. Thus when viewed from the perspective of the whole organism, telomere shortening, rather than acting as a molecular clock in mammals, may be a homeostatic mechanism to prolong lifespan. As yet, there are no reports of extrachromosomal rDNA circles for ageing mammalian cells. It will be interesting to see if mammalian species with longer lifespans do indeed all have relatively longer telomeres. Significantly, SGS1 is the homologue of WRN in man, the gene responsible for Werners syndrome, a premature ageing condition, hence the possibility that a similar derangement of rDNA sequences is contributory to human ageing. Interestingly, the WRN protein has a central domain which is homologous to members of the RecQ family of DNA helicases and has been shown to catalyse DNA unwinding (Gray et al 1997), which is in keeping with the role of SGS 1 in yeast.

In the first instance, the presence of the 3â&#x20AC;&#x2122; G strand extension of the telomere within the telomeric duplex, can be construed as being representative of an intermediate event during the process of recombination. This could theoretically potentiate DNA loss via branch migration and subsequent degradation of single stranded DNA segments. Gain of telomeric DNA is also possible if the 3â&#x20AC;&#x2122; invading strand initiates the priming of de novo DNA synthesis subsequent to strand invasion. These mechanisms are not without precedent . Two alternative mechanisms to initiate DNA replication, one dependent on Escherichia coli RNA polymerase the other dependent on general recombination, have been reported for bacteriophage T4 (Luder and Mosig, 1982). Such mechanisms may contribute to telomere dynamism, particularly in the context of telomerase negative cells, where alternative telomere lengthening (ATL) mechanisms have been reported (Kipling. and Cooke 1990; Starling et al, 1990; Prowse and Greider, 1995). A loop structure also has consequences for telomerase function. How, for example, does telomerase gain access to the terminal sequence? It is not intuitive that having a terminal sequence buried within a D loop and stabilised by bound proteins is a readily accessible substrate. Does this singular structure act as a beacon for the telomerase holoenzyme? The implication is that for telomerase to access the terminus, there may be inherent dynamism in the loop during replication.

B. The nucleolus and ageing Analysis of senescence in Saccharomyces cerevisiae has led to the development of a strikingly simple model to explain senescence in yeast, which may be applicable to ageing in higher organisms. Mutants for the SGS 1 gene, which codes for a yeast Rec Q-like helicase, senesce prematurely and show a characteristic accumulation of extrachromosomal rDNA circles in mother cells following successive asymmetric cell divisions, which leads to cell death (Sinclair and Guarente, 1997). The accumulation of extrachromosomal rDNA circles is accompanied by nucleolar fragmentation (Sinclair et al 1997) and a disruption of silencing complexes at telomeres and HM silent mating type loci. These silencing complexes are composed of Sir proteins and appear to promote longevity. Deletion of component members, such as Sir 3 or Sir 4 results in shortened lifespan. A consequence of their disruption is the relocation of Sir 3 and Sir 4 proteins into the nucleolus (Kennedy et al 1997). Whether this is a direct response to the accumulation of the rDNA circles or events associated with their formation, is not fully determined. The telomeric location of Sir protein complexes and their involvement in regulating telomere length and telomeric silencing, coupled with a role in ageing, is

C. Oxidative damage and ageing Ageing as a consequence of cumulative molecular insults has formed the basis of a number of models over the past 40 years. DNA based models of ageing (Alexander 1967; Ozawa, 1995.1997) operate on the premise that oxidative damage to DNA should increase with age and result in decreased functional capacity, inclusive of the ability to repair oxidative damage. By extension, models which propose that the principal agents of DNA damage are free radicals generated as a by-product of oxidative metabolism would predict that mammals with lower metabolic rates should have an increased lifespan and a decreased rate of accumulation of somatic damage compared to those with higher metabolic rates. Experimental evidence in support of such a hypothesis is widespread, though circumstantial. An inverse correlation between lifespan and metabolic rate has been observed in mammals (Cathcart et al 1984). For exaple, rats who have a higher metabolic rate than man, have a shorter life span and a higher rate of accumulation of free radical engendered DNA damage. These differences are more pronounced in the mouse, whose lifespan is even shorter and metabolic rate higher than

Shiels: Somatic cell nuclear transfer and mammalian ageing that of the rat. Analysis in monkeys again correlates well with this hypothesis (Adelman et al 1988). Rodents do show a comparatively higher age related increase in DNA oxidative damage products, such as 8hydroxy-2’-deoxyguanine (8-OHdG) (Fraga et al 1990: Sohal et al 1994). Significantly, the levels of such DNA damage products can be reduced by calorific restriction, which has been shown to increase longevity in rats and mice (Sohal et al 1994). Dietary restriction would be expected to have an influence on a wide range of genes (Mote et al 1991), including those involved in oxidative metabolism (e.g. superoxide dismutase, catalase) through the generation of cellular stress responses. Indeed, cells from calorifically restricted animals maintain replicative potential longer in vitro than those from ad librum fed controls, which is consistent with a loss of replicative potential in vivo being associated with cumulative oxidative insults (Hass et al, 1993). Studies on cells derived from progeric Werners syndrome patients are also consistent with a model, in which the products of oxidative damage accumulate through lack of sufficient DNA repair. Cells from these patients display hyper-recombination, increased mutation frequency and a propensity for large deletions (Fukuchi et al 1989; Cheng et al 1990). WRN patients also present with a high incidence of rare malignancies (Goto et al 1996), consistent with a defect in DNA repair processes which probably relates to the helicase function of the WRN protein (Gray et al 1997). This is supported by observations on another progeric condition, Cockayne’s syndrome, which similarly displays characteristic defects in DNA repair (van Gool et al 1997). Most DNA repair syndromes, however, do not show features typical of progeria and, in contrast to the above observations, it has been reported that in vitro DNA repair capacity is not affected by age (Kunisada et al 1990). These authors studied the repair capacity of human fetal lung fibroblasts and primary embryo fibroblast cultures from rat lung and skin, for their capacity to repair a reporter plasmid which had been UV irradiated prior to transfection. Neither age-related, nor change as a function of passage number was found in the repair of UV damage in these cells. These data remain to be reconciled with previous observations, as they appear counter-intuitive. The data may reflect elevated stress responses as a consequence of cell culture or, more significantly, reflect DNA repair capacity in relation to the stage of senescence of immortalised cells in vitro or senescing primary cell cultures. It may not be sufficient to extrapolate these observations to in vivo age related changes in DNA repair capacity.

A stronger correlate for models of ageing based on oxidative damage can be found in the examination of mitochondrial DNA (mtDNA). A decrease in mitochondrial respiratory activity and an increase in mitochondrial mutations and fragmentation has been positively associated with increasing age (Linnane et al, 1989,1990; Hayakawa et al 1992). Oxidative damage to mtDNA may have more pronounced effects, as mitochondria in post mitotic cells maintain the capacity to replicate (Menzies and Gold 1971), hence, the potential exists to generate and accumulate deleterious mutations which become fixed in the cell population and lead to respiratory deficiency and degeneration. These ideas have been incorporated into “the redox mechanism of ageing” (Ozawa 1995,1997), which hypothesises a molecular basis for the age related decline in cellular activity, tissue and organ degeneration and age associated deterioration in cognitive performance. Mitochondrial mutations are proposed to arise afresh each generation and accrue with age. These mutations are proposed to correlate directly with oxidative damage and cell death. The level of accumulated mutations is considered to directly equate with age related decline in cellular function. While many aspects of this hypothesis seem intuitive, evidence in support of it still remains largely circumstantial. The pronounced age related accumulation of 8-OHdG in mtDNA relative to nuclear DNA (Hayakawa et al 1992) and a correlation with a decline in the mitochondrial electron transfer chain (Hayakawa et al 1993; Takawasa et al 1993) support the hypothesis.The results of elevated stress resulting from mitochondrial mutations have also been observed clinically in patients with mitochondrial myopathy (Ozawa et al 1995) and in murine models of mitochondrial disease (Esposito et al 1999). Direct confirmation of this hypothesis at an organismal level and its integration with other models of ageing, is a facinating prospect. It is not immediately obvious, for example, how redox theories of ageing can be incorporated into a genetic model of ageing, such as that for the klotho mouse (Kuro-o et al 1997). The latter is especially intriguing as the mouse presents with a progeric syndrome with many features in common with human ageing including infertility, arteriosclerosis, osteoporosis, emphysema and skin atrophy. Surprisingly, these all result from the function of a single mutant gene with similarity to !-glucosidase. Whether the klotho mouse actually represents a true ageing syndrome or is simply a good model for diseases of ageing is undetermined. The investigation of such a genetic model, in the context of current hypotheses of ageing, should prove worthwile given that only a single gene is responsible for the klotho phenotype.

Gene Therapy and Molecular Biology Vol 4, page 15

D. DNA methylation and ageing The rate of loss of DNA 5-methyldeoxycytidine residues appears to be inversely related to lifespan ( Wilson and Jones 1983; Wilson et al, 1987). Losses in genomic 5methyldeoxycytidine content have been observed to correlate with donor age in cultured normal human bronchial epithelial cells and in vivo derived murine genomic DNA (Wilson et al, 1987). Conversely, the level of DNA 5-methyldeoxycytidine appears relatively stable in immortalised cells. Significant losses of DNA 5methyldeoxycytidine residues in old age could alter cellular gene expression and contribute to the physiological decline of the animal. Treatment of cells with agents that induce random hypo-methylation induce premature senescence (Gray et al 1991). This correlation between accelerated DNA demethylation and accelerated ageing, while suggesting that these two phenomena are related, does not indicate direct causation. DNA demethylation during ageing may not be random, and could co-operate with other independent ageing processes to produce a finite lifespan and age associated phenotype. In both instances, accelerated DNA demethylation could advance ageing, though in vivo this may not be a reflection of the overall level of genomic 5 methyldeoxycytidine, but rather the perturbation of function of specific gene(s). Furthermore, methylation is thought to stabilise heterochromatin and loss of methylation with age in vivo, or in vitro , may correlate with the loss of telomeric heterochromatin. As such, age related loss of methylation is not incongruous with other models of ageing, in particular those involving erosion or destabilisation of telomeres.

III. In vivo analysis of ageing. Analysis of in vivo ageing has never been straightforward. Critical testing of the various hypotheses to account for ageing in eukaryotes has generally relied on inter-generational comparisons or between mutant and wild type animals. Many investigations simply extrapolate from in vitro data. Such experiments have produced a wealth of information on the molecular processes involved in ageing, but they cannot be extricated from the influence of endogenous molecular clocks (e.g. telomere length), variation in genetic background and artefacts arising from in vitro senescence phenomena. Consequently, the ability to distinguish between the relative contributions of genetic and structural damage, a reasonable prerequisite for the formulation of an accurate model of in vivo ageing, is not readily addressed in previous analyses. Accordingly, a comprehensive and integrated determination of the contributions of the individual molecular processes to the ageing phenotype has not been achieved.

The development of nuclear transfer (NT) using cultured somatic cells (Campbell et al, 1996; Wilmut et al, 1997; Schnieke et al., 1997; Wells et al , 1997; Ashworth et al., 1998; Signer et al., 1998; Cibelli et al., 1998; Wakayama et al 1998) offers a new analytical approach. It advances the possibility of viewing age related changes, both at the single cell level and at the level of the whole organism, against a uniform genetic background with circumvention of any molecular clock(s) reset in the germline. Importantly, it allows dissection of the relationship between ageing processes at both these levels. Aspects of the nuclear transfer procedure impinge directly upon the central tenets of current theories of ageing which can now be subject to integral analyses. Critical to such analyses is the ability to compare clones at different chronological ages, either in vivo, or in vitro, in order to assay directly age related phenomena. Significantly, this is independent of the age of the progenitor tissue. The capacity to serially derive animals (i.e.: clones of clones) by NT (Wakayama and Yanagimachi, 1999) and to genetically manipulate the cell prior to nuclear transplantation (Schnieke et al 1997), increases the power of the possible investigations. This offers the capability of restoring telomerase activity to previously telomerase negative cells, knocking out or mutating genes implicated in the ageing process, such as klotho, free radical scavengers or nuclear encoded mitochondrial genes, to name but a few. Importantly, nuclear transfer results in the separation of the nucleus from the mitochondria of the progenitor cell during the transplantation procedure. The relative contributions of genomic and mitochondrial damage and how these are manifest in the ageing organism can now be addressed. The obvious sequitur from the use of nuclear transfer as a tool to investigate ageing processes will be to ask if any of these processes are reversible. For example, can telomere erosion be repaired? Can one mitigate the effects of nuclear oxidative damage? Can mitochondrial function be restored?

A. Mammalian clones To date four mammalian species have been used to successfully derive clones by nuclear transplantation; these comprise sheep (Campbell et al, 1995; Wilmut et al, 1997), cattle (Wells et al 1997; Cibelli et al 1998), mice (Wakayama et al 1998) and goats (Baguisi et al 1999). The methodological considerations and general applications of cloning have been reviewed extensively elsewhere (Campbell 1999, Colman 1999). An outline of the nuclear transfer procedure is shown in Figure 2. The creation of Dolly (Wilmut et al, 1997) was of particular significance to studies of ageing in that the progenitor nucleus was not only derived from a somatic cell type, but from a six year old adult. Whilst these authors were unable to confirm whether the cell had a fully differentiated phenotype, subsequent studies using adult cells from cows did 15

Shiels: Somatic cell nuclear transfer and mammalian ageing

F i g u r e 2 . S c h e m a t i c o u t l i n e o f t h e n u c l e a r t r a n s f e r p r o c e d u r e . Oocytes derived from Scottish Blackface (symbolised as a yellow sheep) are enucleated. The donor cell derived from a different sheep breed (symbolised as a red sheep) is placed under the zona pelucida into the perivitelline space. The cell nucleus is introduced into the cytoplast by electrofusion, which also activates the oocyte. The reconstructed embryo is then either cultured in vitro up to blastocyst stage or is transferred into a pseudopregnant intermediate recipient ewe. At day 7 embryos are assessed for development. Late morulae and blastocysts are transferred into final recipients. Pregnancies resulting from nuclear transfer are determined by ultrasound scan at about 60 days after oestrus and development is subsequently monitored at regular intervals.

Gene Therapy and Molecular Biology Vol 4, page 17

Figure 3. Regression analysis of mean TRF lengths in NT sheep and controls. Graph showing the telomere length decline with age for control sheep (solid circles) and NT animals together with the fitted line (solid) and 95% prediction interval for an additional observation at any given age (dashed line).

previously been reported (Von Zglinicki et al , 1995; Zijlmanset al., 1997). The full effect of such oxidative damage, however, is only manifest subsequently, in any clone derived from such cells. The contribution of in vitro culture to telomere erosion in sheep derived by nuclear transfer, superimposed on the age of progenitor tissue, could be gauged from the TRF diminution of OME cells that had undergone up to 27 population doublings (9 passages)in culture. When compared to the mammary gland from which these cells derived and to Dolly, derived in turn from these cells after 3 population doublings, a mean TRF decrease was observed at an average 0.157 kb per population doubling. The immediate implication of these observations is that the extent of TRF shortening can be mitigated, principally by minimising time in culture and the age of donor cells. This is particularly relevant with respect to animal 6LL7, where the use of fetal tissue and minimal culturing yielded an animal where the mean TRF size is not significantly shorter than age matched controls, unlike 6LL3 and 6LL6 where culturing was more prolonged (Figure 3). The most likely explanation for the shorter mean TRF lengths of all three nuclear transfer sheep is that the TRF size observed reflects that of the transferred nucleus. Whether this telomere erosion is reflected in the overall ageing process of these animals is uncertain. It is not known whether the actual physiological age of animals derived by nuclear transfer is accurately reflected by TRF measurement. No physiological progeria has yet been reported in any animal derived by nuclear transfer. Veterinary examination of the cloned animals has confirmed that they are healthy and typical for sheep of their age and breed, despite having a shorter mean TRF length.

use differentiated cells (Wells et al 1999). Adult somatic cells have also been used to derive clones in mice (Wakayama et al 1998).

B. clones.

Observations on ageing in ovine

Initial investigations into ageing in cloned animals has revolved around an examination of telomeres in sheep (Shiels et al 1999a, 1999b). These experiments asked if the generation of animals without germline involvement results in the resetting of telomere lengths and hence any molecular clock which measured time by such lengths. Three cloned animals were examined, whose derivation spanned distinct developmental stages and cell types. These comprised animal 6LL3, (â&#x20AC;&#x153;Dollyâ&#x20AC;?), derived by transfer of a nucleus from ovine mammary epithelial (OME) cells from a 6 year old sheep, 6LL6 from sheep embryonic cells (SEC 1) obtained from day 9 embryos, and 6LL7 which was derived from fibroblasts from a day 25 fetus. All three animals showed apparent telomere diminution as determined by measurement of mean terminal restriction fragment (TRF) lengths. The TRF diminution observed in 6LL3 was the greatest of the three animals and was consistent with the age of her progenitor ovine mammary tissue (6 years old) and significantly the time OME cells derived from that tissue, spent in culture prior to nuclear transfer. The influence of time spent by donor cells in culture is substantial. Telomere shortening due to enhanced damage attributed to reactive oxygen species (ROS)in vitro, has 17

Shiels: Somatic cell nuclear transfer and mammalian ageing Furthermore, 6LL3 has undergone two normal pregnancies and successfully delivered healthy lambs. The telomere hypothesis of ageing (Olovnikov, 1973; Cooke and Smith, 1986; Harley, 1991; Bodnar et al , 1998), however, would predict that animal 6LL3 would reach a critical telomere length sooner than age matched controls. However, ovine TRFs show a large size distribution, from 5-50Kb (Shiels et al 1999a,b), thus it remains to be seen whether a critical length will be reached during the animalâ&#x20AC;&#x2122;s lifetime. Experimental inactivation of murine telomerase only produced a phenotype after five generations (Blasco et al 1997; Lee et al 1998) and similar observations have been made in telomerase deficient yeast cells (Lundblad and Blackburn, 1993). It is noteworthy, in respect of the inverse correlation between telomere length and lifespan, that the TRF spread in sheep appears to fall between those of mouse and man in accord with the hypothesis that longer telomeres mean a shorter lifespan. Clarification of whether telomere erosion is causative for, or an effect of the ageing process is not immediately apparent. The observations on cloned sheep are, however, congruent with the redox theory of ageing (Ozawa 1995,1997) which would predict that the vigour and fecundity of such animals would be physiologically identical to age matched controls.

While these observations indicate a strong correlation between telomere erosion and the timing of cellular senescence, how this will be integrated within a complete description of the chronological ageing phenotype has yet to be determined. Any extrapolation beyond consideration of telomere length as one feature of a mitotic clock would seem to be premature. While telomere erosion can be directly addressed experimentally, it has still to be fully established how such hTERT expressing cells will fare in vivo. A key consideration in this context , will be the effect of any genetic damage acquired in vitro (and for that matter in vivo) which is not accessible to telomerase repair. Another consideration is whether oxidative damage is primarily manifest as telomere erosion, due to telomere sequence content or cellular localisation. A donor nucleus source which is naturally telomerase positive, such as a lymphocyte, might be a better choice to mitigate the effects of telomere erosion without recourse to genetic manipulation. Such a cell type, however, might be considered unsuitable as a donor source, as telomerase positivity is often a characteristic of malignancy. Telomerase positivity coupled with the increased risk of in vitro accrued oxidative damage, means that the chances of neoplastic transformation are increased. Quantification of this risk, however, is not straightforward and requires species specific model systems to address the issue (for a fuller discussion see Colman 1999). Suitable model systems have not yet been established. Parenthetically, it is not known whether outbreeding a cloned animal or inter breeding clones will restore telomere lengths. In the former instance, the presence of the chromosome complement from the naturally derived parent would provide a haploid complement of full length telomeres, while in the latter germline resetting of telomere lengths would be required. It is unknown if germline resetting of telomeres occurs in interbred clones. This is presently being investigated. Another aspect of telomere erosion in mammals that needs clarification is the relationship between this phenomenon and that of extrachromosomal rDNA circle generation as described in yeast (Sinclair et al, 1997). A comparison of Dolly with a relevant control panel of sheep should prove informative in this context, as the nucleolar model would predict that she should at least reflect the age of her progenitor tissue. This is supported by analysis of terminal restriction fragment lengths (Shiels et al, 1999a,b). A final consideration with regard to telomeres is whether there is a causal relationship between age associated demethylation and telomere attrition. If so, ameliorating telomere shortening may mitigate loss of methylation and equally the generation of extrachro-

IV. Cause and effect: testing the models of ageing. Mitigation of the observed effects of both ageing and in vitro senescence do seem feasible. It is this capacity which will allow dissection of the component parts of the ageing process and illustrates the potential of nuclear transfer as a tool to achieve this. Strategies for remedial action can be detailed as follows.

A.Telomere erosion Introduction of telomerase to normal human cells in culture has been reported to significantly increase their lifespan (Bodnar et al, 1998). Ectopic expression of the telomerase catalytic subunit (hTERT) and subsequent activation of telomerase in postsenescent cells has been demonstrated to allow the cells to proliferate beyond crisis (Counter et al 1999). Furthermore, alteration of the carboxyl terminus of hTERT appears not to affect telomerase enzymatic activity, though it prevents telomere maintenance and consequent cell proliferation. Cells expressing hTERT ectopically appear phenotypically normal and exhibit no manifestations of malignant transformation (Jiang et al 1999; Morales et al 1999).

Gene Therapy and Molecular Biology Vol 4, page 19 mosomal rDNA circles. It remains to be determined if this hypothesis is valid.

Molecular gerontology is, if you excuse the pun, rapidly coming of age. Recent advances in nuclear transplantation technology (for a review see Colman 1999) have provided a direct route to the investigation of the individual processes that give rise to the ageing phenotype and for the first time allowed their direct testing in vivo. The manipulation of such processes, using the full panoply of molecular techniques and the powerful genetics available for studies in mice, should allow for a fuller understanding of ageing from the level of the single cell to the whole organism. The potential to mitigate the effects of these processes exists in the laboratory, though the practical application of this technology is not so easily achieved. One immediate application of this technology is the development of Cellular therapies for countering the effects of diseased or degenerating tissues. This would comprise the molecular manipulation of cells in vitro , repairing for example, telomere erosion, replacing deleterious alleles of age important genes, upregulating oxygen free radical scavrengers, or replacement of damaged mitochondria. Such cells could, for example, be used to supplement or replace cells in failing organs. Consequently, the functional replacement of degenerating tissues may no longer fall within the realms of science fiction.

B. Oxidative damage While shortening time spent in culture offers an immediate route to decreasing any telomere erosion due to reactive oxygen species, this should also reduce genome wide oxidative damage. Calorific restriction offers a second means of mitigating oxidation effects, with a proven efficacy in vitro and in vivo (Hass et al, 1993; Sohal et al, 1994). A number of issues pertaining to the source, extent and contribution of such oxidative damage can now be brought into focus. Mitochondrial damage contributing to the overall ageing process, as postulated within â&#x20AC;&#x153;the redox mechanism of ageingâ&#x20AC;? (Ozawa, 1995, 1997) is both amenable to direct investigation and redress. Mitochondria in cloned animals almost entirely derive from the recipient oocyte cytoplasm (Evans et al 1999), hence problems arising as a consequence of the higher mutation rate in mitochondria and their accumulation, particularly in postmitotic cells, are circumvented. If the tenet of this hypothesis is valid then one would expect animals derived by nuclear transplantation from adult cell sources not to show the physiological characteristics of the age of their progenitor, but to reflect that of age control animals. Only genomic damage acquired by the donor nucleus should be transferred to clones and manifested accordingly. The physiological characteristics of Dolly and other sheep clones (Wilmut et al 1997; Shiels et al 1999a,b) and the failure to detect premature ageing in cloned mice (Wakayama et al 1999) are in keeping with such a hypothesis, whereby damage to mitochondria is a key event in the physiological degeneration associated with ageing. Genomic damage inherited by the clones is not easily redressed. As the accumulation of deleterious somatic mutations correlates positively with increasing age, there is a greater likelihood of acquiring damage to developmentally important genes through the use of older donor cell sources. This obviates the practical choice of using nuclei from younger cell sources as donors for nuclear tranplantation. Assessment of the relative consequences of mitochondrial and genomic oxidative damage and nuclearmitochondrial communication within the ageing process, can also be addressed by manipulation of nuclear encoded mitochondrial genes and to some extent, by variation of the recipient cytoplasm. This should allow determination of to what extent accumulation of mutations in the nucleus contributes to age related mitochondrial dysfunction.

Acknowledgements I would like to acknowledge Alan Colman, Ian Garner, Angelika Schnieke, Alex Kind and Keith Campbell for helpful discussions and critical reading of the manuscript

References Adelman R, Saul RL, Ames BN (1 9 8 8 ). Oxidative damage to DNA: relation to species metabolic rate and life span. Proc Natl Acad Sci U S A 85:2706-8 Alexander P (1 9 6 7 ). The role of DNA lesions in processes leading to ageing in mice. S y m p . S o c . E x p . B i o l . 21:29-50. Allsopp RC, Vaziri H, Patterson C, Goldstein S, Younglai EV, Futcher AB, Greider CW, Harley CB (1 9 9 2 ). Telomere length predicts replicative capacity of human fibroblasts. Pr oc . N at. A c ad. Sc i. U SA 8 9 : 10114-10118 Ashworth D, Bishop, M, Campbell, KHS, Colman A, Kind, AJ, Schnieke AE, Blott, S, Griffin H, Haley C, McWhir J, Wilmut I (1 9 9 8 ). DNA microsatellite analysis of Dolly. Nature 394: 329. Austriaco NR Jr, Guarente LP (1 9 9 7 ) Changes in telomere length cause reciprocal changes the lifespan of mother cells in Saccharomyces cerevisiae. P r o c . N a t . A c a d . S c i . U S A 94: 9768-9772. Baguisi A, Behboodi E, Melican DT, Pollock JS, Destrempes MM, Cammuso C, Williams JL,Nims SD, Porter CA, Midura P, Palacios MJ, Ayres SL, Denniston RS, Hayes ML, Ziomek CA, Meade HM, Godke RA, Gavin WG, Overstrom EW, Echelard Y. (1 9 9 9 ) Production of goats by somatic cell nuclear transfer. N a t B i o t e c h n o l . 17: 456-61.

V. Conclusions. 19

Shiels: Somatic cell nuclear transfer and mammalian ageing Blasco MA, Lee HW, Hande MP, Samper E, Lansdorp PM, DePinho RA, Greider CW (1 9 9 7 ) Telomere shortening and tumor formation by mouse cells lacking telomerase RNA. C e l l 91: 25-34.

Gompertz B (1 8 2 5 ). On the nature of the function expressive of the law of human mortality, and on the mode of determining life contingencies. P h i l . T r a n s . R o y . S o c . 115: 513585.

Bodnar AG, Ouellette M, Frolkis M, Holt SE, Chiu CP, Morin GB, Harley CB, Shay JW, Lichtsteiner S, Wright WE (1 9 9 8 ) Extension of life-span by introduction of telomerase into normal human cells. S c i e n c e 279 : 349-352Campbell KHS (1 9 9 9 ) Nuclear equivalwence, Nuclear transfer and the cell cycle C l o n i n g 1: 3-15.

Gottschling DE, Zakian VA (1 9 8 6 ) Telomeric proteins: specific recognition and protection of the natural termini of Oxitricha macronuclear DNA. C e l l 47: 195-205

Campbell KH, McWhir J, Ritchie WA, Wilmut I (1 9 9 6 ) Sheep cloned by nuclear transfer from a cultured cell line. Nature 380: 64-6

Gray MD, Jesch SA, Stein GH (1 9 9 1 ) 5-Azacytidine-induced demethylation of DNA to senescent level does not block proliferation of human fibroblasts. J . C e l l P h y s i o l . 149:477-84

Gray MD, Shen JC, Kamath-Loeb AS, Blank A, Sopher BL, Martin GM, Oshima J, Loeb LA (1 9 9 7 ) The Werner syndrome protein is a DNA helicase. N a t . G e n e t . 17: 100-103.

Cathcart R, Schwiers E, Saul RL, Ames B (1 9 8 4 ) Thymine glycol and thymidine glycol in human and rat urine: apossible assay for oxidative DNA damage. P r o c N a t l Acad Sci U S A 81:5633-7

Griffith JD, Comeau L, Rosenfield S, Stansel RM, Bianchi A, Moss H, de Lange T (1 9 9 9 ) Mammalian telomeres end in a large duplex loop. C e l l 97 503-514.

Cooke HJ, Smith BA (1 9 8 6 ) Variability at the telomeres of the human X/Y pseudoautosomalregion. C o l d S p r i n g H a r b . S y m p . Q u a n t . B i o l . LI, 213-219.

Guarente L (1 9 9 7 ) Link between aging and the nucleolus. G e n e s and Dev. 1 1 : 2449-2455

Cheng R, Murano S, Kurz B, Schmookler-Reis R (1 9 9 0 ) Homologous recombination is elevated in some Werner like syndromes but not during normal in vitro or in vivo senescence of mammalian cells. M u t . R e s 237: 259269

Harley CB, Futcher, AB, Greider CW (1 9 9 0 ) Telomeres shorten during ageing of human fibroblastsNature 345: 458-460.

Harley CB (1 9 9 1 ) Telomere loss: mitotic clock or genetic time bomb? Mutat Res. 256: 271-282.

Hass BS, Hart RW, Lu MH, Lyn-Cook BD (1 9 9 3 ) Effects of calorific restriction in animals on cellular function, oncogene expression, and DNA methylation in vitro. Mutat Res 295: 281-289

Cibelli JB, Stice SL, Golueke PJ, Kane JJ, Jerry J, Blackwell, C, Ponce de Leon FA, Robl JM (1 9 9 8 ). Cloned transgenic calves produced from nonquiescent fetal fibroblasts. S c i e n c e 280: 1256-1258.

Hayakawa M, Hattori K, Sugiyama S, Ozawa T (1 9 9 2 ) Ageassociated oxygen damage and mutations in mitochondrial DNA in human hearts. B i o c h e m B i o p h y s R e s Commun 189: 979-985.

Colman A (1 9 9 9 ) Mammalian somatic cell cloning: progress and applications. Nat. Genet (in press) Counter CM, Hahn WC, Wei W, Caddle SD, Beijersbergen RL, Lansdorp PM, Sedivy JM, Weinberg RA (1 9 9 8 ) Dissociation among in vitro telomerase activity, telomere maintenance, and cellular immortalization. Proc Natl Acad Sci U S A 95: 14723-14728

Hayakawa M, Sugiyama S, Hattori K, Takasawa M, Ozawa T (1 9 9 3 ) Age-associated damage in mitochondrial DNA in human hearts. M o l C e l l B i o c h e m 119: 95-110 Hayflick L, Moorehead PS (1 9 6 1 ). The serial cultivation of human diploid strainsE x p . C e l l . R e s . 25: 585-621.

de Lange T (1 9 9 2 )Human telomeres are attached to the nuclear matrix. EMBO J. 11: 717-724.

Jiang XR, Jimenez G, Chang E, Frolkis M, Kusler B, Sage M, Beeche M, Bodnar AG, Wahl GM, Tlsty TD, Chiu CP (1 9 9 9 ) Telomerase expression in human somatic cells does not induce changes associated witha transformed phenotype. Nat Genet 21: 111-114.

Esposito LA, Melov S, Panov A, Cottrell BA, Wallace DC (1 9 9 9 ) Mitochondrial disease in mouse results in increased oxidative stress. P r o c N a t l A c a d S c i U S A 96: 4820-4825.

Kennedy BK, Gotta M, Sinclair DA, Mills K, McNabb DS, Murthy M, Pak SM, Laroche T, Gasser SM, Guarente L (1 9 9 7 ) Redistribution of silencing proteins from telomeres to the nucleolus is associated with extension of lifespan in S, cerevisiae C e l l 89: 381-391.

Evans MJ,Gurer C, Loike JD, Schnieke AE, Schon EA (1 9 9 9 ) Mitochondrial genotypes in nuclear transfer derived cloned sheep. Nature : in press Frenck RW Jr, Blackburn EH, Shannon KM (1 9 9 8 ) The rate of telomere sequence loss in human leukocytes varies with age. Proc Natl Acad Sci U S A 95: 5607-5610.

Kipling D, Cooke HJ (1 9 9 0 ). Hypervariable ultra-long telomeres in mice. Nature 347: 400-402.

Fraga CG, Shigenaga MK, Park JW, Degan P, Ames BN (1 9 9 0 ) Oxidative damage to DNA during aging: 8hydroxy-2'-deoxyguanosine in rat organ DNA and urine. Proc Natl Acad Sci U S A 87: 4533-4537.

Kunisada T, Miller CD, Schneider EL (1 9 9 0 ) Ultraviolet lightinduced DNA damage in transcribed sequences: nochange in repair with age. Mutat Res 237: 75-81. Kuro-o M, Matsumura Y, Aizawa H, Kawaguchi H, Suga T, Utsugi T, Ohyama Y,

Fukuchi K , Martin G, andMonnat R (1 9 8 9 ) Mutator phenotype of Wernerâ&#x20AC;&#x2122;s syndrome is characterised by extenasive deletions. Proc Natl Acad Sci U S A 86: 5893-5897.

Kurabayashi M, Kaname T, Kume E, Iwasaki H, Iida A, Shiraki-Iida T, Nishikawa S, Nagai R, Nabeshima YI (1 9 9 7 ) Mutations

Gene Therapy and Molecular Biology Vol 4, page 21 of the mouse klotho gene leads to a synderome resembling ageing. Nature 390: 45-51

Prowse, KR, Greider, CW (1 9 9 5 ) Developmental and tissuespecific regulation of mouse telomerase andtelomere length. P r o c . N a t l . A c a d . S c i . U . S . A . 92: 4818-4822.

Lee HW, Blasco MA, Gottlieb GJ, Horner JW 2nd, Greider CW, DePinho RA (1 9 9 8 ). Essential role of mouse telomerase in highly proliferative organs. Nature 3 2 9 : 569-574

Schnieke AE, Kind AJ, Ritchie WA, Mycock K, Scott AR, Ritchie M, Wilmut I, Colman A,Campbell KH (1 9 9 7 ) Human factor IX transgenic sheep produced by transfer of nuclei from transfected fetal fibroblasts. S c i e n c e 278: 2130-2133

Linnane AW, Baumer A, Maxwell RJ, Preston H, Zhang CF, Marzuki S (1 9 9 0 ) Mitochondrial gene mutation: the ageing process and degenerative diseases. B i o c h e m Int 22: 1067-1076.

Shiels PG, Kind AJ, Campbell KH, Waddington D, Wilmut I, Colman A, Schnieke AE (1 9 9 9 a ) Analysis of telomere lengths in cloned sheep. Nature 399: 316-317

Linnane AW, Marzuki S, Ozawa T, Tanaka M (1 9 8 9 ) Mitochondrial DNA mutations as an important contributor to ageingand degenerative diseases. Lancet 1: 642-645

Shiels PG, Kind AJ, Campbell KH, Waddington D, Wilmut I, Colman A, Schnieke AE (1 9 9 9 b ) Analysis of telomere length in Dolly, a sheep derived by nuclear transfer. C l o n i n g 1: 1-7.

Luder A and Mosig G (1 9 8 2 ) Two alternative mechanisms for the initiation forks in bacteriophage t4: priming by RNA polymerase and by recombination. Proc Natl Acad Sci U S A 79: 1101-1105

Signer, EN, Dubrova, YE, Jeffreys, AJ, Wilde, C, Finch, LM, Wells, DN, Misica, PM, Day, TA and Tervit, HR (1 9 9 7 ). DNA fingerprinting Dolly. Nature 394: 329-330.

Lundblad V, Blackburn EH (1 9 9 3 ) An alternative pathway for yeast telomere maintenace rescues est-1 senescence. C e l l 73: 347-360. Menzies RA, Gold PH (1 9 7 1 ) The turnover of mitochondria in a variety of tissues in young adult and aged rats. J . B i o l . C h e m 246: 2425-2429

Sinclair, DA, Guarente L (1 9 9 7 ) Extrachromosomal rDNA circles-a cause of aging in yeast. C e l l 91:1033-1042

Morales CP, Holt SE, Ouellette M, Kaur KJ, Yan Y, Wilson KS, White MA, Wright WE, Shay J (1 9 9 9 ) Absence of cancer-associated changes in human fibroblasts immortalized with telomerase. N a t G e n e t 21: 115118.

Sohal RS, Agarwal S, Candas M, Forster MJ, Lal H (1994) Effect of age and calorific restriction on DNA oxidative damage indifferent tissues of C57BL/6 mice. M e c h A g e i n g D e v 76: 215-224.

Sinclair DA, Mills K, Guarente L (1 9 9 7 ) Accelerated aging and nucleolar fragmentation inyeast sgs 1 mutants. S c i e n c e 277: 1313-1316.

Starling JA, Maule J, Hastie, ND, Allshire, RC (1 9 9 0 ). Extensive telomere repeat arrays in mouse are hypervariable. N u c l e i c A c ids R e s 18 : 6881-6888.

Mote PL, Grizzle JM, Walford RL, Spindler SR (1 9 9 1 ) Influence of age and calorific restriction on expression of hepatic genes for xenobiotic and oxygen metabolizing enzymes in the mouse. J G e r o n t o l 46: 95-100

Takasawa M, Hayakawa M, Sugiyama S,Hattori K, Ito T, Ozawa T (1 9 9 3 ) Age-associated damage in mitochondrial function in rat hearts. Exp Gerontol 28: 269-280.

Moyzis RK, Buckingham JM, Cram LS, Dani M, Deaven LL, Jones MD, Meyne J, Ratliff RL,Wu JR (1 9 8 8 ) A highly conserved repetitive DNA sequence, (TTAGGG)n, present at the telomeres of human chromosomes. P r o c Natl Acad Sci U S A 85: 6622-6

van Gool AJ; van der Horst GTJ, Citterio E, Hoeijmakers J HJ (1 9 9 7 ) Cockayne syndrome: defective repair of transcription? EMBO J. 16: 4155-4162,. von Zglinicki T, Saretzki G, Docke W, Lotze C (1 9 9 5 ) Mild hyperoxia shortens telomeres and inhibits proliferation of fibroblasts: a model for senescence? Exp C e l l R e s 220:186-193.

Oexle K, Zwirner A (1 9 9 7 ). Advanced telomere shortening in respiratory chain disorders. H u m a n M o l . G e n e t . 6 : 905-908.

Wakayama T, Perry AC, Zuccotti M, Johnson KR, Yanagimachi R (1 9 9 8 ) Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei. Nature 394: 369374.

Olovnikov AM (1 9 7 3 ). A theory of marginotomy. The incomplete copying of template margin in enzymic synthesis of polynucleotides and biological significance of the phenomenon. J . T h e o r . B i o l 41:181-190.

Wakayama T, Yanagimachi R (1 9 9 9 ) Cloning the laboratory mouse S e m i n a r s i n C e l l a n d D e v B i o l 10: 253-258.

Osiewacz H.D (1 9 9 7 ) Genetic regulation of agingJ . M o l Med.75: 715-727.

Wells DN, Misica PM, Day TA, Tervit H. (1 9 9 7 ) Production of cloned lambs from an established embryonic cell line: a comparison between in vivo- and in vitro-matured cytoplasts. B i o l . R e p r o d . 57: 385-393.

Ozawa T (1 9 9 5 ) Mechanism of somatic mitochondrial DNA mutations associated with age and diseases. B i o c h i m B i o p h y s A c t a . 1271:177-189. Ozawa T (1 9 9 7 ) Oxidative damage and fragmentation of mitochondrial DNA in cellular apoptosis. B i o s c i e n c e R e p o r t s 17: 237-250.

Wells DN, Misica PM, Tervit HR (1 9 9 9 ) Production of cloned calves following nuclear transfer with cultured adult mural granulosa cells. B i o l R e p r o d 60: 996-1005.

Ozawa T, Sahashi K, Nakase Y, Chance B (1 9 9 5 ) Extensive tissue oxygenation associated with mitochondrial DNAmutations. B i o c h e m B i o p h y s R e s C o m m u n 213: 432-438.

Williamson JR, Raghuraman MK, Cech TR (1 9 8 9 ) Monovalent cation induced structure of telomeric DNA: the G quartet model. C e l l 59: 871-880.

Shiels: Somatic cell nuclear transfer and mammalian ageing Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KH. (1 9 9 7 ) Viable offspring derived from fetal and adult mammalian cells. Nature 385: 810-813. Wilson VL, Jones PA (1 9 8 3 ) DNA methylation decreases in aging but not in immortal cells. S c i e n c e 220: 1055-7 Wilson VL, Smith RA, Ma S, Cutler RG (1 9 8 7 ) Genomic 5methyldeoxycytidine decreases with age. J B i o l C h e m 262: 9948-51 Zijlmans JM, Martens UM, Poon SS, Raap AK, Tanke HJ, Ward RK, Lansdorp PM (1 9 9 7 ) Telomeres in the mouse have large inter-chromosomal variations in the number of T2AG3 repeats. P r o c N a t l A c a d S c i U S A 94: 7423-7428.

Paul Shiels

Gene Therapy and Molecular Biology Vol 4, page 23 Gene Ther Mol Biol Vol 4, 23-31. December 1999.

Semliki Forest virus vectors for in vitro and in vivo applications Review Article Kenneth Lundstrom1 , Christophe Schweitzer1 , J. Grayson Richards1 , Markus U. Ehrengruber2 , Francois Jenck1 , and Cornel Mülhardt1 1

F. Hoffmann-La Roche, Research Laboratories, CH-4070 Basel, Switzerland Brain Research Institute, University of Zurich, CH-8057 Zurich, Switzerland __________________________________________________________________________________________________ 2

Correspondence: Kenneth Lundstrom, Ph.D. Tel: (41-61) 687 8653; Fax: (41-61) 688 4575; E-mail: kenneth.lundstrom@roche.com Abbreviations: SFV, Semliki Forest virus; CTL, cytotoxic T-cell; GFP, green fluorescence protein; GPCRs, G-protein coupled receptors; hNK1R, human neurokinin-1 receptor Key words: Semliki Forest virus, gene therapy, vaccine, LacZ, G-protein coupled receptors Received: 16 December 1999; accepted: 20 December 1999

Summary Rapid virus generation, broad host range, efficient RNA replication in the cytoplasm, and high expression levels are features that have made the use of Semliki Forest virus (SFV) vectors attractive. High-level expression of Gprotein coupled receptors has allowed specific binding and functional studies in a variety of mammalian cell lines. Furthermore, high infection efficiency (75-90%) has greatly facilitated gene expression and localization studies in primary neurons in culture. The establishment of SFV infections of large-scale suspension cultures has resulted in the production of hundreds of milligrams of recombinant receptors now available for structural studies. SFV vectors have been shown to preferentially infect neurons in hippocampal slice cultures, which will facilitate studies on gene expression, transport, and protein localization in neuronal tissue. Injection of replication-deficient SFV vectors into rat brain resulted in local, high-level transient expression in vivo. Recent vector improvements have included the generation of SFV vectors with low-to-moderate transgene expression resulting in more physiological expression levels that are similar to those seen in native tissue. Novel SFV vectors, which have recently been developed, permit prolonged survival of infected host cells.

I. Introduction Recombinant protein expression is an essential part of molecular biology research and drug discovery nowadays. As the sequencing of the human genome is approaching its completion, the requirement for fast and high-level expression of novel gene sequences is rapidly increasing. Several viral and non-viral vector systems have been developed for this purpose. Semliki Forest virus (SFV), a member of the alphavirus family, is an enveloped virus with a singlestranded RNA genome. To generate an expression system the SFV genome was split on two plasmids as cDNA molecules (Liljeström and Garoff, 1991). The cloning vector contains the four nonstructural SFV genes (nsP1-4), which code for the replicase complex, responsible for RNA replication. The subgenomic SFV 26S promoter is 23

located at the 3’ end of the nsP4 gene and drives the expression of the foreign gene of interest inserted immediately downstream of nsP4. The helper vector contains the SFV structural genes encoding the capsid and membrane proteins, required for packaging of infectious particles. RNA molecules are transcribed in vitro from both the cloning and helper vector and transfected into BHK cells either by electroporation or lipofection for in vivo packaging of recombinant SFV particles (Fig. 1). High-titer virus stocks (108-109 infectious particles/ml) harvested 24 h post-transfection are ready for infection of host cells without any further purification or concentration. The second-generation helper vector pSFVHelper2 contains three point mutations in p62 (precursor for the E2 and E3 membrane proteins), which will result in production of conditionally infectious SFV particles (Berglund et al., 1993). These virus stocks require

Lundstrom et al: Semliki Forest virus vectors treatment with !-chymotrypsin to restore the infectivity. The broad host range of SFV allows infection of a wide variety of mammalian cell lines and primary cell cultures (Lundstrom, 1999). Due to the presence of the nonstructural SFV genes in the recombinant SFV RNA introduced into the host cell, extensive RNA replication will occur directly in the cytoplasm. Generation of large quantities of RNA results in high-level expression of foreign genes. However, due to the absence of the SFV structural genes, no further virus particles are generated, which leads to only transient expression from these suicidal vectors.

SFV vectors have been applied for high-level expression of many topologically different proteins. Successful expression of nuclear, cytoplasmic, membrane, and secreted proteins have been reported (Liljestrรถm and Garoff, 1991, Lundstrom 1997). The advantages of using SFV vectors are the rapid production of high-titer virus and the broad host range that allows efficient infection of a variety of mammalian and insect cell lines, primary cell cultures (Lundstrom 1999) as well as neurons in vivo (Lundstrom et al., 1999).

Fig. 1A. Schematic presentation of recombinant SFV particle production. The gene of interest is cloned into the SFV cloning vector and subjected together with SFV-Helper vector to in vitro transcription with SP6 RNA polymerase. RNA molecules are cotransfected into BHK cells for in vivo packaging of recombinant SFV particles.

Fig. 1B. Possible applications of SFV vectors.

Gene Therapy and Molecular Biology Vol 4, page 25 In this review, we describe the versatility of the SFV expression system. We have chosen to focus on SFV vectors but comparable studies have been performed with similar vectors for other types of alphaviruses, namely Sindbis virus and Venezuelan Equine Encephalitis virus. SFV vectors can be efficiently used for receptors previously known to be difficult to express at high levels. The expression has also been expanded to large-scale production, e.g., to facilitate receptor purification for structural studies. Furthermore, the replication-deficient nature of the vectors has allowed in vivo studies in rat brain. Initial studies indicated that there is a great potential for the use of SFV vectors for cancer gene therapy. SFV has also been proven a good candidate as a vector to be used in vaccine production. Finally, we will discuss recent improvements of SFV vectors.

II. Recombinant gene expression Although many types of recombinant proteins have been expressed from SFV vectors, we will focus here on G-protein coupled receptors (GPCRs). Several GPCRs have most efficiently been expressed from SFV vectors (Lundstrom 1999). Metabolic labeling experiments demonstrated high expression levels of the human neurokinin-1 receptor (hNK1R) (Lundstrom et al., 1994). However, SFV-infection caused inhibition of the endogenous gene expression and also triggered the induction of apoptosis, resulting in host cell death within 48 to 72 hours. This was also reflected in the expression pattern measured by specific radioligand binding for different receptors. The maximal expression time and duration of measurable binding activity varied from one receptor to another. For instance, the highest levels of expression for hNK1-R were obtained at 16 hours postinfection, whereas the hamster ! 1b-adrenergic receptor reached maximal activity as late as 40 hours post-infection (Scheer et al., 1999). Saturation assays on isolated membranes revealed extremely high binding activity with B max values up to 80 pmol receptor / mg protein. Binding experiments carried out on whole cells indicated receptor densities of >6 x 10 6 receptors / cell. GPCRs could be efficiently expressed in several different mammalian host cells (BHK, CHO, COS-7, HEK293 cells, etc.). Additionally, high infection rates (75-90%) were obtained in primary rat hippocampal neurons. Two rat odorant receptors, I7 and OR5, as well as the C. elegans odr-10 receptor, were expressed in these neurons and were found to be localized at the plasma membrane by immunofluorescence microscopy (Monastyrskaia et al., 1999). This observation is particularly interesting, since the expression of these olfactory receptors from the same SFV vectors in BHK cells did not target the receptors to the plasma membrane. Functional coupling of GPCRs to G proteins could be demonstrated in CHO, BHK and COS-7 cells by measuring agonist-induced intracellular Ca2+-release (Fura2, FLIPR), inositol phosphate accumulation, cAMP stimulation and GTP" S-binding. Enhanced functional 25

responses could be obtained for the ! 1b-adrenergic receptor expressed from an SFV vector in COS-7 cells after coinfection with SFV expressing the ! 1b-adrenergic receptor and the G!q subunit (Scheer et al., 1999). Interestingly, quadruple infections with SFV vectors for the ! 1badrenergic receptor, G!q as well as G#2 and G" 2 subunits resulted in further increased response to epinephrine stimulation. These results indicate that the decrease in endogenous G-protein levels (due to the inhibition of host cell protein expression) in SFV-infected cells can be compensated by multiple infections. The adaptation of mammalian cell lines (BHK, CHO, HEK293, rat C6 glioma cells) to suspension culture has facilitated large-scale production of recombinant receptors for both drug screening purposes as well as for structural studies. Recently, the hexa-histidine tagged mouse serotonin 5-HT3 receptor was expressed in bioreactors (11.5 liter) to yield large quantities of highly purified receptor (Lundstrom et al., 1997; Hovius et al., 1998). Preliminary data from cryo-EM and gel filtration chromatography confirmed postulations that the 5-HT3 receptor is a homo-pentameric channel. The discovery that the SFV capsid sequence contains a translation enhancement signal has made it possible to further increase transgene expression levels (Sjรถberg et al., 1994). Sjรถberg and co-workers showed that fusion of the fulllength capsid gene to a transferrin receptor resulted in efficient release of the recombinant tranferrin receptor from the fusion by capsid-protein mediated autolytic cleavage. The transferrin receptor yield was approximately 10-fold higher when expressed as a fusion protein to the SFV capsid. We have fused the human neurokinin-1 receptor (hNK1R) to the full-length capsid sequences and achieved 5- to 10-fold higher expression levels. Additionally, the capsid protein is efficiently cleaved from the hNK1R (Fig. 2). This has allowed large-scale production of recombinant hNK1R that has been subjected to solubilization and purification activities with the aim to obtain 2D- and 3Dcrystals for high-resolution structure determination studies.

III. In vivo studies To study the feasibility of SFV vectors for in vivo applications, replication-deficient SFV-LacZ virus was injected into the striatum and amygdala of adult male Wistar rats. Animals were injected with 105 infectious particles and subjected to behavioral studies for 28 days. No differences in body weight, body temperature, feeding behavior, spontaneous exploration, sensorimotor function and muscular capacity were observed between virusinjected rats and control animals (Lundstrom et al., 1999). Histological examinations revealed high local #galactosidase expression levels (Fig. 3). Maximal expression was observed 1-2 days post-injection, and the reporter gene expression decreased with time. Some weak X-gal staining was still visible at 28 days post-injection. However, this was most likely due to the high stability of the recombinant #-galactosidase enzyme. In situ

Lundstrom et al: Semliki Forest virus vectors reduced the infection rate of BHK cells by a factor of 105 (Ohno et al., 1997). Targeted infection could be achieved for human cell lines treated with monoclonal antibodies reacting with cell-surface antigens. Similar approaches have been taken for SFV vectors, where protein A domains have been introduced into various regions of E1 and E2 membrane proteins. However, the second generation SFV vectors possess the additional safety feature of conditional infectivity compared to Sindbis virus (Berglund et al., 1993). This should further enhance the selective advantage of targeting IgG-bound cells vs. non-bound cells by a factor of 105. The development of chimeric SFV vectors is now in progress.

hybridization studies confirmed the transient nature of SFV-mediated gene expression where no LacZ mRNA could be detected any more 4 days post-injection. The SFV infection pattern in rat brain was remarkably neuron-specific. Similar results were found in rat hippocampal slices at 1-5 days after infection with SFV vectors encoding #-galactosidase, green fluorescence protein (GFP), and GFP fused to the ionotropic glutamate receptor 1 subunit (Fig. 4). Upon injection of recombinant SFV into the pyramidal cell layer, the majority of transgene-positive cells were neurons (>90% for GFP), with pyramidal cells, interneurons and granule cells expressing high levels of #-galactosidase and GFP (Ehrengruber et al., 1999).

IV. Gene therapy applications SFV vectors have been used for indirect gene therapy applications to produce infectious recombinant Moloney murine leukemia virus particles in BHK cells (Li and Garoff, 1996). The gag-pol and env genes as well as a recombinant retrovirus genome (LTR-$+-neo-LTR) were introduced into individual SFV vectors and were cotransfected into BHK cells to generate extracellular viruslike particles that possessed reverse-transcriptase activity. Recently, it has been shown that intron-containing sequences could also be efficiently packaged by the SFV expression system (Li and Garoff, 1998). The rapid virus production and relatively high titers obtained (4 x 106 colony forming units/ml) make this an attractive alternative for retrovirus production. The high expression levels achieved from SFV vectors combined with their ability to shut down host cell protein synthesis and to induce apoptosis have increased the interest in cancer gene therapy applications. SFV-LacZ virus infection of prostate tumor cell lines and of biopsies from patients revealed a strong apoptotic effect measured by flow cytometry (Hardy et al., manuscript in preparation). Furthermore, it has been demonstrated that pre-immunization with self-replicating SFV-LacZ RNA could protect mice from tumor challenge (Ying et al., 1999). Therapeutic immunization with SFV-LacZ RNA also prolonged survival of BALB/c mice with established tumors. SFV vectors were tested in vivo in a murine B16 tumor model (Asselin-Paturel et al., 1999). SFV vectors expressing the p35 and p40 subunits of interleukin-12 from two subgenomic promoters (Zhang et al., 1999) were injected into tumors. This led to significant tumor regression and inhibition of tumor blood vessel formation. No sign of toxicity was observed in SFV-IL12-treated mice. The anti-tumor effect could be enhanced by repeated injections and, most encouragingly, no antiviral response to SFV was detected. The broad host range of alphaviruses has been of some concern for using these vectors for gene therapy applications. Attempts to target Sinbis virus vectors have recently been carried out. IgG-binding domains of protein A were introduced into the E2 membrane protein, which

Fig. 2. Metabolic labeling of BHK cells electroporated and infected with SFV vectors. BHK cells co-electroporated with pSFV1-NK1R (lane 1) and pSFVCAP-NK1R-His (lane 3) and RNA from pSFV-Helper2. BHK cells infected with SFVNK1R (lane 2) and SFVCAP-NK1R-His (lane 4) virus. Pulselabeling with [35S]-methionine was carried out after 4 h (electroporations) and 16 h (infections). C, capsid; E1, SFV E1 membrane protein; p62, precursor for SFV E2 and E3 membrane proteins; NK1R, neurokinin-1 receptor; NK1R-His, neurokinin-1 receptor with hexa-histidine tag; CAP-NK1R, capsid-neurokinin-1 receptor fusion.

Gene Therapy and Molecular Biology Vol 4, page 27

Fig. 3. Expression of recombinant #-galactosidase in rat brain. SFV-LacZ virus (105 particles per injection) was injected into adult male Wistar rat brain (striatum and amygdala) and X-gal staining performed at 1 day post-injection. (A) Whole brain section. Arrow indicates injection site, arrowhead spread of virus into the ventricles. The counter-staining seen in cerebellum is nonspecific. (B) Higher magnification, showing expression of #-galactosidase in neuronal processes.

Lundstrom et al: Semliki Forest virus vectors

Fig. 4. Recombinant SFV-mediated gene transfer into pyramidal cells of cultured hippocampal slices from postnatal rats. (A, B, and C) Expression of the GFP reporter gene. Fluorescence illuminations of living slices at 11-16 days in culture and 2 (A and C) and 5 days (B) after injection of 10 4-10 5 infectious SFV-GFP particles into the pyramidal cell layer. (B) GFP-positive CA1 pyramidal cells. (C) Apical dendrites from an infected CA1 pyramidal cell. Note the GFP-positive spines, which are typical of pyramidal cells (arrowheads). (D) SFV-mediated expression of an N-terminal GFP fusion of the ionotropic glutamate receptor 1 subunit. Fluorescence illumination from the CA1 region of a living slice at 29 days in culture and 4 days post-infection. The cDNA for the GFP fusion construct was provided by Drs. Rolf Sprengel and Volker Mack (Max-Planck-Institute for Medical Research, Heidelberg, Germany). Abbreviations: so, stratum oriens; sp, stratum pyramidale; sr, stratum radiatum. Bars: 300 µm (A), 75 µm (B and D), and 25 µm (C).

protection against P815 tumor challenge (Colmenero et al., 1999). In addition to application of recombinant SFV particles, direct injection of nucleic acids have been used for SFV vectors. Injection of self-replicating RNA containing the influenza NP gene into the quadriceps muscle of mice resulted in high humoral responses (Zhou et al., 1994). The advantage of using naked RNA for immunization is the high safety due to the rapid degradation of the injected RNA and the lack of integration into the host genome. Recently, naked RNA vectors expressing the HIV-1 gp41 glycoprotein were injected intramuscularly into mice for the generation of monoclonal antibodies (Giraud et al., 1999). Layered plasmid DNA vectors that utilize an RNA polymerase II expression cassette to drive the transcription of a selfamplifying RNA (replicon) vector have been engineered

V. Vaccine production SFV vectors have been shown to be potential candidates in vaccine production. Intravenous injections of recombinant replication-deficient SFV particles expressing the influenza NP gene led to humoral responses with high antibody titers in BALB/c mice (Zhou et al., 1994). As few as 100 infectious SFV-NP particles induced a strong cytotoxic T-cell (CTL) response and, after one booster injection, a CTL-memory lasting for more than 40 days was generated (Zhou et al., 1995). Pigtail macaques immunized with recombinant SFV particles expressing simian immunodeficiency virus (SIV) gp160 were protected against lethal disease (Mossman et al., 1996). Recently SFV particles expressing P815 tumor antigen generated a strong CTL response and demonstrated

Gene Therapy and Molecular Biology Vol 4, page 29 for SFV (Berglund et al., 1998). These SFV DNA vectors allowed the direct use of plasmid DNA and resulted in higher levels of humoral and cellular immune responses than for conventional DNA plasmids, immunizing animals against normally lethal challenges with influenza virus. In another study, SFV RNA and DNA vectors and recombinant SFV particles were applied to generate monoclonal antibodies against prion proteins (Krasemann et al., 1999).

VI. Vector development Despite the various applications and the high efficiency of SFV vectors, there are still areas where improvements are possible. The extreme transgene expression levels obtained from SFV vectors are not comparable to physiological levels anymore. More moderate expression levels would therefore match better the endogenous gene expression. One approach to achieve this is by site-directed mutagenesis of the subgenomic 26S promoter. It has been shown that a 3-nucleotide insertion in the Sindbis virus subgenomic promoter region resulted in a 100-fold reduction of the promoter activity (Raju and Huang, 1991). Very recently, a series of amino acid substitutions and insertions into the SFV 26S subgenomic promoter resulted in significantly lower reporter gene expression (Lundstrom et al., manuscript in preparation). Depending on the type of mutation in the 26S subgenomic promoter, expression levels of 1%, 3%, and 30% of the wild-type promoter activity were obtained. Despite these reduced expression levels, these mutant viruses still substantially inhibited host cell protein synthesis and, therefore, had cytotoxic effects. Novel SFV vectors with lower cytopathogenicity would be beneficial. Non-cytopathogenic Sindbis vectors have been described (Agapov et al., 1998). This change in phenotype was due to a single amino acid substitution in the nsP2 gene (Pro726Ser). However, titer yields of this virus were rather low and the non-cytopathogenic phenotype was mainly restricted to BHK and Vero cells. Recently, an SFV strain with reduced virulence and only minor inhibition of host cell DNA synthesis was described. Analysis of the nuclear localization signal in the nsP2 gene (Pro647-Arg648-Arg649-Arg650-Val651) revealed a point mutation (Arg649!Asp) in this SFV strain which results in the retention of the nsP2 protein to the cytoplasm and thus prevents interference with host cell gene transcription (Rikkonen, 1996). Site-directed mutagenesis of nsP2 at Arg650 (to Asp) and Ser259 (to Pro) resulted in an SFV mutant with significantly higher transgene expression. Interestingly, in contrast to the conventional SFV vector, the host protein synthesis was not as dramatically reduced (Fig. 5). The reduced cytopathogenicity was observed in various mammalian cell lines (BHK, CHO, and HEK293 cells) and primary rat hippocampal neurons and the survival of host cells was substantially prolonged (Lundstrom et al., manuscript in preparation). Such novel non-cytopathogenic vectors will be extremely useful, not only for expression studies, but also for antisense and ribozyme applications. 29

Fig. 5. Metabolic labeling of BHK cells infected with recombinant SFV vectors. Pulse-labeling of BHK cells infected with SFVCAP-NK1R-His (lane 1) and SFV259-650CAPNK1R-His (lane 2) with [35S]-methionine 16 h post-infection. hNK1R, human neurokinin-1 receptor; SFV 259-650, SFV vector with Ser 259! Pro and Arg650! Asp mutations in nsP2.

VII. Conclusions and future prospects SFV vectors can be used for a variety of applications in modern molecular biology and drug discovery. Rapid high-titer virus production, broad host range, direct cytoplasmic RNA replication and extremely high yields of expressed recombinant protein, have made SFV vectors attractive alternatives to other expression vectors.

Lundstrom et al: Semliki Forest virus vectors Li, K.-J. and Garoff, H. (1996). Production of infectious recombinant Moloney murine leukemia virus particles in BHK cells using Semliki Forest virus-derived RNA expression vectors. Proc. Natl. Acad. Sci. USA 93: 1165811663. Li, K.-J. and Garoff, H. (1998). Packaging of intron-containing genes into retrovirus vectors by alphavirus vectors. Proc. Natl. Acad. Sci. USA 95, 3650-3654. Liljeström, P. and Garoff, H. (1991). A new generation of animal cell expression vectors based on the Semliki Forest virus replicon. Bio/Technology 9: 1356-1361. Lundstrom, K. (1997). Alphaviruses as expression vectors. Curr. Opin. Biotechnol. 8: 578-582. Lundstrom, K. (1999). Alphavirus vectors as tools in neurobiology and gene therapy. J. Receptor & Signal Transd. Res. 19: 673-686. Lundstrom, K., Michel, A., Blasey, H., Bernard, A.R., Hovius, R., Vogel, H. and Surprenant, A. (1997). Expression of ligand-gated ion channels with the Semliki Forest virus expression system. J. Receptor & Signal Transd. Res. 17: 115-126. Lundstrom, K., Mills, A., Buell, G., Allet, E., Adami, N. and Liljeström, P. (1994). High-level expression of the human neurokinin-1 receptor in mammalian cell lines using the Semliki Forest virus expression system. Eur. J. Biochem. 224: 917-921. Lundstrom, K., Richards, J.G., Pink, J.R. and Jenck, F. (1999). Efficient in vivo expression of a reporter gene in rat brain after injection of recombinant replication-deficient Semliki Forest virus. Gene Ther. Mol. Biol. 3: 15-23. Monastyrskaia, K., Goepfert, F., Hochstrasser, R., Acuna, G., Leighton, J., Pink, J.R. and Lundstrom, K. (1999). Expression and intracellular localisation of odorant receptors in mammalian cell lines using Semliki Forest virus vectors. J. Receptor & Signal Transd. Res. 19: 687701. Mossman, S.P., Bex, F., Berglund, P., Arthos, J., O’Neil, S.P., Riley, D., Maul, D.H., Bruck, C., Momin, P., Burny, A., Fultz, P.N., Mullins, J.I., Liljeström, P. and Hoover, E.A. (1996). Protection against lethal simian immunodefieciency virus SIVsmmPBj14 disease by a recombinant Semliki Forest virus gp160 vaccine and by gp120 subunit vaccine. J. Virol. 70: 1953-1960. Ohno, K., Sawai, K., Ijima, Y., Levin, B. and Meruelo, D. (1997). Cell-specific targeting of Sindbis virus vectors displaying IgG-binding domains of protein A. Nat. Biotechnol. 15, 763-767. Raju, R. and Huang, H.V. (1991). Analysis of Sindbis virus promoter recognition in vivo using novel vectors with two subgenomic mRNA promoters. J. Virol. 65: 25012510. Rikkonen, M. (1996). Functional significance of the nucleartargeting and NTP-binding motifs in Semliki Forest virus nonstructural protein nsP2. Virology 218: 352-361. Scheer, A., Björklöf, K., Cotecchia, S. and Lundstrom, K. (1999). Expression of the ! 1b-adrenergic receptor and G protein subunits in mammalian cell lines using the Semliki Forest virus expression system. J. Receptor & Signal Transd. Res. 19: 369-378. Sjöberg, M., Suomalainen, M. and Garoff, H. (1994). A significantly improved Semliki Forest virus expression system based on translation enhancer segments from the viral capsid gene. Bio/Technology 12: 1127-1131.

Additional improvements in vector design will further increase the application range of SFV. In the meanwhile, SFV vectors have been used in animal models for initial gene therapy experiments. The application of SFV vectors to man is soon to come.

Acknowledgements The authors thank Danielle Hermann, Raquel Herrador, Daniel Rotmann and Marie-Therese Zenner for their skillful assistance in the SFV vector construction and virus production, and Lotty Rietschin and Lucette Heeb for the preparation of rat hippocampal slice cultures. We are grateful to Martine Maco and Martine Kapps for their assistance in the neurosurgical injections and to Jürg Messer for his help in imaging of brain sections. This work was supported by the Swiss National Science Foundation (grant no, 31-57’125.99 to M.U.E.).

References Agapov, E.V., Frolov, I., Lindenbach, B.D., Pragai, B.M., Schlesinger, S. and Rice, C.M. (1998). Noncytopathogenic Sindbis RNA vectors for heterologous gene expression. Proc. Natl. Acad. Sci. USA 95: 12989-12994. Asselin-Paturel, C., Lassau, N., Guinebretiere, J.-M., Zhang, J., Gay, F., Bex, F., Hallez, S., Leclere, J., Peronneau, P., MamiChouaib, F. and Chouaib, S. (1999). Transfer of the murine interleukin-12 gene in vivo by a Semliki Forest virus vector induces B16 tumor regression through inhibition of tumor blood vessel formation monitored by Doppler ultrasonography. Gene Ther. 6: 606-615. Berglund, P., Sjöberg, M., Garoff, H., Atkins, G.J., Sheahan, B.J. and Liljeström, P. (1993). Semliki Forest virus expression system: production of conditionally infectious recombinant particles. Bio/Technology 11, 916-920. Berglund, P., Smerdou, C., Fleeton, M.N., Tubulekas, I. and Liljeström, P. (1998). Enhancing immune responses using suicidal DNA vaccines. Nat. Biotechnol. 16: 562-565. Colmenero, P., Liljeström, P. and Jondal, M. (1999). Induction of P815 tumor immunity by recombinant Semliki Forest virus expressing the P1A gene. Gene Ther. 6: 1728-1733. Ehrengruber, M., Lundstrom, K., Schweitzer, C., Heuss, C., Schlesinger, S. and Gähwiler, B.H. (1999). Recombinant Semliki Forest virus and Sindbis virus efficiently infect neurons in hippocampal slice cultures. Proc. Natl. Acad. Sci. USA 96: 7041-7046. Giraud, A., Ataman-Onal, Y., Battail, N., Piga, N., Brand, D., Mandrand, B. and Verrier, B. (1999). Generation of monoclonal antibodies to native human immunodeficiency virus type 1 envelope glycoprotein by immunization of mice with naked RNA. J. Virol. Meth. 79: 75-84. Hovius, R., Tairi, A.-P., Blasey, H., Bernard, A., Lundstrom, K. and Vogel, H. (1998). Characterization of a mouse serotonin 5-HT3 receptor purified from mammalian cells. J. Neurochem. 70: 824-834. Krasemann, S., Jurgens, T. Bodemer, W. (1999). Generation of monoclonal antibodies against prion proteins with an unconventional nucleic acid-based immunization strategy. J. Biotechnol. 73: 119-129.

Gene Therapy and Molecular Biology Vol 4, page 31 Ying, H., Zaks, T.Z., Wang, R.-F., Irvine, K.R., Kammula, U.S., Marincola, F.M., Lettner, W.W. and Restifo, N.P. (1999) Cancer therapy using self-replicating RNA vaccine. Nat. Medicine 5: 823-827. Zhang, J., Asselin-Paturel, C., Bex, F., Bernard, J., Chehimi, J., Willems, F., Caignard, A., Berglund, P., Liljestrรถm, P., Burny, A. and Chouaib, S. (1997). Cloning of human IL12 p40 and p35 DNA into the Semliki Forest virus vector: expression of IL-12 in human tumor cells. Gene Ther. 4: 367-374. Zhou, X., Berglund, P., Rhodes, G., Parker, S.E., Jondal, M. and Liljestrรถm, P. (1994). Self-replicating Semliki Forest virus RNA as a recombinant vaccine. Vaccine 12: 15101513. Zhou, X., Berglund, P., Zhao, H., Liljestrรถm, P. and Jondal, M. (1995). Generation of cytotoxic and humoral immune responses by nonreplicative recombinant Semliki Forest virus. Proc. Natl. Acad. Sci. USA 92: 3009-3013.

Kenneth Lundstrom

Gene Therapy and Molecular Biology Vol 4, page 33 Gene Ther Mol Biol Vol 4, 33-43. December 1999.

Ovine adenovirus vectors promote efficient gene delivery in vivo Research Article

Peter Löser1 , Günter Cichon2 , Gary S. Jennings1 , Gerald W. Both3 , and Christian Hofmann1 1

HepaVec AG für Gentherapie, Berlin, Germany, 2Max Delbrück Center for Molecular Medicine, Berlin, Germany, Division of Molecular Science, CSIRO, North Ryde, New South Wales 2113, Australia __________________________________________________________________________________________________ 3

Correspondence: Dr. Christian Hofmann, Vice President R & D, HepaVec AG für Gentherapie, Robert-Rössle-Str. 10, 13125 Berlin-Buch, Germany, Tel: ++49-30-94892283, Fax: ++49-30-94892913, E-mail: chofmann@hepavec.com Key words: ovine adenovirus vectors, pre-existing antibodies, human ! 1-antitrypsin, viral tropism Received: 8 October 1999; accepted 25 October 1999

Summary The use of vectors derived from human adenoviruses in gene therapy is limited by pre-existing humoral immunity against these vectors in many individuals. We have recently reported the use of a vector derived from ovine adenovirus (OAV) isolate 287 that can transduce cells in vivo (Hofmann et al., 1999). In this report we present data regarding the physical stability of the OAV particles and demonstrate their ability to infect mice preimmunized with a human adenoviral (hAd) vector. The tissue distribution of the vector delivered by several routes of administration is also examined and shown to be different from that observed for vectors derived from human adenoviruses. The construction and rescue of a recombinant virus that expresses the green fluorescent protein gene will further facilitate studies on the tropism of OAV in vivo. modulating proteins (Ilan et al., 1997). However, these strategies cannot be applied to humans due to the fact that most adenovirus vectors are derived from group C subtype 2 and 5 adenoviruses that have infected the vast majority of the population (Horwitz, 1996) and neutralizing antibodies would cause immediate vector inactivation making even a single unimpeded application impossible. For hAd vectors, serotype switching has been proposed as one possible solution to pre-existing immunity (Mastrangeli et al., 1996; Mack et al., 1997). This strategy might indeed allow the application of adenovirus vectors to humans if a rare serotype is used. However, safety concerns regarding transcomplementation of hAd vectors by wild-type hAd infection cannot be overcome with this approach. The problem of pre-existing humoral immunity against the vector also affects other potential gene therapy vectors of human origin such as AAV, which is endemic in 85% of the population (Mayor et al., 1976), or herpes simplex virus, which is endemic in 50 to 95% of the population (Whitley et al., 1996). Therefore, to approach the problem of preexisting humoral immunity to vectors derived from human viruses in a more general way, we and others have suggested the use of recombinant adenoviruses of non-human origin as

I. Introduction A. Rationale for using non-human adenovirus vectors in gene therapy Vectors derived from human adenoviruses (hAd) have been widely used for transfer of potentially therapeutic genes into animals, but successful use of these vectors in humans has been limited by immunological barriers. Transgene expression is often transient due to both a CD8+-dependent T cell response to infected cells which leads to clearance of virus from target tissues (Yang et al, 1994 a, b and 1995), as well as by an immune response against the transgene product (Tripathy et al., 1996; Morral et al., 1997). Moreover, formation of neutralizing antibodies to the vector prevents effective re-application of adenoviral vectors (Smith et al., 1993; Dai et al., 1995). This problem has been overcome in animals by strategies involving immunosupression (Kay et al., 1995; Vilquin et al., 1995; Yang et al., 1996; Kolls et al., 1996; Kay et al., 1997), induction of immune tolerance (Kass-Eisler et al., 1996; Ilan et al., 1998) or co-expression of immuno33

Lรถser et al: Ovine adenovirus vectors gene therapy vectors. Adenoviruses from at least five non-human species, which might have some potential for gene therapy applications, have been used for generation of recombinant vectors (Mittal et al., 1995; Klonjkowski et al., 1997; Zakhartchouk et al., 1998; Reddy et al., 1999; Michou et al., 1999; see also Table 2). Among these, the ovine adenovirus (OAV) isolate 287 (Boyle et al., 1994) has been studied extensively (Vrati et al., 1995, 1996a,b; Khatri et al., 1998; Venktesh et al., 1998) and developed as a vector with the capability to infect mammalian cells in vitro and in vivo (Vrati et al., 1996c; Khatri et al., 1997; Xu et al., 1997; Xu & Both, 1998; Hofmann et al., 1999). In this report we describe recent progress with this novel vector system.

no obvious E3 region and the genomic location of the putative E4 region also differs (Vrati et al., 1995 and 1996b). Homologous sequences to genes coding for structural protein IX and core protein V of human adenoviruses as well as a virus-associated RNA gene are absent in the OAV genome (Vrati et al., 1996b; Venktesh et al., 1998). In addition, OAV uses a primary receptor that is distinct from the Ad5 receptor (Xu and Both, 1998) and lacks an identifiable integrin-binding domain (Vrati et al, 1996a). Recombinant OAVs were initially constructed by insertion of DNA cassettes in either orientation (Xu et al, 1997; Hofmann et al, 1999) into a region of the genome between the pVIII and fiber protein (designated as site I in Figure 1), or into a unique SalI site located near the right hand end (designated as site II in Figure 1). Recently, viruses carrying an expression cassette in site III (base 26575, plasmid OAV600, Xu et al, 1997), which comprises an ApaI/NotI polylinker inserted between the RHE and E4 transcription units (Khatri and Both, 1998), have also been rescued (G. W. Both, unpublished results). For sites I and II, up to 4.3 kb of foreign DNA could be inserted without a compensating deletion and foreign gene insertion did not reduce virus growth. Thus, DNA comprising at least 114% of the wild type viral genome can be packaged in the OAV capsid (Vrati et al., 1996c; Xu et al. 1997). Moreover, deletion of a 2kb sequence between sites II and III that contains apparently redundant ORFs did not effect virus growth, suggesting that OAV may package at least 6.3 kb foreign DNA without further deletions (Xu et al. 1997).

B. OAV287 derived vectors On the basis of phylogenetic analyses of protease and hexon genes OAV287 has been grouped together with several BAV subtypes and Egg Drop Syndrome virus in the proposed genus Atadenovirus (Harrach et al., 1997; Harrach and Benkรถ, 1998) which is distinguished from the other adenovirus genera by base composition and genome arrangement. The organization of the OAV287 genome is shown in Figure 1. As in all adenoviruses the genome is organized into early and late transcription units and homologies to most structural and E2 genes of the genus Mastadenovirus are evident (Vrati et al., 1995; Vrati et al., 1996a). However, unlike the Mastadenoviruses there is no typical E1A/B region,

Figure 1. Organisation of the OAV287 genome. The complete sequence of OAV287 is available (Genbank Accession Number U40839). The locations of the early regions LHE, RHE, E2 and putative E4 are shown in bold type and arrows. Reading frames with homologues in other adenoviruses are named except for the p32 which is unique to Atadenoviruses. Promoter regions (Khatri and Both, 1998) are identified by filled circles. Non-essential sites I-III for insertion of gene cassettes are indicated.

Gene Therapy and Molecular Biology Vol 4, page 35

Figure 2: Heat stability of OAVhaat vs Ad5haat. The respective vectors were diluted to a titer of 1x10 9 per ml, incubated for 30 minutes at the temperature indicated and used in serial dilutions for infection of the permissive cell line (CSL-503 or HEK 293, respectively). Titers lower than 103 infectious particles per ml were considered to be non-infectious.

shown in Figure 2, incubation of Ad5haat at 51°C diminished virus titer by more than one order of magnitude, and incubation at 54°C completely abolished its ability to produce CPE on 293 cells. In contrast, OAVhaat remained stable at up to 54°C and retained some ability to infect CSL503 cells even after incubation at 57°C. Complete inactivation was observed at 60°C. Thus, with regard to heat stability, OAV holds an intermediate position among adenoviruses in that it is more heat stable than hAd vectors but marginally less stable than CELO. This is interesting because protein IX, which is of major importance for Ad5 capsid stability (Ghosh-Choundhury et al., 1987), is absent from both OAV287 and CELO. Thus, other protein-protein interactions or new capsid protein(s) might substitute for pIX in the OAV capsid leading to a more pronounced heat stability. In addition, the heat stability of OAV could be of importance for vector preparation protocols since it may facilitate the purification of OAV-derived vectors from heatsensitive contaminants.

Plasmids to generate OAV recombinants are produced either by direct cloning of the sequence of interest into the OAV genome (Vrati et al., 1996c; Xu et al., 1997 and 1998; Khatri et al., 1998) or by homologous recombination in E.coli BJ5183 (Hofmann et al., 1999). Recombinant viruses are generated by releasing the infectious, linear genome from such plasmids with KpnI digestion, followed by transfection of DNA into the ovine foetal lung cell line, CSL503, which is permissive for OAV replication (Pye, 1989). Lipofectamine was initially used for transfection but more recently cationic lipids related to the series described by Cameron et al, (1999) have produced increased transfection of CSL503 cells (G. W. Both, unpublished results), with a corresponding increase in the ability to rescue recombinant OAVs. OAV grows to useful titers in CSL503 cells although, in contrast to Ad5 growth in 293 cells, about 50% of OAV is released into the medium prior to harvest. Yields of >10,000 opu/cell can routinely be obtained from the cell pellet after CsCl-gradient purification.

B. Inactivation of hAd vector by human serum

II. Results

Pre-existing humoral immunity against hAd vectors might be critical for the use of these vectors in humans. Figure 3 shows the results of an in vitro experiment to examine this. Ad5luc, a hAd5-derived vector expressing the luciferase reporter gene under control of the RSV 3‘LTR, was incubated with a random human serum sample and subsequently used for infection of human liver-derived HuH7 cells at a moi of 100pfu/cell. Pretreatment of the hAd vector with heat-inactivated (30 min at 56°C) serum resulted in a dramatic decrease in luciferase activity (0,8% of control) in these cells. This decrease reflects the influence of pre-existing neutralizing antibodies against hAd5 in the

A. Heat stability of OAV vectors Human adenoviruses become inactivated at temperatures higher than 48°C. In contrast, CELO and related viruses from the avian adenovirus group were shown to be stable at temperatures of up to 56°C (Michou et al., 1999). To check if OAV-derived vectors are more stable than hAd vectors OAVhaat and Ad5haat were incubated for 30 minutes at several temperatures and the remaining virus titer was determined on permissive CSL-503 or HEK293 cells, respectively. As 35

Löser et al: Ovine adenovirus vectors serum, the titer of which (1:320) was determined to be the mean of that within the human population (Hofmann et al., 1999). In addition to neutralizing antibodies, the immune system produces an excess of antibodies after a natural infection that bind to the virus, but do not neutralize it directly. However, these antibodies consequently activate the complement system. Several antibody- and complement-dependent mechanisms generally mediating virus inactivation have been described (Cooper, 1998): i) opsonization and rapid clearance by the reticuloendothelial system, ii) envelopment with antibody and/or complement, thereby masking (hAd)-receptor ligands on the viral surface, iii) antibody and/or complement-dependent viral aggregation. To investigate these possibilities in vitro, we explored the fate of hAd5luc vector in the presence of either untreated serum (measuring the effect of neutralising antibodies + intact complement), or sera, depleted in distinct activation routes (alternative and classical pathway of the complement system). The classical complement pathway can be blocked by EGTA/Mg++-treatment. Proteins of the alternative pathway can selectively be blocked by treatment of sera at 48°C. We observed that blocking of the classical complement pathway by EGTA-Mg++-treatment did not significantly increase the inhibitory effect of heatinactivated serum at 56°C (due only to neutralising antibodies), whereas blocking of the alternative pathway by heating at 48°C or untreated serum (intact complement system) caused a nearly complete inactivation of the hAd vector (Figure 3).

Thus, inactivation of the hAd vector in vitro is predominantly caused by neutralizing antibodies, but the data imply that the complement system increases the inactivation rate of human adenovirus vectors through deposition of C-proteins of the classical pathway on the viral surface. Moreover, in vivo (clinical situation), a rapid clearance of opsonized hAd vectors via the reticuloendothelial system and viral aggregation might be expected in patients with pre-existing anti-hAd antibodies.

C. OAV overcomes pre-existing immunity to hAd vectors To mimic the situation of pre-existing immunity in humans, Balb/C mice were immunized with 5x10 9 pfu of Ad5lacZ or Ad5haat, respectively. At day 31 p.i., a second infection at the same dose was performed: one group of mice received Ad5haat while the second group were given OAVhaat. When immunization at day 0 was performed with Ad5lacZ, injection of OAVhaat, but not Ad5haat, resulted in strong expression of the reporter gene at day 3 after the second injection (Table 1). This confirms, that preexisting immunity to the hAd vector prevented in vivo transduction with the same hAd vector but not with the OAV vector. However, when the vector first injected was Ad5haat, no haat protein was detectable by ELISA after the second injection of either vector. This phenomenon was clearly due to antibodies against haat in Balb/C mice at the time of the second injection (1:10.000 to 1:100.000 as determined by anti-haat ELISA).

Figure 3: Ad5 vector is inactivated by human serum. Ad5luc (5x10 6 pfu ) was incubated at 37°C for 30 min with 50µl of various human sera pre-treated for 30 minutes as indicated and then used to infect of 5x10 4 HuH7 cells in 24 well plates. At 48hours postinfection, cells were lysed and luciferase activity was determined. The graph represents the average of three independent experiments.

Gene Therapy and Molecular Biology Vol 4, page 37 Analysis of RNA in liver at day 4 after second vector application showed clear expression of haat specific transcripts in liver and heart when the second injection was performed with OAVhaat (Figure 4), but not when the second vector was Ad5haat (not shown). Thus, in addition to the pre-existing immunity to the vector, humoral immunity against the transgene product might prevent its function in the organism and must be taken into consideration.

D. Infection and tissue distribution of OAV in vivo To test the ability of OAVhaat to infect cells in vivo, mice were injected by several routes, and haat expression was analysed. As shown in Figure 5, intravenous as well as intraperitoneal application of the vector to Balb/C mice resulted in high levels of reporter gene expression at day 3 post-infection, although individual differences in haat expression were observed after intraperitoneal injection. In addition, intraportal application of 1x109 infectious particles of OAVhaat to C57/bl-6 mice produced hAAT values ranging from 10,3 to 16,1 µg/ml in serum at day 3 post-infection. However, as observed previously (Hofmann et al., 1999), expression was transient, peaking at day 3 to 4 post-infection. To analyse the tissue distribution of the vector, Southern blot analysis was performed with DNA harvested at day 3 from several organs of infected mice. Intravenous injection (Hofmann et al., 1999) resulted in comparable infection of the liver, spleen, heart and kidney and, to a lesser extent, lung tissue. Similarly, intraperitoneal application of OAVhaat resulted in the infection of all tissues investigated (Figure 6a) with no particular preference for any organ. After intraportal application of the vector, the level of OAV DNA in liver was significantly elevated relative to other organs (Figure 6b) but the liver was still not the overwhelming site of accumulation as seen for hAd5 in rodents (Smith et al., 1993; Fang et al., 1994; Kay et al., 1994; Huard et al., 1995).

1. Ad5lacZ (day 0) 2. OAVhaat (day 31) 3,0 µg/ml 2,8 µg/ml 1,9 µg/ml 2,8 µg/ml

Figure 4: Detection of haat-specific transcripts after a second vector application. Mice were immunized with 5x109 infectious particles of Ad5haat and injected with 1x10 9 infectious particles of OAVhaat 31 days later. RNase protection assays using a probe specific for haat RNA were performed with 20µg of total RNA harvested from the tissues indicated at day 4 after second vector administration. In vitro transcribed haat RNA (10 and 25 pg) was used as a standard (st). tRNA was used as negative control. Numbers refer to the same animals.

E. Construction and rescue of an OAV/GFP recombinant The problem with a reporter such as haat is that it does not identify the types of cells infected in vivo. Therefore to extend these studies on the cell tropism of OAV we have constructed a virus (OAV217A) in which the human CMV immediate early promoter/enhancer was used to drive expression of the green fluorescent protein (GFP) gene. The cassette was inserted in the left-to-right orientation in site I (Figure 1). Attempts were made to rescue viruses carrying one of two expression cassettes, one that had an intron and one that did not. Only the virus that lacked the intron was rescued, albeit after several attempts. Using this virus it was confirmed that OAV infects a range of cell types (Khatri et al, 1997), including human prostate (PC3) and cervical carcinoma (HeLa) as well as monkey kidney (COS-7) and mouse prostate (RM-1, Hall et al, 1997) cells (Figure 7). However, subsequent passage of this virus in CSL503 cells and BamHI analysis of its DNA showed that, unlike many other viruses that we have rescued, the foreign gene cassette was unstable. This may explain why the virus was difficult to rescue initially. Nevertheless, useful stocks of passage 3 virus have been purified and will facilitate in vivo studies to determine precisely which cells are infected in particular mouse organs. The HCMV/GFP cassette is being reconstructed in site III such that it will be located between transcription units (Khatri and Both, 1998) rather than interrupting the transcript for the fiber protein.

1. Ad5lacZ (day 0) 2. Ad5haat (day 31) <bg <bg <bg <bg

Table 1: OAV overcomes pre-existing immunity to a hAd vector. Balb/C mice were infected with 5x10 9 pfu of Ad5lacZ at day 0 and reinjected 31 days later with 1x109 infectious particles of OAVhaat or 5x109 pfu of Ad5haat, respectively. haat expression was determined in mouse sera 3 days after the second injection. bg, background (<20 ng/ml).

Lรถser et al: Ovine adenovirus vectors

Figure 5: In vivo expression of haat gene after transduction of mice with OAVhaat. Balb/C or C57/bl-6 mice were infected with 1x10 9 infectious particles of OAVhaat via the route indicated and haat levels in serum samples were determined three days after infection. Each bar represents an individual animal.

Figure 6: Tissue distribution of OAVhaat in mice. Animals were injected with 1x109 infectious particles of OAVhaat or PBS (n.i.) via the (A) intraperitoneal or (B) intraportal routes. DNA was harvested at day 3 post-infection from the tissues indicated and analyzed by Southern blotting using an OAV-specific radiolabelled probe. DNA equivalent to 1 or 5 copies of the virus genome per cell (2,5 or 12,5 pg) was used as a standard. The position of the OAV-specific 2399 bp fragment is shown. Numbers refer to tissues from the same animals.

Gene Therapy and Molecular Biology Vol 4, page 39

Figure 7: Infection of murine RM-1 cells by OAV as monitored by the expression of green fluorescent protein gene. Cells were infected at the multiplicity of infection indicated and examined by fluorescence (left and middle panels) or light microscopy (right panel) at 48hr post-infection.

(Klonjkowski et al., 1997). No such cross-reactivity was observed with OAV287 (Hofmann et al., 1999) which is phylogenetically distant from human adenoviruses (Harrach et al, 1997). In this paper we have shown that OAV vectors are valuable tools for achieving high-level transgene expression in vivo under conditions that are unfavourable for hAd vectors. We observed that both the i.p. and intraportal routes of vector administration led to infection of several organs in mice and to secretion of significant amounts of the transgene product, human ! 1-antitrypsin, into the serum. The problem of transience of gene expression after OAV mediated gene transfer needs further investigation. We were unable to detect residual expression of OAV early and late genes in mouse tissues after local injection of the vector by RT-PCR (P. Lรถser, unpublished results), and the major late promoter of OAV was shown to be only weakly active in semi-permissive BNT cells and silent in non-permissive cells after infection with OAV (Khatri et al., 1997). However, de-novo synthesis of viral gene products was shown to be dispensable for an immune response to hAd vectors (Kafri et al., 1998). Thus, closer inspection of immune infiltrates present in tissues infected by OAV will help to reveal the reason for vector clearance in vivo. On the other hand, gutless adenoviral vectors were reported to remain stable in mouse liver allowing for long-term transgene expression (Morsy et al., 1998). Therefore, construction of analogous OAV vectors similar to those developed for hAd (Kochanek et al, 1996) might help to overcome the problem of short-term expression after OAV mediated gene transfer. Alternatively, OAV vectors may be better suited to problems where short-term gene delivery is sufficient.

III. Discussion Pre-existing neutralizing antibodies against adenoviruses in the vast majority of the human population represent a major hurdle to the use of hAd derived vectors for gene delivery. We have shown here and elsewhere (Hofmann et al., 1999) that pre-incubation of hAd vectors with human serum results in complete inactivation of the vector. In addition, the complement system accelerates hAd vector inactivation, most likely by masking the viral surface with C-proteins. Thus, in vivo, it can be expected that the clearance of opsonized vectors by the reticuloendothelial system will dramatically reduce the efficacy of hAd-vectors in patients with pre-existing immunity. The antibody titer of the serum used was similar to that induced in mice after infection with hAd vectors. The unsuccessful readministration of hAd vectors to mice immunised with a first injection of the same hAd vector very likely predicts the outcome in humans. Since other gene therapy vectors such as AAV and HSV-1 are also derived from viruses which commonly infect humans, we and others favour the use of nonmammalian (Hofmann et al., 1995) or non-human viral vectors for use in human gene therapy approaches (Mittal et al., 1995; Klonjkowski et al., 1997; Zakhartchouk et al., 1998; Reddy et al., 1999; Hofmann et al., 1999; Michou et al., 1999). Adenoviral vectors of non-human origin developed so far are summarised in Table 2. Although most of these vectors were created for vaccination purposes they might also be useful as gene therapy vectors. However, these vectors are all derived from viruses belonging to the Mastadenoviruses and some cross-reactivity of antibodies in human sera was observed with canine adenovirus-derived vectors 39

Löser et al: Ovine adenovirus vectors Adenovirus

bovine adenovirus type 3

ovine adenovirus 287

canine adenovirus type 2

porcine adenovirus type 3

chicken embryo lethal orphan adenovirus type 1

abbreviation

BAV-3

OAV-287

CAV-2

PAV-3

CELO

genus

mastadenovirus

atadenovirus

mastadenovirus

aviaadenovirus

first publication of recombinants

1995

1996

1997

1999

number of publications

phylogenetic relationship to Ad5

distant

insertion of transgene in:

site I, II or III

right end

non-essential or deleted regions

E3 (1249 bp)

RHE (2000 bp)

E3 (600 bp)

right end (3600 bp)

human cell lines successfully infected (published to date)

non

HepG2, MRC5, MCF-7, T47D-2, HAT29, PC-3, HeLa, COS-7

293, HeLa, HIB, myocytes

293, A549

HepG2, A549, HeLa, primary fibroblasts

animal models

cotton rat

mouse

chicken embryo

none

chicken embryo

Table 2: Data on non-human adenovirus recombinants published to date.

animal models. The prospect that OAV-derived vectors may replace or supplement their hAd counterparts warrants further development of this vector system to broaden its potential application in the field of gene delivery.

The tissue distribution of OAV is significantly different from that observed for hAd vectors which mainly infect the liver after systemic application in rodents (Smith et al., 1993, Fang et al., 1994, Kay et al., 1994, Huard et al., 1995). We found nearly equal amounts of vector DNA after i.p. (this paper) or intravenous (Hofmann et al., 1999) vector delivery. Moreover, even after local injection into the portal vein of C57/bl-6 mice the vector is only moderately enriched in the liver and is still found in all tissues examined. This is consistent with evidence (Xu and Both, 1998) that OAV vectors use a primary receptor that is distinct from CAR, the Ad5 receptor (Bergelson et al., 1997, Tomko et al 1997). The full spectrum of cells that are infected by OAV remains to be determined but it is likely that there will be some cells that are better infected by OAV compared with hAd vectors and vice versa. The availability of an OAV/GFP recombinant will greatly facilitate these studies. In summary, we have shown that OAV vectors are valuable tools for achieving high-level gene expression in animals. Further studies are in progress to extend the investigations of this novel vector system to other

IV. Experimental Procedures A. Cells and viruses Human embryonic kidney 293 cells, permissive for E1-deleted human adenoviruses and HuH7 (human hepatoma) cells were cultured in Dulbecco’s modified Eagles medium (GibcoBRL) with 2mM glutamine (Sigma, Deisenhofen, Germany) and 10% foetal calf serum (Roche Diagnostics, Mannheim, Germany) at 5% CO2. CSL503 cells (foetal ovine lung, permissive for OAV) were grown under the same conditions but in 15% foetal calf serum. RM1 cells were grown in DMEM with additives (Cat # 12100-103; Life Technologies) plus 10% foetal bovine serum (Hall et al, 1997). Ad5luc (a generous gift of M. Hillgenberg, Berlin) contains a hCMV IE promoter-driven luciferase gene. The generation of OAVhaat in which expression of the human ! 1-antitrypsin (haat) cDNA is driven by the RSV 3’LTR has been described (Hofmann et 40

Gene Therapy and Molecular Biology Vol 4, page 41 al., 1999). Ad5haat, which contains the identical haat gene expression cassette, was a generous gift of Mark Kay, Stanford. To construct OAV217A containing the HCMV/GFP cassette we used a GFP gene that was modified by Dr. Shinichi Aota (Biomolecular Engineering Research Insitute, Japan) to optimise expression in mammalian cells. The gene was bluntcloned into the XhoI/SmaI sites of plasmid pCI (Promega Corp, Madison WI) and the promoter/gene cassette was excised by BglII/BamHI digestion and blunt-cloned into the XbaI site of pGem11zf (Promega Corp, Madison WI). A clone with a 5’ ApaI and 3’ NotI site was selected and the insert was cloned into these sites in pOAV200 for virus rescue (Vrati et al, 1996b). Subsequently, the cassette was further subcloned and modified by AflII digestion and blunt end ligation to remove the intron provided in pCI. The virus was rescued after transfection of CSL503 cells as described previously (Vrati et al, 1996b) except that cationic lipids were used (Cameron et al, 1999) in place of lipofectamine. Viruses were grown on permissive cell lines and purified as described (Sandig et al., 1996). Virus titers were determined by an end point dilution assay on permissive cell lines. Particle/infectious unit ratios for Ad5 recombinants and OAV/haat were <40:1.

150µl) was injected at a rate of 100µl per minute. After removal of the tube the portal vein was clamped for one minute to allow closure. For determination of haat gene expression, blood samples were collected from the external jugular vein of mice and used in an enzyme-linked immunosorbent assay as described (Cichon and Strauss, 1998). Antibody titers to haat were determined according to Morral et al. (1997). Detection of hAd-specific antibodies was performed as described (Hofmann et al., 1999). For Southern blotting and RNase protection assay animals were sacrificed, organs of interest were frozen immediately and homogenized in liquid nitrogen and DNA and RNA were isolated separately from the same tissue piece using standard methods. For Southern blotting, genomic DNA (20µg) was digested with EcoRI, which releases a 2399 bp fragment from the OAV genome. After separation on a 1% agarose gel and transfer to a nylon membrane, hybridisation was performed using a probe spanning bp 1968 to 3408 of the OAV genome. A specific OAV EcoRI fragment (2,5 or 12.5 pg, equivalent to 1 or 5 copies per cell, respectively) was used as a standard. RNase protection assays were carried out with total RNA (20µg) following standard procedures. A radiolabelled RNA fragment of 362 bases comprising the EcoNI fragment of the hAAT gene was used as a probe. In vitro transcribed haat RNA (10 or 25 pg, respectively) served as a standard.

B. Treatment of vectors by human serum and heat

Acknowledgement We thank V. Sladek, E. Bennetts and K. Smith for excellent technical assistance and Dr. Z. Xu for providing the image in Figure 7.

Ad5luc was incubated with 50 µl of human serum for 30 minutes at 37°C. Serum was either untreated, treated to remove complement at 56°C for 30 min, or heated at 48°C for 30 min. EGTA/Mg-treated serum contained 10mM EGTA and 7mM MgCl 2. Serumtreated virus was then used to infect HuH7 cells at a moi of 100. At 48hr post infection, cells were harvested and luciferase activity was determined as described previously (Löser et al., 1996). HuH7 cells infected with Ad5luc incubated for 30 minutes with PBS served as positive control. For heat treatment, 1x106 pfu of either OAVhaat or Ad5haat were incubated for 30 minutes at 4, 42, 45, 48, 51, 54, 57 and 60°C, respectively, and virus titer was subsequently determined on permissive cells using an end point dilution assay.

References Bergelson JM, Cunningham JA, Droguett G, Kurt-Jones EA, Krithivas A, Hong JS, Horwitz MS, Crowell RL, and Finegold RW (1997) Isolation of a common receptor for coxsackie B viruses and adenoviruses 2 and 5. Science 275, 1320-1323. Boyle DB, Pye AD, Kocherhans R, Adair BM, Vrati S, and Both GW (1994) Characterisation of australian ovine adenovirus isolates. Vet Microbiol 41, 281-291. Cameron FH, Moghaddam MJ, Bender VJ, Whittaker RG, Mott M, and Locket TJ (1999) A transfection compound series based on a versatile Tris linkage Bioch Bioph Acta 1417, 37-50 Cichon G and Strauss M (1998) Transient immunosuppression with 15-deoxyspergualin prolongs reporter gene expression and reduces humoral immune response after adenoviral gene transfer. Gene Therapy 5, 85-90. Cooper NR (1998) Complement-dependent virus neutralization. In Rother K, Till GO, and Hänsch GM (eds) The Complement System Springer Verlag Berlin Heidelberg New York, 302-309. Dai Y, Schwarz EM, Gu D, Zhang W-W, Sarvetnick N, and Verma IM (1995) Cellular and humoral immune response to adenoviral vectors containing factor IX gene: Tolerization of factor IX and vector antigens allows for long-term

C. Animal procedures, antibodies and detection of adenovirus-mediated gene transfer Female Balb/C or C57/bl-6 mice aged 6 to 8 weeks (Charles River, Germany) received intravenous or intraperitoneal injections of adenoviral vectors as indicated. For portal vein injection a mid-line incision of 2 cm was made below the region of the liver. The intestines were carefully displaced to the right and the ileum extended to display the portal vein. A capillary tube connected to a syringe was inserted about 1 cm into the portal vein and vector suspension (a maximum of

Löser et al: Ovine adenovirus vectors expression (1995). Proc Natl Acad Sci USA 92, 14011405. Fang B, Eisensmith RC, Li XH, Finegold MJ, Shedlovsky A, Dove W, and Woo, SL (1994) Gene therapy for phenylketonuria: phenotypic correction in a genetically deficient mouse model by adenovirusmediated hepatc gene transfer. Gene Therapy 1, 247254. Ghosh-Choundhury G, Haj-Ahmad Y, and Graham FL (1987) Protein IX, a minor component of the human adenovirus capsid, is essential for the packaging of full length genomes. EMBO J 6, 1733-1739. Hall SJ, Mutchnik SE, Chen SH, Woo SLC, and Thompson TC (1997) Adenovirus mediated herpes simplex virus thymidine kinase gene and gancyclovir therapy leads to systmic against spontaneous and induced metastasis in an orthotopic mouse model of prostate cancer. Int J Cancer 70, 183-187. Harrach B, Meehan BM, Benkö M, Adair BM, and Todd D (1997) Close phylogenetic relationship between egg drop syndrome virus, bovine adenovirus serotype 7, and ovine adenovirus strain 287. Virology 229, 302306. Harrach B and Benkö M (1998) Phylogenetic analysis of adenovirus sequences. Proof of the necessity of establishing a third genus in the adenoviridae family. In Wold WSM (ed) Adenovirus methods and protocols Humana Press, 309-339. Hofmann C, Sandig V, Jennings GS, Rudolph M, Schlag P, and Strauss M (1995) Efficient gene transfer into human hepatocytes by baculovirus vectors. Proc Natl Acad Sci USA 92, 10099-10103. Hofmann C, Löser P, Cichon G, Arnold W., Both GW, and Strauss M (1999) Ovine adenovirusvectors overcome preexisiting humoral immunity against human adenoviruses in vivo. J Virol 73, 6930-6936. Horwitz MS (1996) Adenoviruses. In Fields BN, Knipe DM, and Howley PM (ed) Fields Virology Raven Press, 2149-2197. Huard J, Lochmuller H, Ascadi G, Jani A, Massie B, and Karpati G (1995) The route of administration is a major determinant of the transduction efficiency of rat tissues by adenoviral recombinants. Gene Therapy 2, 107-115. Ilan Y, Droguett G, Roy Chowdury N, Li Y, Sengupta K, Thummala NR, Davidson A, Roy Chowdury J, and Horwitz MS (1997) Insertion of the adenoviral E3 region into a recombinant viral vector prevents antiviral humoral and cellular immune responses and permits long-term gene expression. Proc Natl Acad Sci USA 94, 2587-2592. Ilan Y, Sauter B, Roy Chowdhury N, Reddy BVN, Thummala NR, Droguett G, Davidson A, Ott M, Horwitz MS, and Roy Chowdhury J (1998) Oral tolerization to adenoviral proteins permits repeated adenovirusmediated gene therapy in rats with pre-existing immunity to adenoviruses. Hepatology 27, 13681376. Kafri T, Morgan D, Krahl T, Sarvetnick N, Sherman L, and Verma IM (1998) Cellular immune response to adenoviral vector infected cells does not require de

novo viral gene expression: Implications for gene therapy. Proc Natl Acad Sci USA 95, 11377-11382. Kass-Eisler A, Leinwand L, Gall J, Bloom B, and Falck-Pedersen E (1996) Circumventing the immune response to adenovirus-mediated gene therapy. Gene Therapy 3, 154162. Kay MA, Landen CN, Rothenberg SR, Taylor LA, Lleland F, Wiehle S, Fang B, Bellinger D, Finegold M, Thompson AR, Read M, Brinkhous KM, and Woo SLC (1994) In vivo hepatic gene therapy: complete albeit transient correction of factor IX deficiency in hemophilia B dogs. Proc Natl Acad Sci USA 91, 2353-2357. Kay MA, Holterman AX, Meuse L, Gown A, Ochs HD, Linsley PS, and Wilson CB (1995). Long-term hepatic adenovirusmediated gene expression in mice following CTLA4Ig administration. Nat Genet 11, 191-197. Kay MA, Meuse L, Gown AM, Linsley P, Hollenbaugh D, Aruffo A, Ochs HD, and Wilson CB (1997) Transient immunomodulation with anti-CD40 ligand antibody and CTLA4Ig enhances persistence and secondary adenovirusmediated gene transfer into mouse liver. Proc Natl Acad Sci USA 94, 4686-4691. Khatri A, Xu ZZ, and Both GW (1997) Gene expression by atypical recombinant ovine adenovirus vectors during abortive infection of human and animal cells in vitro. Virology 239, 226-237. Khatri A, and Both GW (1998) Identification of transcripts and promoter regions of ovine adenovirus OAV287. Virology 245, 128-141. Klonjkowski B, Gilardi-Hebenstreit P, Hadchouel J, Randrianarison V, Boutin S, Yeh P, Perricaudet M, and Kremer EJ (1997) A recombinant E1-deleted canine adenoviral vector capable of transduction and expression of a transgene in human-derived cells and in vivo. Hum Gene Ther 8, 2103-2115. Kochanek S, Clement PR, Mitani K, Chen H-H, Chan S, and Caskey CT (1996) A new adenoviral vector: Replacement of all viral coding sequences with 28 kb of DNA independently expressing both full-lenght dystrophin and "-galactosidase. Proc Natl Acad Sci USA 93, 57315736. Kolls JK, Lei D, Odom G, Nelson S, Summer WR, Gerber MA, and Shellito JE (1996) Use of transient CD4 lymphocyte depletion to prolong transgene expression of E1-deleted adenoviral vectors. Hum Gene Ther 7, 489-497. Löser P, Sandig V, Kirillova I, and Strauss M (1996) Evaluation of HBV promoters for use in hepatic gene therapy Biol Chem Hoppe-Seyler 377, 187-193. Mack CA, Song WR, Carpenter H, Wickham TJ, Kovesdi I, Harvey BJ, Magovern CJ, Isom OW, Rosengart T, FalckPedersen E, Hackett NR, Crystal RG, and Mastrangeli A (1997) Circumvention of anti-adenovirus neutralizing immunity by administration of an adenoviral vector of an alternate serotype. Hum Gene Ther 8, 99-109. Mastrangeli A, Harvey BG, Yao J, Wolff G, Kovesdi I, Crystal RG, and Falck-Pedersen E (1996). “Sero-switch“ adenovirus-mediated in vivo gene transfer: Circumvention of anti-adenovirus humoral immune defenses against repeat adenovirus vector administration by changing the adenovirus serotype. Hum Gene Ther 7, 79-87.

Gene Therapy and Molecular Biology Vol 4, page 43 Mayor HD, Drake S, Stahmann J, and Mumford DM (1976) Antibodies to adeno-associated satellite virus and herpes simplex in sera from cancer patients and normal adults. Am J Obstet Gynecol 126, 100-104. Michou A-I, Lehrmann H, Saltyk M, and Cotten M (1999) Mutational analysis of the avian adenovirus CELO, which provides a basis for gene delivery vectors. J Virol 73, 1399-1410. Mittal SK, Prevec L, Graham FL, and Babiuk LA (1995) Development of a bovine adenovirus type 3-based expression vector. J Gen Virol 76, 93-102. Morral N, O’Neal W, Zhou H, Langston C, and Beaudet A (1997) Immune responses to reporter proteins and high viral dose limit duration of expression with adenoviral vectors: Comparison of E2a wild type and E2a deleted vectors. Hum Gene Ther 8, 1275-1286. Morsy MA, Gu M, Motzel S, Zhao J, Lin J, Su Q, Allen H, Franlin L, Parks RJ, Graham FL, Kochanek S, Bett AJ, and Caskey CT (1998) An adenoviral vector deleted for all viral coding sequences results in enhanced safety and extended expression of a leptin transgene. Proc Natl Acad Sci USA 95, 7866-7871. Pye D (1989) Cell lines for growth of sheep viruses. Austr Vet J 66, 231-232. Reddy PS, Idamkanti N, Hyun B-H, Tikoo SK, and Babiuk LA (1999) Development of porcine adenovirus-3 as an expression vector. J Gen Virol 80, 563-570. Sandig V, Löser P, Lieber A, Kay MA, and Strauss M (1996) HBV-derived promoters direct liver-specific expression of an adenovirally transduced LDL receptor gene. Gene Therapy 3, 1002-1009. Smith TGA, Mehaffey MG, Kayda DB, Saunders JM, Yei S, Trapnell BC, McClelland A, and Kaleko M (1993) Adenovirus mediated expression of therapeutic plasma levels of human factor IX in mice. Nat Med 5, 397-402. Tomko RO, Xu RL, and Philipson L (1997) HCAR and MCAR: The human and mouse cellular receptors for subgroup C adenoviruses and group B coxsackieviruses. Proc Natl Acad Sci USA 94, 33523356. Tripathy SK, Black HB, Goldwasser E, and Leiden JM (1996) Immune response to transgene encoded proteins limit the stability of gene expression after injection of replication-defective adenovirus vectors. Nat Med 2, 545-550. Venktesh A, Watt F, Xu ZZ, and Both GW (1998) Ovine adenovirus (OAV287) lacks a virus-associated RNA gene. J Gen Virol 79, 509-516. Vilquin J-T, Guérette B, Kinoshita I, Roy B, Goulet M, Gravel C, Roy R, and Tremblay JP (1995) FK506 immunosuppression to control the immune reactions triggered by first-generation adenovirus-mediated gene transfer. Hum Gene Ther 6, 1391-1401. Vrati S, Boyle DB, Kockerhans R, and Both GW (1995) Sequence of ovine adenovirus 100k hexon assembly, 33k, pVIII and fiber genes: early region E3 is not in the expected location. Virology 209, 400-408. Vrati S, Brookes DE, Boyle DB, and Both GW (1996a) Nucleotide sequence of ovine adenovirus tripartite

leader sequence and homologous of Iva2, DNA polymerase and terminal proteins. Gene 177, 35-41. Vrati S, Brookes PS, Khatri A, Boyle DB, and Both GW (1996b) Unique genome arrangement of an ovine adenovirus: Identification of new proteins and proteinase cleavage sites. Virology 220, 186-199. Vrati S, Macavoy ES, Xu ZZ, Smole C, Boyle DB, and Both GW (1996c) Construction and transfection of ovine adenovirus genomic clones to rescue modified viruses. Virology 220, 200-203. Whitley RJ (1996) Herpes simplex viruses. In Fields BN, Knipe DM, and Howley PM (ed) Fields Virology Raven Press, 2297-2342. Xu ZZ, Hyatt A, Boyle DB, and Both GW (1997) Construction of ovine adenovirus recombinants by gene insertion or deletion of related terminal region sequences. Virology 238, 62-71. Xu ZZ, and Both GW (1998) Altered tropism of an ovine adenovirus carrying the fiber protein cell binding domain of human adenovirus type 5. Virology 248, 156-163. Yang Y, Nunes FA, Berensci K, Furth EE, Gonczol E, and Wilson JM (1994a) Cellular immunity to viral antigens limits E1-deleted adenoviruses for gene therapy. Proc Natl Acad Sci USA 91, 4407-4411. Yang Y, Ertl HCJ, and Wilson JM (1994b) MHC class I restricted cytotoxic T lymphocytes to viral antigens destroy hepatocytes in mice infected with E1-deleted recombinant adenoviruses. Immunity 1, 433-442. Yang Y, Li Q, Ertl HCJ, and Wilson JM (1995) Cellular and humoral immune responses to viral antigens create barriers to lung-directed gene therapy with recombinant adenoviruses. J Virol 69, 2004-2015. Yang Y, Greenough K, and Wilson JM (1996) Transient immune blockade prevents formation of neutralizing antibody to recombinant adenovirus and allows repetaed gene transfer to mouse liver. Gene Therapy 3, 412-420. Zakhartchouk AN, Reddy PS, Baxi M, Baca-Estrada ME, Methali M, Babiuk LA, and Tikoo SK (1998) Construction and characterization of E3-deleted bovine adenovirus type 3 expressing full lenght and truncated form of bovine herpesvirus type 1 glycoprotein gD. Virology 250, 220229.

Christian Hofmann

Gene Therapy and Molecular Biology Vol 4, page 45 Gene Ther Mol Biol Vol 4, 45-58. December 1999.

Efficient expression of ribozyme and reduction of stromelysin mRNA in cultured cells and tissue from rabbit knee via Adeno-associated Virus (AAV) Research Article

Elisabeth Roberts*, Piruz Nahreini*, Kristi Jensen, Ira von Carlowitz, Karyn Bouhana, Stephen Hunt III**, Thale Jarvis, Larry Couture***, and Dennis Macejak Ribozyme Pharmaceuticals Inc., 2950 Wilderness Place, Boulder, Colorado 80301, USA *these authors contributed equally; ** Parke-Davis Pharmaceutical Research, Division of Warner-Lambert Company, Ann Arbor, MI 48105; *** Present address: City of Hope National Medical Center, 1500 E. Duarte Rd., Duarte, CA 91010; A preliminary account of this study was presented at Cold Spring Harbor symposium on Gene therapy on September 25th, 1996.

__________________________________________________________________________________________________ Correspondence: Dennis Macejak, Ribozyme Pharmaceuticals Inc., 2950 Wilderness Place, Boulder, Colorado 80301; Tel: (303) 546-8153; Fax: (303)-449-6995; E-mail: macejad@rpi.com. Key Words: hammerhead ribozymes, Adeno-associated virus, gene delivery, osteoarthritis, stromelysin, nerve growth factor receptor Received: 26 October 1999; accepted 3 November 1999

Summary The potential use of Adeno-associated virus (AAV) as an efficient gene delivery vector is increasingly being recognized in human gene therapy. We have investigated the utility of recombinant AAV (rAAV) vectors for the delivery and expression of hammerhead ribozymes targeted against a cellular mRNA encoding a matrix metalloproteinase, stromelysin. Stromelysin expression has been linked to the pathogenesis of human osteoarthritis. We have constructed several rAAV backbone plasmids containing single or multiple hammerhead ribozymes expression cassettes under the control of either the tRNA, U1, or U6 promoter, and have used these plasmids to generate rAAV. These rAAV vectors also contain a selectable marker, the truncated nerve growth factor receptor (NGFR) driven by the cytomegalovirus immediate early gene promoter. rAAV expressing stromelysin-specific ribozyme transduced ex vivo cultured rabbit synovial fibroblasts (RSFs) with a greater than 95% efficiency. Stable ribozyme expression can readily be detected throughout the life span of RSFs in culture. Furthermore, ribozyme mediated knockdown of stromelysin mRNA was detected in RSFs infected by a rAAV containing the tRNA-based transcription unit.

(Werb, Alexander, and Adler, 1992); however, abnormal expression has been implicated in a wide range of human diseases such as atherosclerosis, arthritis (Murphy and Hembry, 1992), glomerulonephritis, corneal ulceration, periodontitis, encephalomyelitis, and tumor metastasis (Stetler-Stevenson, Aznavoorian, and Liotta, 1993). Most MMPs (except for MMP-11 and MT-MMP) are secreted into the extracellular matrix as proenzymes. Stromelysin (MMP-3) may be a key mediator in arthritic diseases. It degrades proteoglycans and a wide range of other matrix components (Woessner, 1991) and activates the proenzyme forms of collagenase (Suzuki et al., 1990), gelatinases (Miyazaki et al., 1992) (Ogata, Enghild, and Nagase,1992), and other MMPs, leading to initiation of a proteolytic cascade. Synovial fibroblasts derived from

I. Introduction Human arthritic disorders are the major cause of chronic disability among adults (Vincenti, Clark, and Brinckerhoff, 1994). Several etiological factors have been reported to play roles in the initiation and progression of degenerative cartilage disorders. For example, a group of proteolytic enzymes, including several members of the matrix metalloproteinase (MMP) family, are strongly implicated in the pathology of arthritis (Vincenti, Clark, and Brinckerhoff, 1994) (Cawston, 1996). MMPs degrade extracellular matrix components such as collagens, gelatins, proteoglycan, and fibronectin (Woessner, 1991). MMPs normally play an important role in embryogenesis, wound healing, and tissue remodeling

Roberts et al: Gene therapy for arthritis using AAV osteoarthritic or rheumatoid synovium produce high levels of stromelysin upon stimulation (Brinckerhoff and Auble, 1990). In addition, there is a significant up-regulation of stromelysin and other MMPs in articular tissues from patients with osteo- or rheumatoid arthritis (Hembry et al., 1995) (Okada et al., 1992). Thus targeted inhibition of one or more of these proteolytic activities may be a valid therapeutic approach for arthritis. Ribozymes are RNA-based enzymes that have the ability to cleave RNA molecules in a sequence-specific manner. Sequence specificity comes from the base-pairing of ribozyme sequences with nucleotides spanning the cleavage site of the target RNA. Cleaved RNA is rapidly degraded in the cell, resulting in a decrease in expression of the encoded protein. Because of their sequencespecificity, ribozymes can be utilized as a therapeutic agent to down-regulate a specific RNA in the background of other cellular RNAs (Rossi, 1999). We have previously reported that a chemically stabilized synthetic hammerhead ribozyme targeting nucleotide position 1049 in the human stromelysin mRNA significantly reduced the target RNA levels upon intra-articular injection (Flory et al., 1996). In this study, we have used recombinant AAV (rAAV) for the expression of hammerhead ribozymes targeting stromelysin mRNA at the same nucleotide position. Because ribozymes can function as RNA molecules they can be synthesized in a variety of transcription units besides RNA polymerase II mRNAs. However, the success of ribozyme efficacy is dependent upon at least three parameters: ribozyme intracellular localization, ribozyme cleavage activity, and ribozyme RNA levels. Ribozyme localization is somewhat dependent upon the inherent characteristics of the particular transcription unit used and of the RNA elements contained within the transcript (Bertrand et al., 1997; Thompson et al., 1995b). Ribozyme cleavage activity is effected by flanking sequences contained within the transcript (Chowrira, Pavco, and McSwiggen, 1994; Thompson et al., 1995a) and ribozyme RNA levels are limited by promoter strength and ribozyme RNA stability (Thompson et al., 1995a) (Thompson et al., 1995b) (Rossi, 1999). Thus, we have designed and constructed a variety of ribozyme transcription units that maintain catalytic activity and have the potential to accumulate to high levels in the nucleus or to utilize endogenous splicing machinery to promote ribozyme/target RNA hybridization. We have previously reported the construction of a modified tRNA promoter for the expression of an HIV specific hammerhead ribozyme (Thompson et al., 1995a). In this report, we compare the efficacy of ribozymes expressed from modified tRNA, U6 snRNA, and U1 snRNA transcription units. Ribozyme expression via retroviral and adenoviral vectors has been reported in several studies (Macejak et al., 1999) (Thompson et al., 1995a), however the utility of AAV for ribozyme expression has not been explored

extensively. Retroviruses and adeno-associated virus (AAV), by virtue of integrating into chromosomal DNA, are attractive gene delivery vehicles for the treatment of human disorders in which a long-term therapy is essential for an effective treatment. The attractive features of AAV as a vector are nonpathogenicity, low immunogenicity, stable and efficient expression of transgenes from the integrated or episomal form, infection of non-dividing cells, broad host range, generation of high titer (8 x 108 IU/ml), and physically stable virions. Recently, AAV-mediated expression of ribozymes targeting a mutated rhodopsin mRNA was shown to slow the rate of photoreceptor degeneration in a transgenic rat model (Lewin et al., 1998). AAV is a small non-pathogenic human parvovirus whose genome (4681 bases) is a single-stranded DNA of either polarity, flanked by inverted terminal repeats (ITRs). AAVITRs are 145 bases in length and function as the sole cisacting elements essential for chromosomal excision (rescue), integration, replication, and encapsidation of nascent viral DNA (Muzyczka, 1992). AAV infects a variety of mammalian cells with a broad host range; however, some human megakaryocytic cell lines are refractory to AAV infection, presumably because they lack the putative AAV receptor (Ponnazhagan et al., 1996). AAV can establish a lytic or latent infection in mammalian cells in the presence or absence of a helper virus, respectively. Productive AAV infection is dependent on a helper DNA virus, which is usually adenovirus; however, herpes virus and vaccinia virus can substitute for the helper functions of adenovirus (Schlehofer, Ehrbar, and zur Hausen, 1986) (Thomson et al., 1994). In the absence of a helper virus, the AAV genome preferentially integrates into a defined region of human chromosome 19 (q13.3-qter), and establishes a stable latent infection (Kotin et al., 1990) (Samulski et al., 1991). In this study, we demonstrate the utility of rAAV as a vector for ribozyme expression in primary synovial cells of the rabbit knee, and report the feasibility of rAAV-mediated gene therapy for arthritic disorders.

II.Results A. rAAV-LacZ transduction of synoviocytes and chondrocytes. Although AAV is known to infect a variety of mammalian cell lines, some cells, such as megakaryocytic MB-O2 and MO7e cell lines, are refractory to AAV infection (Ponnazhagan et al., 1996). Therefore, we first tested whether AAV can infect primary cells from the intra-articular lining of rabbit knees. Primary synovial fibroblasts from rabbit knee tissue (RSFs) were cultured ex vivo and were infected with rAAV containing a !-galactosidase gene driven by the cytomegalovirus immediate early promoter. rAAV-mediated LacZ expression was readily detected in RSF cells (Figure 1). The LacZ expression was detectable 30 hrs post-infection and remained stable during the entire life span of primary RSF cells in culture (5 â&#x20AC;&#x201C; 7 passages, data not shown).

Gene Therapy and Molecular Biology Vol 4, page 47

Figure 1: rAAV Transduction of Synoviocytes and Chondrocytes. !-galactosidase expression in uninfected or rAAV-LacZ infected cells. Synovial fibroblasts (primary culture isolated from rabbit knee), chondrocyte sarcoma (human cell line SW1353), and cartilage explants (rabbit femoral groove tissue) are shown, rAAV-LacZ contains the LacZ gene encoding !-galactosidase.

Rabbit cartilage explants infected with rAAV also demonstrated LacZ expression (Figure 1). It is not clear, however, whether the infected cells in the cartilage explants were chondrocytes or another cell type(s). In addition, rAAV-mediated LacZ expression can readily be detected in the chondrocyte sarcoma cell line SW1353 (Figure 1). Together, these observations indicate that rAAV-LacZ can infect the ex vivo cultured RSFs and cells of rabbit cartilage explants.

Several ribozyme transcription units were cloned into the backbone rAAV plasmid, and the resulting plasmids were used to generate rAAV particles. The rAAVs used in this study are shown in Figure 2 and transcription units are shown in Figure 3A. The modified tRNA transcription unit was originally developed to increase the copy number and maintain catalytic activity of ribozyme expressed inside cells (Thompson et al., 1995a). To generate ribozyme with minimal flanking sequence we chose the U6 snRNA promoter since the U6 promoter is extragenic except for the "G" at +1. We have developed a U6+1-Rz (called "U6C", see Figure 3A) that contains a 5'/3' stem interaction for stability, analogous to that in the improved "TRZ-motif" (Thompson et al., 1995a). Certain splicosomal RNAs interact with pre-mRNA molecules during normal pre-mRNA processing and the sequences which mediate these interactions in the context of a snRNP are known. The first snRNP to contact a pre-mRNA contains U1 RNA which hybridizes to the 5' splice site. We reasoned that replacing the U1 RNA sequences that hybridize to the 5' splice site with a ribozyme (Figure 3A) could promote hybridization of a U1-ribozyme snRNP to a target cleavage site in the nucleus.

B. rAAV vectors and ribozyme transcription units The backbone plasmid, in which all ribozyme transcription units were inserted, contains a modified nerve growth factor receptor (NGFR) gene, driven by the cytomegalovirus early promoter, as a selection marker. Because of a deletion in the c-terminus of the cDNA, NGFR is biologically inactive when expressed from rAAV-infected cells. However, rAAV-mediated expression of this altered NGFR can be readily detected on the membrane of transduced cells by Fluorescence Activated Cell Scanning (FACS; see Figure 5A).

Roberts et al: Gene therapy for arthritis using AAV

Figure 2: Schematic Representation of rAAV Vectors. Recombinant viruses contain the ribozyme (Rz) transcription units as noted. vAT22 is the "empty" (contains no ribozyme transcription unit) control virus derived from the backbone plasmid pAT22. AAV-ITR, AAV inverted terminal repeat sequence; " - LacZ, sequence derived from the LacZ gene to increase genome size for efficient packaging; MCS, multiple cloning site; CMV, cytomegalovirus immediate early gene promter; NGFR, truncated nerve growth factor receptor cDNA.

C. Efficient transduction and expression of a stromelysin-specific hammerhead ribozyme in RSFs

To confirm that cleavage activity was maintained by each ribozyme within the U1-, U6-, or tRNA-derived RNA transcripts, ribozyme RNAs were synthesized in vitro to contain the flanking sequences predicted for each transcription unit. The chimeric ribozyme RNAs were then tested for cleavage activity against a RNA containing the target site. All ribozyme transcripts contained comparable cleavage activity in vitro that was partially reduced compared to a chemically synthesized ribozyme with no extraneous sequence (Figure 3B). To further characterize these ribozyme transcripts within cells, we analyzed their stability in 293 cells following Actinomycin D treatment (Figure 3C). The U6- and tRNA-derived ribozyme RNAs had a similar stability with a half-life of 1-1.5 h. On the other hand, no reduction in the level of U1-derived ribozyme RNA was detected over 4h. In addition, we observed that the U1-ribozyme was immunoprecipitable with anti-trimethyl-G or anti-SM antibodies (our unpublished results), indicating that the U1-ribozyme transcript goes through a maturation/modification process analogous to authentic U1 RNA.

Since rAAV could infect RSFs and the ribozyme chimeric transcripts retained cleavage activity, a series of rAAV containing one or more expression cassettes encoding ribozymes targeting stromelysin mRNA were prepared (Figure 2). Although ribozyme expression was detected in 293 cells, the promoter activity of the ribozyme transcription units in primary RSFs was uncertain. Thus, we investigated ribozyme expression in rAAV-infected RSFs. RSF cells were infected (moi = 20) with rAAV containing either single or multiple transcription units (see Figure 2). Northern analyses demonstrate that each transcript was expressed in RSFs (data from initial experiments not shown, but see Figure 5B). We then investigated the duration of ribozyme expression. RSF cells were infected with a subset of rAAV (containing either U6 or both U1 and U6 transcription units). Cells were harvested and total RNA purified for Northern analysis at different cell passages post-infection. Cells were infected at passage 3, and analyzed for RNA 48 hours after infection, prior to passaging (Figure 4, P3 lane).

Gene Therapy and Molecular Biology Vol 4, page 49

Figure 3: Ribozyme Transcription Units. A) Predicted RNA secondary structures of ribozyme transcripts used. Rz, site of inserted ribozyme. B) In Vitro cleavage activity of chimeric ribozyme transcripts shown in A. C) Stability of ribozyme transcripts in 293 cells, following treatment with Actinomycin D.

Roberts et al: Gene therapy for arthritis using AAV

Figure 4: Expression of Ribozyme Throughout the Life Span of rAAV-Infected RSF Cells. Northern analysis of total RNA from RSFs infected with rAAVs as noted. P3, P5, P6, passage 3, 5, 6, respectively. Arrows denote U1-derived (U1-Rz) or U6derived (U6-Rz) ribozyme transcripts.

Figure 5: Transduction, rAAV-Mediated Ribozyme Expression and Reduction in Stromelysin RNA Levels in RSF Cells. A) FACS analysis of NGFR expression in cells infected with rAAVs as noted. Proportion of cells transduced is noted for each rAAV infection.

Gene Therapy and Molecular Biology Vol 4, page 51 Figure 5 (Continued): Transduction, rAAVMediated Ribozyme Expression and Reduction in Stromelysin RNA Levels in RSF Cells. B) Northern analysis of total RNA from RSFs infected with rAAVs as indictated. Arrows denote ribozyme (Rz) transcripts. C) RNase Protection analysis of stromelysin RNA in RSFs infected with rAAVs as indicated. A, B, and C were all performed with same populations of rAAVinfected cells.

treatment in RSFs (our unpublished results). RSFs were infected with rAAV, induced with IL-1 for 10 hours beginning 36 hours post-infection, and harvested 46 hours post-infection. Transduction efficiency with each virus was greater than 95%, with the exception that the vAT43 infection gave 89% transduction efficiency (Figure 5A). The cell surface expression of NGFR was heterogenous over the RSF population, particularly with vAT23, as evidenced by the broad shifted peak. Northern analysis demonstrated ribozyme expression with each of the infections (Figure 5B). The steady state levels of ribozyme were greatest in the tRNA-ribozyme chimera, followed by the U6- and U1-derived RNAs. The tRNA-ribozyme levels were reduced in the trimeric expression construct (vAT44) compared to the monomer cassette (vAT42), indicating possible promoter interference.

Ribozyme expression from U6 and U1 transcription units is readily detectable during the entire life span of RSF cells in culture (6 passages; Figure 4). Ribozyme expression from the U1 and U6 promoters significantly increased upon passaging. Transduction efficiencies as determined by NGFR expression using FACS analysis were similar (>90%) with each of the rAAVs tested.

D. Effect of stromelysin specific ribozyme expression on target RNA The inflammatory cytokine IL-1 is implicated in the initiation and progression of arthritic disorders in human and animal models (Cawston, 1996). IL-1 induced stromelysin mRNA expression peaks at 8-12 hours post51

Roberts et al: Gene therapy for arthritis using AAV

Figure 6: rAAV-Mediated Ribozyme Expression and Effect on Stromelysin RNA Levels with Active, Attenuated, or Irrelevant Ribozyme. A) Northern analysis of total RNA from RSFs infected with rAAVs as indicated. B) RNase Protection analysis of stromelysin RNA in RSFs infected with rAAVs as noted. Sets of cells not treated with IL-1 (no IL1) or pretreated for 10h (+IL-1) prior to infection are indicated.

cells, but this effect was lost in cells infected with virus containing the trimeric ribozyme cassette (vAT44).

The level of stromelysin RNA was induced upon rAAV infection alone 3-4 fold over that of Il-1 treatment alone (data not shown). Thus, in terms of ribozyme efficacy, cells infected with the empty vector, vAT22, is the appropriate control for comparison. The level of stromelysin target RNA was slightly reduced in RSFs expressing U1- or U6-derived ribozymes compared to RSFs infected with the control vector (Figure 5C, vAT23 or vAT36 versus vAT22; p < 0.05). The greatest reduction (>60%) in stromelysin RNA was observed in RSF expressing tRNA-ribozyme (vAT42 versus vAT22; p < 0.05). Interestingly, the U1- and U6-derived ribozyme appeared to have an additive effect in vAT43 infected

E. Reduction in target RNA is due to a ribozyme mechanism. To confirm that the reduction in stromelysin RNA observed was due to ribozyme cleavage activity and not due to over-expression of the tRNA motif itself, we compared target RNA levels in RSFs infected with rAAV encoding either a ribozyme targeting an irrelevant sequence (vAT30), the ribozyme targeting site 1049 of stromelysin (vAT42), or an attenuated version of the stromelysin ribozyme (vAT46); all genes were driven by the tRNA promoter within an otherwise 52

Gene Therapy and Molecular Biology Vol 4, page 53 identical rAAV vector. The attenuated control contains two base substitutions in the ribozyme catalytic core domain. This attenuated analog is still capable of binding the target site, but has reduced cleavage activity. Previous work demonstrated the greatly reduced catalytic effects of base substitutions in the ribozyme core on ribozyme activity in vitro (Ruffner, Stormo, and Uhlenbeck, 1990) as well as in cell culture and in vivo (Jarvis et al., 1996) (Flory et al., 1996). The active and attenuated versions of the stromelysin-specific ribozyme and the irrelevant HIVspecific ribozyme were each readily detectable in these cells by Northern analysis (Figure 6A). The probe used in this Northern blot was designed to hybridize to tRNA sequences, such that the same probe will detect all three ribozyme transcripts as well as endogenous tRNA. After normalization to endogenous tRNA, it appears that the irrelevant (HIV-specific) ribozyme (vAT30) level is the greatest, yet no reduction in the level of stromelysin RNA is observed with rAAV encoding irrelevant ribozyme (Figure 6B). As observed previously, RSF infected with rAAV containing active ribozyme (vAT42) significantly reduced stromelysin mRNA levels as compared to either control virus (Figure 6B; p < 0.05). This reduction was not dependent upon IL-1 treatment since vAT42-infected RSFs without IL-1 also displayed reduced target RNA levels compared to RSFs infected with either vector control vAT22 or vAT30 (p < 0.05). The attenuated version of the stromelysin-specific ribozyme (vAT46) did not show a statistically significant decrease in stromelysin mRNA levels relative to the control vector vAT22 in the presence or absence of IL-1. Statistical analysis was done with Kruskal-Wallis One Way Analysis of Variance on Ranks to test for differences between groups. When significant, post-hoc analysis was done using the Dunnett's test and p values < 0.05 were considered significant. These results confirm that the decrease in stromelysin mRNA observed in vAT42 infected cells is largely due to a ribozyme cleavage-dependent mechanism.

tissues. Therefore, MMPs are very attractive therapeutic targets whose abnormal levels of expression can potentially be controlled via genetic and non-genetic interventions. Besides T cells and macrophages, synovial fibroblasts and chondrocytes are directly involved in initiating the arthritic disease, primarily by overexpression and secretion of proteases into the intra-articular region of the joint. Cytokines secreted by T cells and macrophages are known to induce synovial fibroblasts to express and secret augmented levels of MMPs into the joint cavity (Burger et al., 1998). Natural inhibitors of MMPs, such as tissue inhibitor of metalloproteases (TIMPs), normally function to keep the activities of these proteases within physiological homeostasis during the course of wound healing, tissue remodeling, and embryogenesis (Nagase, 1996) (Brown, 1997). However, the perturbation of this balance in favor of increased MMPs expression and secretion would eventually lead to extracellular matrix destruction associated with arthritis and tumor metastasis. Because arthritic disorders, for the most part, are chronic in nature, a long-term therapeutic genetic intervention may be essential to halt the disease process. AAV vectors are attractive in this regard because they are nonpathogenic parvoviruses of human origin, which integrate into chromosomal DNA of a host cell and stably express the therapeutic gene during the entire life span of transduced cells. Second, AAV vectors are less dependent on the proliferative nature of the target cells for efficient transduction. This is supported by many reports indicating that these vectors efficiently infect dividing and non-dividing cells (Kaplitt et al., 1994) (Ponnazhagan et al., 1997) (Koeberl et al., 1997). This is critical for gene delivery to synovial fibroblasts and chondrocytes, which are quiescent cells in vivo. In this report, rAAV-mediated expression of !galactosidase was readily detectable in ex vivo infected cultures of RSFs and chondrocytes. More importantly, stromelysin-specific ribozymes driven by U1-, U6,- and tRNA-based promoters were efficiently expressed in RSFs. Such expression cannot be assumed as we have previously observed that a similar U1-ribozyme transcription unit was not expressed from a recombinant adenovirus in primary human or rat smooth muscle cells (Macejak et al., 1999). Because the ribozyme transcripts are small RNA molecules (relative to most mRNAs) we were able to express more than one transcription unit from a single rAAV vector. Although the three ribozyme transcription units used here are thought to direct nuclear localization (Bertrand et al., 1997) (Thompson et al., 1995a), their subnuclear localization may be unique. The U1-ribozyme RNA was extremely stable (Figure 3C), contained a trimethyl-G â&#x20AC;&#x153;capâ&#x20AC;? and was bound by SM proteins, consistent with this ribozyme transcript being part of a snRNP. Expression of multiple ribozyme transcription units from one vector may enable the ribozyme transcripts to attack the target RNA at different points along its maturation/transport pathway. Consistent with this hypothesis we observed an additive effect with the U1- and U6-derived ribozymes expressed in combination compared to their sole expression

III. Discussion Arthritis is a chronic debilitating disease affecting 1640 million people in the USA (Lawrence et al., 1989). There is a strong correlation between the levels of matrix metalloproteinase (MMP) expression and secretion from the intra-articular cell lining of the joint, and the initiation and progression of the extracellular matrix degeneration seen in human arthritic disorders, such as osteoarthritis and rheumatoid arthritis. In addition, MMPs are directly implicated in tissue remodeling and tumor metastasis in the angiogenic phase of tumor growth. For example, MMP2 is shown to trigger the migration of breast epithelial cells via specific cleavage of laminin-5 (Ln-5) (Brooks et al., 1996) (Giannelli et al., 1997). The cleaved form of Ln-5 is commonly discovered in tumors and in tissues undergoing remodeling, but not in quiescent

Roberts et al: Gene therapy for arthritis using AAV (Figure 5C). The situation does not appear to be that simple, however, since expression of all three ribozyme transcription units in combination did not generate the most reduction in target RNA. This may be due in part to the lower level of expression of all three ribozyme transcripts compared to their individual expression levels (Figure 5B). Interestingly, the levels of U1- and U6-derived ribozyme increased in rAAV-infected RSFs upon passaging (Figure 4). There are several possible explanations for this observation. The first is related to the fact that the conversion of ss-DNA of rAAV genome to a double stranded form is essential for the expression of a transgene. This rate-limiting step is reported to be more efficient in dividing cells as compared to growth-arrested cells. Second, genomic integration of rAAV may enhance transgene expression and the formation of double-stranded AAV genomes is indispensable for a stable chromosomal integration. In addition, it is possible that the enhancement of ribozyme expression in dividing RSFs might be due to an increase in the pool of episomally ds-form of viral genome. The steady state levels of tRNA-ribozyme were much greater than that of either the U1- or U6-derived ribozymes (Figure 5B). Coordinately, the pooled level of stromelysin mRNA in RSF cells infected with rAAV-encoding the tRNA-ribozyme chimera were significantly reduced (6080%, p < 0.05) as compared to the levels in RSF infected with the empty vector or a vector expressing an irrelevant ribozyme (Figure 5C and Figure 6B). Since the expression of NGFR was heterogeneous among the infected RSF populations, it is likely that ribozyme expression from cell to cell is coordinately heterogenous. Thus, in some cells ribozyme transcript levels may not be sufficient to impact stromelysin RNA, while in those cells with the highest ribozyme expression stromelysin RNA may be virtually absent. The reduction of stromelysin mRNA was observed to be largely due to a ribozyme cleavage mechanism, as the attenuated version of this ribozyme did not show a statistically significant reduction in stromelysin RNA levels compared to the empty vector control (Figure 6B). Recently, Bertrand et al. (Bertrand et al., 1997) analyzed the activity of ribozymes in cells and observed inhibition of a target gene with an mRNAribozyme chimera, but not with tRNA, U1, or U6 ribozyme transcripts. However, the tRNA- and U6-derived ribozymes that we tested here were significantly different from those reported by Bertrand et al. and were optimized for cleavage activity and intracellular stability. It is uncertain whether a 60-80% reduction in stromelysin RNA by the tRNA-ribozyme would be sufficient to halt the disease process in vivo. Surprisingly, rAAV infection alone induced the expression of stromelysin RNA in cultured RSFs in the absence of IL-1 treatment. The reason for this is not clear; however, several vectors (rAAV and others) which do not target stromelysin

have a similar effect in cell culture, as do other manipulations such as transfection. The induction of stromelysin mRNA by the transduction itself may be a transient event, and it may subside after a few cell passages. Because RSF cells undergo senescence after 5-7 passages, it is difficult to test this hypothesis. If stromelysin induction due to rAAV infection occurs in vivo, the impact of this induction would need to be resolved before considering rAAV as a viable vector strategy for therapeutic intervention in arthritis. Alternatively, additional methods to deliver the tRNA-ribozyme to RSFs in vivo can be pursued. In any case, our observations underscore the utility of AAV as a useful vector for the stable, long-term expression of ribozymes in primary mammalian cell lines.

Acknowledgments We are grateful to Dr. Jim Dahlberg, Dr. William Marzloff, Dr. Jude Samulski, and Dr. Arun Srivastava for the generous gifts of plasmids. We would like to acknowledge Dr. Dave DiGiusto for his help on FACS analysis. We thank Dr. Jim Thompson, Dave Ayers, Jennifer Chase, Tim McKenzie, and Saiphone Web for valuable technical information, and Dr. Jennifer Sandberg for assistance with statistical analysis. We are grateful to Dr. A. Srivastava for his helpful comments on rAAV technology during the course of this study.

IV. Materials and Methods A. Cells and tissues. Rabbit synovial fibroblasts (RSFs) were surgically removed from the intra-articular lining of rabbit knees and cultured in Dulbecco’s modified Eagle’s medium (DMEM) plus 10% fetal bovine serum (FBS) and 100 µg/ml of penicillin and streptomycin. The human chondrocyte sarcoma cell line SW1353 was obtained from ATCC. Monolayer cultures of 293 cells were grown and maintained in DMEM plus 10% FBS.

B. Ribozyme transcription unit plasmids. The mouse U6 promoter was amplified from plasmid pMU6 (obtained from J. Dahlberg, Univ. of Wisconsin) via PCR with oligonucleotides (AAGTCGACCGACGC CGCCATCTCTA and AAGGAATTCGAGTGCCCAAA CAAGGCTTTTCTCCAAGGG) and cloned into the SalI and EcoRI sites of a modified pGEM5Z (Promega). The modified plasmid contained EcoRI and HindIII sites I inserted between the SalI and NcoI sites of the original plasmid. Oligonucleotides (AGCTTGAGTTCGAGTGTTTTTGC and CATGGCAAAAACACTCGAACTCA) encoding an RNA polymerase III termination signal were designed to generate HindIII/NcoI sites upon annealing and were inserted into the modified 5Z plasmid containing the U6 promoter. Oligonucleotides encoding the ribozyme and an unstructured spacer region (AATTCAAGCACAAACACAACGAAGGAACTGATGAGGCCG AAAGGCCGAAAGATGGCACACACACACAACAA and AGCTTTGTTGTGTGTGTGTGCCATCTTTCGGCCTTTCGGCCT CATCAGTTCCTTCGTTGTGTTTGTGCTTG) were designed to generate EcoRI/HindIII ends upon annealing and were then inserted into the modified U6 promoter plasmid. The final plasmid was designated “pU6C1049”. The human U1 promoter was amplified

Gene Therapy and Molecular Biology Vol 4, page 55 from plasmid HU1L (obtained from William Marzloff, Univ. of North Carolina) via PCR with oligonucleotides (GAGCGAATTCAGATCTAGGGCGACTTCTATGTAG and GAGCAAGCTTACGCGTCCTAGGTTCGGGCTCTGCCCCG AC). The plasmid HU1L was also amplified with oligonucleotides (AAACCTAGGATACTTACTGCA GGGGAGATACCATGATCACGAAGG and AAAGGATCCTC TAACTTTTAGCACCACAAAAATCAGGGGAAAGCGCGAA CGCAGTCCCCCACTACCACAAATTATGCAGTCGAGTTTC CCACATTTG) to generate sequences encoding a truncated U1 RNA and a mouse U1 RNA processing signal. These two PCR products were cloned into the EcoRI/BamHI sites in a pGEM4 plasmid (Promega) generating AvrII/PstI sites at the transcriptional start site of the U1 gene. Oligonucleotides encoding the ribozyme (CTAGGGAAGGAACTGATGA GGCCGAAAGGCCGAAAGATGGCCTGCA and GGCCATCT TTCGGCCTTTCGGCCTCATCAGTTCCTTCC) were designed to generate AvrII/PstI ends upon annealing and were inserted into the pGEMU1 plasmid. Oligonucleotides encoding the ribozyme and a spacer region (GAAGGAACTGATGAGGCCGAAAG GCCGAAAGATGGCAAAACTGTTCTGTTTTTG and GATCCAAAAACAGAACAGTTTTGCCATCTTTCGGCCTTT CGGCCTCATCAGTTCCTTCGC) were annealed to generate SacII/BamHI ends and cloned into the TRZ parent plasmid (Thompson et al., 1995a). Oligonucleotides were purchased from The Midland Certified Reagent Company. All ribozyme transcription unit inserts were verified by sequencing.

within the "LacZ stuffer fragment. The presence of the correct MCS was confirmed by sequence analysis. The entire expression cassette was then transferred to pAT19 using NotI to generate pAT22, such that the cassette is flanked by the AAV terminal repeats. Transcription units encoding the stromelysin-specific ribozyme driven by either a modified human U1 sn RNA, mouse U6 sn RNA, or human tRNA promoter (see above) were cloned into the MCS of pAT22 to generate pAT 23, pAT 36, and pAT 42, respectively. The U1- and U6-derived ribozyme transcription units were both cloned into the MCS of pAT22 to generate plasmid pAT43. All three of the U1-, U6-, and tRNA-derived ribozyme transcription units were cloned into the MCS of pAT22 to generate pAT44. For each of these plasmids, the entire expression cassette (for ribozyme and NGFR expression) was transferred to pAT19 using NotI, such that the cassette is flanked by the AAV terminal repeats. Each ribozyme transcription unit insert was verified by sequencing and rAAV plasmids were demonstrated to be functional in an AAV rescuereplication assay performed in 293 cells.

D. In vitro cleavage assay. Ribozyme plasmid inserts were amplified with oligonucleotides to generate T7 RNA polymerase promoter templates. Ribozyme RNAs were then transcribed in vitro (Ambion Megashortscript) and gel purified. Ribozyme RNA was renatured in cleavage buffer (10mM MgCl2, 50mM Tris-HCl, pH7) for 5 min at 65˚C, followed by 10 min at 37˚C, prior to incubation with a 32P-labeled RNA substrate containing the cleavage site. The cleavage reaction was performed under single turnover conditions in ribozyme excess and analyzed by denaturing PAGE.

C. rAAV plasmids. pLK-1b was a kind gift from Dr. Arun Srivastava (Indiana University School of Medicine, Indianapolis, IN). pLK-1b contains the bacterial LacZ gene with the SV40 poly(A) signal in the 3’ untranslated region, driven by the Cytomegalovirus immediate early (CMV) promoter. This CMV-LacZ cassette is flanked by AAV-termini within the pEMBL vector backbone. The recombinant and helper AAV plasmids, pSub201 and pAAV/Ad, respectively, were kindly provided by Dr. R.J. Samulski, University of North Carolina, Chapel Hill, N.C. pSub201 was digested to completion with Xho I and ligated to XhoI-NotI-XhoI linkers (adapters), generating the plasmid pAT19.

E. Ribozyme RNA stability analysis. Plasmids encoding the various ribozyme transcription units were transfected (2 µg DNA/ well of a 6-well plate) into 293 cells by calcium phosphate precipitation. After an overnight transfection, media was replaced and the cells were incubated an additional 24 h. Cells were then incubated in media containing 5µg/ml Actinomycin D. At various times, cells were lysed in guanidinium isothiocyanate, and total RNA was purified by phenol/chloroform extraction and isopropanol precipitation. RNA was analyzed by Northern blot and the levels of specific ribozyme RNAs were quantified by phosphorimager analysis (Molecular Dynamics PhosphorImager). The level of each RNA at time zero was set to 100%.

Due to the palindromic nature of AAV-ITRs which are prone to deletion in the bacterial host, most of the cloning steps were conducted in a vector backbone devoid of of AAV-ITRs, and the final version of the expression cassette for the selective marker and ribozyme was subsequently subcloned into pAT19 containing AAV-ITRs. A truncated cDNA of the NGFR gene (lacking the region encoding the cytolasmic domain) was subcloned from plasmid E2N as a 1.1 kb SacII-EcoRV fragment into pGEM5Z downstream of the CMV promoter. NotI linkers were then ligated to the purified NsiI – FspI fragment containing the CMV-NGFr-poly (A) expression cassette, and the fragment was subcloned into pBluescript-SK+ at the NotI site. Several restriction sites (SmaI through KpnI) from the pBluescript multiple cloning site (MCS) were deleted. To increase the size of rAAV genome for efficient packaging, a 1.6 kb stuffer fragment originating from the bacterial LacZ gene was inserted immediately upstream of the CMV promoter at the SalI site. An MCS for ribozyme insertion was cloned into the EcoRV site

F. Viral production and titer determination. For large scale preparation of rAAV, 293 cells were cotransfected using a calcium phosphate method (Morris, Labrie, and Mathews, 1994) with the helper plasmid pAAV/Ad and each recombinant AAV plasmid using 100 µg of each plasmid per 500 cm2 plate. Transfection was carried out for 12-16 hrs, followed by adenovirus dl312 infection (moi 5-10). Cells were processed for rAAV purification when significant adenovirus cytopathic effect was visible microscopically (65-72 hrs p.i). The cell pellet was frozen and thawed 3-4 times at -70oC and 37 oC, respectively, and the cell lysate was centrifuged to collect the supernatant. Virions in this crude preparation were precipitated in 10% polyethylene glycol, and the pellet was resuspended in Buffer C (50 mM Hepes, 10 mM EDTA, 150 mM NaCl pH 7.5 - 8.0). The rAAV particles in this preparation

Roberts et al: Gene therapy for arthritis using AAV were further purified by two CsCl equilibrium step gradient centrifugations (1.33 / 1.55 g/cm3 densities). The rAAV band (1.42-1.45 g/cm3 density) located at the junction between 1.33 and 1.5 g/cm3 CsCl densities, just below the visible band of adenovirus (1.33 g/cm3 density), was collected and dialyzed against PBS containing 1% sucrose. rAAVs are designated vAT corresponding to the plasmid used for co-transfection (e.g., pAT22 yields vAT22). The rAAV for LacZ expression was designated vLK-1b. The copy number of rAAV genomes was determined using a previously described slot-blot hybridization method (Kube and Srivastava, 1997). Infectious units per ml were determined by FACS analysis of cell surface NGFR expression.

complimentary to the poly(U) and ribozyme but not to the U6 promoter. Thus the probe should hybridize to ribozyme transcripts synthesized from the U1, U6, or tRNA promoters with similar efficiencies. Following hybridization, filters were washed and subjected to both autoradiography and phosphorimager analysis. Ribozyme RNA samples were normalized to endogenous tRNA.

J. RNase Protection Analysis (RPA). A glyceraldehyde-3-phosphate dehydrogenase (GAPDH) RNA probe was used in each of these experiments as an internal control. Template DNA plasmids were transcribed using T7 RNA polymerase in the presence of [#32P]-CTP for 1 hr at 37° C. The reaction was terminated by the addition of DNase I and the volume was increased to 200 µl with TE. Unincorporated ribonucleotides were removed using a G-50 column equilibrated in TE. The stromelysin probe contains sequences corresponding to nucleotide positions 949 - 1201, based on the rabbit (RABSTROM) sequence (accession #M25664). The integrity of the RNA probe was checked by autoradiography following electrophoresis on a 6% denaturing polyacrylamide gel. Each probe (1x106 cpm) was added to crude lysate from 100,000 cells (in lysis buffer, see RNA purification above), and incubated overnight at 37° C. RNase mixture (500ml) containing RNase A (500 U) and RNase T1 (1,000 U) was then added and the reaction was incubated for an additional 45 minutes at 37° C. The reaction was terminated by the addition of SDS and proteinase K and incubation proceeded for 45 minutes at 37°C. The RNA was precipitated with isopropanol and resuspended in loading buffer. The sample was heated to 95°C for 5 min, chilled on ice, and electrophoresed on a 7 M urea / 6% polyacrylamide sequencing gel. The levels of stromelysin and GAPDH RNAs were quantified by phosphorimager analysis.

G. X-gal staining of rAAV-infected cells. Cell culture media was aspirated and cells were fixed for 5 minutes at 4 °C with 0.74% formaldehyde / 0.05% glutaraldehyde in PBS lacking magnesium. The fixative was then aspirated and X-Gal Staining Solution (5 mM each of ferroisothiocyanate and ferrus-isothiocyanate containing 1 mg/mL XGal and 2 mM MgCl2) was added. Cells were incubated at 37 °C until color develops.

H. Fluorescence Activated Cell Scanning (FACS) Analysis. RSF cells were harvested 48 hours after virus or mock infection and washed twice with FACS Buffer (phosphatebuffered saline containing 1.0% BSA and 0.1% azide). 500 cells / µl were incubated for 30 minutes at 4 °C with primary antibody (mouse anti-human NGFR IgG). Cells were then washed twice with FACS Buffer and incubated for 30 minutes at 4 °C with secondary antibody (FITC-conjugated sheep anti-mouse IgG). Cells were washed twice with FACS Buffer and resuspended in FACS Buffer containing 1 µg/µl propidium iodide for analysis. Stained cell preparations were analyzed using a Becton Dickinson FACScan instrument.

References Bertrand, E., Castanotto, D., Zhou, C., Carbonnelle, C., Lee, N. S., Good, P., Chatterjee, S., Grange, T., Pictet, R., Kohn, D., Engelke, D., and Rossi, J. J. (1997) The expression cassette determines the functional activity of ribozymes in mammalian cells by controlling their intracellular localization. RNA 3, 75-88. Brinckerhoff, C. E., and Auble, D. T. (1990) Regulation of collagenase gene expression in synovial fibroblasts. Annals New York Acad Sci 580, 355-74.

I. Northern analysis. Virus or mock infected cells were lysed at 1x107 cells/mL in Lysis Buffer (4 M guanidine thiocyanate, 25 mM sodium citrate pH 7.0, and 0.5% sarcosyl) to which 0.1 volume of 3 M sodium acetate was added and the mixture was vortexed. Two sequential extractions with chloroform/isoamyl alcohol (25:1) were performed, and RNA in the aqueous phase was precipitated with isopropanol. The pellet was washed with 100% ethanol and resuspended in RNase-free water. Northern analysis was performed as described (Thompson et al., 1995a). Briefly, purified RNA (10 µg per sample) was denatured in loading dye by heating at 95 °C for 3 min and electrophoresed on a 6% acrylamide / 7 M Urea gel in Tris-Borate-EDTA (TBE) buffer. The separated RNA was then transferred to a Genescreen Plus membrane, prehybridized for 1 hour with Church buffer (5% SDS, 0.2 M sodium phosphate) + 0.5% casein, and hybridized overnight at 65 °C with a 32[P]-labelled RNA probe (1 x 106 cpm/ml) in the same buffer. The ribozyme probe was obtained by in vitro T7 transcription using the pGEM5Z-based plasmid pU6C1049 described above, and contains sequences

Brooks, P. C., Strömblad, S., Sanders, L. C., von Schalscha, T. L., Aimes, R. T., Stetler-Stevenson, W. G., Quigley, J. P., and Cheresh, D. A. (1996) Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alpha v beta 3. Cell 85, 683-93. Brown, P. D. (1997) Matrix metalloproteinase inhibitors in the treatment of cancer. Medical Oncology 14, 1-10. Burger, D., Rezzonico, R., Li, J. M., Modoux, C., Pierce, R. A., Welgus, H. G., and Dayer, J. M. (1998) Imbalance between interstitial collagenase and tissue inhibitor of metalloproteinases 1 in synoviocytes and fibroblasts upon direct contact with stimulated T lymphocytes: involvement of membrane-associated cytokines. Arthritis Rheum 41, 1748-59. Cawston, T. E. (1996) Metalloproteinase inhibitors and the prevention of connective tissue breakdown. Pharmacol Therapeutics 70, 163-82.

Gene Therapy and Molecular Biology Vol 4, page 57 Chowrira, B. M., Pavco, P. A., and McSwiggen, J. A. (1994) In vitro and in vivo comparison of hammerhead, hairpin, and hepatitis delta virus self-processing ribozyme cassettes. J Biol Chem 269, 25856-64. Flory, C. M., Pavco, P. A., Jarvis, T. C., Lesch, M. E., Wincott, F. E., Beigelman, L., Hunt, S. W. r., and Schrier, D. J. (1996) Nuclease-resistant ribozymes decrease stromelysin mRNA levels in rabbit synovium following exogenous delivery to the knee joint. Proc Natl Acad Sci USA 93, 754-8. Giannelli, G., Falk-Marzillier, J., Schiraldi, O., Stetler-Stevenson, W. G., and Quaranta, V. (1997) Induction of cell migration by matrix metalloprotease-2 cleavage of laminin-5. Science 277, 225-8.

2/progelatinase A complex by stromelysin. Biochem Biophys Res Commun 185, 852-9. Morris, G. F., Labrie, C., and Mathews, M. B. (1994) Modulation of transcriptional activation of the proliferating cell nuclear antigen promoter by the adenovirus E1A 243-residue oncoprotein depends on proximal activators. Mol Cell Biol 14, 543-53. Murphy, G., and Hembry, R. M. (1992) Proteinases in rheumatoid arthritis. J Rheumatol. Suppl 32, 61-4. Muzyczka, N. (1992) Use of adeno-associated virus as a general transduction vector for mammalian cells. Curr Topics Microbiol Immunol 158, 97-129. Nagase, H. (1996). Matrix Metalloproteinases. In â&#x20AC;&#x153;Zinc Metalloproteases in Health and Diseaseâ&#x20AC;? (N. M. Hooper, Ed.), pp. 153-196. Taylor and Francis. Ogata, Y., Enghild, J. J., and Nagase, H. (1992) Matrix metalloproteinase 3 (stromelysin) activates the precursor for the human matrix metalloproteinase 9. J Biol Chem 267, 3581-4. Okada, Y., Shinmei, M., Tanaka, O., Naka, K., Kimura, A., Nakanishi, I., Bayliss, M. T., Iwata, K., and Nagase, H. (1992) Localization of matrix metalloproteinase 3 (stromelysin) in osteoarthritic cartilage and synovium. Lab Investig 66, 680-90. Ponnazhagan, S., Mukherjee, P., Yoder, M. C., Wang, X. S., Zhou, S. Z., Kaplan, J., Wadsworth, S., and Srivastava, A. (1997) Adenoassociated virus 2-mediated gene transfer in vivo: organ-tropism and expression of transduced sequences in mice. Gene 190, 20310. Ponnazhagan, S., Wang, X. S., Woody, M. J., Luo, F., Kang, L. Y., Nallari, M. L., Munshi, N. C., Zhou, S. Z., and Srivastava, A. (1996) Differential expression in human cells from the p6 promoter of human parvovirus B19 following plasmid transfection and recombinant adeno-associated virus 2 (AAV) infection: human megakaryocytic leukaemia cells are nonpermissive for AAV infection. J Gen Virol 77 ( Pt 6), 1111-22.

Hembry, R. M., Bagga, M. R., Reynolds, J. J., and Hamblen, D. L. (1995) Immunolocalisation studies on six matrix metalloproteinases and their inhibitors, TIMP-1 and TIMP-2, in synovia from patients with osteo- and rheumatoid arthritis. Annals Rheum Diseases 54, 25-32. Jarvis, T. C., Wincott, F. E., Alby, L. J., McSwiggen, J. A., Beigelman, L., Gustofson, J., DiRenzo, A., Levy, K., Arthur, M., Matulic-Adamic, J., Karpeisky, A., Gonzalez, C., Woolf, T. M., Usman, N., and Stinchcomb, D. T. (1996) Optimizing the cell efficacy of synthetic ribozymes. Site selection and chemical modifications of ribozymes targeting the protooncogene c-myb. J Biol Chem 271, 29107-12. Kaplitt, M. G., Leone, P., Samulski, R. J., Xiao, X., Pfaff, D. W., O'Malley, K. L., and During, M. J. (1994) Long-term gene expression and phenotypic correction using adeno-associated virus vectors in the mammalian brain. Nature Genetics 8, 148-54. Koeberl, D. D., Alexander, I. E., Halbert, C. L., Russell, D. W., and Miller, A. D. (1997) Persistent expression of human clotting factor IX from mouse liver after intravenous injection of adeno-associated virus vectors. Proc Natl Acad Sci USA 94, 1426-31. Kotin, R. M., Siniscalco, M., Samulski, R. J., Zhu, X. D., Hunter, L., Laughlin, C. A., McLaughlin, S., Muzyczka, N., Rocchi, M., and Berns, K. I. (1990) Site-specific integration by adeno-associated virus. Proc Natl Acad Sci USA 87, 2211-5. Kube, D. M., and Srivastava, A. (1997) Quantitative DNA slot blot analysis: inhibition of DNA binding to membranes by magnesium ions. Nucleic Acids Res 25, 3375-6. Lawrence, R. C., Hochberg, M. C., Kelsey, J. L., McDuffie, F. C., Medsger, T. A. J., Felts, W. R., and Shulman, L. E. (1989) Estimates of the prevalence of selected arthritic and musculoskeletal diseases in the United States. J Rheumatol 16, 427-41. Lewin, A. S., Drenser, K. A., Hauswirth, W. W., Nishikawa, S., Yasumura, D., Flannery, J. G., and LaVail, M. M. (1998) Ribozyme rescue of photoreceptor cells in a transgenic rat model of autosomal dominant retinitis pigmentosa. Nat Med 4, 967-71. Macejak, D. M., Lin, H., Webb, S., Chase, C., Jensen, K., Jarvis, T. C., Leiden, J. M., and Couture, L. (1999) Adenovirusmediated expression of a ribozyme to c-myb mRNA inhibits smooth muscle cell proliferation and neointima formation in vivo. J Virol 73, 7745-7751. Miyazaki, K., Umenishi, F., Funahashi, K., Koshikawa, N., Yasumitsu, H., and Umeda, M. (1992) Activation of TIMP-

Rossi, J. J. (1999) Ribozymes, genomics and therapeutics. Chem Biol 6, R33-37. Ruffner, D. E., Stormo, G. D., and Uhlenbeck, O. C. (1990) Sequence requirements of the hammerhead RNA self-cleavage reaction. Biochemistry 29, 10695-702. Samulski, R. J., Zhu, X., Xiao, X., Brook, J. D., Housman, D. E., Epstein, N., and Hunter, L. A. (1991) Targeted integration of adeno-associated virus (AAV) into human chromosome 19 [published erratum appears in EMBO J 1992 Mar;11(3):1228]. EMBO J 10, 3941-50. Schlehofer, J. R., Ehrbar, M., and zur Hausen, H. (1986) Vaccinia virus, herpes simplex virus, and carcinogens induce DNA amplification in a human cell line and support replication of a helpervirus dependent parvovirus. Virology 152, 110-7. Stetler-Stevenson, W. G., Aznavoorian, S., and Liotta, L. A. (1993) Tumor cell interactions with the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol 9, 541-73. Suzuki, K., Enghild, J. J., Morodomi, T., Salvesen, G., and Nagase, H. ( 1990) Mechanisms of activation of tissue procollagenase by matrix metalloproteinase 3 (stromelysin). Biochemistry 29, 10261-70. Thompson, J. D., Ayers, D. F., Malmstrom, T. A., McKenzie, T. L., Ganousis, L., Chowrira, B. M., Couture, L., and Stinchcomb, D. T. (1995a) Improved accumulation and activity of ribozymes expressed from a tRNA-based RNA polymerase III promoter. Nucleic Acids Res 23, 2259-68.

Roberts et al: Gene therapy for arthritis using AAV Thompson, J. D., Macejak, D., Couture, L., and Stinchcomb, D. T. (1995b) Ribozymes in gene therapy. Nat Med 1, 277-8. Thomson, B. J., Weindler, F. W., Gray, D., Schwaab, V., and Heilbronn, R. (1994) Human herpesvirus 6 (HHV-6) is a helper virus for adeno-associated virus type 2 (AAV-2) and the AAV-2 rep gene homologue in HHV-6 can mediate AAV-2 DNA replication and regulate gene expression. Virology 204, 304-11. Vincenti, M. P., Clark, I. M., and Brinckerhoff, C. E. (1994) Using inhibitors of metalloproteinases to treat arthritis. Easier said than done? [see comments]. Arthritis Rheumatism 37, 1115-26. Werb, Z., Alexander, C. M., and Adler, R. R. (1992) Expression and function of matrix metalloproteinases in development. Matrix. Supplement 1, 337-43. Woessner, J. F. J. (1991) Matrix metalloproteinases and their inhibitors in connective tissue remodeling. FASEB Journal 5, 2145-54.

Gene Therapy and Molecular Biology Vol 4, page 59 Gene Ther Mol Biol Vol 4, 59-74. December 1999.

Glial cell line-derived neurotrophic factor (GDNF) gene therapy in an aged rat model of Parkinson's disease Review Article

Bronwen Connor and Martha C. Bohn Children's Memorial Institute For Education and Research, Department of Pediatrics, Northwestern University, Chicago, IL 60614. ______________________________________________________________________________________________________ Correspondence: Martha C. Bohn, Ph.D., Children's Memorial Institute For Education and Research, 2300 Children's Plaza # 209, Chicago, IL 60614; Tel: (773)-868 8052; Fax: (773)-868 8066; E-mail: m-bohn@nwu.edu Abbreviations: GDNF, glial cell line-derived neurotrophic factor; BDNF, brain-derived neurotrophic factor; DA, dopaminergic; 6-OHDA, 6-hydroxydopamine; Ad, adenoviral; SN, substantia nigra pars compacta; CNS, central nervous system; MFB, medial forebrain bundle; MPTP, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine; MPP+, 1-methyl-4-phenyl pyridinium; MAO-B, monoamine oxidase; TH, tyrosine hydroxylase; AAV, adeno-associated virus; PSP, persephin; NTN, neurturin; IR, immunoreactivity Key Words: Glial cell line-derived neurotrophic factor, GDNF, gene therapy, ex-vivo, Parkinson's disease, neurotrophic factors, dopaminergic neurons, striatum, 6-OHDA, neurodegenerative disorders, brain-derived neurotrophic factor Received: 22 September 1999; accepted: 29 September 1999

Summary The chronic delivery of neuroprotective factors to specific regions of the CNS via gene therapy offers a new strategy for the treatment of neurodegenerative disorders. The neurotrophic factor, glial cell line-derived neurotrophic factor (GDNF) is a potent dopaminergic (DA) trophic factor that ameliorates the behavioral and histological consequences of lesioning DA neurons in rodent and primate models of Parkinson’s disease. GDNF gene therapy may therefore have potential use in the treatment of Parkinson's disease. We have observed previously that an adenoviral (Ad) vector harboring a GDNF minigene protects DA neurons from degeneration in the young rat brain. However, as Parkinson’s disease occurs primarily in aged populations, we have also studied the effects of GDNF gene delivery in an aged rat model of Parkinson’s disease. In the aged (20 month) Fischer 344 rat, an Ad vector was used to deliver GDNF either to the DA cell bodies in the SN or to the DA terminals in the striatum. One week following gene delivery, the neurotoxin 6-hydroxydopamine (6-OHDA) was injected unilaterally into the striatum to cause progressive degeneration of DA neurons. Injection of GDNF vector into either the striatum or SN provided significant cell protection against 6-OHDA. However, only striatal injections of Ad GDNF protected against the development of behavioral and neurochemical changes that occur in the DA-depleted brain. The results of this study are reviewed here and the behavioral and cellular effects of GDNF gene delivery into striatal versus mesencephalic sites are discussed.

I. Introduction While neurotrophic factors are promising therapeutic agents in the treatment of neurodegenerative disorders, the delivery of these factors to the central nervous system (CNS) provides an interesting challenge. The identification of factors with potent dopaminergic (DA) trophic activities in vitro led to the concept of using neurotrophic factors as therapeutic agents for Parkinson’s disease. The most potent of these factors is glial cell linederived neurotrophic factor (GDNF) (Lin et al., 1993; Lin

et al., 1994). There is much in vivo evidence demonstrating the effectiveness of GDNF protein in ameliorating neurodegeneration and maintaining behavioral function in animal models of Parkinson's disease. However, neurotrophic factors are labile substances that are unable to cross the bloodbrain barrier in significant amounts. Therefore, the therapeutic use of factors such as GDNF in the treatment of Parkinson's disease will require the development of methods for delivering these factors to specific regions of the CNS in a continuous and regulatable manner that is safe, minimally invasive and does not result in side effects. 59

Connor and Bohn: GDNF gene therapy for Parkinson's disease Parkinsonâ&#x20AC;&#x2122;s disease is a neurodegenerative disorder characterized by the progressive degeneration of DA neurons in the substantia nigra pars compacta (SN) that innervate the striatum. This degeneration results in the loss of DA terminals in the caudate and putamen leading to the clinical symptoms of bradykinesia, rigidity and resting tremor. Because Parkinson's disease is progressive, it is likely that long-term trophic support for degenerating DA neurons will be required. Repeated injections of recombinant neurotrophic factors into the human brain are unlikely to be practical, and are likely to elicit deleterious side effects over the long-term. In contrast, genetic approaches are ideal for delivery of neurotrophic factors to the CNS. Through the use of gene therapy techniques, increased levels of neurotrophic factor biosynthesis in the CNS might prevent neuronal cell death and enhance neuronal function. Furthermore, the development of vectors harboring genes driven by specific cellular promoters would provide the potential of limiting expression of a transgene to one typically defined cell population in the CNS, as well as regulating transgene expression. Several research groups, using adenoviral (Ad) or adeno-associated viral vectors expressing GDNF, have reported that viral delivery of a GDNF gene protects DA neurons in the young adult rat brain from 6-OHDAinduced degeneration (Bilang-Bleuel et al., 1997; ChoiLundberg et al., 1997; Choi-Lundberg et al., 1998; Mandel et al., 1997). However, Parkinsonâ&#x20AC;&#x2122;s disease is a neurodegenerative disorder that primarily affects the aging population. Age-related reductions in neurotransmitter synthesis and alterations in receptor levels may contribute to ongoing degenerative and behavioral deficits in Parkinson's disease. Multiple functional changes of the nigrostriatal DA system, similar to those observed in the Parkinsonian brain, have been described during normal aging in both animals and humans (Felten et al., 1992; Hubble, 1998; Morgan and Finch, 1988; Roth and Joseph, 1994). Compensatory events may be observed in the aged brain due to age-dependent degenerative changes occurring in the nigrostriatal and mesolimbic pathways. Alternatively, the aged brain may be unable to produce compensatory changes in response to either age-dependent or lesion-induced neuronal degeneration, worsening lesion-induced degenerative alterations in DA neuronal morphology and function in the aged brain (Demarest et al., 1980; Schallert, 1988; Unnerstall and Long, 1996; Zigmond et al., 1993). Therefore, while Ad GDNF has been observed to exhibit neuroprotective effects within the young rat brain, we believed it to be important to examine whether chronic biosynthesis of GDNF, achieved by Admediated delivery of GDNF, is able to protect DA neurons and maintain DA function in an aged rat model of Parkinson's disease. Here, we review the use of Ad to deliver a GDNF minigene to the aged (20 month) rat brain, prior to a partial, progressive lesion of the nigrostriatal pathway. The results show that Ad GDNF not only protects DA neurons from degeneration, but that functional

consequences in DA target neurons following a lesion of the nigrostriatal pathway are prevented by Ad GDNF in a brain region specific manner.

II. Animal models of Parkinson's disease Since Parkinson's disease does not occur naturally in animals, several well characterized models have been developed for stimulating the neuropathological and neurological features of Parkinson's disease in laboratory animals (reviewed in Bankiewicz et al., 1993). The most common animal model of Parkinson's disease involves the intracerebral injection of the catecholamine neurotoxin, 6hydroxydopamine (6-OHDA), resulting in a reduction of dopaminergic phenotypic markers (i.e. tyrosine hydroxylase) and the selective death of DA neurons. 6-OHDA is a dopamine analog that is specifically taken up by the highaffinity dopamine transporter. 6-OHDA undergoes autooxidation, generating hydroxyl radical, hydrogen peroxide and superoxide anion, which causes damage to various cellular components. Injection of 6-OHDA into either the striatum, SN or the medial forebrain bundle (MFB), which contains dopaminergic axons from the ventral tegmental area and SN, results in specific loss of dopaminergic neurons and fibers. The neurotoxin MPTP (1-methyl-4-phenyl-1,2,3,6tetrahydropyridine) also causes Parkinsonian symptoms in humans, non-human primates and mice, following systemic administration. MPTP is oxidized to the toxic molecule, MPP+ (1-methyl-4-phenyl pyridinium) by the monoamine oxidase system (MAO-B). MPP+ enters dopaminergic neurons by high-affinity through the dopamine transporter and interferes with ATP production by inhibiting complex 1 of the mitochondrial electron transport chain. The standard animal model of Parkinson's disease, based on unilateral injection of 6-OHDA into either the SN or the MFB, has been used extensively in Parkinson's disease research. Injection of 6-OHDA in either the SN or the MFB results in the rapid death of dopaminergic neurons, within 48 hours. However, due to the rapid and near-complete degeneration of nigral DA neurons, these lesion models do not closely reflect the clinical picture in which DA neurons die over a prolonged time. These models also are not optimal for studies on neuroprotection or regeneration of the nigrostriatal system (reviewed in Bjorklund et al., 1997). In contrast, intrastriatal delivery of 6-OHDA results in the rapid destruction of dopaminergic terminals in the striatum, followed by progressive degeneration of parent DA cell bodies in the SN over a period of several weeks (Sauer and Oertel, 1994). This degeneration is preceded and accompanied by cellular atrophy and a partial loss of tyrosine hydroxylase (TH) expression. This progressive lesion model yields an animal model which closely resembles the early stages of Parkinson's disease in humans, in which a portion of the nigrostriatal projection remains intact. These remaining neurons serve as a substrate for regeneration and functional recovery. The progressive lesion model is an ideal model for studying the neuroprotective or restorative properties of trophic factors (Bilang-Bleuel et al., 1997; Choi-Lundberg et al., 1997; Choi-Lundberg et al., 1998; Connor et al., 1999; Horger et al., 1998; Mandel et al., 1997; Milbrandt et al., 60

Gene Therapy and Molecular Biology Vol 4, page 61 1998; Rosenblad et al., 1999; Rosenblad et al., 1998; Sauer et al., 1995; Shults et al., 1996). Unilateral physical and chemical lesions of the nigrostriatal pathway result in an imbalance in the level of DA and DA receptors between the two striatae, producing impairment in DA-dependent behavioral function. Lesioninduced behavioral impairment and the effects of experimental interventions can be quantified by several means (reviewed in (Schallert, 1995; Schwarting and Huston, 1996). Specifically, unilaterally lesioned animals exhibit rotational behavior in response to amphetamine or DA agonists, such as apomorphine, that is readily quantifiable (Schwarting and Huston, 1996; Ungerstedt and Arbuthnott, 1970; Figure 1A). Specifically, in animals with a unilateral lesion of the nigrostriatal DA

system, the injection of drugs that act to release DA, such as amphetamine, will induce rotational behavior towards the denervated striatum due to an imbalance in striatal DA levels. Lesioned animals will turn away from the hemisphere where there is greater amphetamine-stimulated DA release and greater DA receptor stimulation (Ungerstedt, 1971). In contrast, injection of DA agonists, such as apomorphine, in animals exhibiting a < 90% loss of DA in the lesioned hemisphere, will induce rotational behavior away from the denervated striatum due to compensatory upregulation and supersensitivity of DA receptors in the unlesioned striatum (Ungerstedt and Arbuthnott, 1970). In addition, unilaterally lesioned animals exhibit deficits in contralateral limb use in several spontaneous behaviors (Olsson et al., 1995; Schallert, 1995; Figure 1B).

Figure 1A. - Diagram representing the behavioral and neurochemical changes that occur in the DA-depleted brain. In rats with a unilateral lesion of the nigrostriatal DA system, injection of amphetamine will induce rotational behavior towards the lesioned hemisphere. The lesioned rat turns away from the hemisphere where there is greater amphetamine-stimulated DA release and greater DA receptor stimulation. In the unlesioned striatum, injection of amphetamine also results in the induction of the transcription factor, c-fos in striato-nigral cells. However in the lesioned striatum, due to a loss of functional DA terminals and a reduction in striatal DA levels, amphetamine-induced Fos expression is reduced. Dashed lines indicates the degeneration and loss of the nigrostriatal pathway connecting the SN and the ST. Hatched area represents a loss of DA.

Figure 1B. - Rat exhibiting preferential use of the ipsilateral forelimb following a unilateral 6-OHDA-induced lesion of the nigrostriatal pathway. Spontaneous exploratory forelimb use is a non-drug induced test of forelimb locomotor function which has been shown to correlate with DA depletion in the lesioned hemisphere. Following a unilateral 6-OHDA lesion of the nigrostriatal pathway, rats preferentially use the forelimb ipsilateral to the side of the lesion to initiate and terminate weight-shifting movements during rearing and exploration along vertical surfaces.

Connor and Bohn: GDNF gene therapy for Parkinson's disease

III. Neurotrophic factors for dopaminergic neurons The discovery and characterization of neurotrophic factors that promote the survival, neurite outgrowth and phenotypic differentiation of DA neurons has been an area of intense research in recent years. Early studies of DA neurons in culture showed glia from different brain regions and glial cell lines produce factors that influence the survival and differentiation of embryonic DA neurons in culture (Engele and Bohn, 1991; Engele et al., 1991; Rousselet et al., 1988). These studies prompted the search for specific DA neurotrophic factors. To date, more than 20 neurotrophic factors have been identified for DA neurons (Table 1). These include members of several growth factor families, operating through different intracellular signalling mechanisms, including the TGF-! superfamily, the neurotrophins, cytokines and mitogenic growth factors. The first neurotrophic factor shown to act directly on DA neurons was brain-derived neurotrophic factor (BDNF) (Hyman et al., 1991). Subsequently, GDNF was purified from the glial cell line, B49, and shown to be a very potent DA factor that enhances survival and neurite outgrowth of embryonic DA neurons in vitro (Lin et al., 1993; Lin et al., 1994). Recently, three additional members of the GDNF family have been

identified - neurturin (Kotzbauer et al., 1996), persephin (Milbrandt et al., 1998) and artemin (Baloh et al., 1998). These factors have been shown to exhibit various degrees of trophic support to DA neurons. A wealth of in vivo evidence has been accumulated in animal models of Parkinson's disease, supporting the potential therapeutic use of several DA neurotrophic factors, in particular GDNF, in the treatment of Parkinson's disease. A summary of these studies is shown in Table 2. GDNF has been shown to ameliorate the behavioral and pathological consequences of lesioning DA neurons in rodent and primate models of Parkinson's disease when administered to the adult nigrostriatal DA system (reviewed in Bjorklund et al., 1997; Gash et al., 1998). Specifically, in animal models of Parkinson's disease, single or continuous injection of recombinant GDNF protein has been shown to rescue injured or axotomized DA neurons when administered before or shortly after insult, and to preserve injured atrophic DA neurons during chronic neurodegeneration (Beck et al., 1995; Bowenkamp et al., 1995; Gash et al., 1996; Hoffer et al., 1994; Kearns and Gash, 1995; Lapchak et al., 1997; Sauer et al., 1995; Tomac et al., 1995). In addition, GDNF stimulates regenerative growth or axonal sprouting after partial lesions of the DA system and stimulates metabolism and function of lesioned DA neurons (Lindner et al., 1995; Rosenblad et al., 1998; Shults et al., 1996).

Table 1. Dopaminergic Neurotrophic Factors (updated from Bohn and Choi-Lundberg, 1997) TGF-! Superfamily GDNF Family GDNF Neurturin Persephin Artemin

(Lin et al., 1993; Lin et al., 1994) (Kotzbauer et al., 1996) (Milbrandt et al., 1998) (Baloh et al., 1998)

Others TGF-!-1 TGF-!-2, 3 GDF-5 Activin A

(Krieglstein et al., 1995) (Krieglstein et al., 1995; Poulsen et al., 1994) (Krieglstein et al., 1995) (Krieglstein et al., 1995)

Neurotrophins BDNF NT-3 NT-4/5

(Hyman et al., 1991) (Hyman et al., 1994) (Hyman et al., 1994)

Mitogenic Growth Factors TGF-" aFGF and bFGF EGF Insulin IGF-I IGF-2 PDGF Midkine

(Alexi and Hefti, 1993) (Beck et al., 1993; Engele and Bohn, 1991; Ferrari et al., 1989; Otto and Unsicker, 1990) (Casper et al., 1991) (Knusel et al., 1990) (Beck et al., 1993) (Liu and Lauder, 1992) (Othberg et al., 1995) (Kikuchi et al., 1993)

Cytokines CNTF IL-1! IL-6 IL-7 Cardiotrophin-1

(Hagg and Varon, 1993; Magel et al., 1993) (Akaneya et al., 1995) (Hama et al., 1991; von Coelln et al., 1995) (von Coelln et al., 1995) (Pennica et al., 1995)

Gene Therapy and Molecular Biology Vol 4, page 63 term. In contrast, genetic approaches are ideal for delivering GDNF to the CNS. Several laboratories have studied viral vector mediated GDNF gene delivery in animal models of Parkinson's disease (Table 2). In young rats, we have observed that injection of an Ad vector harboring human GDNF into the SN one week prior to a progressive 6OHDA lesion significantly protects DA neurons in the lesioned SN (Choi-Lundberg et al., 1997; Choi-Lundberg et al., 1998). In addition, intrastriatal injection of Ad GDNF in this model was effective in protecting DA cell bodies and also prevented the onset of DA-dependent behaviors that occurred in control rats as a consequence of the unilateral 6-OHDA lesion (Choi-Lundberg et al., 1998). Similar effects of GDNF gene delivery have been confirmed by other groups using either adenoviral or adeno-associated viral vectors (Table 2). In addition, our laboratory has recently observed that injection of Ad GDNF into the SN 1 week after 6-OHDA induced damage has commenced, not only rescues DA neurons, but also protects against alterations in DA target cells in the striatum (Kozlowski et al, Submitted).

Comparable findings for BDNF have been reported in both the 6-OHDA lesioned rat and the MPTP primate model in which BDNF improved DA levels or DAdependent behavior in the absence of an effect on the density of DA fibers in the striatum (Altar et al., 1994; Galpern et al., 1996; Levivier et al., 1995; Lucidi-Phillipi et al., 1995; Tsukahara et al., 1995; Yoshimoto et al., 1995). In addition, both neurturin and persephin have been reported to protect mature DA neurons from cell death induced by 6-OHDA in the absence of striatal reinnervation and DA-dependent behavioral improvement (Akerud et al., 1999; Horger et al., 1998; Milbrandt et al., 1998; Rosenblad et al., 1999; Tseng et al., 1998). While many studies have demonstrated the efficacy of DA neurotrophic factors, such as GDNF, in animal models of Parkinson's disease, these studies have utilized infusion of protein into the striatum, near the SN or into the lateral ventricles. In addition, large quantities of proteins are infused in these studies (typically 10Âľg or more), and repeated injections are typically required to maintain effects. Repeated injections of recombinant neurotrophic factors into the human brain are unlikely to be practical, and are likely to elicit deleterious side effects over the long

Table 2 Delivery of neurotrophic factors in animal models of Parkinsonâ&#x20AC;&#x2122;s disease (selected references) Paradigm

Delivery site, Factor

Biological Effects

Reference

Neurotrophic factor protein infusion 6-OHDA - complete

Intranigral, GDNF

! in nigral DA levels Improvement in behavioral function

Hoffer et al., 1994

6-OHDA - partial striatal

Intranigral, BDNF / NT-3

! in striatal DA metabolites Improvement in behavioral function No change in striatal DA levels No reinnervation of lesioned striatum

Altar et al., 1994

6-OHDA - complete

Intranigral, GDNF

! survival of DA cell bodies in SN Improvement in behavioral function No reinnervation of lesioned striatum

Bowenkamp et al., 1995

MPTP - mouse model

Intrastriatal, GDNF Intranigral, GDNF

! in nigral and striatal DA / metabolite levels Protection of striatal innervation (intrastriatal only)

Tomac et al., 1995

6-OHDA - complete

Intranigral, GDNF

Protection of DA cell bodies in SN Improvement in behavioral function

Beck et al., 1995

MPTP - monkey model

Intrathecal, BDNF

Protection of DA cell bodies in SN Improvement in behavioral function

Tsukahara et al., 1995

6-OHDA - complete

Intranigral, GDNF Intraventricular, GDNF

! in striatal DA / metabolite levels Protection of DA uptake sites Improvement in behavioral function

Opacka-Juffry et al., 1995

6-OHDA - complete or partial

Intranigral, GDNF

Protection of DA cell bodies in SN

Kearns and Gash, 1995

6-OHDA - partial striatal

Intranigral ,GDNF

Protection of DA cell bodies in SN

Sauer et al., 1995

6-OHDA - partial striatal

Intrastriatal ,GDNF

Protection of DA cell bodies in SN Protection of striatal innervation Improvement in behavioral function

Shults et al., 1996

Connor and Bohn: GDNF gene therapy for Parkinson's disease 6-OHDA - complete

Intranigral, GDNF

Protection of DA cell bodies in SN No reinnervation of lesioned striatum No improvement in behavioral function

Winkler et al., 1996

MPTP - monkey model

Intracerebral, GDNF

! DA cell body size ! fiber density ! in nigral and striatal DA levels Improvement in behavioral function

Gash et al., 1996

6-OHDA - complete

Intraventricular, GDNF

! in nigral DA / metabolite levels Protection of DA cell bodies in SN Improvement in behavioral function

Bowenkamp et al., 1997

6-OHDA - complete

Intranigral, GDNF

Protection of DA cell bodies in SN Did not restore TH-IR in nigral neurons

Lu and Hagg, 1997

6-OHDA - complete

Intranigral ,GDNF

Partial restoration of nigral DA levels Improvement in behavioral function

Hoffman et al., 1997

6-OHDA - complete

Intraventricular, GDNF Intranigral, GDNF

! expression of TH in SN Improvement in behavioral function Intranigral GDNF prevented lesion-induced increase in dynorphin A

Lapchak et al., 1997

6-OHDA - partial striatal

Intranigral, NTN

Protection of DA cell bodies in SN

Horger et al., 1998

6-OHDA - complete

Intranigral, GDNF Intraventricular, GDNF

Prevented loss of DA re-uptake sites Prevented loss of striatal DA / metabolites Prevented loss of DA cell bodies in SN Improvement in behavioral function

Sullivan et al., 1998

6-OHDA - partial striatal

Intranigral, PSP

Protection of DA cell bodies in SN

Milbrandt et al., 1998

MPTP - mouse model

Intrastriatal ,GDNF

Protection of nigral and striatal DA levels Improvement in behavioral function

Cheng et al., 1998

6-OHDA - partial striatal

Intrastriatal, GDNF

Restored DA uptake sites Protection of DA cell bodies in SN Improvement in behavioral function

Rosenblad et al., 1998

6-OHDA - partial striatal

Intrastriatal, GDNF / NTN Intraventricular, GDNF / NTN

GDNF - Protection of DA cell bodies in SN - No reinnervation of lesioned striatum - No improvement in behavioral function NTN - Partial protection of DA cell bodies in SN (intrastriatal only) - No reinnervation of lesioned striatum - No improvement in behavioral function

Rosenblad et al., 1999

MPTP - monkey model

Intraventricular, GDNF

! in nigral DA / metabolite levels Improvement in behavioral function

Gerhardt et al., 1999

MPP+ - rat model

Fibroblasts, BDNF

Protection of DA cell bodies in SN

Frim et al., 1994

6-OHDA - complete

BHK, GDNF

TH-IR fiber ingrowth to BHK-GDNF capsules No improvement in behavioral function

Lindner et al., 1995

6-OHDA - partial SN

Astrocytes, BDNF

No effect on TH-IR fibers in ipsilateral striatum Improvement in behavioral function

Yoshimoto et al., 1995

6-OHDA - partial striatal

Fibroblasts, BDNF

Prevented loss of nerve terminals Protection of DA cell bodies in SN

Levivier et al., 1995

6-OHDA - complete

Fibroblasts, BDNF

! in nigral DA / metabolite levels Did not induce fiber sprouting No improvement in behavioral function

Lucidi-Phillipi et al., 1995

Ex vivo gene delivery

Gene Therapy and Molecular Biology Vol 4, page 65 MPP+ - rat model

Fibroblasts, BDNF

! in nigral DA levels

Galpern et al., 1996

6-OHDA - complete

BHK, GDNF

Protection of DA cell bodies in SN Improvement in behavioral function Did not prevent loss of striatal DA

Tseng et al., 1997

6-OHDA - complete

BHK, NTN

Protection of DA cell bodies in SN No improvement in behavioral function

Tseng et al., 1998

6-OHDA - complete

BHK, GDNF

Both GDNF and NTN protected DA cell bodies in SN GDNF induced TH-IR, sprouting or hypertrophy of DA neurons

Akerud et al., 1999

BHK, NTN

In vivo gene delivery 6-OHDA - partial striatal

Intranigral, Ad GDNF

Protection of DA cell bodies in SN

Choi-Lundberg et al., 1997

6-OHDA - partial striatal

Intrastriatal, Ad GDNF

Protection of DA cell bodies in SN Protected innervation of striatum Improvement in behavioral function

Bilang-Bleuel et al., 1997

6-OHDA - complete

Intranigral, Ad GDNF

Restored nigral DA levels Improvement in behavioral function

Lapchak et al., 1997

MPTP - mouse model

Intrastriatal, Ad GDNF

! in striatal DA levels

Kojima et al., 1997

6-OHDA - partial striatal

Intranigral, AAV GDNF

Protection of DA cell bodies in SN

Mandel et al., 1997

6-OHDA - partial striatal

Intrastriatal, Ad GDNF

Protection of DA cell bodies in SN Improvement in behavioral function

Choi-Lundberg et al., 1998

1997; Choi-Lundberg et al., 1998; Rosenblad et al., 1998; Shults et al., 1996). In order to fully examine the differential effect of injecting GDNF either near DA cell bodies in the SN or at dopaminergic terminals in the striatum, we compared the behavioral and cellular effects of GDNF gene delivery into striatal and mesencephalic sites in aged (20 month) Fischer-344 rats with progressive lesions of the nigrostriatal pathway (Connor et al., 1999). In this study, a subpopulation of DA neurons was prelabelled by bilateral intrastriatal injection of the retrograde tracer fluorogold (FG) so that the fate of those DA neurons that projected specifically to the lesion site could be assessed without having to rely on DA phenotypic markers. Following FG injection, one group of rats was injected unilaterally in the striatum, while a second group of rats was injected unilaterally into the SN with Ad vectors encoding either GDNF, a mutant form of GDNF lacking bioactivity (muGDNF) or LacZ. An additional group of rats underwent surgery, but received no vector injection into either the striatum or the SN. One week later, rats received a unilateral intrastriatal injection of 6-OHDA on the same side as the vector injection and at the same coordinates as the FG injection. Thirty-five days after lesioning, we observed that injections of Ad GDNF into either the striatum or the SN provided significant cell protection against 6-OHDA (Connor et al., 1999). As shown in Figure 2, Ad GDNF injected in the SN protected an average of about 55% of

IV. The differential effects of GDNF gene delivery in the striatum and SN of the aged Parkinsonian rat In order to extend our initial observations in the young rat, we recently examined whether Ad GDNF was able to restore compensatory events, protect DA neurons and maintain DA function in the aged rat brain following a partial lesion of the nigrostriatal pathway. It has also become apparent that the site of GDNF administration is important in determining the degree of functional recovery observed in the damaged or degenerating nigrostriatal DA system. Previous studies have demonstrated that, while repeated injections of GDNF near cell bodies in the SN following a partial 6-OHDA lesion of the nigrostriatal pathway protects or restores DA neuronal function, intranigral injection of GDNF does not prevent impairment of behavioral function in the unilaterally lesioned Parkinsonian animal (Sauer et al., 1995; Winkler et al., 1996). This suggests that long-lasting functional recovery in the intrastriatal 6-OHDA lesion model may require reinnervation and restoration of DA neurotransmission in the denervated striatum. Supporting this proposal, several studies examining the effect of intrastriatal injections of GDNF have reported that, in addition to preventing the loss of dopaminergic neurons in the lesioned SN, GDNF injected into the striatum also induces partial axonal regeneration or reinnervation of the denervated striatum and prevents behavioral impairment caused by 6-OHDAinduced depletion of striatal DA (Bilang-Bleuel et al., 65

Connor and Bohn: GDNF gene therapy for Parkinson's disease the FG labeled DA neurons while Ad GDNF injected in the striatum protected an average of about 65% of FG labeled neurons. In contrast, only an average of about 30% of the neurons remained in the control groups (Connor et al., 1999). Overall, injections of Ad GDNF

into either the striatum or the SN were similarly effective in protecting DA neurons in the lesioned SN of the aged rat brain. This indicates that the site of GDNF injection (striatum versus SN) does not affect the degree of neuronal protection seen in the lesioned SN.

Figure 2A: Injection of Ad GDNF in either the striatum or the SN prevents the degeneration of FG-positive DA neurons in the SN five weeks after 6-OHDA. - Five weeks following intrastriatal injection of 6-OHDA many large FG-positive cells (i.e.: DA neurons arrows) remain in the SN on the unlesioned side (A & B) and on the lesioned side in rats injected with Ad GDNF in either the striatum (E & F) or the SN (G & H). In contrast, fewer large FG-positive neurons, but many small secondarily labeled FG-positive cells (microglia and other non-neuronal cells - *) were observed in the lesioned SN in control rats (C & D). Scale bars: A, C, E & G = 100Âľm; B, D, F & H = 50Âľm. Reprinted with permission from Connor et al, 1999. Copyright 1999 Stockton Press.

Gene Therapy and Molecular Biology Vol 4, page 67

Figure 2B: (Cont.) - A significant increase in the percentage of FGpositive DA neurons was observed in the lesioned SN in rats injected with Ad GDNF in either the striatum or the SN compared to control groups (* p # 0.01). There was no significant difference in the percentage of FG-positive neurons between rats injected with Ad GDNF in the striatum versus the SN (p > 0.05). Reprinted with permission from Connor et al, 1999. Copyright 1999 Stockton Press.

Figure 3: Striatal injections of Ad GDNF prevents the development of amphetamine-induced ipsilateral rotational behavior observed in control rats after 6-OHDA lesioning. Rats were injected with dl-amphetamine (5mg/kg i.p), placed in a rotation chamber and their behavior recorded for 60 minutes. Baseline amphetamine rotation tests were performed 7 days before lesioning (-7) and the results used to assign the side of 6-OHDA lesioning. Five weeks after lesioning, a significant reduction in ipsilateral rotational behavior was observed in rats injected with Ad GDNF in the striatum compared to control groups and rats injected with Ad GDNF in the SN (* p # 0.05). In contrast, rats injected with Ad GDNF into the SN exhibited a significant increase in ipsilateral rotational behavior 35 days after lesioning compared to control groups (** p # 0.05). Reprinted with permission from Connor et al, 1999. Copyright 1999 Stockton Press.

The differential effects of the vector placed into these two sites became evident however, when we examined DA dependent behaviors and cellular changes in DA target neurons. Only striatal injections of Ad GDNF protected against the development of behavioral impairment characteristic of unilateral DA-dependent deficits (Connor

et al., 1999). Thirty-five days after lesioning, rats injected with Ad GDNF in the striatum exhibited a significant reduction in amphetamine-induced rotational asymmetry compared to control groups and rats injected with Ad GDNF in the SN (Figure 3, Connor et al., 1999). In addition, rats injected with Ad GDNF in striatum exhibited a significant

Connor and Bohn: GDNF gene therapy for Parkinson's disease reduction in the preferential use of the ipsilateral forelimb 21 days and 28 days after 6-OHDA lesioning compared to control groups and rats injected with Ad GDNF in the SN (Connor et al., 1999). Maintenance of DA-dependent behavioral function requires the presence of functional DA terminals and/or preservation of striatal DA levels. Our observations suggest that when GDNF biosynthesis is increased near the DA terminals prior to 6-OHDA delivery, the terminals are protected against degeneration and striatal DA levels persist. In contrast, when GDNF biosynthesis is increased near DA cell bodies in the aged rat brain, levels are sufficient to inhibit cell death, but insufficient to increase striatal DA levels or to stimulate terminal sprouting. This apparently results in the development of behaviors indicative of unilateral DA deficiency. These behavioral observations are further supported by morphological data on the response of striatal target neurons and DA fiber density. In the unlesioned striatum, injection of indirect DA agonists, such as amphetamine results in the expression of Fos protein (Figure 1A; Graybiel et al., 1990; Robertson et al., 1989). Amphetamine is thought to induce Fos in the striatum primarily by its ability to release endogenous DA which, via activation of DA D1 receptors, triggers a transduction cascade culminating in the expression of the c-fos gene in striato-nigral cells (Berretta et al., 1992; Konradi et al., 1994; Paul et al., 1992; Robertson et al., 1992). Therefore in the striatum, amphetamine-induced Fos expression requires the presence of functional DA terminals and is used as a marker of DA-mediated postsynaptic responses (Cenci et al., 1992; Labandeira-Garcia et al., 1996; Morgan and Curran, 1991). We observed that striatal, but not SN, injections of Ad GDNF increased amphetamineinduced Fos expression in the lesioned striatum above that in the unlesioned contralateral striatum, indicating protection of DA terminals (Connor et al., 1999). As shown in Figure 4, a reduction in the percentage of Fosimmunopositive neurons, reflecting a decrease in functional DA terminals was seen in both control groups and rats injected with Ad GDNF in the SN at the lesion site and 700Âľm posterior to the lesion site (Connor et al., 1999). In contrast, rats injected with Ad GDNF in the striatum exhibited a significant increase in the percentage of Fos-immunopositive neurons at both the lesion site and 700Âľm posterior to the lesion site compared to control groups and rats injected with Ad GDNF in the SN (Figure 4; Connor et al., 1999). This suggests that increased levels of GDNF near the terminals of DA neurons increased DA levels available to target neurons. This increase could result from either an increase in DA release per terminal or an increase in the number of DA terminals. In support of this latter possibility, we observed that striatal, but not SN, injections of Ad GDNF reduced tyrosine hydroxylase fiber (TH) loss in the lesioned striatum. The area of THimmunoreactive fiber denervation (lesion size) was significantly decreased in rats injected with Ad GDNF in the striatum compared to control groups (Connor et al., 1999), suggesting that Ad GDNF injection in the striatum

protects TH-immunopositive fibers or stimulates neuronal sprouting of fibers in the denervated striatum. In a partial 6-OHDA lesion model, GDNF could exert its neurotrophic effects both at the level of DA cell bodies in the SN or at the level of DA axon terminals in the striatum. Injection of Ad GDNF into the striatum will produce GDNF expression at both the terminals and the somata of the nigrostriatal pathway via retrograde transport (Choi-Lundberg et al., 1998). In contrast, Ad GDNF injected into the SN will produce GDNF mainly near DA somata. Striatal injections of Ad GDNF prior to lesioning may have prevented the initial loss of DA terminals in the striatum, or, though less likely, may have interfered with the up-take of 6-OHDA so that cells were not susceptible to degeneration. Our results suggest that increased levels of striatal GDNF biosynthesis prevents DA neuronal loss and protects DA terminals from 6-OHDAinduced damage, thereby maintaining nigrostriatal function in the aged rat brain (Connor et al., 1999; Figure 5).

V. Conclusion GDNF gene therapy has a potential use in the clinical treatment of Parkinson's disease. As Parkinson's disease is a progressive disorder, we feel that both protective and rescue paradigms are important approaches to this disease. The protective GDNF paradigm aims at preventing further loss of DA neurons and function, while the rescue paradigm aims at reversing damage to DA neurons. To date, our results indicate that, with regard to preventing DA neuronal loss and protecting DA terminals from degeneration, the striatum is a more desirable site for GDNF therapy than the SN. This finding is relevant for the application of gene therapy to patients with Parkinson's disease. However, further technological advances are required to realize the potential of gene therapy for Parkinson's disease. Currently, there is no clinically safe vector available that provides long-term, stable gene expression in the CNS in absence of cytotoxic effects. Therefore, the development of new generation vectors will hopefully lead to stable gene expression and a reduction in host cellular and humoral responses.

References Akaneya, Y., Takahashi, M., and Hatanaka, H. (1995) Interleukin-1 beta enhances survival and interleukin-6 protects against MPP+ neurotoxicity in cultures of fetal rat dopaminergic neurons. Exp Neurology 136: 44-52. Akerud, P., Alberch, J., Eketjall, S., Wagner, J., and Arenas, E. (1999) Differential effects of glial cell line-derived neurotrophic factor and neurturin on developing and adult substantia nigra dopaminergic neurons. J Neurochem. 73: 70-78. Alexi, T., and Hefti, F. (1993) TGF-alpha selectively increases dopaminergic cell survival in ventral mesencephalic cultures. Neurosci 55: 903-918. Altar, C. A., Boylan, C. B., Fritsche, M., Jones, B. E., Jackson, C., Wiegand, S. J., Lindsay, R. M., and Hyman, C. (1994) Efficacy of brain-derived neurotrophic factor and neurotrophin-3 on neurochemical and behavioral deficits associated with partial nigrostriatal dopamine lesions. J Neurochem. 63: 1021-1032.

Gene Therapy and Molecular Biology Vol 4, page 69

Figure 4A: The level of amphetamine-induced Fos expression is significantly increased in rats injected with Ad GDNF in the striatum. -Rats were injected with DL-amphetamine (5mg/kg i.p) 2 hours prior to sacrifice. In the unlesioned striatum, amphetamineinduced Fos expression is observed in striato-nigral cells (A). In contrast, a reduction in the level of Fos-immunoreactivity is observed in the lesioned striatum in both control rats (B) and rats injected with Ad GDNF in the SN (D). In rats injected with Ad GDNF in the striatum (C), the level of Fos-immunoreactivity in the lesioned striatum is greatly increased compared to the unlesioned hemisphere. Scale bar = 100Âľm. Reprinted with permission from Connor et al, 1999. Copyright 1999 Stockton Press.

Figure 4B - Rats injected with Ad GDNF in the striatum exhibit a significant increase in the percentage of Fos-immunoreactive cells at both the lesion site and 700Âľm posterior to the lesion site compared to rats injected with Ad GDNF in the SN and control groups (* p # 0.05). Reprinted with permission from Connor et al, 1999. Copyright 1999 Stockton Press.

Connor and Bohn: GDNF gene therapy for Parkinson's disease

Figure 5: The differential effects of Ad GDNF in the striatum and the SN of the aged Parkinsonian rat. The site of GDNF administration is important in determining the degree of functional recovery observed in the damaged or degenerating nigrostriatal DA system. We compared the behavioral and cellular effects of GDNF gene delivery into striatal and mesencephalic sites in aged (20 month) rats with progressive lesions of the nigrostriatal pathway. Ad GDNF was injected either at dopaminergic terminals in the striatum (A) or near dopaminergic cell bodies in the SN (B). One week following gene delivery, the neurotoxin 6-OHDA was injected unilaterally into the striatum to cause progressive degeneration of DA neurons. Five weeks after lesioning, we observed that injections of Ad GDNF into either the striatum (A) or the SN (B) significantly protected DA cell bodies in the lesioned SN. However, only striatal injections of Ad GDNF (A) protected against the development of behavioral deficits characteristic of unilateral DA depletion. Furthermore, we observed that striatal (A), but not SN (B), injections of Ad GDNF prevented the loss of DA terminals as indicated by a reduction in tyrosine hydroxylase fiber loss and an increase in amphetamine-induced striatal Fos expression. This indicates that increased levels of striatal (A), but not nigral (B), GDNF biosynthesis prevents DA neuronal loss and protects DA terminals from 6-OHDA-induced damage, thereby maintaining DA function in the aged rat. Dead axons and terminals are denoted by dashed lines. Down-regulation of tyrosine hydroxylase and loss of DA neurons is indicated by smaller cell bodies and hatched areas. Baloh, R. H., Tansey, M. G., Lampe, P. A., Fahrner, T. J., Enomoto, H., Simburger, K. S., Leitner, M. L., Araki, T., Johnson, E. M., and Milbrandt, J. (1998) Artemin, a novel member of the GDNF ligand family, supports peripheral and central neurons and signals through the GFRalpha3-RET receptor complex. Neuron 21: 1291-1302.

insulin-like growth factor-1, and basic fibroblast growth factor on mesencephalic dopaminergic neurons developing in culture. Neurosci 52: 855-866. Beck, K. D., Valverde, J., Alexi, T., Poulsen, K., Moffat, B., Vandlen, R. A., Rosenthal, A., and Hefti, F. (1995) Mesencephalic dopaminergic neurons protected by GDNF from axotomy-induced degeneration in the adult brain. Nature 373: 339-341.

Bankiewicz, K., Mandel, R. J., and Sofroniew, M. V. (1993) Trophism, transplantation and animal models of Parkinson's disease. Exp Neurology 124: 140-149.

Berretta, S., Robertson, H. A., and Graybiel, A. M. (1992) Dopamine and glutamine agonists stimulate neuron-specific

Beck, K. D., Knusel, B., and Hefti, F. (1993) The nature of the trophic action of brain-derived neurotrophic factor, des (1-3)-

Gene Therapy and Molecular Biology Vol 4, page 71 expression of Fos-like protein in the striatum. J Neurophys 68: 767-777.

striatum and substantia nigra of the aged parkinsonian rat. Gene Ther 6, 1936-1951.

Bilang-Bleuel, A., Revah, F., Colin, P., Locquet, I., Robert, J.-J., Mallet, J., and Horellou, P. (1997) Intrastriatal injection of an adenoviral vector expressing glial cell-line derived neurotrophic factor prevents dopaminergic neuron degeneration and behavioral impairment in a rat model of Parkinson's disease. Proc Natl Acad Sci USA 94: 88188823.

Demarest, K. T., Riegle, G. D., and Moore, K. E. (1980) Characteristics of dopaminergic neurons in the aged male rat. Neuroendo 31: 222-227. Engele, J., and Bohn, M. C. (1991) The neurotrophic effects of fibroblast growth factors on dopaminergic neurons in vitro are mediated by mesencephalic glia. J Neurosci 11: 30703078.

Bjorklund, A., Rosenblad, C., Winkler, C., and Kirik, D. (1997) Studies on neuroprotective and regenerative effects of GDNF in a partial lesion model of Parkinson's disease. Neurobiol Disease 4: 186-200.

Engele, J., Schubert, D., and Bohn, M. C. (1991) Conditioned media derived from glial cell lines promote survival and differentiation of dopaminergic neurons in vitro: Role of mesencephalic glia. J Neurosci Res 30: 359-371.

Bohn, M. C., and Choi-Lundberg, D. (1997). Gene Therapies for Parkinson's disease. In: Gene Therapy for Neurological Disorders and Brain Tumors, E. A. Chiocca and X. O. Breakfield, Eds. (Totowa, NJ: Humana Press Inc), pp. 377395.

Felten, D., Felten, S., Steece-Collier, K., Date, I., and Clemens, J. (1992) Age-related decline in the dopaminergic nigrostriatal system: the oxidative hypothesis and protective strategies. Ann Neurol 32: S133-S136. Ferrari, G., Minozzi, M.-C., Toffano, G., Leon, A., and Skaper, S. B. (1989) Basic fibroblast growth factor promotes the survival and development of mesencephalic neurons in culture. Dev Biol 133: 140-147.

Bowenkamp, K. E., Hoffman, A. F., Gerhardt, G. A., Henry, M. A., Biddle, P. T., Hoffer, B. J., and Granholm, A., E, (1995) Glial cell line-derived neurotrophic factor supports survival of injured midbrain dopaminergic neurons. J Comp Neurol 355: 479-489.

Frim, D. M., Uhler, T. A., Galpern, W. R., Beal, M. F., Breakefield, X. O., and Isacson, O. (1994) Implanted fibroblasts genetically engineered to produce brain-derived neurotrophic factor prevent 1-methyl-4-phenylpyridinium toxicity to dopaminergic neurons in the rat. Proc Natl Acad Sci USA 91: 5104-5108.

Bowenkamp, K. E., Lapchak, P. A., Hoffer, B. J., Miller, P. J., and Bickford, P. C. (1997) Intracerebroventricular glial cell line-derived neurotrophic factor improves motor function and supports nigrostriatal dopamine neurons in bilaterally 6hydroxydopamine lesioned rats. Exp Neurology 145: 104117.

Galpern, W. R., Frim, D. M., Tatter, S. B., Altar, C. A., Beal, M. F., and Isacson, O. (1996) Cell-mediated delivery of brainderived neurotrophic factor enhances dopamine levels in an MPP+ rat model of substantia nigra degeneration. Cell Transplant 5: 225-232.

Casper, D., Mytilineou, C., and Blum, M. (1991) EGF enhances the survival of dopaminergic neurons in rat embryonic mesencephalon pimary cell culture. J. Neurosci Res 30: 372381.

Gash, D. M., Zhang, Z., and Gerhardt, G. (1998) Neuroprotective and neurorestorative properties of GDNF. Ann Neurol 44: S121-S125.

Cenci, M., Kalen, P., Mandel, R., Wictorin, K., and Bjorklund, A. (1992) Dopaminergic transplants normalize amphetamine and apomorphine-induced Fos expression in the 6hydroxydopamine lesioned striatum. Neurosci 46:,943-957.

Gash, D. M., Zhang, Z. M., Ovadia, A., Cass, W. A., Yi, A., Simmerman, L., Russell, D., Martin, D., Lapchak, P. A., Collins, F., Hoffer, B. J., and Gerhardt, G. A. (1996) Functional recovery in Parkinsonian monkeys treated with GDNF. Nature 380: 252-255.

Cheng, F.-C., Ni, D.-R., Wu, M.-C., Kuo, J.-S., and Chia, L.-G. (1998) Glial cell line-derived neurotrophic factor protects against 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-induced neurotoxicity in C57BL/6 mice. Neurosci Lett 252: 87-90.

Gerhardt, G., Cass, W. A., Huettl, P., Brock, S., Zhang, Z., and Gash, D. M. (1999) GDNF improves dopamine function in the substantia nigra but not the putamen of unilateral MPTPlesioned rhesus monkeys. Brain Res 817: 163-171.

Choi-Lundberg, D. L., Lin, Q., Chang, Y.-N., Chiang, Y. L., Hay, C. M., Mohajeri, H., Davidson, B. L., and Bohn, M. C. (1997) Dopaminergic neurons protected from degeneration by GDNF gene therapy. Science 275: 338-341.

Graybiel, A. M., Moratalla, R., and Robertson, H. A. (1990) Amphetamine and cocaine induces drug-specific activation of the c-fos gene in striosome-matrix compartments and limbic subdivisions of the striatum. Proc Natl Acad Sci USA 87: 6912-6916.

Choi-Lundberg, D. L., Lin, Q., Schallert, T., Crippens, D., Davidson, B. L., Chang, Y.-N., Chiang, Y. L., Qian, J., Bardwaj, L., and Bohn, M. C. (1998) Behavioral and cellular protection of rat dopaminergic neurons by an adenoviral vector encoding glial cell line-derived neurotrophic factor. Exp Neurology 154: 261-275.

Hagg, T., and Varon, S. (1993) Ciliary neurotrophic factor prevents degeneration of adult rat substantia nigra dopaminergic neurons in vivo. Proc Natl Acad Sci USA 90: 6315-6319.

Connor, B., Kozlowski, D. A., Schallert, T., Tillerson, J. L., Davidson, B. L., and Bohn, M. C. (1999) Differential effects of glial cell line-derived neurotrophic factor (GDNF) in the

Hama, T., Kushima, Y., Miyamoto, M., Kubota, M., Takei, M., and Hatanaka, H. (1991) Interleukin-6 improves the survival

Connor and Bohn: GDNF gene therapy for Parkinson's disease of mesencephalic catecholamines and septal cholinergic neurons from post-natal two-week old rats in cultures. Neurosci 40:,445-452.

Neurturin, a relative of glial cell-line derived neurotrophic factor. Nature 384: 467-470. Kozlowski, D.A., Connor, B., Tillerson, J.L., Schallert, T., and Bohn, M.C. (Submitted) Delivery of a GDNF gene into the substantia nigraafter a progressive 6-OHDA lesion maintains functional nigrostriatal connections.

Hoffer, B. J., Hoffman, A., Bowenkamp, K., Huettl, P., Hudson, J., Martin, D., Lin, L.-F. H., and Gerhardt, G. A. (1994) Glial cell line-derived neurotrophic factor reverses toxin-induced injury to midbrain dopaminergic neurons in vivo. Neurosci Lett 182: 107-111.

Krieglstein, K., Suter-Crazzolara, C., Fischer, W. H., and Unsicker, K. (1995) TGF-beta superfamily members promote survival of midbrain dopaminergic neurons and protect them against MPP+ toxicity. EMBO J 14: 736-742.

Hoffman, A. F., van Horne, C. G., Eken, S., Hoffer, B. J., and Gerhardt, G. A. (1997) In vivo microdialysis studies of somatodendritic dopamine release in the rat substantia nigra: effects of unilateral 6-OHDA lesions and GDNF. Exp Neurology 147: 130-141.

Krieglstein, K., Suter-Crazzolara, C., Hotten, G., Pohl, J., and Unsicker, K. (1995) Trophic and protective effects on growth/differentiation factor 5, a member of the transforming growth factor family-beta superfamily, on midbrain dopaminergic neurons. J Neurosci Res 42: 724-732.

Horger, B. A., Nishimura, M. C., Armanini, M. P., Wang, L.-C., Poulsen, K. T., Rosenblad, C., Kirik, D., Moffat, B., Simmons, L., Johnson, E., Milbrandt, J., Rosenthal, A., Bjorklund, A., Vandlen, R. A., Hynes, M. A., and Phillips, H. S. (1998) Neurturin exerts potent actions on survival and function of midbrain dopaminergic neurons. J Neurosci 18: 4929-4937.

Labandeira-Garcia, J., Rozas, G., Lopez-Martin, E., Liste, I., and Guerra, M. (1996) Time course of striatal changes induced by 6-hydroxydopamine lesion of the nigrostriatal pathway, as studied by combined evaluation of rotational behavior and striatal Fos expression. Exp Brain Res 108: 69-84.

Hubble, J. P. (1998) Aging and the basal ganglia. Neurol Clinical 16: 649-657.

Lapchak, P., Araujo, D., Hilt, D., Sheng, J., and Jiao, S. (1997) Adenoviral vector-mediated GDNF gene therapy in a rodent lesion model of late stage Parkinson's disease. Brain Res 777: 153-160.

Hyman, C., Hofer, M., Barde, Y.-A., Juhasz, M., Yancopoulos, G. D., Squinto, S. P., and Lindsay, R. M. (1991) BDNF is a neurotrophic factor for dopaminergic neurons of the substantia nigra. Nature 350: 230-232.

Lapchak, P. A., Miller, P. J., Collins, F., and Jiao, S. (1997) Glial cell line-derived neurotrophic factor attenuates behavioral deficits and regulates nigrostriatal dopaminergic and peptidergic markers in 6-hydroxydopamine-lesioned adult rats: comparison of intraventricular and intranigral delivery. Neurosci 78: 61-72.

Hyman, C., Juhasz, M., Jackson, C., Wright, P., Ip, N. Y., and Lindsay, R. M. (1994) Overlapping and distinct actions of the neurotrophins BDNF, NT-3 and NT-4/5 on cultured dopaminergic and GABAergic neurons of the ventral mesencephalon. J Neurosci 14: 335-347.

Levivier, M., Przedborski, S., Bencsics, C., and Kang, U. J. (1995) Intrastriatal implantation of fibroblasts genetically engineered to produce brain-derived neurotrophic factor prevents degeneration of dopaminergic neurons in a rat model of Parkinson's disease. J Neurosci 15: 7810-7820.

Kearns, C., and Gash, D. (1995) GDNF protects nigral dopaminergic neurons against 6-hydroxydopamine in vivo. Brain Res 672: 104-111. Kikuchi, S., Muramatsu, H., Muramatsu, T., and Kim, S. U. (1993) Midkine, a novel neurotrophic factor, promotes survival of mesencephalic neurons in culture. Neurosci Lett 160: 9-12.

Lin, L.-F. H., Doherty, D. H., Lile, J. D., Bektesh, S., and Collins, F. (1993) GDNF: a glial cell line-derived neurotrophic factor for midbrain dopaminergic neurons. Science 260: 1130-1132.

Knusel, B., Michel, P. P., Schwaber, J. S., and Hefti, F. (1990) Selective and nonselective stimulation of central cholinergic and dopaminergic development in vitro by nerve growth factor, basic fibroblast growth factor, epidermal growth factor, insulin and the insulin-like growth factors I and II. J Neurosci 10: 558-570.

Lin, L.-F. H., Zhang, T. J., Collins, F., and Armes, L. G. (1994) Purification and initial characterization of rat B49 glial cell line-derived neurotrophic factor. J Neurochem. 63: 758-768. Lindner, M. D., Winn, S. R., Baetge, E. E., Hammang, J. P., Gentile, F. T., Doherty, E., McDermott, P. E., Frydel, B., Ullman, M. D., Schallert, T., and Emmerich, D. F. (1995) Implantation of encapsulated catecholamine and GDNFproducing cells in rats with unilateral dopamine deletions and Parkinsonian symptoms. Exp Neurology 132: 62-76.

Kojima, H., Abiru, Y., Sakajiri, K., Watabe, K., Ohishi, N., Takamori, M., Hatanaka, H., and Yagi, K. (1997) Adenovirus-mediated transduction with human glial cell linederived neurotrophic factor gene prevents 1-methyl-4-phenyl1,2,3,6-tetrahydropyridine-induced dopamine depletion in striatum of mouse brain. Biochem Biophys Res Comm 238: 569-573.

Liu, J. P., and Lauder, J. M. (1992) S-100beta and insulin-like growth factor-II differentially regulate growth of developing serotonin and dopamine neurons in vitro. J Neurosci Res 33: 248-256.

Konradi, C., Cole, R. L., Heckers, S., and Hyman, S. E. (1994) Amphetamine regulates gene expression in rat striatum via transcription factor CREB. J Neurosci 14: 5623-5634.

Lu, X., and Hagg, T. (1997) Glial cell line-derived neurotrophic factor prevents death, but not reductions in tyrosine hydroxylase, of injured nigrostriatal neurons in adult rats. J Comp Neurol 388: 484-494.

Kotzbauer, P. T., Lampe, P. A., Heuckeroth, R. O., Golden, J. P., Creedon, D. J., Johnson, E. M., and Milbrandt, J. (1996)

Gene Therapy and Molecular Biology Vol 4, page 73 Lucidi-Phillipi, C. A., Gage, F. H., Shults, C. W., Jones, K. R., Reichardt, L. F., and Kang, U. J. (1995) Brain-derived neurotrophic factor-transduced fibroblasts: production of BDNF and effects of grafting to the adult rat brain. J Comp Neurol 354: 361-376.

factor receptor/gp130 signaling complex. J Biol Chem 270: 10915-10922. Poulsen, K. T., Armanini, M. P., Klein, R. D., Hynes, M. A., Phillips, H. S., and Rosenthal, A. (1994) TGF beta 2 and TGF beta 3 are potent survival factors for midbrain dopaminergic neurons. Neuron 13: 1245-1252.

Magel, E., Burnham, P., Varon, S., and Louis, J.-C. (1993) Convergent regulation by ciliary neurotrophic factor and dopamine of tyrosine hydroxylase expression in cultures of rat substantia nigra. Neurosci 52: 867-881.

Robertson, G. S., Vincent, S. R., and Fibiger, H. C. (1992) D1 and D2 dopamine receptors differentially regulate c-fos expression in striatonigral and striatopallidal neurons. Neurosci 49: 285-296.

Mandel, R., Spratt, S., Snyder, R., and Leff, S. (1997) Midbrain injection of recombinant adeno-associated virus encoding rat glial cell line-derived neurotrophic factor protects nigral neurons in a progressive 6-hydroxydopamine-induced degeneration model of Parkinson's disease in rats. Proc Natl Acad Sci USA 94: 14083-14088.

Robertson, H. A., Peterson, M. R., Murphy, K., and Robertson, G. S. (1989) D1-dopamine receptor agonists selectively activate striatal c-fos independent of rotational behavior. Brain Res 503: 346-349. Rosenblad, C., Kirik, D., Devaux, B., Moffat, B., Phillips, H. S., and Bjorklund, A. (1999) Protection and regeneration of nigral dopaminergic neurons by neurturin or GDNF in a partial lesion model of Parkinson's disease after administration into the striatum or the lateral ventricle. Eur J Neurosci 11: 1554-1566.

Milbrandt, J., de Sauvage, F. J., Fahrner, T. J., Baloh, R. H., Leitner, M. L., Tansey, M. G., Lampe, P. A., Hauckeroth, R. O., Kotzbauer, P. T., Simburger, K. S., Golden, J. P., Davies, J. A., Vejsada, R., Kato, A. C., Hynes, M., Sherman, D., Nishimura, M., Wamg, L.-C., Vandlen, R., Moffat, B., Klein, R. D., Poulsen, K., Gray, C., Garces, A., Henderson, C. E., Phillips, H. S., and Johnson, E. M. (1998) Persephin, a novel neurotrophic factor related to GDNF and neurturin. Neuron 20: 245-253.

Rosenblad, C., Martinez-Serrano, A., and Bjorklund, A. (1998) Intrastriatal glial cell line-derived neurotrophic factor promotes sprouting of spared nigrostriatal dopaminergic afferents and induces recovery of function in a rat model of Parkinson's disease. Neurosci 82: 129-137.

Morgan, D., and Finch, C. (1988) Dopaminergic changes in the basal ganglia: a generalized phenomenon of aging in mammals. Ann NY Acad Sci 515: 145-160.

Roth, G. S., and Joseph, J. A. (1994) Cellular and molecular mechanisms of impaired dopaminergic function during aging. Ann NY Acad Sci 719: 129-135.

Morgan, J. I., and Curran, T. (1991) Stimulus-transcription coupling in the nervous system: involvement of the inducible proto-oncogenes fos and jun. Ann Rev Neurosci 14: 421451.

Rousselet, A., Felter, L., Chamak, B., and Prochiantz, A. (1988) Rat mesencephalic neurons in culture exhibit different morphological traits in the presence of media conditioned on mesencephalic or striatal astroglia. Dev Biol 129: 495-504.

Olsson, M., Nikkhah, G., Bentlage, C., and Bjorklund, A. (1995) Forelimb akinesia in the rat Parkinson model: differential effects of dopamine agonists and nigral transplants as assessed by a new stepping test. J Neurosci 15: 3863-3875.

Sauer, H., and Oertel, W. H. (1994) Progressive degeneration of nigrostriatal dopamine neurons following intrastriatal terminal lesions with 6-hydroxydopamine a combined retrograde tracing and immunocytochemical study in the rat. Neurosci 59: 401-415.

Opacka-Juffry, J., Ashworth, S., Hume, S. P., Martin, D., Brooks, D. J., and Blunt, S. B. (1995) GDNF protects against 6OHDA nigrostriatal lesion: in vivo study with microdialysis and PET. NeuroReport 7: 348-352.

Sauer, H., Rosenblad, C., and Bjorklund, A. (1995) Glial cell line-derived neurotrophic factor but not transforming growth factor beta-3 prevents delayed degeneration of nigral dopaminergic neurons following striatal 6-hydroxydopamine lesion. Proc Natl Acad Sci USA 92: 8935-8939.

Othberg, A., Odin, P., Ballagi, A., Ahgren, A., Funa, K., and Lindvall, O. (1995) Specific effects of platelet derived growth factor (PDGF) on fetal rat and human dopaminergic neurons in vitro. Exp Brain Res 105: 111-122.

Schallert, T. (1988). Aging-dependent emergence of sensorimotor dysfunction in rats recovered from dopamine sustained early in life. In: Central determinants of age-related decline in motor function, J. A. Joseph, Ed.: Ann NY Acad Sci pp. 108-119.

Otto, D., and Unsicker, K. (1990) Basic FGF reverses chemical and morphological deficits in the nigrostriatal system of MPTP-treated mice. J Neurosci 10: 1912-1921. Paul, M. L., Graybiel, A. M., David, J. C., and Robertson, H. A. (1992) D1-like and D2-like dopamine receptors synergistically activate rotation and c-fos expression in the dopamine-depleted striatum in a rat model of Parkinson's disease. J Neurosci 12: 3729-3742.

Schallert, T. (1995) Models of neurological defects and defects in neurological models. Behav Brain Sci 18: 68-69. Schwarting, R., and Huston, J. (1996) The unilateral 6hydroxydopamine lesion model in behavioral brain research analysis of functional deficits, recovery and treatments. Prog Neurobiol 50: 275-331.

Pennica, D., Shaw, K. J., Swanson, T. A., Moore, M. W., Shelton, D. L., Zioncheck, K. A., Rosenthal, A., Taga, T., Paoni, N. F., and Wood, W. I. (1995) Cardiotrophin-1: Biological activities and binding to the leukemia inhibitory

Connor and Bohn: GDNF gene therapy for Parkinson's disease Shults, C. W., Kimber, T., and Martin, D. (1996) Intrastriatal injection of GDNF attenuates the effects of 6hydroxydopamine. NeuroReport 7: 627-631. Sullivan, A. M., Opacka-Juffry, J., and Blunt, S. B. (1998) Longterm protection of the rat nigrostriatal dopaminergic system by glial cell line-derived neurotrophic factor against 6hydroxydopamine in vivo. Eur J Neurosci 10: 57-63. Tomac, A., Lindqvist, E., Lin, L.-F. H., Ogren, S. O., Young, D., Hoffer, B. J., and Olson, L. (1995) Protection and repair of the nigrostriatal dopaminergic system by GDNF in vivo. Nature 373: 335-339. Tseng, J. L., Baetge, E. E., Zurn, A. D., and Aebischer, P. (1997) GDNF reduces drug-induced rotational behavior after medial forebrain bundle transection by a mechanism not involving striatal dopamine. J Neurosci 17: 325-333. Tseng, J. L., Bruhn, S. L., Zurn, A. D., and Aebischer, P. (1998) Neurturin protects dopaminergic neurons following medial forebrain bundle axotomy. NeuroReport 9: 1817-1822. Tsukahara, T., Takeda, M., Shimohama, S., Ohara, O., and Hashimoto, N. (1995) Effects of brain-derived neurotrophic factor on 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridineinduced parkinsonism in monkeys. Neurosurg 37: 733-741. Ungerstedt, U. (1971) Striatal dopamine release after amphetamine or nerve degeneration revealed by rotational behavior. Acta Physiol Scand 82: 49-68. Ungerstedt, U., and Arbuthnott, G. W. (1970) Quantitative recording of rotational behavior in rats after 6hydroxydopamine lesions of the nigrostriatal dopamine system. Brain Res 24: 485-493. Unnerstall, J. R., and Long, M. M. (1996) Differential effects of the intraventricular administration of 6-hydroxydopamine on the induction of type II beta-tubulin and tyrosine hydroxylase mRNA in the locus coeruleus of the aging Fischer 344 rat. J Comp Neurol 364: 363-381. von Coelln, R., Unsicker, K., and Krieglstein, K. (1995) Screening of interleukins for survival-promoting effects on cultured mesencephalic dopaminergic neurons from embryonic rat brain. Dev Brain Res 89: 150-154. Winkler, C., Sauer, H., Lee, C. S., and Bjorklund, A. (1996) Short-term GDNF treatment provides long-term rescue of lesioned nigral dopaminergic neurons in a rat model of Parkinson's disease. J Neurosci 16: 7206-7215. Yoshimoto, Y., Lin, Q., Collier, T. J., Frim, D. M., Breakefield, X. O., and Bohn, M. C. (1995) Astrocytes retrovirally transduced with BDNF elicit behavioral improvement in a rat model of Parkinson's disease. Brain Res 691: 25-36. Zigmond, M., Abercrombie, E., Berger, T., Grace, A., and Stricker, E. (1993). Compensatory responses to partial loss of dopaminergic neurons: studies with 6-hydroxydopamine. In: Current Concepts in Parkinson's Disease Research, J. Schneider and M. Gupta, Eds.: Hogrefe and Huber, pp. 99140.

Gene Therapy and Molecular Biology Vol 4, page 75 Gene Ther Mol Biol Vol 4, 75-82. December 1999.

Structural insights into NF-!B/I!B signaling Review Article

Gourisankar Ghosh, De-Bin Huang, and Tom Huxford Department of Chemistry and Biochemistry, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 920930359 USA. ______________________________________________________________________________________________________ Correspondence: Gourisankar Ghosh, Ph.D., Department of Chemistry and Biochemistry, University of California, San Diego 9500 Gilman Drive, Urey Hall, La Jolla, CA. 92093, USA. Tel: (619)-822-0469; Fax: (619)-534-7042; E-mail: gghosh@ucsd.edu Abbr e v i a t i o ns: R HR , rel homology region; TAD, transcriptional activation domain; IKK, I!B kinase complex; ARD, ankyrin repeat-containing domain ; PEST sequence, amino acids sequence rich in proline, glutamic acid, serine and threonine; SRD, aminoterminal signal receiving domains; NLS, nuclear localization signals. Key Words: NF!B/I!B signaling, apoptosis, mitogen-induced signaling, transcription factors, Received: 7 August 1999; accepted: 15 September 1999

Summary The vital cellular activities o f growth, differentiation, reaction t o s t i m u l i , and apoptosis are controlled by the coordinated expression of a vast number of genes. Regulation of gene expression occurs primarily at the level of transcription. The process begins as one of a multitude of mitogeninduced signaling events triggers the intricate and exquisitely regulated signal transduction pathways. These lead ultimately to the recruitment of transcription factors on specific promoter/enhancer sites. One example of how complex cell signaling can lead to the temporal activation of transcription in a specific set of genes is illustrated by the transcription factor NF-!B . S i n c e i t s d i s c o v e r y a l m o s t a decade and a half ago, NF- !B has fascinated researchers because o f the complexity o f the NF-!B signaling pathway. Several recently determined crystal structures of a number of NF- !B complexes have given a new dimension of understanding of the biochemistry behind NF-!B . W e w i l l r e v i e w these structures in light of their functions. combinations has been observed in cells. Polypeptides of the Rel/NF-!B family can be divided into two subgroups based on the presence or absence of a transcriptional activation domain (TAD). Two family members, p50 and p52, do not contain distinct TADs and are therefore categorized as belonging to class I. Evidence exists suggesting that the homodimers of p50 and p52 and the p50/p52 heterodimer can function as repressors of gene transcription (Franzoso et al, 1992). The other three Rel/NF!B family members, p65, c-Rel, and RelB, constitute the class II subgroup. NF-!B dimers containing one or two of these polypeptides act as activators by virtue of the presence of at least one transactivation domain. The two most abundant and biologically well characterized of the Rel/NF!B dimers are the p50 homodimer and p50/p65 heterodimer. In most resting cells, Rel/NF-!B dimers with transactivation potential are localized in the cytoplasm in complex with inhibitor proteins I!B", I!B#, and I!B$ (Whiteside et al, 1997; Li and Nabel, 1997). These macromolecules belong to the larger I!B family of transcription factor inhibitor proteins which also includes

I. Introduction Members of the Rel/NF-!B transcription factor family are involved in diverse biological activities ranging from the regulation of inflammatory and immune responses to development and apoptosis (Ghosh et al, 1998; Baldwin, 1996; Baeuerle and Henkel, 1994; Baeuerle and Baltimore, 1996). Rel/NF-!B proteins modulate transcription by binding distinct DNA target sites that are collectively known as !B DNA sequences (Baeuerle and Henkel, 1994). In mammals, five homologous polypeptides, p50, p52, p65 (or RelA), c-Rel, and RelB constitute the Rel/NF-!B family (Figure 1a) (Ghosh et al, 1998). Common to each of the Rel/NF-!B polypeptides is an amino-terminal region of approximately 300 amino acids with high sequence homology known as the rel homology region (RHR). This segment is responsible for nuclear translocation and !B DNA binding. Similar to most other site-specific DNA-binding transcription factors, the Rel/NF-!B proteins function as dimers. The Rel/NF-!B dimerization domain is also contained within the RHR. Interestingly, the existence of many, but not all, of the possible homo- and heterodimer 75

Ghosh et al: NF-!B/I!B signaling I!B% , Bcl-3, p105, and p100 (Figure 1b) (Verma et al, 1995). The p105 and p100 proteins contain the RHR of p50 and p52, respectively, in their amino-termini and inhibit NF!B proteins by a different mechanism than do I!B", -#, and -$. The inhibitory functions of I!B% and Bcl-3 are at present not entirely clear. Various physiological and environmental signals promote nuclear translocation of NF-!B proteins by removing inhibitory I!B", I!B#, and I!B$ from NF-!B/I!B complexes (Figure 2a). Each of these three I!B proteins

contains two conserved serines within a homologous segment in their amino-terminal signal receiving domains (SRD). These serines are phosphorylated by the multisubunit I!B kinase complex (IKK) in a signal dependent manner (DiDonato et al, 1997; Lee et al, 1998). Specific lysine residues of I!B located near the phosphorylated serines are then polyubiquinated, marking I!B for rapid degradation by the proteosome (Baeuerle and Henkel, 1994).

F i g u r e 1 . Domain organization of the NF-!B and I!B protein families. a) The Rel/NF-!B transcription factor family polypeptides are organized according to class I and class II depending on the presence or absence of transactivation domains. The Rel homology region is indicated with the aminoterminal domain in red and dimerization domain in green. Other structural elements of interest are labeled. b ) The I!B family of transcription factor inhibitor proteins are aligned according to ankyrin repeat-containing domains. The aminoterminal signal receiving domain and carboxyterminal PEST sequence are indicated. The p105 and p100 polypeptides are special cases which contain I!B%-like carboxy-terminal domains as well as their own amino-terminal rel homology regions.

Gene Therapy and Molecular Biology Vol 4, page 77

F i g u r e 2 . I!B" regulates NF-!B transcriptional activity by dual mechanisms. a) A host of cellular stimuli activate signal transduction cascades which lead to the phosphorylation and subsequent proteolysis of NF-!B-associated I!B". Free NF!B then enters the nucleus where it binds gene enhancers and activates transcription. b ) Shortly after NF-!B induction (1 ), newly synthesized I!B" (2 ) accumulates in the cell cytoplasm (3 ). This free I!B" then enters the nucleus (4 ) where it removes NF-!B from target genes and shuttles it back to the cytosol (5 ).

Ghosh et al: NF-!B/I!B signaling fold, finer analysis of the domain structure shows deviation from the canonical immunoglobulin domain. This domain also represents the region of highest sequence variation among different Rel/NF-!B family members. The differences stem primarily from a non-homologous insertion of varying lengths within the RHR amino-terminal domain. The insert is largest in the p50 and p52 subunits where it forms a two "-helix bundle connected by a large loop. In p65 and c-Rel, on the other hand, the insert is smaller forming only a single small alpha-helix. This helical insert may play a role in transcriptional activation by mediating interactions with other transcription factors or coactivators. The overall structure of the core of the RHR amino-terminal domain shows distant similarity to the DNA binding domains of the tumor suppressor p53 and the STAT family transcription factors. The RHR amino- and carboxy-terminal domains are joined by a 10 amino acid linker. In the DNA bound conformation, these two domains contact each other. However, interdomain interactions are not conserved between RHR polypeptides and depend highly on the target DNA sequences. In the absence of any structural information for free RHR (in the absence of DNA) one can only speculate if these two domains interact. High proteolytic sensitivity of the linker and the variations of interactions between two domains in the DNA bound forms suggest that in the absence of DNA the two domains are flexible about the linker polypeptide.

Removal of I!B proteins from the NF-!B/I!B complex activates nuclear localization signals (NLS) on the NF-!B subunits. This allows for rapid nuclear translocation of the active NF-!B dimer. Interestingly, among the host of genes regulated by NF-!B activation is the gene encoding I!B". As a result of this arrangement, a pool of newly synthesized I!B" begins to accumulate in the cell cytosol shortly after NF-!B induction (Figure 2b) (Baldwin, 1996). In the absence of NF-!B binding partner, this newly synthesized I!B" translocates to the nucleus where it binds to and dissociates pre-formed NF-!B/DNA complexes (Ghosh et al, 1998). Nuclear NF-!B/I!B" complexes are then shuttled back to the cytoplasm, restoring the pre-induction state (Arenzana-Seisdedos et al, 1997). Therefore, I!B" plays an important role in the regulation of NF-!B activity by regulating NF-!B both in the cytoplasm and in the nucleus. High resolution x-ray crystal structures containing the RHR of various Rel/NF-!B dimers in complex with DNA and the structure of the NF-!B p50/p65 heterodimer/I!B" complex have significantly extended our knowledge of the chemistry behind the NF-!B and I!B proteins.

II. Structure of the RHR High resolution x-ray crystal structures have been determined containing the RHR of four of the five NF-!B family members namely, p50 (Ghosh et al, 1995; M端ller et al, 1995; Cramer et al, 1997), p52 (Chen et al, 1998), p65 (Cramer et al, 1997; Chen Y-Q et al, 1998), and c-Rel (Chen Y-Q, unpublished data). As expected from their high degree of sequence similarity, these structures are also highly homologous. The RHR is comprised of two independent structural modules or domains and a flexible 25 amino acid segment at the carboxy-terminus (Figure 3a). This carboxy-terminal flexible segment contains a cluster of basic amino acids responsible for nuclear localization of NF-!B proteins. We refer this segment as the NLS polypeptide. Ordered polypeptide structure is not observed for this NLS polypeptide in the absence of its I!B" protein binding partner (Ghosh et al, 1995; Cramer et al, 1997; Huang et al, 1997). The two structured domains are folded into immunoglobulin-like folds. Like other immunoglobulin domains, the roughly 100 amino acids long carboxy-terminal domain exhibits only beta structure whereby two beta sheets form a globular beta-sandwich. This immunoglobulin-like domain is solely responsible for subunit association and so is also commonly referred to as the dimerization domain. Residues from the carboxy-terminal dimerization domain also participate in non-specific binding to the sugar-phosphate backbone of target DNA. The larger, approximately 180 amino acids long aminoterminal domain determines the sequence specificity of DNA binding by mediating base-specific DNA contacts. Although this domain also bears an apparent immunoglobulin-like

III. Dimerization of NF-!B proteins Rel/NF-!B dimers are formed by the stacking of two symmetrical beta sheets, one from each immunoglobulinlike carboxy-terminal dimerization domain, onto one another (Huang et al, 1997). The residues that participate in intersubunit contacts are highly homologous among the family members. In fact, seven of twelve dimer forming residues are identical while the other five represent conservative substitutions. Immunochemical and other in vivo and in vitro experiments suggest that Rel/NF-!B proteins form dimers in a combinatorial fashion. Interestingly, the Rel/NF-!B polypeptides show a drastic specificity for homo- and heterodimer formation. The most dramatic example is the inability of the RelB polypeptide to self associate or to form heterodimers with c-Rel and p65 subunits (Ryseck et al, 1995). In another case of NF-!B dimer selectivity, the p50/p65 heterodimer forms with greater stability than do either homodimers of p50 or p65. Against the backdrop of such a global similarity the slight energetic differences at the dimer interfaces invoke interesting regulatory interactions that probably dictate differential dimerization affinity. Alanine scanning mutagenesis of p50 dimer interface residues shows that only four residues, Tyr267, Leu269, Asp302, and Val310, contribute

Gene Therapy and Molecular Biology Vol 4, page 79 significantly to the binding energy of p50 dimerization (Sengchanthalangsy LL, unpublished data). All four of these residues are identical in RelB. Results from the mutagenesis experiments, therefore, suggest that identical residues play different roles in the context of different partners.

targets are collectively known as !B DNA which display strong sequence similarity. Most of the sequences are pseudosymmetric. The identifiable feature of!B DNA is the strong conservation of flanking G:C base pairs. The G:C base pair at the 5' end is even more conserved. Crystal structures of NF-!B/DNA complexes have revealed the detailed mechanism of DNA complex formation.

IV. NF-!B DNA binding There are over 100 DNA sequences that NF-!B dimers have been known to recognize (Baldwin, 1996). The DNA

F i g u r e 3 . Ribbon diagrams of NF-!B complex xray crystal structures. a) The Rel homology regions of the NF-!B p50/p65 heterodimer bound to the Ig !B DNA target site. The p50 subunit is represented in green, the p65 subunit is shown in red, the DNA double helix is depicted in cyan and magenta. Each subunit displays the characteristic Rel/NF-!B folds with two immunoglobulin-like domains connected by a short, flexible linker. b ) The x-ray crystal structure of the I!B"/NF-!B p50/p65 heterodimer complex. The color scheme for NF-!B is consistent with part a, I!B" is represented in purple. Note the large change in conformation made by the amino-terminal domain of p65 in the presence of the I!B" inhibitor. The p50 amino-terminal domain is modeled into this figure in its DNA-bound conformation and is not present in the x-ray crystal structure.

Ghosh et al: NF-!B/I!B signaling amino-terminal DNA specificity region of a monomer, in the absence of proper nucleotide sequence in the DNA half site, may move about to find an alternate mode of DNA interaction without sacrificing too much binding energy. Consequently, Rel/NF-!B dimers are capable of binding DNA using multiple conformations. This unusual mechanism of DNA target selection allows the relatively few Rel/NF-!B family dimers to regulate a large number of !B DNA enhancers.

All NF-!B dimers recognize !B DNA targets in an overall similar mode. The most visible feature of these complexes are: (i ) dimers interact with DNA targets through loops connecting secondary structures (beta strands). Five such loops are contributed by each subunit. (i i ) Both domains and the flexible linker polypeptide interact with DNA targets. (i i i ) One turn of the DNA major groove is contacted intimately. (i v ) Symmetrical homodimers recognize both symmetric and asymmetric DNA targets asymmetrically. The recognition loop, which arises from the amino-terminal domain, provides three to four residues for sequence specific DNA recognition. A cluster of basic residues from a second loop of the amino-terminal domain binds the sugar phosphate backbone from the minor groove side of the DNA. The interdomain connector loop contributes one basic residue for sequence specific DNA binding and other residues for backbone interactions. Two loops from the dimerization domain primarily recognize the sugar phosphate backbone of the DNA targets. These structures also show that each monomer recognizes a half site of DNA with sequence specificity. p50 subunit optimally binds to a five base pair half site of sequence 5'GGGAA-3', whereas p65 prefers a four base pair 5'-GGAA-3' half site. The two half sites are separated by a non-contacted central base pair. This central base pair is often A:T. The ideal target sites of the p50 homodimer and the p65 homodimer are the 11-mer nucleotide sequence (5'GGGGAATTTCC-3') and 9 base pair sequence (5'GGAATTTCC-3'), respectively. The p50/p65 heterodimer preferably interacts with 10 base pair targets. In p50 the first three G:C base pairs are recognized by four conserved amino acids, Arg54, Arg56, Glu60, and His64, of which the first base pair is contacted by His64. Whereas Arg54, Arg56, and Glu60 are identical throughout the Rel/NF-!B polypeptides, His64 is replaced by an Alanine in p65. This helps to explain why p65 recognizes and binds to a shorter DNA target. The pseudodyad axis of DNA passing through the central base pair aligns with the dyad axis of the dimerization domain. Not all the physiological DNA targets, however, follow these general rules of !B DNA sequence consensus. Some of the known physiological targets, such as the Pelement of the IL-4 promoter, has only partial sequence homology to that of the consensus !B DNA. NF-!B dimers are also capable of binding to these targets with affinities comparable to "ideal" !B DNA (Rao et al, 1997). The crystal structure of p65 homodimer bound to the P-elementlike target (Chen et al, 1998) shows that one subunit recognizes the non-consensus half site with an altered conformation with no sequence specific interactions with DNA. NF-!B p65 homodimer thus can bind to DNA targets with as few as only four base pairs. The modular architecture of the DNA binding domains of Rel/NF-!B factors enables them to bind DNA with such diversity. The

V. Structure of I!B proteins The I!B family proteins contain either six or seven ankyrin repeats in a centrally located ankyrin repeatcontaining domain (ARD) (Verma et al, 1995). Ankyrin repeats were first identified as a repeated 33 amino acid element in the erythrocyte membrane protein ankyrin and have since been identified in many proteins of diverse biological functions (Bork et al, 1986). The ankyrin domain in I!B is flanked by two segments. In three of the I!B family proteins, I!B", I!B# and I!B$, these flanking segments are homologous and likely to perform similar functions. The amino-terminal segments in I!B", -# and -$ contain two serines that are phosphorylated by the I!B kinase. The carboxy-terminal segments are rich in the amino acids proline, glutamic acid, serine and threonine (PEST). The acidic PEST sequence is common to many proteins which display rapid turnover in the cell (Rogers et al, 1986). Serines and threonines of the I!B PEST sequences are constitutively phosphorylated by casein kinase II (McElhinny, 1996). Whereas, the role of the ankyrin domain and the carboxy-terminal PEST sequence have been shown to be essential for interactions with NF-!B dimers (Ernst et al, 1995; Malek et al, 1998), it is not yet clear what role, if any, the amino-terminal signal response domain plays in NF-!B recognition. The other I!B family members, specifically I!B% , p105, p100, and Bcl-3, exhibit significant differences from this domain arrangement. The three dimensional structures of the ankyrin repeat domains from several different proteins have been solved (Gorina and Pavletich, 1996; Luh et al, 1997; Venkataramani et al, 1998; Batchelor et al, 1998). Their overall structures are similar in that each 33 amino acid sequence forms the repeating ankyrin structural unit. This structural motif consists of two alpha helices, followed by a loop of variable length, and a short beta turn. Each repeat forms a layer in the stacked, approximately cylindrical domain. The beta turn in each repeat is projected in an orientation roughly perpendicular from the helices and extends like a finger. Ankyrin domains are curved displaying two distinct surfaces. The helical parts lie on the concave face and the stacked fingers form the convex surface. The recent solution by two laboratories of the I!B" crystal structure in complex with the NF-!B p50/p65

Gene Therapy and Molecular Biology Vol 4, page 81 heterodimer (Jacobs and Harrison, 1998; Huxford et al, 1998) reveals its structural similarity to other ankyrin repeat-containing structures (Figure 3b). The six ankyrin repeats of I!B" are stacked with approximately 10 Ă&#x2026; spacing between them with a slight superhelical twist. The helical bundle in each layer interacts with the helices in the layer above and below. A conserved set of residues present in all the inner helices are critical for maintaining the structure of ankyrin repeat domains. Residues in the finger regions vary among I!B and other ankyrin proteins. Therefore, these segments are likely to be responsible for the discriminatory interactions with NF-!B dimers. In complex with NF-!B, the highly charged PEST sequence of I!B" assumes an extended conformation devoid of any distinct secondary structure.

conformations in the context of differing protein binding partners. The NLS polypeptide of p50 also lies in close proximity to I!B" without making any direct interactions. It is important to note that Latimer et al. have shown that the SRD of I!B" is essential for masking the p50 NLS (Latimer et al, 1998). The SRD is absent in both structures which may explain the lack of direct interactions between the p50 NLS polypeptide and I!B". Whereas, I!B" ankyrin repeats one and two are engaged in mediating interactions with the NLS polypeptides of NF-!B, the sixth repeat and the acidic carboxy-terminal PEST sequence of I!B" contact the p65 amino-terminal domain primarily through electrostatic interactions (Huxford et al, 1998). A field of negatively charged residues from these segments of I!B interacts with the positively charged DNA binding face of the p65 amino-terminal domain. This interaction is augmented by interdomain contacts between negatively charged residues of p65 dimerization domain with its amino-terminal domain. Therefore, both intra- and intermolecular interactions involving the amino-terminal domain of p65 serve to force the NF-!B p65 subunit into a closed conformation. Adoption of this conformation renders NF-!B incapable of binding to DNA as key charged specificity and affinitydetermining amino acid side chains are buried in an electrostatic sea.

VI. Interactions between I!B" and NF!B p50/p65 heterodimer Results from biochemical experiments suggest that the discriminatory and affinity-determining interactions between I!B" and NF-!B p50/p65 heterodimer are highly complex (Ernst et al, 1995; Malek et al, 1998; Latimer et al, 1998). The complicated nature of these interactions can be explained by the fact that each of the independent structural elements of both I!B" and NF-!B participate in the extensive proteinprotein interface. Two interacting segments of I!B"; the ankyrin domain and the PEST sequence, do not form a single structural unit but are rather flexible with respect to each other. Similarly, five flexibly linked structural units of the heterodimer, the dimerization domains and the carboxyterminal NLS polypeptides of p50 and p65 and the aminoterminal domain of p65, are engaged in I!B" recognition. Results from the biochemical experiments strongly imply that the amino-terminal domain of p50 is not essential for I!B" recognition. The x-ray crystal structures of the I!B"/NF-!B p50/p65 complex determined independently in two laboratories support the model generated from biochemical experiments. The combined results of these crystallographic analyses of the complex also help to explain genetic and biochemical experiments. The most important insights that these structures provide involve the mechanisms of NF-!B cytoplasmic retention and its dissociation from !B DNA in the nucleus. Upon binding to I!B", the NLS polypeptide of p65 forms two successive alpha helices which contact the first two ankyrin repeats of I!B" (Jacobs and Harrison, 1998). Three of the central four basic residues comprising the NLS of p65 mediate direct salt bridges with acidic residues of I!B". A similar type I basic NLS sequence from SV40 has been shown to interact with nuclear transport protein "karyopherin in an extended conformation (Conti et al, 1998). The free NLS sequence can thus adopt different

VII. Conclusion and future direction While tremendous progress has been made over the last few years in our understanding of the NF-!B signaling pathway and transcriptional regulation, a number of questions remain to be answered. Chief among these are: How does I!B" discriminate for p65 and c-Rel-containing NF-!B dimers? What is the role of the p50 amino-terminal domain in the NF-!B/I!B" complex? Does the I!B" signal response domain contribute to NF-!B binding? What is the nature of the I!B kinase/I!B" interaction? What controls I!B" proteolytic processing? What is the exact mode for nuclear trafficking of NF-!B and I!B"? How does NF-!B interact with other transcriptional coactivators and ultimately influence basal transcription machinery? And, what roles do the remaining I!B family members play? Further biochemical and biophysical testing of the elements involved coupled with careful structural analyses will be required before questions like these can be properly addressed.

Acknowledgements Work performed in this laboratory is supported by grants from the National Institutes of Health/NCI, the University of California AIDS Research Program (UARP), and California Career Research Program. G.G. is a Alfred P. Sloan fellow. T.H. is supported by a dissertation award from UARP. The authors wish to thank Devin Drew and Chris Phelps for help preparing the manuscript.

Ghosh et al: NF-!B/I!B signaling

References

Lee FS, Peters RT, Dang LC and Maniatis T (1 9 9 8 ). MEKK1 activates both I!B kinase " and I!B kinase #. P r o c N a t l Acad Sci USA 95, 9319-9324. Li Z and Nabel GJ (1 9 9 7 ). A new member of the I!B protein family, I!B$, inhibits RelA (p65)-mediated NF-!B transcription. M o l C e l l B i o l 17, 6184-6190. Luh FY, Archer SJ, Domaille PJ, Smith BO, Owen D, Brotherton DH, Raine AR, Xu X, Brizuela L, Brenner SL and Laue ED (1 9 9 7 ). Structure of the cyclin-dependent kinase inhibitor p19 INK4d . Nature 389, 999-1003. Malek S, Huxford T and Ghosh G (1 9 9 8 ). I!B" functions through direct contacts with the nuclear localization signals and the DNA binding sequences of NF-!B. J B i o l C h e m 273, 25427-25435. McElhinny JA, Trushin SA, Bren GD, Chester N and Paya CV (1 9 9 6 ). Casein kinase II phosphorylates I!B" at S-283, S289, S-293, and T-291 and is required for its degradation. M o l C e l l B i o l 16, 899-906. Müller CW, Rey FA, Sodeoka M, Verdine GL and Harrison SC (1 9 9 5 ). Structure of the NF-!B p50 homodimer bound to DNA. Nature 373, 311-317. Rao A, Luo C and Hogan PG (1 9 9 7 ). Transcription factors of the NF-AT family: regulation and function. Annu R e v Immunol 15, 707-747. Rogers S, Wells R and Rechsteiner M (1 9 8 6 ). Amino acid sequences common to rapidly degraded proteins: the PEST hypothesis. S c i e n c e 234, 364-368. Ryseck RP, Novotny J and Bravo R (1 9 9 5 ). Characterization of elements determining the dimerization properties of RelB and p50. M o l C e l l B i o l 15, 3100-3109. Venkataramani R, Swaminathan K and Marmorstein R (1 9 9 8 ). Crystal structure of the CDK4/6 inhibitory protein p18INK4c provides insights into ankyrin-like repeat strucutre/function and tumor-derived p16INK4 mutations. N a t S t r u c t B i o l 5, 74-81. Verma IM, Stevenson JK, Schwarz EM, Van Antwerp D and Miyamoto S (1 9 9 5 ). Rel/NF-!B/I!B family: intimate tales of association and dissociation. Ge ne s D e v 9, 2723-2735. Whiteside ST, Epinat JC, Rice NR and Israël A (1 9 9 7 ). I!B", a novel member of the I!B family, controls RelA and c-Rel NF-!B activity. EMBO J. 16, 1413-1426.

Arenzana-Seisdedos F, Turpin P, Rodriguez M, Thomas D, Hay RT, Virelizier JL and Dargemont C (1 9 9 7 ). Nuclear localization of I!B" promotes active transport of NF-!B from the nucleus to the cytoplasm. J C e l l S c i 110, 369378. Baeuerle PA and Baltimore D (1 9 9 6 ) NF-!B: ten years after. C e l l 87 , 13-20. Baeuerle PA and Henkel T (1 9 9 4 ). Function and activation of NF-! B in the immune system. A n n u R e v I m m u n o l 12, 141-179. Baldwin AS Jr. (1 9 9 6 ). The NF-!B and I!B proteins: new discoveries and insights. A n n u R e v I m m u n o l 14, 649683. Batchelor AH, Piper DE, de la Brousse FC, McKnight SL, and Wolberger C (1 9 9 8 ). The structure of GABP"/#: an ETS domain-ankyrin repeat heterodimer bound to DNA. S c i e n c e 279, 1037-1041. Bork P. (1 9 9 3 ). Hundreds of ankyrin-like repeats in functionally diverse proteins: mobile modules that cross phyla horizontally? P r o t e i n s 17 , 363-374 Chen FE, Huang D-B, Chen Y-Q and Ghosh G (1 9 9 8 ). Crystal structure of p50/p65 heterodimer of transcription factor NF!B bound to DNA. Nature 391, 410-413. Chen Y-Q, Ghosh S, and Ghosh G (1 9 9 8 ). A novel DNA recognition mode by the NF-!B p65 homodimer. Nat S t r u c t B i o l 5, 67-73. Conti E, Uy M, Leighton L, Blobel G and Kuriyan J (1 9 9 8 ). Crystallographic analysis of the recognition of a nuclear localization signal by the nuclear import factor karyopherin ". C e l l 94, 193-204. Cramer P, Larson CJ, Verdine GL and Müller CW (1 9 9 7 ). Structure of the human p52 homodimer-DNA complex at 2.1 Å resolution. EMBO J 16, 7078-7090. DiDonato JA, Hayakawa M, Rothwarf DM, Zandi E and Karin M (1 9 9 7 ). A cytokine-responsive I!B kinase that activates the transcription factor NF-!B. Nature 388, 548-554. Ernst MK, Dunn LL and Rice NR (1 9 9 5 ). The PEST-like sequence of I!B" is responsible for inhibition of DNA binding but not for cytoplasmic retention of c-Rel or RelA homodimers. M o l C e l l B i o l 15 , 872-882. Franzoso G, Bours V, Park S, Tomita-Yamaguchi M, Kelly K and Siebenlist U (1 9 9 2 ). The candidate oncoprotein Bcl-3 is an antagonist of p50/NF-!B-mediated inhibition. Nature 359, 339-342. Ghosh G, Van Duyne G, Ghosh S and Sigler PB (1 9 9 5 ). Structure of NF-!B p50 homodimer bound to a !B site. Nature 3 7 3 , 303-310. Ghosh S, May MJ and Kopp EB (1 9 9 8 ). NF-!B and Rel proteins: evolutionarily conserved mediators of immune responses. Annu Rev Immunol 16, 225-260. Gorina S and Pavletich NP (1 9 9 6 ). Structure of the p53 tumor suppressor bound to the ankyrin and SH3 domains of 53BP2. S c i e n c e 274, 1001-1005. Huang D-B, Huxford T and Ghosh G (1 9 9 7 ). The role of DNA in the mechanism of NF-!B dimer formation: crystal structures of the dimerization domains of the p50 and p65 subunits. Structure 5, 1427-1436. Huxford T, Huang D-B, Malek S and Ghosh G (1 9 9 8 ). The crystal structure of the I!B"/NF-!B complex reveals mechanisms of NF-!B Inactivation. C e l l 95, 759-770. Jacobs MD and Harrison SC (1 9 9 8 ). Structure of an I!B"/NF!B complex. C e l l 95 , 749-758. Latimer M, Ernst MK, Dunn .L, Drutskaya M and Rice NR (1 9 9 8 ). The N-terminal domain of I!B" masks the nuclear localization signals(s) of p50 and c-Rel homodimers. M o l C e l l B i o l 18, 2640-2649.

Gourisankar Ghosh 82

Gene Therapy and Molecular Biology Vol 4, page 83 Gene Ther Mol Biol Vol 4, 83-98. December 1999.

Mammalian c-Jun N-terminal kinase pathway and STE20-related kinases Review Article

Yi-Rong Chen and Tse-Hua Tan* Department of Immunology, Baylor College of Medicine, Houston, Texas 77030, USA __________________________________________________________________________________________________ *Correspondence: Department of Immunology, M929, One Baylor Plaza, Houston, TX 77030; Tel: (713) 798-4665; Fax: (713) 798-3033; E-mail: ttan@bcm.tmc.edu Key words: c-Jun, N-terminal Kinase, STE20-related Kinases, JNK signaling, MEKK1, MAPK Received: 6 August, 1999; accepted 11 October, 1999

Summary The c-Jun N-terminal kinases (JNKs) belong to a subgroup of mitogen-activated protein kinases (MAPKs) that are activated by environmental stress, proinflammatory cytokines, and mitogenic stimuli in mammalian cells. Studies on the JNK pathway in mammalian cells demonstrate that JNK regulates the transcriptional activities of many transcription factors, and that JNK is required for the regulation of cell proliferation and apoptosis. Studies on jnk-deficient mice reveal that JNK is involved in the response to immunological stimuli and in embryonic morphogenesis. JNK, as other MAPKs, is regulated by a kinase cascade. JNK activation is mediated by dual phosphorylation on the motif, Thr-ProTyr. To date, this phosphorylation is known to be mediated by the MAPK kinases (MAP2Ks), MKK4 and MKK7. MKK4 and MKK7 are activated by MEKK1 and other MAPK kinase kinases (MAP3Ks). The MAPK kinase kinase kinases (MAP4Ks) including HPK1, GCK, and homologous kinases, which have a kinase domain related to yeast STE20, can activate the JNK signaling cascade. These mammalian STE20-related MAP4Ks may be involved in integrating various stimuli to the JNK cascade. The signaling specificity of mammalian JNK pathway may be controlled by scaffold proteins that interact with kinases at different levels in the pathway.

I. Introduction Mitogen-activated protein kinases (MAPKs) are important mediators for intracellular signaling in cells (Schaeffer and Weber, 1999). Mammalian MAPKs consist of three major groups including extracellular signal-regulated kinases (ERKs), c-Jun N-terminal kinases (JNKs, also known as stress-activated protein kinases, SAPKs), and p38-MAPKs (Schaeffer and Weber, 1999). All of MAPKs share a common character: they are activated by phosphorylation at a Thr-X-Tyr motif (X is Glu in ERKs, Pro in JNKs, and Gly in p38-MAPKs) in kinase subdomain VIII (Schaeffer and Weber, 1999). The major targets for MAPK kinases are transcription factors that regulate gene expression. MAPKs are involved in signaling induced by various extracellular or intracellular stimuli. Currently, the JNK pathway is one of the known cellular signaling pathways that respond to the most diverse stimuli (Ip and Davis, 1998; Schaeffer and Weber, 1999). JNK is activated by mitogenic signals such as epidermal growth factor (Minden et al., 1994b), lymphocyte activation signals (Su

et al., 1994; Sakata et al., 1995; Berberich et al., 1996; Chen et al., 1996a; Chen et al., 1996b), and oncogenic Ras (Derijard et al., 1994). JNK is also activated by pro-inflammatory cytokines (TNF-! and IL-1) (Kyriakis et al., 1994; Sluss et al., 1994), lipopolysaccharide (Hambleton et al., 1996), G protein-coupled receptors (Collins et al., 1996; Coso et al., 1996), shear stress (Li et al., 1996), osmotic shock (GalchevaGargova et al. , 1994), protein synthesis inhibitors (Kyriakis et al., 1994), and apoptotic stimuli such as growth factor withdrawal (Xia et al., 1995), heat shock (Kyriakis et al., 1994; Zanke et al., 1996), ceramides (Westwick et al., 1995), DNA-damaging chemicals (Saleem et al., 1995; Zanke et al., 1996), UV radiation (Derijard et al., 1994; Chen et al., 1996b; Zanke et al., 1996), and " radiation (Kharbanda et al., 1995; Chen et al., 1996a; Chen et al., 1996b). The diversity of JNKactivating stimuli imply that mammalian cells may be equipped with multiple upstream regulators that link various cellular signals to the JNK pathway, and the accumulated experimental evidence proves that is the case. To date, the JNK pathway consists of JNKs and various MAP2Ks, MAP3Ks, and MAP4Ks (Figure 1). The subtle regulation of 83

Chen and Tan: c-Jun N-terminal & STE20-related kinases the JNK pathway by its regulators in conjunction with other signaling pathways may allow JNK to regulate a variety of cellular functions. In this review, we will discuss the known components in the JNK pathway and how the emerging mammalian scaffold proteins may control signaling diversity and specificity in this signaling pathway.

II. c-Jun N-terminal kinases The human JNKs are encoded by three genes jnk1, jnk2, and jnk3 (Derijard et al., 1994; Kallunki et al., 1994; Sluss et al., 1994; Gupta et al., 1996). The corresponding genes have also been identified in rats (Kyriakis et al., 1994). JNK3 is preferentially expressed in neuronal tissues, while JNK1 and JNK2 are widely expressed in many tissues. Ten isoforms of JNK, generated by alternative splicing of the transcripts from the three genes, have been identified (Gupta et al., 1996). The protein

products of the JNK isoforms have molecular weights of 46 kDa or 55 kDa. The 55 kDa JNK isoforms contain a Cterminal extension, a result of alternative splicing, which distinguishes them from the 46 kDa isoforms (Gupta et al., 1996). No apparent functional differences exist among the 46 kDa and 55 kDa isoforms encoded by the same JNK gene (Gupta et al., 1996). An additional alternative splicing exists in the kinase domains of JNK1 and JNK2, but not in JNK3 (Gupta et al., 1996). The alternative splicing in the kinase domains of JNK1 and JNK2 changes the specificity of interaction between JNKs and their substrates (Gupta et al., 1996), suggesting that JNK isoforms may target different substrates in vivo. The JNK binding sites are different from the sites of phosphorylation on the substrates (Kallunki et al., 1996). Deletion of the binding site prevents phosphorylation of the substrate by JNK (Kallunki et al., 1996). However, a substrate lacking a JNK-binding site can also be phosphorylated through association with a protein containing the JNK-binding region (Kallunki et al., 1996).

Figure 1. The mammalian JNK signaling pathway. Currently known MAPKs, MAP2Ks, MAP3Ks, and MAP4Ks in the JNK pathway are illustrated schematically. PAKs are capable of activating JNK; however, the direct link between PAKs and the JNK signaling module has not been established. The activation of the JNK pathway is known to be mediated by adaptor molecules, p21 small G proteins, or TNFreceptor-associated factors (TRAFs). The signaling specificity among the components is not presented in this figure.

The known substrates for JNK family members include the transcription factors c-Jun (Hibi et al., 1993;

Derijard et al., 1994; Kyriakis et al., 1994; Gupta et al., 1996), JunD (Gupta et al., 1996), ATF-2 (Gupta et al., 1995; van 84

Chen and Tan: c-Jun N-terminal & STE20-related kinases Dam et al., 1995; Whitmarsh et al., 1995), ATFa (Bocco et al., 1996), Elk-1 (Cavigelli et al., 1995; Whitmarsh et al., 1995; Zinck et al., 1995), Sap-1a (Janknecht and Hunter, 1997), GABP!, GABP#_(Hoffmeyer et al., 1998), and the tumor suppressor p53 (Milne et al., 1995; Alder et al., 1997). Generally, phosphorylation of these factors by JNK increases their transcriptional activity. The physiological functions of JNK have been examined by genetic analysis. The jnk1-/-, jnk2-/-, and jnk3-/single mutant mice have no global abnormality (Yang et al., 1997b; Dong et al., 1998; Yang et al., 1998). The T cells in jnk1-/- and jnk2-/- mice preferentially differentiate into Th2 rather than Th1 cells (Dong et al., 1998; Yang et al., 1998). The jnk1-/- T cells also hyper-proliferate and exhibit decreased activation-induced apoptosis (Dong et al., 1998). Excitotoxicity-induced apoptosis in the hippocampus is absent in jnk3-/- mice in comparison to normal mice (Yang et al., 1997b). The jnk1/jnk3 and jnk2/jnk3 deficient mice also develop normally (Kuan et al., 1999); however, jnk1/jnk2 deficient mice are embryonically lethal and have severe dysregulation of apoptosis in the brain (Kuan et al., 1999). These results indicate that JNK1 and JNK2 may have overlapping functions, and are important in regulation of immune response and embryonic development. JNK3 may have its unique functions in the neuronal tissues. These studies also provide animal models which support the accumulated evidence on the role of JNK in apoptotic signaling in mammalian cells (Ip and Davis, 1998).

III. MAP2Ks in the JNK pathway The activation of JNK is dependent on the phosphorylation on Thr-183 and Tyr-185. MKK4 (also known as SEK1 or JNKK1) is a physiological activator of JNK (Sanchez et al., 1994; Derijard et al., 1995; Lin et al., 1995). MKK4 phosphorylates and activates JNK in vitro and in vivo (Sanchez et al., 1994; Derijard et al., 1995; Lin et al., 1995). However, recombinant wild-type JNK proteins are phosphorylated at Tyr, Ser and Thr residues in the presence of recombinant MKK4, whereas a kinaseinactive JNK is phosphorylated predominantly on Tyr (Sanchez et al., 1994). This suggests that recombinant MKK4 does not have apparent dual specificity toward JNK. It is possible that the phosphorylation on Thr-183 is caused by the proline-directed kinase activity of JNK itself, occurring after MKK4-mediated Tyrphosphorylation. Another possibility is that MKK4 obtains dual-specific kinase activity only after activation by upstream kinases. Two isoforms of MKK4 have been reported through the differential usage of translation initiation sites (Derijard et al., 1995; Lin et al., 1995). MKK4 has been found to be mutated or deleted in some tumor cells, suggesting that it may be a tumor

suppressor gene (Teng et al., 1997; Su et al., 1998). Homologous deletion in mkk4 genes is embryonically lethal in mice, indicating that MKK4 is essential for embryonic development (Nishina et al., 1997; Yang et al., 1997a). Studies in mkk4-/-/rag-/- chimaeric mice reveal that MKK4 protects thymocytes from apoptosis mediated by CD95 and CD3 (Nishina et al., 1997), and is required for maintenance of a normal peripheral lymphoid compartment but not for lymphocyte development (Swat et al., 1998). Mkk4-/- T cells derived from mkk4-/-/rag-/- chimaeric mice are defective in heat shock and anisomycin-induced JNK activation, but normal in osmotic shock-induced JNK activation (Nishina et al., 1997). These results indicate that MKK4 is one but not the only activator of JNK in mammalian cells. Recently, a novel kinase MKK7 (also named as JNKK2) has been cloned and found to specifically activate JNK, but not p38-MAPK or ERK (Moriguchi et al., 1997; Tournier et al., 1997; Wu et al., 1997; Yao et al., 1997). MKK7 is related to MKK4 and belongs to the mammalian MAPK kinase superfamily (Tournier et al., 1997; Yao et al., 1997). MKK7 is also closely related to the Drosophila protein kinase hemipterous (HEP) (Tournier et al., 1997; Yao et al., 1997), which is the activator of Drosophila JNK (DJNK). Both MKK4 and MKK7 are widely expressed in human and murine tissue, whereas the relative abundance of each MKK differs among tissues (Tournier et al., 1997; Yao et al., 1997). Both MKK4 and MKK7 mediate signals from the same panel of extracellular stimuli (Wu et al., 1997); however, studies show that they are preferentially activated by different MAP3Ks (Hirai et al., 1998; Merritt et al., 1999; Tournier et al., 1999). Furthermore, the MKK7 gene can encode six isoforms of protein products through alternative splicing of the mRNA transcripts (Tournier et al., 1999). These MKK7 isoforms respond differently to extracellular stimuli and upstream kinases (Tournier et al., 1999). The differential regulation of MKK4 and MKK7 isoforms by their upstream activators needs to be further examined.

IV. MAP3Ks in the JNK pathway Multiple upstream MAPK/ERK kinase kinases or MAP kinase kinase kinases (MEKK or MAP3Ks) have been reported to activate the JNK pathway via MKK4 and/or MKK7 (Figure 1). These MEKK-like kinases include MEKK1-4, ASK1/MAPKKK5, MAPKKK6, TAK1, Tpl2/Cot, MLK2/MST, MLK3/SPRK/PTK1, MUK/DLK/ZPK, and LZK.

A. MEKKs MEKK1 is the first identified MAP3K that activates JNK (Minden et al., 1994a; Yan et al., 1994). MEKK1 was cloned on the basis of its homology with the yeast STE11 and Byr2 kinases (Lange-Carter et al., 1993; Xu et al., 1996). To date, four kinases have been cloned and named MEKK1-4 (Lange86

Gene Therapy and Molecular Biology Vol 4, page 87 Carter et al., 1993; Blank et al., 1996; Gajewski and Thompson, 1996; Xu et al., 1996; Ellinger-Ziegelbauer et al., 1997; Gerwins et al., 1997; Takekawa et al., 1997). The four MEKKs (ranging from 69.5-195 kDa in size) have homologous kinase domains in the C-termini of the proteins; however, their N-terminal domains have little homology. MEKK1 and MEKK4 can interact with GTPbinding proteins Cdc42 and Rac (Fanger et al., 1997; Gerwins et al., 1997). MEKK1 also binds to Ras in a GTPdependent manner (Russell et al., 1995). All four MEKKs (MEKK1-4) activate the JNK pathway (Lange-Carter et al., 1993; Blank et al., 1996; Gajewski and Thompson, 1996; Xu et al., 1996; Ellinger-Ziegelbauer et al., 1997; Gerwins et al., 1997; Takekawa et al., 1997). Besides the JNK pathway, MEKKs also regulate other cellular signaling pathways. MEKK1, MEKK2, and MEKK3 activate the ERK pathway (Lange-Carter et al., 1993; Blank et al., 1996; Ellinger-Ziegelbauer et al., 1997), and also activate the NF-$B through the I$B kinases (IKKs) (Lee et al., 1997; Zhao and Lee, 1999). MEKK3 and MEKK4 have been shown to activate the p38-MAPK pathway through MKK6 (Takekawa et al., 1997; Deacon and Blank, 1999).

B. TAK1 TGF-# activated kinase 1 (TAK1) was identified by its ability to rescue STE11 mutants in S cerevisiae (Yamaguchi et al., 1995). TAK1 is a 579 amino acid protein with the kinase domain in its N-terminus (Yamaguchi et al., 1995). The C-terminal region has no distinct domain structures but interacts with TAK1 binding protein (TAB) 1 and 2 (Yamaguchi et al., 1995; Shibuya et al., 1996). Association of TAK1 and TAB1 enhances the kinase activity of TAK1 (Shibuya et al., 1996). TAK1 is activated by TGF-# (Yamaguchi et al., 1995), interleukin 1 (Ninomiya-Tsuji et al., 1999), ceramide, and UV-C treatments (Shirakable et al., 1997). TAK1 activates JNK and p38, but has no effect on ERK (Yamaguchi et al., 1995; Wang et al., 1997). TAK1 also indirectly activates IKK activity and NF-$B transcriptional activity (Ninomiya-Tsuji et al., 1999). Xenopus TAK1 and TAB1 are important in the dorsoventral patterning of early embryos (Shibuya et al., 1998). Ectopic expression of TAK1 induces apoptosis in early Xenopus embryos (Shibuya et al., 1998).

C. ASK1/MAPKKK5 and ASK2/MAPKKK6 Apoptosis signal-regulating kinase 1 (ASK1, also named MAPKKK5) was identified by a polymerase chain reaction (PCR)-based strategy (Wang et al., 1996; Ichijo et al., 1997). ASK1 consists of 1375 amino acids with a molecular weight about 155 kDa, and the kinase domain of ASK1 is in the middle part of the protein (Wang et al., 1996; Ichijo et al., 1997). ASK1 has been shown to activate JNK and p38-MAPK through MKK4 and MKK3 respectively (Wang et al., 1996; Ichijo et al., 1997). ASK1

is activated by TNF-!, and dominant-negative ASK1 suppresses TNF-!-induced apoptosis (Ichijo et al., 1997). ASK1 is also activated by Daxx, a Fas-binding protein, and is activated by Fas ligation (Chang et al., 1998). ASK1 has also been shown to be involved in apoptosis induced by oxidative stress (Gotoh and Cooper, 1998; Saitoh et al., 1998), microtubule-interfering agents (Wang et al., 1998a), and genotoxic chemicals (Chen et al., 1999b). Overexpression of ASK1 is capable of inducing apoptosis in transfected cells (Ichijo et al., 1997). An ASK1-related kinase kinase, MAPKKK6 (murine homologue is ASK2), was identified by yeast two-hybrid screen using ASK1 as a bait (Wang et al., 1998b). MAPKKK6 also interacts with ASK1 when coexpressed in 293 cells (Wang et al., 1998b). Overall, MAPKKK6 is 45% homologous to ASK1, and the kinase domain of MAPKKK6 is 82% identical to that of ASK1. The catalytic domain of MAPKKK6 shares 37, 42, 43, and 42% identity to MEKK1, MEKK2, MEKK3, and MEKK4, respectively (Wang et al., 1998b). In contrast to ASK1, which is a strong JNK and p38MAPK activator, MAPKKK6 only weakly activates JNK1 but does not activate p38-MAPK or ERK (Wang et al., 1998b).

D. Tpl-2/Cot Tumor progression locus 2 (Tpl-2) was originally identified as a proto-oncogene that is involved in T lymphomas induced by Moloney murine leukemia virus (Patriotis et al., 1993). Tpl-2 is about 90% identical to the human cot gene, which was first identified by its transforming ability (Miyoshi et al., 1991). The Tpl-2 kinase domain shares approximately 30-35% identity to other JNK-activating MAP3Ks; however, the overall similarity between Tpl-2 and other MAP3Ks is low. Expression of Tpl-2 in mammalian cells activates ERK and JNK through the direct phosphorylation of MEK-1 and MKK4, respectively (Salmeron et al., 1996). Tpl-2 has also been shown to participate in CD3 and CD28-induced NF-$B activation through NF-$B-inducing kinase (NIK) and the IKKs (Lin et al., 1999). Tpl-2 also activates the nuclear factor of activated T cells (NF-AT) and induces IL-2 expression in T-cell lines (Tsatsanis et al., 1998).

E. MLKs The mixed lineage kinase (MLK) family is a group of kinases consisting of MLK1 (Dorow et al., 1993), MLK2/mammalian STE20-like (MST) (Dorow et al., 1993; Katoh et al., 1995), MLK3/src-homology 3 (SH3) domaincontaining proline-rich kinase (SPRK)/protein tyrosine kinase 1 (PTK-1) (Rena et al., 1996; Teramoto et al., 1996; Tibbles et al., 1996), MAPK-upstream kinase (MUK)/dual leucine zipper-bearing kinase (DLK)/leucine-zipper protein kinase (ZPK) (Fan et al., 1996; Hirai et al., 1996), and leucine zipper-bearing kinase (LZK) (Sakuma et al., 1997). This group of kinase is characterized by their catalytic domains which show structural features of both tyrosine- and 87

Chen and Tan: c-Jun N-terminal & STE20-related kinases serine/threonine-specific protein kinase. MLKs contain an SH3 motif at the N-terminus and proline-rich regions at the C-terminus. MLKs also have Leu/Ile-zipper motifs near the C-terminus. These motifs may allow MLKs to dimerize or interact with other molecules. The effect of MLK1 on the MAPK pathways is unknown. MLK2/MST activates the JNK pathway through both MKK4 and MKK7 (Hirai et al., 1998). However, MLK2/MST activates recombinant MKK7 more effectively than recombinant MKK4 (Hirai et al., 1998). MLK2/MST weakly activates p38-MAPK and ERK (Hirai et al., 1997). MLK3 activates the JNK and p38-MAPK pathway via MKK4 and MKK3/6, respectively, but has no effect on the ERK pathway (Tibbles et al., 1996), MUK preferentially activates the JNK pathway (Hirai et al., 1996), and utilizes MKK7 but not MKK4 as a substrate (Merritt et al., 1999). LZK has been shown to activate JNK and induce c-Jun phosphorylation in transfected cells (Sakuma et al., 1997), but its activity toward the other MAPK pathways is unclear. MLK1, 2, 3, and MUK contain potential binding sites (CRIB motifs) for Cdc42 and Rac. MLK2 and MLK3 have been shown to interact with Cdc42 and Rac proteins (Teramoto et al., 1996; Nagata et al., 1998).

V. STE20 related Kinases The MAP kinase modules in S. cerevisiae are controlled by a MAP4K named STE20. Several kinases containing a kinase domain that is homologous to STE20 have recently been identified in mammalian cells (Figures 2 and 3A). A phylogenetic analysis on the protein sequences of these mammalian STE20-related kinases reveals that these kinases are divided into several subgroups (Figure 4), which roughly correspond to their structures and biochemical properties. The HPK1/GCK subgroup and HGK/NIK have been shown to regulate the JNK pathway through MAP3Ks. Therefore, they could be classified as MAP4Ks in mammalian cells.

A. PAK subgroup p21-activated kinases (PAK1-4) are characterized by their ability to bind to the Ras-related small G-proteins, Rac1 and Cdc42, through their CRIB domains (Bagrodia et al., 1995b; Martin et al., 1995; Abo et al., 1998) (Figure 2). The binding of PAKs to GTP-bound Rac1 or Cdc42 results in the autophosphorylation and activation of the kinase (Manser et al., 1994; Bagrodia et al., 1995b; Martin et al., 1995; Teo et al., 1995). Rac1 and Cdc42, as well as their direct guanine nucleotide exchange factors (GEF), Ost and Dbl, respectively, were shown to stimulate the JNK pathway via MEKK1 (Bagrodia et al., 1995a; Coso et al., 1995; Minden et al., 1995; Olson et al., 1995; Brown et al., 1996). However, PAKs do not behave as MAP4Ks in the JNK pathway, since PAKs have no (Yablonski et al.,

1998; Zhou et al., 1998) or only modest (Bagrodia et al., 1995a; Polverino et al., 1995; Frost et al., 1996) effect on JNK activation. In addition, the direct interaction between PAKs and MAP3Ks has not been shown. It is found that PAK binding to Rac1 is dispensible for Rac1-induced activation of JNK1 (Westwick et al., 1997). Since Cdc42 and Rac can directly interact with MEKK1, MEKK4, MLK2 and MLK3, PAKs may not be the direct link between small G proteins and the JNK kinase cascade. Neutrophil PAK1 and PAK2 have been shown to phosphorylate the p47phox subunit of NADPH oxidase in response to a chemotactic peptide fMetLeuPhe (Knaus et al., 1995). Recombinant PAK also phosphorylate p67phox subunit of NADPH oxidase (Ahmed et al., 1998). Therefore, it is possible that PAKs indirectly regulate JNK activation by inducing chronic oxidative stress through NADPH oxidase in cells.

B. GCK/HPK1 subgroup The second subgroup of STE20-related kinases includes germinal center kinase (GCK), hematopoietic progenitor kinase 1 (HPK1), kinase homologous to STE20 (KHS)/GCK related kinase (GCKR), and GCK-like kinase (GLK). These kinases are around 96-97 kDa, and share a similar structural configuration. They have the STE20-like catalytic domain in their N-terminus, at least 2 proline-rich binding domains, which can bind to the Src homology 3 (SH3) domain, in their middle region. A domain distantly related to part of murine citron protein (citron homology domain, CNH domain) is located in the C-terminus of these kinases (Figures 2 and 3). This subgroup of STE20-related kinases have been shown to be strong and specific JNK activators. Germinal center kinase (GCK) was first found to be expressed in B lymphocytes residing in the germinal center region of lymphoid follicles (Katz et al., 1994), but later found to be ubiquitously expressed in many tissues. It is a potent and specific activator of JNK but not of ERK1 or p38-MAPK (Pombo et al., 1995). GCK-mediated JNK activation was blocked by a dominant-negative MKK4/SEK construct which indicates that GCK also activates MKK4/SEK (Pombo et al., 1995); however, a direct interaction between GCK and MKK4 has not been detected. GCK has been shown to interact with TNF receptor-associated factor-2 (TRAF2) and with MEKK1 (Yuasa et al., 1998). Therefore, GCK may link the TNF receptor complex to the JNK pathway through MEKK1. GCK interacts with small G protein Rab8 (Ren et al., 1996); however, the biological significance of this interaction is unclear. Hematopoietic progenitor kinase 1 (HPK1) is preferentially expressed in hematopoietic cells, especially in lymphocytes (Hu et al., 1996; Kiefer et al., 1996). HPK1 does not contain CRIB motif and does not interact with Rac1 or Cdc42 (Hu et al., 1996). HPK1 contains four proline-rich domains (putative SH3 domain-binding sites) (Figure 2). HPK1 has been shown

Gene Therapy and Molecular Biology Vol 4, page 89 to interact with the adaptor molecules Grb2, Crk, CrkL and Nck (Anafi et al., 1997; Oehrl et al., 1998; Ling et al., 1999). These adaptor proteins are involved in signaling induced by receptor-linked tyrosine kinases through their SH2 domains, which bind to phosphorylated tyrosine residues. Fas receptor signaling results in the caspasemediated-cleavage of HPK1 at aspartic acid residue 385, which leads to an increase of HPK1 kinase activity and a decrease in its binding to Crk and Grb2 (Chen et al., 1999a). Several pieces of evidence show that HPK1 is regulated by tyrosine kinases. HPK1 is activated by tyrosine phosphatase inhibitors and is tyrosine-

phosphorylated after epidermal growth factor stimulation and T-cell receptor ligation (Anafi et al., 1997; Ling et al., 1999). The interaction between HPK1 and adaptor proteins may recruit HPK1 to surface receptor-tyrosine kinase complexes. HPK1 preferentially activates JNK but not ERK or p38MAPK (Hu et al., 1996; Kiefer et al., 1996). HPK1 interacts with MEKK1 in vivo and directly phosphorylates its regulatory region in vitro (Hu et al., 1996). HPK1 also interacts with MLK3 and TAK1 (Kiefer et al., 1996; Zhou et al., 1999). HPK1 is upstream of TAK1 in TGF-# induced JNK activation, and the interaction between HPK1 and TAK1 is enhanced by TGF-# treatment (Zhou et al., 1999).

Figure 2. Structures of mammalian STE20-related kinases. The protein sequences of the STE20-related kinases were analyzed using the Web-based SMART program (simple modular architecture research tool, EMBL). Human kinase sequences are used in these analyses, except PASK (rat), SLK (murine), and TAO1 (rat).

Gene Therapy and Molecular Biology Vol 4, page 89

Figure 3. Sequence comparison among mammalian STE20-related kinases. (A) Alignment among the kinase domains of STE20-related kinases. The percentages of identity vary from 95% to 36%. Human kinase sequences are used in these analyses, except PASK (rat), SLK (murine), and TAO1 (rat).

Chen and Tan: c-Jun N-terminal & STE20-related kinases

Figure 3 (Continued). (B) Alignment of citron homology domains of HPK1, GCK, KHS, GLK, HGK and murine citron protein. The sequences were analyzed using the PILEUP program (version 7.2; Wisconsin Genetics Computer Group). The output was processed using the BoxShade 3.21 program. Identical residues are shown in black shade, and conserved residues are shown in gray shade. Gaps introduced into the sequences to optimize the alignment are illustrated with dots.

KHS/GCKR is a kinase closely related to GCK and HPK1 (Shi and Kehrl, 1997; Tung and Blenis, 1997). KHS/GCKR also preferentially activates JNK but not p38MAPK and ERK (Shi and Kehrl, 1997; Tung and Blenis, 1997). KHS/GCKR is activated by TNF-! and UV irradiation (Shi and Kehrl, 1997). A KHS/GCKR dominant-negative mutant or antisense suppresses TNF-!, TRAF2, and UV-induced JNK activation (Shi and Kehrl, 1997). KHS/GCKR also physically interacts with TRAF2 (Shi and Kehrl, 1997). KHS/GCKR interacts with the SH3 domains of Crk and CrkL, but not with the SH3 domains of Grb2 or Nck, through its proline rich domains (Oehrl et al., 1998). KHS/GCKR is constitutively active in chronic myeloid leukemia (CML) cells and interacts with oncoprotein Bcr-Abl (Shi et al., 1999). KHS/GCKR is

activated by Bcr-Abl in a Ras-dependent manner (Shi et al., 1999). A dominant-negative KHS/GCKR blocks Bcr-Ablinduced JNK activation (Shi et al., 1999). GLK is widely expressed in many tissues (Diener et al., 1997). GLK preferentially activates JNK, but not p38-MAPK or ERK, when co-expressed in mammalian cells (Diener et al., 1997). GLK-induced JNK activation is blocked by a dominant-negative mutant of MKK4 or MEKK1 (Diener et al., 1997). GLK phosphorylates recombinant MEKK1 (Diener et al. , 1997). These data suggest that GLK regulates the JNK pathway through MEKK1 and MKK4. To date, GLK is known to be regulated by UV irradiation and TNF-! (Diener et al., 1997). Since GLK also contains proline-rich motifs, like other kinases in this family (Figure 2), it may be regulated through interaction with SH3 domain-containing molecules.

Gene Therapy and Molecular Biology Vol 4, page 91

C. HGK, LOK, SLK, and TAO1 Human HPK/GCK like kinase (HGK) is a 133.4 kDa protein (Yao et al., 1999). Its murine counterpart, Nckinteracting kinase (NIK), is 98% identical to HGK, except for an insertion containing two proline-rich motifs in the middle region of the kinase (Su et al., 1997) (Figure 2). A longer form of human HGK that contains these prolinerich motifs was also detected in brain tissue by RT-PCR (Yao et al., 1999). However, the short form of HGK appears to be the predominant form in other human tissues including liver, skeletal muscle and placenta (Yao et al., 1999). Murine HGK/NIK activates JNK and MKK4 when co-expressed in cells, and interacts with MEKK1 (Su et al., 1997). Murine HGK/NIK strongly interacts with the SH3 domain of Nck, but not with other molecules that contain SH3 domains, such as Grb2 and phospholipase C- " (Su et al., 1997). Therefore, it may link Nck-mediated signals to the JNK pathway through MEKK1 and MKK4. Human HGK/NIK-induced JNK activation can be blocked by a dominant-negative mutant of MKK4, MKK7, or TAK1, but not by a dominant-negative MEKK1 mutant (Yao et al., 1999). The difference between human and murine HGK/NIK (long versus short forms) is intriguing and needs to be further examined. Although HGK/NIK contains a citron-homology (CNH) domain (Figure 2), this domain shares low homology to the CNH domains in HPK1, GCK, KHS, and GLK (Figure 3B). HGK/NIK also contains a coiled-coil domain that is not found in any other kinase of the GCK/HPK1 subgroup (Figure 2). A phylogenetic analysis reveals that HGK does not belong to the HPK1/GCK subgroup (Figure 4). Lymphocyte-oriented kinase (LOK) is a 130 kDa kinase which contains a STE20-like kinase domain in its N-terminus followed by a proline-rich region that is a potential SH3 domain binding site, and a long coiled-coil structure at its C-terminus (Kuramochi et al., 1997; Kuramochi et al., 1999) (Figure 2). LOK, unlike HPK1/GCK subgroup, only weakly activates the JNK and p38-MAPK pathways by itself (Kuramochi et al., 1997). LOK may have a specific function in lymphocytes; however, the physiological stimuli and downstream effectors for LOK are unknown (Kuramochi et al., 1997). Thousand and one amino acid protein kinase 1 (TAO1) was identified by a PCR-based method from a rat brain cDNA library (Hutchison et al., 1998). TAO1 is preferentially expressed in brain and testis (Hutchison et al., 1998). TAO1 contains a kinase domain which shares 40-50 % homology with other STE20-related kinases, and a C-terminal regulatory domain which contains coiled-coil structures (Hutchison et al., 1998) (Figure 2). TAO1 activates recombinant MKK4, MKK3, and MKK6 in vitro; however, it preferentially interacts with MKK3 and only activates MKK3 when co-expressed in mammalian cells (Hutchison et al., 1998).

STE20-like kinase (SLK) is a 141 kDa kinase (Itoh et al., 1997; Pytowski et al., 1998). The N-terminal catalytic domain of SLK shares 70% identity with LOKâ&#x20AC;&#x2122;s kinase domain, and 40-45% identity with the kinase domains of other STE20related kinases. The C-terminus of SLK shares 40% homology with LOK, 20% homology with HGK and TAO1, and no homology with other STE20-related kinases. The specificity of SLK toward the known MAPK pathways is unknown.

D. MST/SOK subgroup This subgroup includes mammalian STE20 like 1 (MST1)/kinase regulated by stress 2 (Krs2) (Creasy and Chernoff, 1995a; Taylor et al., 1996), MST2/Krs1 (Creasy and Chernoff, 1995b; Taylor et al., 1996), MST3 (Schinkmann and Blenis, 1997), STE20/oxidant stress response kinase-1 (SOK-1)/yeast Sps1/Ste20-related kinase 1 (YSK1) (Pombo et al., 1996; Osada et al., 1997), and proline-alanine-rich Ste20related kinase (PASK) (Ushiro et al., 1998). These kinases are approximately 45-55 kDa and share 55-90% identity in their STE20-like kinase domain. These kinases have not been shown to activate any of the MAPK pathways, however, their abilities to affect MAPK signaling in the presence of MAPKinducing agents have not been tested. MST1/Krs2 and MST2/Krs1 are activated by extreme heat shock (55Â°C), arsenite, okadaic acid, and staurosporine (Taylor et al., 1996). MST1/Krs2 is also activated by in vitro incubation with purified PP2A (Creasy and Chernoff, 1995a). SOK-1/YSK1, as its name indicates, is activated by oxidative stress such as H2O2 (Pombo et al., 1996). The stimuli that activate MST3 are unknown.

VI. Scaffold Proteins Since JNK is regulated by such diverse upstream pathways, the mechanism by which cells ensure signaling specificity is intriguing. In S. Cerevisiae, the MAPK kinase pathways are coordinated by scaffold proteins (Herskowitz, 1995; Whitmarsh and Davis, 1998). The Ste5p protein binds the components of the MAPK module that control mating (Herskowitz, 1995; Whitmarsh and Davis, 1998). Ste5p interacts with the MAP3K Ste11, the MAP2K Ste7, and the MAPKs Fus3p and Kss1p through different regions of the protein. Pbs2p is a scaffold protein that coordinates an osmoregulatory MAPK pathway (Herskowitz, 1995; Whitmarsh and Davis, 1998). In contrast with Ste5p, Pbs2p is not only a scaffold, but also a protein kinase component (MAP2K) of this MAPK pathway. In mammalian cells, the understanding of scaffold proteins is limited. Recent studies indicate that mammalian cells do contain proteins which serve as scaffolds. One such protein which facilitates signaling of the JNK pathway has been identified. JNK-interacting protein 1 (JIP1) was isolated by a two-hybrid screen for proteins that bind to JNK. JIP1 preferentially binds to JNK, but not to ERK or p38-MAPK (Whitmarsh et al., 91

Chen and Tan: c-Jun N-terminal & STE20-related kinases 1998). JIP1 was first characterized as a JNK inhibitor, because it inhibits the nuclear translocation of JNK and suppresses JNK-mediated functions including transformation and apoptosis (Dickens et al., 1997). However, further characterization of JIP1 has shown that JIP1 interacts with multiple components of the JNK cascade (Whitmarsh et al., 1998). JIP1 binds to MKK7 but not MKK4, which activates both JNK and p38-MAPK. JIP1 also selectively interacts with the MLK family of MAP3K. MKK7 and MLK bind to regions on JIP1 that are distinct from the JNK binding site (Whitmarsh et al., 1998). HPK1 also interacts with JIP1; however, whether this interaction is direct or mediated through MLK family members is uncertain (Whitmarsh et al., 1998). These results strongly suggest that JIP1 serves as a scaffold protein in mammalian cells like Ste5p in yeast cells. Mammalian MAP kinase pathways are much more complicated than those of yeast. It is unknown whether each kinase cascade is coordinated by a specific scaffold protein. Several pieces of evidence suggest that some mammalian kinases may serve as scaffold proteins, similar

to Pbs2p in yeast. For example, MEKK1 is capable of interacting with JNK (Xu and Cobb, 1997), MKK4 (Su et al., 1997), NIK (murine HGK) (Su et al., 1997), and HPK1 (Hu et al., 1996) through different regions of the MEKK1 protein. These properties enable MEKK1 to serve as a scaffold protein. However, whether MEKK1 can bind to all of these kinases simultaneously and the contribution of this binding to signaling specificity is unknown. Several pieces of evidence suggest that certain stimuli may use distinct kinase cascades to activate JNK through signal-specific scaffold proteins. For example, TAK1-induced JNK activation can be blocked by a dominant-negative mutant of MKK4 or MKK7, suggesting that both MKK4 and MKK7 have the potential to mediate TAK1-induced JNK activation (Zhou et al., 1999). However, TGF-#-induced JNK activation, which is mediated by TAK1, is blocked by the dominant-negative mutant of MKK4, but not by a MKK7 mutant (Zhou et al., 1999). These data suggest that TGF-# signaling specifically uses the TAK1-MKK4-JNK cascade, but not the TAK1-MKK7-JNK cascade, implying a signal-specific scaffold may be involved.

Figure 4. Relationship between members of mammalian STE20-related kinases. The phylogenetic analysis was performed using the multiple alignment server provided by the DNA DataBank of Japan (malign@nig.ac.jp). Human kinase sequences are used in these analyses, except PASK (rat), SLK (murine), and TAO1 (rat).

Gene Therapy and Molecular Biology Vol 4, page 93 is implicated in the reorganization of the actin cytoskeleton and in the formation of filopodia. EMBO J., 17, 6527-6540.

VII. Conclusion The c-Jun N-terminal kinase pathway is activated by a variety of extracellular and intracellular stimuli. The biochemical mechanisms by which these stimuli converge upon and regulate the JNK pathway are intriguing. The discovery of multiple upstream JNK regulators, especially at the MAP3K and MAP4K levels, suggests that these kinases may connect the JNK signaling module to the upstream signals. However, the involvement and requirement of these kinases in JNK activation by specific stimuli remains unclear. To date, the studies on these kinases rely upon transient transfection assay, or on the examination of the activation of endogenous kinases by certain stimuli. The dominant-negative kinase mutants are useful to determine the possible involvement of these kinases in response to a specific stimulus; however, the establishment of genetically deficient animals or cell lines would be extremely critical to elucidate the biochemical and physiological importance of these JNK activators. In addition, revealing the further upstream regulators that link MAP3Ks and MAP4Ks to stimuli will be important. JNK isoforms appear to have different substrate specificity. MKK4 and MKK7 also seem to have different substrate specificity, and are differentially regulated by upstream activators. The expanding molecules in the MAP3K and MAP4K levels further create complexity and diversity in the signaling specificity. The identification of scaffold-like proteins, such as JIP1, in mammalian cells indicates that the signaling molecules in the MAPK pathways may form complexes through interactions with scaffold proteins. Different complexes, with distinct components, may mediate different upstream signals and regulate distinct downstream effectors. Although the emergence of the scaffold model provides an explanation for the control of signaling specificity, many questions arise. Does the formation of these complexes occur before or after the stimulation? Is the specificity of these scaffold proteins stringent, or is it relatively flexible? Are these signaling complexes stable, or can the components in different complexes be exchanged rather freely? The answers to these questions will greatly enhance the understanding of cellular signaling.

Acknowledgements T.-H. Tan is supported by NIH grants R01-AI42532 and R01-AI38649, as well as a Leukemia Society of America Scholar award. Y.-R. Chen is supported by a postdoctoral fellowship (DAMD17-99-1-9507) from the Department of Defense Prostate Cancer Research Program.

References Abo A, Qu J, Cammarano MS, Dan C, Fritsch A, Baud V, Belisle B and Minden A (1998) PAK3, a novel effector for Cdc42Hs,

Ahmed S, Prigmore E, Govind S, Veryard C, Kozma R, Wientjes FB, Segal AW and Lim L (1998) Cryptic Rac-binding and p21Cdc42HS/Rac-activated kinase phosphorylation sites of NADPH oxidase component p67phox. J. Biol. Chem., 273, 15693-15701. Alder V, Pincus MR, Minamoto T, Fuchs SY, Bluth MJ, Brandt-Rauf PW, Friedman FK, Robinson RC, Chen JM, Wang XW, Harris CC and Ronai Z (1997) Conformation-dependent phosphorylation of p53. Proc. Natl. Acad. Sci. USA, 94, 16861691. Anafi M, Kiefer F, Gish GD, Mbamalu G, Iscove NN and Pawson T (1997) SH2/SH3 adaptor proteins can link tyrosine kinases to a Ste20-related protein kinase, HPK1. J. Biol. Chem., 272, 2780427811. Bagrodia S, Derijard B, Davis RJ and Ceroine RA (1995a) Cdc42 and PAK-mediated signaling leads to Jun kinase and p38 mitogen-activated protein kinase activation. J. Biol. Chem., 270, 27995-27998. Bagrodia S, Taylor SJ, Creasy CL, Chernoff J and Cerione RA (1995b) Identification of a mouse p21Cdc42/Rac activated kinase. J. Biol. Chem., 270, 22731-22737. Berberich I, Shu G, Siebelt F, Woodgett JR, Kyriakis JM and Clark EA (1996) Cross-linking CD40 on B cells preferentially induces stress-activated protein kinases rather than mitogen-activated protein kinases. EMBO J., 15, 92-101. Blank JL, Gerwins P, Elliot EM, Sather S and Johnson GL (1996) Molecular cloning of mitogen-activated protein/ERK kinase kinase (MEKK) 2 and 3. J. Biol. Chem., 271, 5361-5368. Bocco JL, Bahr A, Goetz J, Hauss C, Kallunki T, Kedinger C and Chatton B (1996) In vivo association of ATFa with JNK/SAP kinase activation. Oncogene, 12, 1971-1980. Brown JL, Stower L, Baer M, Trejo J, Coughlin S and Chant J (1996) Human Ste20 homologue hPAK1 links GTPase to the JNK MAP kinase pathway. Curr. Biol., 6, 598-605. Cavigelli M, Dolfi F, Claret F-X and Karin M (1995) Induction of cfos expression through JNK-mediated TCF/Elk-1 phosphorylation. EMBO J., 14, 5957-5964. Chang HY, Nishitoh H, Yang X, Ichijo H and Baltimore D (1998) Activation of apoptosis signal-regulating kinase 1 (ASK1) by the adaptor protein Daxx. Science, 281, 1860-1863. Chen Y-R, Meyer CF, Ahmed B, Yao Z and Tan T-H (1999a) Caspase-mediated cleavage and functional changes of hematopoietic progenitor kinase 1 (HPK1). Oncogene, 18, 73707377. Chen Y-R, Meyer CF and Tan T-H (1996a) Persistent activation of cJun N-terminal kinase 1 (JNK1) in " radiation-induced apoptosis. J. Biol. Chem., 271, 631-634. Chen Y-R, Wang X, Templeton D, Davis RJ and Tan T-H (1996b) The role of c-Jun N-terminal kinase (JNK) in apoptosis induced by ultraviolet C and " radiation: Duration of JNK activation may determine cell death and proliferation. J. Biol. Chem., 271, 31929-31936. Chen Z, Seimiya H, Naito M, Mashima T, Kizaki A, Dan S, Imaizumi M, Ichijo H, Miyazono K and Tsuruo T (1999b) ASK1

Chen and Tan: c-Jun N-terminal & STE20-related kinases mediates apoptotic cell death induced by genotoxic stress. Oncogene, 18, 173-180.

p46SAPK and p38MAPK but not ERK2. J. Biol. Chem, 271, 24788-24793.

Collins LR, Minden A, Karin M and Brown JH (1996) G!12 stimulates c-Jun NH2-terminal kinase through the small G proteins Ras and Rac. J. Biol. Chem., 271, 173-17353.

Fanger GR, Johnson NL and Johnson GL (1997) MEK kinases are regulated by EGF and selectively interact with Rac/Cdc42. EMBO J., 16, 4961-4972.

Coso OA, Chiariello M, Yu J-C, Teramoto H, Crespo P, Xu N, Miki T and Gutkind JS (1995) The small GTP-binding proteins Rac1 and Cdc42 regulate the activity of the JNK/SAPK signaling pathway. Cell, 81, 1137-1146.

Frost JA, Xu S, Hutchison MR, Marcus S and Cobb MH (1996) Activations of Rho family small G proteins and p21-activated protein kinases on mitogen-activated protein kinase family members. Mol. Cell. Biol., 16, 3707-3713.

Coso OA, Teramoto H, Simonds WF and Gutkind JS (1996) Signaling from G protein-coupled receptors to c-Jun kinase involves #" subunits of heterotrimeric G proteins acting on a Ras and Rac1-dependent pathway. J. Biol. Chem., 271, 3963-3966.

Gajewski TF and Thompson CB (1996) Apoptosis meets signal transduction: elimination of a BAD influence. Cell, 87, 589-592.

Creasy CL and Chernoff J (1995a) Cloning and characterization of a human protein kinase with homology to Ste20. J. Biol. Chem., 270, 21695-21700.

Gerwins P, Blank JL and Johnson GL (1997) Cloning of a novel mitogen-activated protein kinase kinase kinase MEKK4, that selectively regulates the c-Jun amino terminal kinase pathway. J. Biol. Chem., 272, 8288-8295.

Creasy CL and Chernoff J (1995b) Cloning and characterization of a member of the MST subfamily of Ste20-like kinases. Gene, 167, 303-306. Deacon K and Blank JL (1999) MEK kinase 3 directly activates MKK6 and MKK7, specific activators of the p38 and c-Jun NH2-terminal kinases. J. Biol. Chem., 274, 16604-16610. Derijard B, Hibi M, Wu I-H, Barrett T, Su B, Deng T, Karin M and Davis RJ (1994) JNK1: A protein kinase stimulated by UV light and Ha-Ras that binds and phosphorylates the c-Jun activation domain. Cell, 76, 1025-1037. Derijard B, Raingeaud J, Barrett T, Wu I-H, Han J, Ulevitch RJ and Davis RJ (1995) Independent human MAP kinase signal transduction pathways defined by MEK and MKK isoforms. Science, 267, 682-685. Dickens M, Rogers JS, Cavanagh J, Raitano A, Xia Z, Halpern JR, Greenberg ME, Sawyers CL and Davis RJ (1997) A cytoplasmic inhibitor of the JNK signal transduction pathway. Science, 277, 693-696. Diener K, Wang XS, Chen C, Meyer CF, Bray J, Zukowski M, Tan T-H and Yao Z (1997) Activation of the c-Jun Nterminal kinase pathway by a novel protein kinase related to human germinal center kinase. Proc. Natl. Acad. Sci. USA, 94, 9687-9692. Dong C, Yang DD, Wysk M, Whitmarsh AJ, Davis RJ and Flavell RA (1998) Defective T cell differentiation in the absence of jnk1. Science, 282, 2092-2095. Dorow DS, Devereux L, Dietzsch E and De Kretser T (1993) Identification of a new family of human epithelial protein kinases containing two leucine/isoleucine-zipper domains. Eur. J. Biochem., 15, 701-710. Ellinger-Ziegelbauer H, Brown K, Kelly K and Siebenlist U (1997) Direct activation of the stress-activated protein kinase (SAPK) and extracellular signal-regulated protein kinase (ERK) pathways by an inducible mitogen-activated protein kinase/ERK kinase kinase 3 (MEKK) derivative. J. Biol. Chem., 272, 2668-2674. Fan G, Merrit SE, Kortenjann M, Shaw PE and Holzman LB (1996) Dual leucine zipper-bearing kinase (DLK) activates

Galcheva-Gargova Z, Derijard B, Wu I-H and Davis R (1994) An osmosensing signal transduction pathway in mammalian cells. Science, 265, 806-808.

Gotoh Y and Cooper J (1998) Reactive oxygen species- and dimerization-induced activation of apoptosis signal-regulating kinase 1 in tumor necrosis factor-! signal transduction. J. Biol. Chem., 273, 17477-17482. Gupta S, Barrett T, Whitmarsh AJ, Cavanagh J, Sluss HK, Derijard D and Davis RJ (1996) Selective interaction of JNK protein kinase isoforms with transcription factors. EMBO J., 15, 2760-2770. Gupta S, Campbell D, Derijard B and Davis RJ (1995) Transcription factor ATF2 regulation by the JNK signal transduction pathway. Science, 267, 389-393. Hambleton J, Weistein SL, Lem L and DeFranco AL (1996) Activation of c-Jun N-terminal kinase in bacterial lipopolysaccharide-stimulated macrophages. Proc. Natl. Acad. Sci. USA, 93, 2774-2778. Herskowitz I (1995) MAP kinase pathways in yeast: for mating and more. Cell, 80, 187-197. Hibi M, Lin A, Smeal T, Minden A and Karin M (1993) Identification of an oncoprotein- and UV-responsive protein kinase that binds and potentiates the c-Jun activation domain. Genes & Dev., 7, 2135-2148. Hirai S-i, Izawa M, Osada S-i, Spyrou G and Ohno S (1996) Activation of the JNK pathway by distantly related protein kinases, MEKK and MUK. Oncogene, 12, 641-650. Hirai S-i, Katoh M, Terada M, Kyriakis JM, Zon LI, Rana A, Avruch J and Ohno S (1997) MST/MLK2, a member of the mixed lineage kinase family, directly phosphorylates and activates SEK1, an activator of c-Jun N-terminal kinase/stress-activated protein kinase. J. Biol. Chem., 1997, 15167-15173. Hirai S-i, Noda K, Moriguchi T, Nishida E, Yamashita A, Deyama T, Fukuyama K and Ohno S (1998) Differential activation of two JNK activators, MKK7 and SEK1, by MKN28-derived nonreceptor serine/threonine kinase/mixed lineage kinase 2. J. Biol. Chem., 273, 7406-7412. Hoffmeyer A, Avots A, Flory E, Weber CK, Serfling E and Rapp UR (1998) The GABP-responsive element of the interleukin-2 enhancer is regulated by JNK/SAPK-activating pathways in T lymphocytes. J. Biol. Chem., 273, 10112-10119.

Gene Therapy and Molecular Biology Vol 4, page 95 Hu MC-T, Qiu WR, Wang X, Meyer CF and Tan T-H (1996) Human HPK1, a novel human hematopoietic progenitor kinase that activates the JNK/SAPK kinase cascade. Genes & Dev., 10, 2251-2264. Hutchison M, Berman KS and Cobb MH (1998) Isolation of TAO1, a protein kinase that activates MEKs in stressactivated protein kinase cascades. J. Biol. Chem., 273, 28625-28632. Ichijo H, Nishida E, Irie K, ten Dijke P, Saitoh M, Moriguchi T, Takagi M, Matsumoto K, Miyazono K and Gotoh Y (1997) Induction of apoptosis by ASK1, a mammalian MAPKKK that activates SAPK/JNK and p38 signaling pathways. Science, 275, 90-94. Ip YT and Davis RJ (1998) Signal transduction by the c-Jun Nterminal kinase (JNK)-from inflammation to development. Curr. Opin. Cell Biol., 10, 205-219. Itoh S, Kameda Y, Yamada E, Tsujikawa K, Mimura T and Kohama Y (1997) Molecular cloning and characterization of a novel putative STE20-like kinase in Guinea pigs. Arch. Biochem. Biophys., 340, 201-207. Janknecht R and Hunter T (1997) Activation of the Sap-1a transcription factor by the c-Jun N-terminal kinase (JNK) mitogen-activated protein kinase. J. Biol. Chem., 272, 42194224. Kallunki T, Deng T, Hibi M and Karin M (1996) c-Jun can recruit JNK to phosphorylate dimerization partners via specific docking interactions. Cell, 87, 929-939. Kallunki T, Su B, Tsigelny I, Sluss HK, Derijard B, Moore G, Davis R and Karin M (1994) JNK2 contains a specificitydetermining region responsible for efficient c-Jun binding and phosphorylation. Genes & Dev., 8, 2996-3007. Katoh M, Hirai M, Sugimura T and Terada M (1995) Cloning and characterization of MST, a novel (putative) serine/threonine kinase with SH3 domain. Oncogene, 10, 1447-1451. Katz P, Whalen G and Kehrl JH (1994) Differential expression of a novel protein kinase in human B lymphocytes. Preferential localization in the germinal center. J. Biol. Chem., 269, 16802-16809. Kharbanda S, Saleem A, Shafman T, Emoto Y, Taneja N, Rubin E, Weichselbaum R, Woodgett J, Avruch J, Kyriakis J and Kufe D (1995) Ionizing radiation stimulates a Grb2-mediated association of the stress-activated protein kinase with phosphatidylinositol 3-kinase. J. Biol. Chem., 270, 1887118874. Kiefer F, Tibbles LA, Anafi M, Janssen A, Zanke BW, Lassam N, Pawson T, Woodgett JR and Iscove NN (1996) HPK1, a hematopoietic protein kinase activating the SAPK/JNK pathway. EMBO J., 15, 7013-7025. Knaus U, Morris S, Dong H-J, Chernoff J and Bokoch GM (1995) Regulation of human leukocyte p21-activated kinases through G protein-coupled receptors. Science, 269, 221-223. Kuan CY, Yang DD, Samanta RDR, Davis RJ and Flavell RA (1999) The jnk1 and jnk2 protein kinases are required for regional specific apoptosis during brain development. Neuron, 22, 667-676.

Kuramochi S, Matsuda Y, Okamoto M, Kitamura F, Yonekawa H and Karasuyama H (1999) Molecular cloning of the human gene STK10 encoding lymphocyte-oriented kinase, and comparative chromosomal mapping of the human, mouse, and rat homologues. Immunogenetics, 49, 369-375. Kuramochi S, Moriguchi T, Kuida K, Endo J, Semba K, Nishida E and Karasuyama H (1997) LOK is a novel mouse STE20-like protein that is expressed predominantly in lymphocytes. J. Biol. Chem., 272, 22679-22684. Kyriakis JM, Banerjee P, Nikolakaki E, Dai T, Rubie EA, Ahmad MF, Avruch J and Woodgett JR (1994) The stress-activated protein kinase subfamily of c-Jun kinases. Nature, 369, 156-160. Lange-Carter CA, Pleiman CM, Gardener AM, Blumer KJ and Johnson GL (1993) A divergence in the MAP kinase regulatory network defined by MEK kinase and Raf. Science, 260, 315-319. Lee FS, Hagler J, Chen ZJ and Maniatis T (1997) Activation of the I$B! kinase complex by MEKK1, a kinase for the JNK pathway. Cell, 88, 213-222. Li Y-S, Shyy JY-J, Li S, Lee J, Su B, Karin M and Chien S (1996) The Ras-JNK pathway is involved in shear-induced gene expression. Mol. Cell. Biol., 16, 5947-5954. Lin A, Minden A, Martinetto H, Claret F-X, Lange-Carter C, Mercurio F, Johnson GL and Karin M (1995) Identification of a dual specificity kinase that activates the Jun kinases and p38Mpk2. Science, 268, 286-290. Lin X, Cunningham ETJ, Mu Y, Geleziunas R and Greene WC (1999) The proto-oncogene Cot kinase participates in CD3/CD28 induction of NF-$B acting through the NF-$B-inducing kinase and I$B kinase. Immunity, 10, 271-280. Ling P, Yao Z, Meyer CF, Wang XS, Oehrl W, Feller SM and Tan TH (1999) Interaction of HPK1 with the adaptor proteins Crk and CrkL leads to the synergistic activation of c-Jun N-terminal kinase (JNK). Mol. Cell. Biol., 19, 1359-1368. Manser E, Leung T, Salihuddin H, Zhao Z-s and Lim L (1994) A brain serine/threonine protein kinase activated by Cdc42 and rac1. Nature, 367, 40-46. Martin GA, Bollag G, McCormick F and Abo A (1995) A novel serine kinase activated by rac1/Cdc42Hs-dependent autophosphorylation is related to PAK65 and STE20. EMBO J., 14, 1970-1978. Merritt SE, Mata M, Nihalani D, Zhu C, Hu X and Holzman LB (1999) The mixed lineage kinase DLK utilizes MKK7 and not MKK4 as substrate. J. Biol. Chem., 274, 10195-10202. Milne DM, Campbell LE, Campbell DG and Meek DW (1995) p53 is phosphorylated in vitro and in vivo by an ultraviolet radiationinduced protein kinase characteristic of the c-Jun kinase, JNK1. J. Biol. Chem., 270, 5511-5518. Minden A, Lin A, Claret F-X, Abo A and Karin M (1995) Selective activation of the JNK signaling cascade and c-Jun transcription activity by the small GTPases Rac and Cdc42Hs. Cell, 81, 11471157. Minden A, Lin A, McMahon M, Lange-Carter C, Derijard B, Davis RJ, Johnson GL and Karin M (1994a) Differential activation of ERK and JNK mitogen-activated protein kinases by Raf-1 and MEKK. Science, 266, 1719-1722.

Chen and Tan: c-Jun N-terminal & STE20-related kinases Minden A, Lin A, Smeal T, Derijard B, Cobb M, Davis R and Karin M (1994b) c-Jun N-terminal phosphorylation correlates with activation of the JNK subgroup but not the ERK subgroup of mitogen-activated protein kinases. Mol. Cell. Biol., 14, 6683-6688. Miyoshi J, Higashi T, Mukai H, Ohuchi T and Kakunaga T (1991) Structure and transforming potential of the human cot oncogene encoding a putative protein kinase. Mol. Cell. Biol., 11, 4088-4096. Moriguchi T, Toyoshima F, Masuyama N, Hanafusa H, Gotoh Y and Nishida E (1997) A novel SAPK/JNK kinase MKK7, stimulated by TNF-! and cellular stresses. EMBO J., 16, 7045-7053. Nagata K-i, Puls A, Futter C, Aspenstrom P, Schaefer E, Nakata T, Hirokawa N and Hall A (1998) The MAPK kinase kinase kinase MLK2 co-localizes with activated JNK along microtubules and associates with kinesin subfamily motor KIF3. EMBO J., 17, 149-158. Ninomiya-Tsuji J, Kishimoto K, Hiyama A, Inoue J-I, Cao Z and Matsumoto K (1999) The kinase TAK1 can activate the NIKI$B as well as the MAP kinase cascade in the IL-1 signaling pathway. Nature, 398, 252-256. Nishina H, Fischer KD, Radvanyl L, Shahinian A, Hakem R, Rubie EA, Bernstein A, Mak TW, Woodgett JR and Penninger JM (1997) Stress-signaling kinase SEK1 protects thymocytes from apoptosis mediated by CD95 and CD3. Nature, 385, 350-353. Oehrl W, Kardinal C, Ruf S, Adermann K, Groffen J, Feng G-S, Blenis J, Tan T-H and Feller SM (1998) The germinal center kinase (GCK)-related protein kinases HPK1 and KHS are candidates for highly selective signal transducers of Crk family adapter proteins. Oncogene, 17, 1893-1902. Olson MF, Ashworth A and Hall A (1995) An essential role for Rho, Rac, and Cdc42 GTPases in cell cycle progression through G1. Science, 269, 1270-1272. Osada S-i, Izawa M, Saito R, Mizuno K, Suzuki A, Hirai S-i and Ohno S (1997) YSK1, a novel mammalian protein kinase structurally related to Ste20 and SPS1, but is not involved in the know MAPK pathways. Oncogene, 14, 2047-2057. Patriotis C, Makris A, Bear SE and Tsichlis PN (1993) Tumor progression locus 2 (Tpl-2) encodes a protein kinase involved in the progression of rodent T-cell lymphomas and in T-cell activation. Proc. Natl. Acad. Sci. USA, 90, 2251-2255. Polverino A, frost J, Yang P, Hutchison M, Neiman AM, Cobb MH and Markus S (1995) activation of mitogen-activated protein kinase in cell -free extract of Xenopus oocytes. J. Biol. Chem., 270, 26067-26070.

Pytowski B, Hicklin DJ, Kornhaber G, Dellaratta DV and Witte L (1998) Identification and initial characterization of mSLK, a murine member of the Ste20 family of kinases. Arch. Biochem. Biophys., 15, 310-319. Ren M, Zeng J, Lemos-Chiarandini D, Rosenfeld M, Adesnik M and Sabatini DD (1996) In its active form, the GTP-binding protein rab8 interacts with a stress-activated protein kinase. Proc. Natl. Acad. Sci. USA, 93, 5151-5155. Rena A, Gallo K, Godowski P, Hirai S, Ohno S, Zon L, Kyriakis JM and Avrich J (1996) The mixed lineage kinase SPRK phosphorylates and activates the stress-activated protein kinase activator, SEK-1. J. Biol. Chem., 271, 19025-19028. Russell M, Lange-Carter C and Johnson GL (1995) Direct interaction between Ras and the kinase domain of mitogen-activated protein kinase kinase kinase (MEKK1). J. Biol. Chem., 270, 1175711760. Saitoh M, Nishitoh H, Fujii M, Takedo K, Tobiume K, Sawada Y, Kawabata M, Miyazono K and Ichijo H (1998) Mammalian thioredoxin is a direct inhibitor of apoptosis signals-regulating kinase. EMBO J., 17, 2596-2606. Sakata N, Patel HR, Terada N, Aruffo A, Johnson GL and Gelfand EW (1995) Selective activation of c-Jun kinase mitogen-activated protein kinase by CD40 on human B cells. J. Biol. Chem., 270, 30823-30828. Sakuma H, Ikeda A, Oka S, Kozutsumi Y, Zanetta J-P and Kawasaki T (1997) Molecular cloning and functional expression of a cDNA encoding a new member of mixed lineage protein kinase from human brain. J. Biol. Chem., 272, 28622-28629. Saleem A, Dattaa R, Yuan Z-M, Kharban da S and Kufe D (1995) Involvement of stress-activated protein kinase in the cellular response to 1-#-D-arabinofuranosylcytosine and other DNAdamaging agents. Cell Growth Diff., 6, 1651-1658. Salmeron A, Ahmad TB, Calile GW, Pappin D, Narsimhan RP and Ley SC (1996) Activation of MEK1 and SEK-1 by Tpl-2 protooncoprotein, a novel MAP kinase kinase kinase. EMBO J., 15, 817-826. Sanchez I, Hughes RT, Mayer BJ, Yee K, Woodgett JR, Avruch J, Kyriakis JM and Zon LI (1994) Role of SAPK/ERK kinase-1 in the stress-activated pathway regulating transcription factor c-Jun. Nature, 372, 794-798. Schaeffer HJ and Weber MJ (1999) Mitogen-activated protein kinases: specific messages from ubiquitous messengers. Mol. Cell. Biol., 19, 2435-2444. Schinkmann K and Blenis J (1997) Cloning and characterization of a human STE20-like protein kinase with unusual cofactor requirements. J. Biol. Chem., 272, 28695-28703.

Pombo CM, Bonventre JV, Molnar A, Kyriskis J and Force T (1996) Activation of a human STE20-like kinase by oxidant stress defines a novel stress response pathway. EMBO J., 15, 4537-4546.

Shi C-S and Kehrl JH (1997) Activation of stress-activated protein kinase/c-Jun N-terminal kinase, but not NF-$B, by the tumor necrosis factor (TNF) receptor 1 through a TNF receptorassociated factor 2- and germinal center kinase related-dependent pathway. J. Biol. Chem., 272, 32102-32107.

Pombo CM, Kehrl JH, Irma S, Woodgett JR, Force T and Kyriakis JM (1995) Activation of the SAPK pathway by the human STE20 homogue germinal centre kinase. Nature, 377, 750-754.

Shi C-S, Tuscano JM, Witte ON and Kehrl JH (1999) GCKR links the Bcr-Abl oncogene and Ras to the stress-activated protein kinase pathway. Blood, 93, 1338-1345. Shibuya H, Iwata H, Masuyama N, Gotoh Y, Yamaguchi K, Irie K, Matsumoto K, Nishida E and Ueno N (1998) Role of TAK1 and

Gene Therapy and Molecular Biology Vol 4, page 97 TAB1 in BMP signaling in early Xenopus development. EMBO J., 17, 1019-1028. Shibuya H, Yamaguchi K, Shirakabe K, Tonegawa A, Gotoh Y, Ueno N, Irie K, Nishida E and Matsumoto K (1996) TAB1, an activator of the TAK1 MAPKKK in TGF-# signal transduction. Science, 272, 1179-1182. Shirakable K, Yamaguchi K, Shibuya H, Irie K, Matsuda S, Moriguchi T, Gotoh Y, Matsumoto K and Nishida E (1997) TAK1 mediates the ceramide signaling to stress-activated protein kinase/c-Jun N-terminal kinase. J. Biol. Chem., 272, 8141-8144.

Tibbles LA, Ing YL, Kiefer F, Chan J, Iscove N, Woodgett JR and Lassam NJ (1996) MLK-3 activates the SAPK/JNK and p38/RK pathways via SEK1 and MKK3/6. EMBO J., 15, 7026-7035. Tournier C, Whitmarsh AJ, Cavanagh J, Barrett T and Davis R (1999) The MKK7 gene encodes a group of c-Jun NH2-terminal kinase kinases. Mol. Cell. Biol., 19, 1569-1581. Tournier C, Whitmarsh AJ, Vavanagh J, Barret T and Davis RJ (1997) Mitogen-activated protein kinase kinase 7 is an activator of the c-Jun NH2-terminal kinase. Proc. Natl. Acad. Sci. USA, 94, 7337-7342.

Sluss HK, Barrett T, Derijard B and Davis RJ (1994) Signal transduction by tumor necrosis factor mediated by JNK protein kinases. Mol. Cell. Biol., 14, 8376-8384.

Tsatsanis C, Pariotis C, Bear SE and Tsichlis PN (1998) The Tpl-2 protooncoprotein activates the nuclear factor of activated T cells and induces interleukin 2 expression in T cell lines. Proc. Natl. Acad. Sci. USA, 95, 3827-3832.

Su B, Jacinto E, Hibi M, Kallunki T, Karin M and Ben-Neriah Y (1994) JNK is involved in signal integration during costimulation of T lymphocytes. Cell, 77, 727-736.

Tung RM and Blenis J (1997) A novel human SPS1/STE20 homologue, KHS, activates Jun N-terminal kinase. Oncogene, 14, 653-659.

Su GH, Hilgers W, Shekher MC, Tang DJ, Yeo CJ, Hruban RH and Kern SE (1998) Alterations in pancreatic, biliary, and breast carcinomas support MKK4 as a genetically targeted tumor suppressor gene. Cancer Res., 58, 2339-2342.

Ushiro H, Tsutsumi T, Suzuki K, Kayahara T and Nakano K (1998) Molecular cloning and characterization of a novel Ste20-related protein kinase enriched in neurons and transporting epithelia. Arch. Biochem. Biophys., 355, 233-240.

Su Y-C, Han J, Xu S, Cobb M and Skolnik EY (1997) NIK is a new Ste20-related kinase that binds NCK and MEKK1 and activates the SAPK/JNK cascade via a conserved regulatory domain. EMBO J., 16, 1279-1290.

van Dam H, Wilhelm D, Herr I, Steffen A, Herrlich P and Angel P (1995) ATF-2 is preferentially activated by stress-activated protein kinases to mediate c-Jun induction in response to genotoxic agents. EMBO J., 14, 1798-1811.

Swat W, Fujikawa K, Ganiatsas S, Yang D, Xavier RJ, Harris NL, Davidson L, Ferrini R, Davis RJ, Labow MA, Flavell RA, Zon LI and Alt FW (1998) SEK1/MKK4 is required for maintenance of a normal peripheral lymphoid compartment but not for lymphocyte development. Immunity, 8, 625-634.

Wang TH, Wang H-S, Ichijo H, Giannakakou P, Foster JS, Fojo T and Wimalasena J (1998a) Microtubule-interfering agents activate c-Jun N-terminal kinase/stress-activated protein kinase through both Ras and apoptosis signal-regulated kinase pathways. J. Biol. Chem., 273, 4928-4936.

Takekawa M, Posas F and Saito H (1997) A human homolog of the yeast Ssk2/Ssk22 MAP kinase kinase kinases, MTK1, mediates stress-induced activation of the p38 and JNK pathways. EMBO J., 16, 4973-4982.

Wang W, Zhou G, Hu MC-T, Yao Z and Tan T-H (1997) Activation of the hematopoietic progenitor kinase-1 (HPK-1)-dependent, stress-activated c-jun N-terminal kinase (JNK) pathway by transforming growth factor # (TGF-#)-activated kinase (TAK1), a kinase mediated of TGF-# signal transduction. J. Biol. Chem. , 272, 22771-22775.

Taylor LK, Wang H-CR and Erikson RL (1996) Newly identified stress-responsive protein kinases, Krs-1 and Krs-2. Proc. Natl. Acad. Sci. USA, 1996, 10099-10104. Teng DH, Perry WL, 3rd, Hogan JK, Baumgard M, Bell R, Berry S, Davis T, Frank D, Frye C, Hattier T, Hu R, Jammulapati S, Janecki T, Leavitt A, Mitchell JT, Pero R, Sexton D, Schroeder M, Su PH, Swedlund B, Kyriakis JM, Avruch J, Bartel P, Wong AK, Tavtigian SV, et al. (1997) Human mitogen-activated protein kinase kinase 4 as a candidate tumor suppressor. Cancer Res., 57, 4177-4182. Teo M, Manser E and Lim L (1995) Identification and molecular cloning of a p21cdc42/rac1-activated serine/threonine kinase that is rapidly activated by thrombin in platelets. J. Biol. Chem., 270, 26690-26697. Teramoto H, Coso OA, Miyata H, Igishi T, Miki T and Gutkind JS (1996) Signaling from the small GTP-binding proteins Rac1 and Cdc42 to the c-Jun N-terminal kinase/stressactivated protein kinase pathway: A role for mixed lineage kinase 3/protein-tyrosine kinase 1, a novel member of the mixed lineage kinase family. J. Biol. Chem., 271, 2722527228.

Wang XS, Diener K, Jannuzzi D, Trollinger D, Tan T-H, Lichenstein H, Zukowski M and Yao Z (1996) Molecular cloning and characterization of a novel protein kinase with a catalytic domain homologous to mitogen-activated protein kinase kinase kinase. J. Biol. Chem., 271, 31607-31611. Wang XS, Diener K, Tan T-H and Yao Z (1998b) MAPKKK6, a novel mitogen-activated protein kinase kinase kinase, that associates with MAPKKK5. Biochem. Biophy. Res. Commun., 253, 33-37. Westwick JK, Beilawska AE, Dbaibo G, Hannun YA and Brenner DA (1995) Ceramide activates the stress-activated protein kinases. J. Biol. Chem., 270, 22689-22692. Westwick JK, Lambert QT, Clark GJ, Symons M, Aelst LV, Restell RG and Der CJ (1997) Rac regulation of transformation, gene expression, and actin organization by mutiple, PAK-independent pathways. Mol. Cell. Biol., 17, 1324-1335. Whitmarsh AJ, Cavanagh J, Tournier C, Yasuda J and Davis RJ (1998) A mammalian scaffold complex that selectively mediates MAP kinase activation. Science, 281, 1671-1674.

Chen and Tan: c-Jun N-terminal & STE20-related kinases Whitmarsh AJ and Davis RJ (1998) Structural organization of MAP-kinase signaling modules by scaffold proteins in yeast and mammals. Trend. Biochem. Sci, 23, 481-485.

activated protein kinase kinase (MKK7). J. Biol. Chem., 272, 32378-32383.

Whitmarsh AJ, Shore P, Sharrocks AD and Davis RJ (1995) Integration of MAP kinase signal transduction pathways at the serum response element. Science, 269, 403-407.

Yao Z, Zhou G, Wang XS, Brown A, Diener K, Gan H and Tan T-H (1999) A novel mammalian STE20-related protein kinase, HGK, that specifically activates the c-Jun N-terminal kinase signaling pathway. J. Biol. Chem., 274, 2118-2125.

Wu Z, Wu J, Jacinto E and Karin M (1997) Molecular cloning and characterization of human JNKK2, a novel Jun NH2terminal kinase-specific kinase. Mol. Cell. Biol., 17, 74077416.

Yuasa T, Ohno S, Kehrl JH and Kyriakis JM (1998) Tumor necrosis factor signaling to stress-activated protein kinase (SAPK)/Jun NH2-terminal kinase (JNK) and p38. J. Biol. Chem., 273, 22681-22692.

Xia Z, Dickens M, Raingeaud J, Davis RJ and Greenberg ME (1995) Opposing effects of ERK and JNK-p38 MAP kinase on apoptosis. Science, 270, 1326-1331.

Zanke BW, Boudreau K, Rubie E, Winnett E, Tibbles LA, Zon L, Kyriakis J, Liu F-F and Woodgett JR (1996) The stress-activated protein kinase pathway mediates cell death following injury induced by cis-platinum, UV irradiation or heat. Curr. Biol. , 6, 606-613.

Xu S and Cobb MH (1997) MEKK1 binds directly to the c-Jun N-terminal kinases/stress-activated protein kinases. J. Biol. Chem., 272, 32056-32060. Xu S, Robbins DJ, Christerson LB, English JM, Vanderbilt CA and Cobb MH (1996) Cloning of rat MEK kinase 1 cDNA reveals an endogenous membrane-associated 195-kDa protein with a large regulatory domain. Proc. Natl. Acad. Sci. USA, 93, 5295. Yablonski D, Kane LP, Qian D and Weiss A (1998) A NckPAK1 signaling module is required for T-cell receptormediated activation of NFAT, but not of JNK. EMBO J., 17, 5647-5657. Yamaguchi K, Shirakabe K, Shibuya H, Kenji I, Oishi I, Ueno N, Taniguchi T, Nishida E and Matsumoto K (1995) Identification of a member of the MAPKKK family as a potential mediator of TGF-# signal transduction. Science, 270, 2008-2011. Yan M, Dai T, Deak JC, Kyriakis JM, Zon LI, Woodgett JR and Templeton DJ (1994) Activation of stress-activated protein by MEKK1 phosphorylation of its activator SEK1. Nature, 372, 798-800.

Zhao Q and Lee F (1999) Mitogen-activated protein kinase/ERK kinase kinase 2 and 3 activate nuclear factor-$B through I$B kinase-! and I$B kinase-#. J. Biol. Chem., 274, 8355-8358. Zhou G, Lee SC, Yao Z and Tan T-H (1999) Hematopoietic progenitor kinase 1 is a component of transforming growth factor #-induced c-Jun N-terminal kinase signaling cascade. J. Biol. Chem., 274, 13133-13138. Zhou K, Wang Y, Gorski JL, Nomura N, Collard J and Bokoch GM (1998) Guanine nucleotide exchange factors regulate specificity of downstream signaling from Rac and Cdc42. J. Biol. Chem., 273, 16782-16786. Zinck R, Cahill MA, Kracht M, Sachsenmaier C, Hipskind RA and Nordheim A (1995) Protein synthesis inhibitors reveal different regulation of mitogen-activated protein kinase and stressactivated protein kinase pathways that converge on ELK-1. Mol. Cell. Biol., 15, 4930-4938.

Yang D, Tournier C, Wysk M, Lu H-T, Davis RJ and Flavell RA (1997a) Targeted disrurtion of the MKK4 gene causes embryonic death, inhibition of c-Jun N-terminal kinase activation, and defects in AP-1 transcription. Proc. Natl. Acad. Sci. USA, 94, 3004-3009. Yang DD, Conze D, Whitmarsh AJ, Barret T, Davis RJ, Rincon M and Flavell RA (1998) Differentiation of CD4+ T cells to Th1 cells requires MAP kinase JNK2. Immunity, 9, 575585. Yang DD, Kuan C-Y, Whitmarsh A, Rincon M, Zheng TS, Davis RJ, Rakic P and Flavell RA (1997b) Absence of excitotoxicity-induced apoptosis in the hippocampus of mice lacking the jnk3 gene. Nature, 389, 865-870. Yao Z, Diener K, Wang XS, Zukowski M, Matsumoto G, Zhou G, Tibbles LA, Sasaki T, Nishina H, Tan T-H, Woodgett J and Penninger JM (1997) Activation of stress-activated protein kinase/c-Jun N-terminal protein by a novel mitogen-

Tse-Hua Tan

Gene Therapy and Molecular Biology Vol 4, page 99 Gene Ther Mol Biol Vol 4, 99-107. December 1999.

Nucleocytoplasmic trafficking and glucocorticoid receptor function Review Article

Robert J.G. Haché1,2,3, Joanne G.A. Savory1 and Yvonne A. Lefebvre1,2 Departments of Medicine1, Biochemistry, Microbiology and Immunology2, The Loeb Health Research Institute at the Ottawa Hospital, University of Ottawa, 725 Parkdale Ave., Ottawa, Ontario, Canada, K1Y 4K9 ______________________________________________________________________________________________________ 3

Correspondence: Robert J.G. Haché, Ph.D., Director, Hormones, Growth and Development Program, The Loeb Health Research Institute at the Ottawa Hospital, University of Ottawa, 725 Parkdale Ave., Ottawa, Ontario, Canada, K1Y 4K9. Tel: (613)-798-5555 ext. 6283; Fax: (613)-761-5036; E-mail: rhache@lri.ca Abbreviations: GR, glucocorticoid receptor; DBD, DNA-binding domain; LBD, ligand-binding domain; NL or NLS, nuclear localization signal; hsp, heat shock protein; WT, wild type. Key words: glucocorticoid receptor, nuclear import, steroid agonists, chaperone, importin, nuclear export signal, nuclear localization signal, steroid hormone receptors, transcription factor Received: 8 August 1999; accepted: 17 August 1999

Summary The glucocorticoid receptor (GR) is a ligand activated transcription factor that redistributes between nucleus and cytoplasm in response to the addition and withdrawal of steroidal ligands. Localization of the receptor in the cell is dynamic and changes in GR localization reflect the shifting of equilibria between several competing cellular pathways. Since the naïve receptor is transformed from a transcriptionally inert cytoplasmic factor to a potent sequence-specific, DNA-bound transcriptional regulator, delimiting the controls on receptor localization is seminal to understanding how receptor activity may be manipulated or controlled within the cell. A number of recent reports have begun to reveal that the controls on GR trafficking are more sophisticated than previously expected and point to an important role for trafficking controls in the regulation of the steroid response.

I. Introduction The glucocorticoid receptor (GR) is encoded by one of six genes for steroid hormone receptors. Steroid receptors are ligand-activated transcription factors of the nuclear hormone receptor superfamily with highly similar cys-4 type zinc fingered DNA binding domains. Receptor ligand binding domains are less well conserved and determine specific responses to steroid hormone agonists and antagonists. Steroid receptors are distinguished from other members of the nuclear receptor family by their association in the absence of ligand with a chaperone complex anchored to the receptors by hsp90, but also including a number of other heat shock proteins and immunophilins (Pratt and Toft, 1997). Association of unliganded receptors with the chaperone complex is dynamic and assembly of the mature complexes follows an ordered pathway (Pratt and Toft, 1997; Smith, 1998). For GR, hsp association appears to be a

prerequisite for steroid binding and thus functions as an important control point for steroid signaling (Bresnick et al., 1989; Picard et al., 1990). The modular structure of GR and the functional activities localized within the receptor are summarized in Figure 1. The central DNA binding domain is preceded by an extended amino terminus that contains a ligandindependent transcriptional activation function (Bocquel et al., 1989; Hollenberg and Evans, 1988). The amino terminus of GR also contains several serine/threonine phosphorylation sites whose phosphorylation is modulated through the cell cycle and which may have some effects on the transcriptional regulatory potential of GR as well as on receptor stability and subcellular trafficking (Bodwell et al., 1998; Hsu and DeFranco, 1995; Hsu et al., 1992; Hu et al., 1997; Munck and Holbrook, 1984; Webster et al., 1997). A second, steroid-dependent transcriptional activation function is located within the C-terminal ligand binding domain of

Haché et al: Glucocorticoid receptor nucleocytoplasmic trafficking the receptor (Bocquel et al., 1989; Danielian et al., 1992; Hollenberg and Evans, 1988). The hsp90 binding surface of GR is also localized to the C-terminal region of the receptor (Dalman et al., 1991; Howard et al., 1990). Ligand binding induces receptor transformation, characterized by the dissociation of GR from the chaperone complex and, in the presence of hormone agonist, leads to the free transcriptionally active form of the receptor (Beato et al., 1996). Function of GR involves the cyclic redistribution of the receptor between the nucleus and cytoplasm. This proposed cycle is shown in its simplest form in Figure 2. In the absence of steroidal stimulus, naïve GR is localized predominantly to the cytoplasm under most circumstances and is complexed with the chaperone complex (Picard and Yamamoto, 1987; Sackey et al., 1996). Upon ligand binding, the chaperone complex is dissociated and the receptor dimerizes and is rapidly transferred to the nucleus (Cidlowski et al., 1990; Picard and Yamamoto, 1987; Sackey et al., 1996; Wikstrom et al., 1987). The precise sequence of these events remains to be completely elucidated. Upon arrival in the nucleus the activated receptors bind to specific response elements and regulate gene transcription through the recruitment of transcriptional coregulatory proteins that promote changes in chromatin structure and stimulate the basal transcriptional machinery

(Beato and Sanchez-Pacheco, 1996; Beato et al., 1996; Glass et al., 1997; Wolffe, 1997). Upon loss of ligand or withdrawal of stimulus, GR becomes reassociated with the chaperone complex and becomes slowly redistributed to the cytoplasm (Haché et al., 1999; Qi et al., 1989; Sackey et al., 1996). A series of recent studies have shown that this simple visual model for the cycling and recycling of GR likely occurs above a complex regulatory network in which constitutive receptor trafficking is subject to the push and pull of multiple regulatory signals.

II. Nuclear-cytoplasmic transport signals within GR Despite intensive study into the regulatory control of nucleo-cytoplasmic trafficking of GR and other steroid receptors, our knowledge of the signals that mediate the transfer of GR across the nuclear membrane is relatively modest. GR has been shown to have two independent nuclear localization signals, NL1 in the hinge region of the receptor between the DBD and LBD and NL2 within the LBD itself (Picard and Yamamoto, 1987). NL2 is a unique feature of GR, as the other steroid receptors lack similar activities in their LBDs.

Figure 1. Schematic depiction of functional motifs within the glucocorticoid receptor. The position of the DNA and ligand binding domains of rat GR are shown integrated into the schema depicting the A-E domain organization employed for nuclear hormone receptors, while location of transcriptional activation functions (TAF’s), nuclear localization signals (NLS’s) dimerization motifs and the hsp90 binding determinants of GR are summarized below.

100

Gene Therapy and Molecular Biology Vol 4, page 101

Figure 2. Simplified model for the response of glucocorticoid receptor to steroid hormone. Prior to steroid treatment, naĂŻve GR is localized to the cytoplasm in a complex with heat shock proteins and immunophilins that is anchored by hsp90. Upon steroid treatment, liganded GR dissociates from the hsp-immunophilin complex, transfers to the nucleus and activates transcription as a dimer from specific DNA response elements. Steroid binding is a transient event. Upon loss of steroid, the receptor becomes recycled into a similar or identical hsp complex that is able to respond to subsequent hormonal challenge.

The core of the nuclear localization sequence that is NL1 is a series of basic amino acids (rGR 510-RKTKKKIK517) that resemble classical basic NLSs such as that of SV40 T antigen and which appear to mediate the binding of GR to the importin ! NLS binding proteins (Savory et al., 1999). Several amino acids N-terminal to this NL1 core are two additional small groups of basic amino acids that may also contribute to the import of GR into the nucleus (Tang et al., 1997). Mutations within these two additional basic motifs decrease the level of nuclear occupancy of GR (Tang et al., 1997). However the mutations examined to date have also compromised DNA binding by GR. This may be significant as a separate study showed that other mutations in the GR DBD that impair DNA binding without affecting these basic motifs led to a similar decrease in the nuclear occupancy of the liganded receptor (Sackey et al., 1996). In addition, substitutions within the NL1 core were sufficient to abrogate the nuclear import of GR fragments also lacking NL2 (Savory et al., 1999). Thus the significance of the contribution of these two additional basic motifs to the nuclear import of GR remains to be confirmed. Study of the NL2 nuclear import signal in the ligand binding domain of GR has been slowed by itâ&#x20AC;&#x2122;s overlap with the chaperone binding region of GR and an apparent strict dependence on bound steroid (Dalman et al., 1991; Howard et al., 1990; Picard and Yamamoto, 1987; Savory et al., 1999). Indeed, the minimal NL2 signal described to date overlaps completely with the minimal GR ligand binding domain (rGR aa 540-795) (Picard and Yamamoto, 1987).

101

However, the minimal LBD of GR is sufficient for transfer of the receptor to the nucleus and also mediates the nuclear transport of normally cytoplasmic proteins such as "galactosidase in hybrid experiments (Picard and Yamamoto, 1987). NL2 appears to be strictly dependent upon the binding of steroid agonists to GR (Savory et al., 1999). The treatment of cells with the GR antagonist RU486, which binds to GR and induces chaperone dissociation and the transfer of WT receptor to the nucleus, has been shown to be unable to promote the nuclear transfer of GRs in which NL1 has been inactivated (Savory et al., 1999). These results suggest that NL2 is highly dependent upon the three dimensional structure of the GR LBD bound to steroid agonist and is likely to overlap with the positioning of the Cterminal !-helix of GR that is also crucial for the AF-2 activity of receptor. Withdrawal of steroid agonist leads to the inactivation of NL2 and the rapid return of NL1- GR to the cytoplasm (Savory et al., 1999). NL2 appears to mediate the nuclear uptake of GR through a pathway that is different from the importin !dependent pathway that is likely to be employed by NL1. The GR LBD lacks an obvious basic motif that might become exposed upon steroid binding. In addition, mutation of the GR NL1 has been shown to prevent the binding of agonist-bound receptor to an importin ! homologue in vitro and in vivo (Savory et al., 1999). Third, a recent kinetic analysis of NL2-mediated nuclear import of GR identified clear differences in the kinetics of nuclear import and the levels of nuclear occupancy of WT GR and GR with a mutation inactivating NL1 (Savory et al., 1999). While no

Haché et al: Glucocorticoid receptor nucleocytoplasmic trafficking experimental information presently exists to substantiate hypothetical rationales for requirement for NL2 in the function of GR, it may be expected that NL2 imparts a selected capacity for the nuclear import of GR under particular or peculiar physiological circumstances that meets a requirement for GR function that is lacking for the other steroid receptors. By contrast to the picture emerging for the nuclear import of GR, there is little information on the nature or position of signals that mediate the nuclear export of GR. The export of GR from the nucleus however, appears to be an active process that is ATP-dependent and is likely to be signal-mediated. In the only study performed to date, separate deletion of the N-terminus and LBD had no apparent effect on the export of GR from the nucleus (Yang et al., 1997). This suggests either that a nuclear export signal for GR is encoded within the receptor DBD or that the N-terminus and LBD contain distinct export signals. The export of many and perhaps even most proteins from the nucleus has been shown to be accomplished through a protein family called exportins, that are related to the importin " nuclear importers (Ullman et al., 1997). GR contains at least 5 hydrophobic motifs that exhibit similarity to the motifs recognized by the exportins and thus may be potential export signals (R. Haché, unpublished observation). Further, nuclear localization of GR is promoted by the treatment of cells with leptomycin B, a specific inhibitor of CRM1-mediated nuclear import, suggesting that the export of GR from the nucleus may involve the exportin CRM1 (Savory et al., 1999).

III. Nuclear import of GR in response to steroid treatment Interestingly, the import of GR into the nucleus may be mediated by two separate pathways that are distinguished by a dependence upon hsp90 and association of the receptor with the cytoskeleton. Several recent studies have shown that inhibition of the nuclear transport of GR by pharmacological agents and the divalent anion molybdate can be reversed by treatment of cells with agents that induce the depolarization of the cytoskeletal network. In the first instance it was shown that geldanamycin, a compound that inhibits ATP hydrolysis by hsp90 and which blocks the maturation of steroid receptor complexes inhibited the transfer of GR to the nucleus when added to cells following steroid treatment, but prior to the dissociation of GR from the chaperone complex (Czar et al., 1997; Galigniana et al., 1999; Galigniana et al., 1998). However, in cells in which the cytoskeleton had been disrupted by pretreatment with colcemid or cytochalasin D, geldanamycin no longer inhibited the transfer of GR to the nucleus (Galigniana et al., 1998). The implications of this work have been that GR may reach the nucleus primarily by tracking

102

along the cytoskeleton to the nuclear pore in a manner that involves the continued contact of the receptor with hsp90, whereas in the absence of the cytoskeleton or hsp90 the GR is free to diffuse through the cytoplasm until it encounters the nuclear import machinery. Similarly, other studies have suggested that other components of the GR chaperone complex, such as hsp56/FKBP56 may also be required for the nuclear import of GR following ligand binding (Czar et al., 1995). However, while provocative, these studies are based mainly on using agents targeting the chaperone complex to block nuclear import of GR following steroid treatment. Therefore, to date, the alternative possibility that these agents extend the contact between GR and the chaperone complex or otherwise alter the complex in a manner that leads to the promotion or stabilization of an interaction between GR and the cytoskeleton that inhibits nuclear import, can not yet be completely excluded.

IV. Localization and trafficking of the unliganded receptor GR exchanges rapidly between inactive, chaperoneassociated forms and the freed, transcriptionally active state. Further, the liganded receptor shuttles rapidly between nucleus and cytoplasm despite the near complete visualization of liganded GR in the nucleus (Madan and DeFranco, 1993). These results are similar to those obtained with other steroid receptors (Dauvois et al., 1993; GuiochonMantel et al., 1991), and reflect the rapid trafficking of liganded GR between heterokaryon nuclei, which could only happen if the liganded receptor were rapidly shuttling between nucleus and cytoplasm. By contrast it has long been hypothesized that unliganded GR is localized to the cytoplasm because the association of GR into the chaperone complex masks the nuclear localization signals on GR in much the same way that I#B has been shown to retain NF#B in the cytoplasm by blocking NLS exposure (Beg et al., 1992; Hutchison et al., 1993; Pratt, 1993). Recent studies have provided results that indicate that chaperone association is not equivalent to I#B association, and that the trafficking of unliganded, chaperone-associated GR may be constitutive. The first indication that the association of GR into the chaperone complex is not sufficient to trap GR in the cytoplasm was obtained in studies examining the movement of the receptor in the cell following the withdrawal of ligand treatment. For example, while steroid is rapidly lost from GR upon hormone withdrawal (Munck and Holbrook, 1984) and reassembly into hsp90-containing chaperone complexes occurs within minutes (Haché et al., 1999), the relocalization or redistribution of GR to the cytoplasm upon the withdrawal of steroid treatment occurs only over a period of many hours (Haché et al., 1999; Madan and DeFranco, 1993; Sackey et al., 1996). Moreover, we have demonstrated in cells withdrawn from the steroid antagonist RU486, that the GR

Gene Therapy and Molecular Biology Vol 4, page 103 remains entirely localized to the nucleus for periods of up to 48 h following the withdrawal of the antagonist (Sackey et al., 1996). However, chaperone association, steroid binding capacity and transcriptional activation potential of these GRs was recovered within minutes of withdrawal of the antagonist (Haché et al., 1999). But do the ligand-withdrawn, chaperone associated GRs continue to shuttle or traffic across the nuclear membrane? Apparently so, as both RU486 withdrawn GRs and the almost completely nuclear GRs withdrawn from cortisol treatment for only 1 h were found to continue to transfer efficiently between heterokaryon nuclei in cell fusion experiments (Haché et al., 1999). Thus, reassembly of GR into the chaperone complex following ligand withdrawal does not appear to be a barrier to the re-import of GR from the cytoplasm into the nucleus. Further, it has been shown that GRs can reassemble into the chaperone complex without first exiting the nucleus (Liu and DeFranco, 1999). Thus is appears likely that bi-directional transport of GR across the nuclear membrane is not markedly impeded by chaperone association. Not unexpectedly, given the dependence of NL2 on steroidal ligand, maintenance of ligand withdrawn GRs in the nucleus appeared to be entirely dependent upon NL1 and correlated with the binding of the GR-chaperone complex with importin ! (Savory et al., 1999). Molybdate is a divalent metal ion that binds to hsp90 and which is known to artificially stabilize the GR-chaperone complex (Leach et al., 1979). In vitro, addition of molybdate to chaperone associated GR prevented the NL1-dependent binding of GR to importin ! (Savory et al., 1999), while micro-injection of molybdate in to tissue culture cells blocked the re-import of hormone-withdrawn GR into the nucleus (Yang and DeFranco, 1996). Unlike WT GR, GRs directed to the nucleus entirely under the control of NL2 redistributed rapidly to the cytoplasm following withdrawal of steroid while reassociating indistinguishably with the chaperone complex (Savory et al., 1999). If the GR that reassociates into a chaperone complex continues to traffic extensively between nucleus and cytoplasm, is it tenable to continue to suggest that the naïve GR-chaperone complex is statically localized to the cytoplasm? It has been established for several years now that although PR and ER! are constitutively nuclear proteins, the naïve, chaperone associated forms of both receptors exchange continuously between nucleus and cytoplasm. $ number of studies provide experimental support for the notion that it is also likely that naïve GR traffics continuously between nucleus and cytoplasm despite its primary localization to the cytoplasm. Nonetheless, the evidence favoring trafficking of naïve GR remains indirect. For example, it has been noticed in several cell lines, that overexpression of GR leads to increased accumulation of the naïve, hsp-associated receptor 103

to the nucleus (Martins et al., 1991; Sanchez et al., 1990). Indeed, the nuclear accumulation of such GRs appears to be NL1-dependent (Savory et al., 1999). Addition of leptomycin B to cells cultured in the absence of steroidal ligands, which is expected to decrease GR export from the nucleus promoted the nuclear accumulation of previously cytoplasmic naïve GRs (Savory et al., 1999). It has also been reported that addition of a nuclear retention signal to the N-terminus of GR, that is unable to promote the nuclear import of heterologous proteins, shifts otherwise cytoplasmic naïve receptor almost complete to the nucleus (Haché et al., 1999). While a direct demonstration of the entry of naïve WT GR to the nucleus under normal culture conditions remains to be accomplished, experiments designed to conclusively demonstrate trafficking of the naïve cytoplasmic GR now seem likely to yield a positive result.

V. An expanded model for the nucleocytoplasmic exchange of GR: A role for retention mechanisms and the control transfer rates When brought together, recent developments in our understanding of the movement of GR in the cell suggest a more dynamic movement of GR about the cell than has been previously appreciated (detailed in Figure 3). In the first instance, it now appears that chaperone associated GRs are not statically localized in the cells, but are exchanged continuously between nucleus and the cytoplasm. At physiological levels of expression in most cells, naïve GR is seen as being predominantly cytoplasmic. However, this localization is likely overlying the continuous exchange of the receptor between nucleus and cytoplasm. That increasing the level of expression of GR is sufficient to promote the accumulation of the naïve receptor in the nucleus suggests that nuclear accumulation may occur through a saturation of retention mechanisms that otherwise promote the maintenance of GR in the cytoplasm. What these retention mechanism might be is not clear, but they may possibly include interactions between the GRchaperone complex and the cytoskeleton. At the far end the steroid response, the interaction of chaperone-associated, steroid-withdrawn GRs with the nuclear matrix, that correlates directly with the slow redistribution of these GRs to the cytoplasm in the face of continuous nucleocytoplasmic exchange supports active nuclear retention as being important for slowing or preventing the redistribution of shuttling receptors to the cytoplasm. However, it remains possible that direct regulation of the rates of nuclear import and export may contribute in important ways to the localization of naïve and ligand withdrawn GR in the cell. Upon ligand binding, GR dramatically changes its interactions with the chaperone complex through a process

HachĂŠ et al: Glucocorticoid receptor nucleocytoplasmic trafficking that results in a receptor that may only remain loosely associated with hsps including hsp90 and hsp56. This liganded GR dimerizes and accumulates rapidly on DNA in the nucleus. DNA binding appears to be a strong determinant for nuclear localization of GR, as mutations in the receptor that impair DNA binding lead to a form of liganded GR that becomes only modestly more nuclear than cytoplasmic (Sackey et al., 1996). In this context, that the rapid reassociation of GRs with hsps upon loss of ligand does not result in immediate effects on the nuclear localization of GR, reinforces the apparent importance of the transfer of the GR from DNA to the nuclear matrix upon release of ligand from the receptor. With time however, the interactions that maintain the nuclear occupancy of the steroid-withdrawn GRs appear to be slowly reversed, with a concomitant slow redistribution of the receptor to the cytoplasm. One intriguing open question is why GRs withdrawn from RU486 are apparently unable to relocalize to the cytoplasm for extended periods or time (Sackey et al., 1996). Thus it would seem that RU486-withdrawn GRs may

associate indefinitely with the nuclear matrix or some other subnuclear compartment, while the association of agonistwithdrawn GRs with the same or alternative compartments becomes slowly reversed over a period of several hours. One possibility is that differences in phosphorylation patterns within the N-terminus of GR may promote the long term interaction of the GR-chaperone complexes with the nuclear matrix. Certainly it has been reported that GR becomes differentially phosphorylated in response to steroid agonists and RU486 (Hsu et al., 1992). Further, both the differential phosphorylation and long term nuclear retention of RU486-withdrawn GRs may have something to do with differences that have been observed in the subnuclear targeting of the liganded receptors (Htun et al., 1996). However, as the GRs in cells withdrawn from the cell cycle and receptors in actively growing cells return to the cytoplasm with similar kinetics following the withdrawal of agonists, it would seem unlikely that the redistribution of the shuttling, unliganded hsp-reassociated GRs is dependent of specific effects of the cell cycle on the receptor (Hsu et al., 1992).

Figure 3. Expanded model for the subcellular distribution and nucleocytoplasmic trafficking of GR. Schematic presentation of the events identified in the localization and trafficking of GR prior to, during and following steroid treatment. NaĂŻve, hsp associated receptor is hypothesized to traffic continuously between nucleus and cytoplasm (1.), but may be preferentially retained in the cytoplasm through active retention (2.) or an imbalance in nuclear import and export rates (3.). Upon steroid treatment, liganded GRs dissociate from the chaperone complex (4.), dimerize and accumulate on DNA in the cell nucleus to regulate transcription (5.), all the while continuing to traffic rapidly between nucleus and cytoplasm (6.) Following loss of ligand, the shuttling free GRs reassociate with the chaperone complex (7.) and localize to the nuclear matrix (8.) while continuing to shuttle between nucleus and cytoplasm (9.). Over an extended period of time, the receptor-chaperone complex reorients in some way and relocalizes to the cytoplasm (8.).

104

Gene Therapy and Molecular Biology Vol 4, page 105 In addition to changes in the subcytoplasmic and subnuclear retention, changes in the actual rates of nuclear import and export of liganded and unliganded GRs could also influence receptor localization in response to the binding and release of ligand. Thus while association of GR with molecular chaperones does not appear to block binding to importin !, it remains possible that access to the NLS and binding to importin ! may be reduced relative to free GR. Decreased affinity for importin ! upon chaperone association would be expected to result in a slower rate of nuclear import. If the rate of nuclear export were maintained, or even increased at the same time, the longterm result would be localization or redistribution of the shuttling receptor to the cytoplasm prior to hormone treatment and following hormone withdrawal. There is also evidence that is beginning to emerge that localization of GR may be regulated through the control of export rates. For example, using digitonin permeabilized cells it has been observed that the nuclear export of GR, hsp90, hsp70, but not other proteins is specifically inhibited by incubation with tyrosine kinase inhibitors (DeFranco et al., 1991; Yang et al., 1997). These results suggest that some component of the GR export pathway may be dependent upon tyrosine phosphorylation. To date however, there is little evidence that GR itself may be phosphorylated on tyrosine. Conversely the treatment of cells with the serine-threonine phosphatase inhibitor okadaic acid or appears to prevent the re-transfer of steroid-withdrawn GRs to the nucleus (Galigniana et al., 1999). This inhibition is lifted upon co-treatment with agents that depolarize the cytoskeleton. A similar inhibition of the maintenance of GR in the nucleus was observed earlier upon over expression of the serine-threonine kinase v-mos (Qi et al., 1989). These data suggest that nuclear import or the cytoplasmic retention of GR may be directly regulated by serine-threonine kinases and phosphatases, an interesting parallel to the potential dependence of GR export and/ or nuclear retention by tyrosine kinases.

VII. Concluding remarks The trafficking and localization of GR as it has been described in this review reflects the situation observed in asynchronously growing cells and cells that have be withdrawn from the cell cycle by serum starvation. Further progress in understanding how localization and trafficking of GR in the cell affects the responses of the receptor to steroid will require more precise delimitation of the molecular events that control subcellular movements of the receptor including a description of the mechanisms mediating the nuclear export of GR, factors influencing movement of GR across the nuclear membrane, and the continued identification of factors that determine the specific retention of GR in subcellular compartments.

105

Acknowledgements We would like to thank our colleagues in the HachĂŠ and Lefebvre Laboratories for their valuable comments on the manuscript. R.J.G.H. is a Scientist of the Medical Research Council of Canada.

References Beato, M., and Sanchez-Pacheco, A. (1996). Interaction of steroid hormone receptors with the transcription initiation complex. Endocr. Rev. 17, 587-609. Beato, M., Truss, M., and Chavez, S. (1996). Control of transcription by steroid hormones. Ann. N Y Acad. Sci. 784, 93-123. Beg, A. A., Ruben, S. M., Scheinman, R. I., Haskill, S., Rosen, C. A., and Baldwin, A. S., Jr. (1992). I kappa B interacts with the nuclear localization sequences of the subunits of NF-kappa B: a mechanism for cytoplasmic retention. Genes Dev. 6, 18991913. Bocquel, M. T., Kumar, V., Stricker, C., Chambon, P., and Gronemeyer, H. (1989). The contribution of the N- and Cterminal regions of steroid receptors to activation of transcription is both receptor and cell-specific. Nucleic Acids Res. 17, 2581-2595. Bodwell, J. E., Webster, J. C., Jewell, C. M., Cidlowski, J. A., Hu, J. M., and Munck, A. (1998). Glucocorticoid receptor phosphorylation: overview, function and cell cycledependence. J. Steroid Biochem. Mol. Biol. 65, 91-99. Bresnick, E. H., Dalman, F. C., Sanchez, E. R., and Pratt, W. B. (1989). Evidence that the 90-kDa heat shock protein is necessary for the steroid binding conformation of the L cell glucocorticoid receptor. J. Biol. Chem. 264, 4992-4997. Cidlowski, J. A., Bellingham, D. L., Powell-Oliver, F. E., Lubahn, D. B., and Sar, M. (1990). Novel antipeptide antibodies to the human glucocorticoid receptor: recognition of multiple receptor forms in vitro and distinct localization of cytoplasmic and nuclear receptors. Mol. Endocrinol. 4, 1427-1437. Czar, M. J., Galigniana, M. D., Silverstein, A. M., and Pratt, W. B. (1997). Geldanamycin, a heat shock protein 90-binding benzoquinone ansamycin, inhibits steroid-dependent translocation of the glucocorticoid receptor from the cytoplasm to the nucleus. Biochemistry 36, 7776-7785. Czar, M. J., Lyons, R. H., Welsh, M. J., Renoir, J. M., and Pratt, W. B. (1995). Evidence that the FK506-binding immunophilin heat shock protein 56 is required for trafficking of the glucocorticoid receptor from the cytoplasm to the nucleus. Mol. Endocrinol. 9, 1549-1560. Dalman, F. C., Scherrer, L. C., Taylor, L. P., Akil, H., and Pratt, W. B. (1991). Localization of the 90-kDa heat shock proteinbinding site within the hormone-binding domain of the glucocorticoid receptor by peptide competition. J. Biol. Chem. 266, 3482-3490. Danielian, P. S., White, R., Lees, J. A., and Parker, M. G. (1992). Identification of a conserved region required for hormone dependent transcriptional activation by steroid hormone receptors. EMBO J 11, 1025-1033. Dauvois, S., White, R., and Parker, M. G. (1993). The antiestrogen ICI182780 disrupts estrogen receptor nucleocytoplasmic shuttling. J. Cell Sci. 106, 1377-1388.

Haché et al: Glucocorticoid receptor nucleocytoplasmic trafficking DeFranco, D. B., Qi, M., Borror, K. C., Garabedian, M. J., and Brautigan, D. L. (1991). Protein phosphatase types 1 and/or 2A regulate nucleocytoplasmic shuttling of glucocorticoid receptors. Mol. Endocrinol. 5, 1215-1228. Galigniana, M. D., Housley, P. R., DeFranco, D. B., and Pratt, W. B. (1999). Inhibition of glucocorticoid receptor nucleocytoplasmic shuttling by okadaic acid requires intact cytoskeleton. J Biol Chem 274, 16222-16227. Galigniana, M. D., Scruggs, J. L., Herrington, J., Welsh, M. J., Carter-Su, C., Housley, P. R., and Pratt, W. B. (1998). Heat shock protein 90-dependent (geldanamycin-inhibited) movement of the glucocorticoid receptor through the cytoplasm to the nucleus requires intact cytoskeleton. Mol. Endocrinol. 12, 1903-1913. Glass, C. K., Rose, D. W., and Rosenfeld, M. G. (1997). Nuclear receptor coactivators. Curr. Opin. Cell Biol. 9, 222-232. Guiochon-Mantel, A., Lescop, P., Christin-Maitre, S., Loosfelt, H., Perrot-Applanat, M., and Milgrom, E. (1991). Nucleocytoplasmic shuttling of the progesterone receptor. EMBO J. 10, 3851-3859. Haché, R. J. G., Tse, R., Reich, T., Savory, J. G. A., and Lefebvre, Y. A. (1999). Nucleocytoplasmic trafficking of steroid-free glucocorticoid receptor. J. Biol. Chem. 274, 1432-1439. Hollenberg, S. M., and Evans, R. M. (1988). Multiple and cooperative trans-activation domains of the human glucocorticoid receptor. Cell 55, 899-906. Howard, K. J., Holley, S. J., Yamamoto, K. R., and Distelhorst, C. W. (1990). Mapping the HSP90 binding region of the glucocorticoid receptor. J. Biol. Chem. 265, 11928-11935. Hsu, S. C., and DeFranco, D. B. (1995). Selectivity of cell cycle regulation of glucocorticoid receptor function. J. Biol. Chem. 270, 3359-3364. Hsu, S. C., Qi, M., and DeFranco, D. B. (1992). Cell cycle regulation of glucocorticoid receptor function. EMBO J. 11, 3457-3468. Htun, H., Barsony, J., Renyi, I., Gould, D. L., and Hager, G. L. (1996). Visualization of glucocorticoid receptor translocation and intranuclear organization in living cells with a green fluorescent protein chimera. Proc. Natl. Acad. Sci. U S A 93, 4845-4850. Hu, J. M., Bodwell, J. E., and Munck, A. (1997). Control by basal phosphorylation of cell cycle-dependent, hormoneinduced glucocorticoid receptor hyperphosphorylation. Mol. Endocrinol. 11, 305-311. Hutchison, K. A., Scherrer, L. C., Czar, M. J., Stancato, L. F., Chow, Y. H., Jove, R., and Pratt, W. B. (1993). Regulation of glucocorticoid receptor function through assembly of a receptor-heat shock protein complex. Ann. N Y Acad. Sci. 684, 35-48. Leach, K. L., Dahmer, M. K., Hammond, N. D., Sando, J. J., and Pratt, W. B. (1979). Molybdate inhibition of glucocorticoid receptor inactivation and transformation. J. Biol. Chem. 254, 11884-11890. Liu, J., and DeFranco, D. B. (1999). Chromatin recycling of glucocorticoid receptors: implications for multiple roles of heat shock protein 90. Mol Endocrinol 13, 355-65. Madan, A. P., and DeFranco, D. B. (1993). Bidirectional transport of glucocorticoid receptors across the nuclear envelope. Proc. Natl Acad. Sci. U S A 90, 3588-3592.

106

Martins, V. R., Pratt, W. B., Terracio, L., Hirst, M. A., Ringold, G. M., and Housley, P. R. (1991). Demonstration by confocal microscopy that unliganded overexpressed glucocorticoid receptors are distributed in a nonrandom manner throughout all planes of the nucleus. Mol. Endocrinol. 5, 217-225. Munck, A., and Holbrook, N. J. (1984). Glucocorticoid-receptor complexes in rat thymus cells. Rapid kinetic behavior and a cyclic model. J. Biol. Chem. 259, 820-831. Picard, D., Khursheed, B., Garabedian, M. J., Fortin, M. G., Lindquist, S., and Yamamoto, K. R. (1990). Reduced levels of hsp90 compromise steroid receptor action in vivo. Nature 348, 166-168. Picard, D., and Yamamoto, K. R. (1987). Two signals mediate hormone-dependent nuclear localization of the glucocorticoid receptor. EMBO J. 6, 3333-3340. Pratt, W. B. (1993). The role of heat shock proteins in regulating the function, folding, and trafficking of the glucocorticoid receptor. J. Biol. Chem. 268, 21455-21458. Pratt, W. B., and Toft, D. O. (1997). Steroid receptor interactions with heat shock protein and immunophilin chaperones. Endocr. Rev. 18, 306-360. Qi, M., Hamilton, B. J., and DeFranco, D. (1989). v-mos Oncoproteins affect the nuclear retention and reutilization of glucocorticoid receptors. Mol. Endocrinol. 3, 1279-1288. Sackey, F. N. A., Haché, R. J. G., Reich, T., Kwast-Welfeld, J., and Lefebvre, Y. A. (1996). Determinants of subcellular distribution of the glucocorticoid receptor. Mol. Endocrinol. 10, 1191-1205. Sanchez, E. R., Hirst, M., Scherrer, L. C., Tang, H. Y., Welsh, M. J., Harmon, J. M., Simons, S. S., Jr., Ringold, G. M., and Pratt, W. B. (1990). Hormone-free mouse glucocorticoid receptors overexpressed in Chinese hamster ovary cells are localized to the nucleus and are associated with both hsp70 and hsp90. J. Biol. Chem. 265, 20123-20130. Savory, J. G. A., Hsu, B., Laquian, I. R., Giffin, W., Reich, T., Haché, R. J. G., and Lefebvre, Y. A. (1999). Discrimination between NL1 and NL2 mediated nuclear localization of the glucocorticoid receptor. Mol. Cell. Biol. 19, 1025-1037. Smith, D. F. (1998). Sequence motifs shared between chaperone components participating in the assembly of progesterone receptor. Biol. Chem. 379, 283-288. Tang, Y., Ramakrishnan, C., Thomas, J., and DeFranco, D. B. (1997). A role for HDJ-2/HSDJ in correcting subnuclear trafficking, transactivation, and transrepression defects of a glucocorticoid receptor zinc finger mutant. Mol. Biol. Cell 8, 795-809. Ullman, K. S., Powers, M. A., and Forbes, D. J. (1997). Nuclear export receptors: from importin to exportin. Cell 90, 967-970. Webster, J. C., Jewell, C. M., Bodwell, J. E., Munck, A., Sar, M., and Cidlowski, J. A. (1997). Mouse glucocorticoid receptor phosphorylation status influences multiple functions of the receptor protein. J. Biol. Chem. 272, 9287-93. Wikstrom, A. C., Bakke, O., Okret, S., Bronnegard, M., and Gustafsson, J. A. (1987). Intracellular localization of the glucocorticoid receptor: evidence for cytoplasmic and nuclear localization. Endocrinology 120, 1232-42. Wolffe, A. P. (1997). Chromatin remodeling regulated by steroid and nuclear receptors. Cell Res. 7, 127-142. Yang, J., and DeFranco, D. B. (1996). Assessment of glucocorticoid receptor-heat shock protein 90 interactions in

Gene Therapy and Molecular Biology Vol 4, page 107 vivo during nucleocytoplasmic trafficking. Mol. Endocrinol. 10, 3-13. Yang, J., Liu, J., and DeFranco, D. B. (1997). Subnuclear trafficking of glucocorticoid receptors in vitro: chromatin recycling and nuclear export. J. Cell Biol. 137, 523-538.

107

Gene Therapy and Molecular Biology Vol 4, page 109 Gene Ther Mol Biol Vol 4, 109-118. December 1999.

Ribozyme-dependent inactivation of lacZ mRNA in E. coli: a feasibility study to set up a rapid in vivo system for screening HIV-1 RNA-specific ribozymes Research Article

Maria Fe C. Medina and Sadhna Joshi Department of Medical Genetics and Microbiology, Faculty of Medicine, University of Toronto, Toronto, Ontario M5S 3E2, Canada __________________________________________________________________________________________________ Correspondence: Sadhna Joshi, Department of Medical Genetics and Microbiology, Faculty of Medicine, University of Toronto, 150 College St. # 212, Toronto, Ontario M5S 3E2, Canada. Tel: (416)-978-2499; Fax: (416)-638-1459; E-mail: sadhna.joshi.sukhwal@utoronto.ca Key Words: Ribozyme, lacZ, mRNA, HIV-1 RNA, bacterial indicator cell system, !-galactosidase, env coding region, pGEM4Z-based plasmid, hammerhead ribozyme Received: 15 August 1999; accepted: 30 August 1999

Summary Ribozymes are potentially useful tools with widespread applications in gene therapy of several diseases. In order to assess the in vivo cleavage efficiency of human immunodeficiency virus (HIV)-1 RNA-specific ribozymes, a bacterial indicator cell system could be developed in which the degree of inhibition of !-galactosidase activity would correlate with ribozyme activity. The suitability of this indicator cell system was assessed using a ribozyme targeted against the env coding region within the HIV-1 RNA. To this end, a pGEM4Z-based plasmid was engineered wherein oligodeoxynucleotides containing a hammerhead ribozyme and its target site were cloned in frame within the lacZ coding region that encodes for the " fragment of !-galactosidase. Extra nucleotides were included in the insert to ensure that the lacZ open reading frame was not interrupted due to a frameshift or nonsense mutation. In E. coli indicator cells harbouring this plasmid, ribozyme-mediated cleavage of the target site provided in cis and the subsequent loss of !-galactosidase activity should correlate with ribozyme activity. However, frameshift mutations were observed upon sequence analysis of plasmid DNA isolated from the selected light blue to white colonies. Because these mutations affected the production of the !-galactosidase " fragment, a direct correlation between !-galactosidase and ribozyme activities could not be established in vivo. Thus, in clones which demonstrated visibly lower !-galactosidase activities than the control, the effect of the frameshift mutations on lacZ mRNA translation can not be discounted. In clones expressing ribozymes but displaying dark blue colour, it is possible that lacZ mRNAs were cleaved but that the !-galactosidase substrates used were sensitive enough to allow detection of proteins translated from residual lacZ mRNA transcripts. The use of alternative !-galactosidase substrates with less sensitivity may enable the use of the proposed indicator cell system.

I. Introduction Hammerhead ribozymes are small, catalytic RNA molecules first identified in the avocado sunblotch viroid as well as in the satellite RNAs of lucerne transient streak and tobacco ringspot viruses (reviewed by Vaish, 1998). The hammerhead ribozyme catalytic and substrate binding domains have been well characterized (Haseloff and Gerlach, 1988; Uhlenbeck, 1987). Hammerhead ribozymes may be targeted against any given RNA (reviewed by Birikh et al, 1997) provided that the ribozyme catalytic domain is flanked by antisense sequences to allow

ribozyme binding to the target RNA. The cleavage site within the target RNA must be immediately preceded by NUH (Ruffner et al, 1990), with N being any nucleotide and H being any nucleotide except G. Cleavage results in a 5' product with a 5' hydroxyl group and a 3' product with a 2', 3' cyclic phosphate. AIDS is caused by HIV, a retrovirus with an RNA genome. During its life cycle, HIV produces numerous mRNAs which are all potential targets for designing ribozymes (reviewed by Joshi and Joshi, 1996). Monomeric hammerhead ribozymes have been developed and tested against several sites within the HIV-1 RNA (reviewed by 109

Medina and Joshi: Screening HIV-1 RNA-specific ribozymes Macpherson et al, 1999); however, virus breakthrough was eventually observed in each case (reviewed by Ramezani and Joshi, 1999). Ribozymes with increased catalytic activity have been selected via in vitro selection/evolution (reviewed by Pan 1997). However, the in vivo cleavage activity of these ribozymes may be less than what is anticipated from results in vitro. The in vitro cleavage activity of HIV-1 RNA-specific ribozymes has been shown not to correlate with their in vivo cleavage activity in human cell lines (Koseki et al, 1999; Crisell et al 1993; Ramezani and Joshi, 1996; Ventura et al, 1994; Domi et al, 1996; Kuwabara et al, 1999). A ribozyme targeted against the HIV-1 5' leader sequence, although active in vitro, was less active upon testing in HeLa and H9 cells (Koseki et al, 1999). A ribozyme against the first coding exon of the HIV-1 tat which possessed short flanking sequences performed better in vitro than ribozymes with longer flanking sequences (Crisell et al, 1993). However, upon testing in Jurkat cells, the opposite was the case. Similarly, a ribozyme targeted against the HIV-1 env coding region cleaved poorly in vitro, but demonstrated the highest inhibition against viral replication in the MT4 cell line (Ramezani and Joshi, 1996). On the other hand, ribozymes targeted against the HIV-1 R region (Ventura et al, 1994) or 5â&#x20AC;&#x2122; leader sequence (Domi et al, 1996) were catalytically inactive in vitro but were found to be active in a cellular environment. A dimeric maxizyme possessing a 2-bp common stem loop II demonstrated weak activity in vitro against the HIV-1 tat coding region, but in transiently transfected HeLa cells expressing a chimeric HIV-1 LTR and luciferase gene, luciferase activity was inhibited by up to 90% (Kuwabara et al, 1999). Thus, selection of ribozymes on the basis of their in vitro activity alone may eliminate molecules with increased therapeutic potential in vivo. In vivo systems are therefore required for screening ribozymes with increased/altered catalytic activities. The development of such screening systems should greatly accelerate ribozyme applications, for example in gene therapy. Ribozymes have been shown to be active in bacterial cells. A ribozyme targeted against the A2 coding region of RNA coliphage SP was tested in E. coli. Cells expressing this ribozyme produced less progeny phage than those expressing the inactive ribozyme (Inokuchi et al, 1994). Ribozyme cleavage of HIV-1 RNA target sites have also been demonstrated in bacterial cells. RNA containing the IN coding region of HIV-1 and a ribozyme targeted against it were expressed under control of the T7 promoter in bacteria producing T7 RNA polymerase (Sioud and Drlica, 1991). Upon induction, integrase mRNA could not be detected by analyzing RNA extracted from bacteria expressing the active ribozyme. However, it was present when an inactive ribozyme was expressed. Induction of target RNA synthesis prior to ribozyme induction led to the detection of one of the cleavage products. The amount of integrase protein produced in vivo was also shown to be decreased by Western blot analysis. Ribozymes targeted against the RT and pro coding regions within the HIV-1

RNA were also tested in E. coli expressing an RNA containing HIV-1 pro and RT coding regions (Ramezani et al, 1997). Trans cleavage of HIV-1 RNA was demonstrated by semi-quantitative RT-PCR and HIV-1 RT activity assay. However, although ribozyme activity against HIV-1 RNA could be demonstrated in both of these studies (Sioud and Drlica, 1991; Ramezani et al, 1997), the assays used were rather time consuming, and thus would not allow the fastest possible screening of ribozyme activity in vivo. We were interested in designing an E. coli based indicator cell system for rapid initial screening of active ribozymes without performing extensive biochemical characterizations. In the proposed bacterial indicator cell system (Figs. 1, 2), a ribozyme and its target site were cloned in frame within the lacZ open reading frame (ORF) present in the plasmid pGEM4Z, which gives rise to the " fragment of !-galactosidase. Accordingly, the lacZ transcript would contain the ribozyme and its target site in cis. Ribozymemediated cleavage of the target RNA would prevent its translation and thus production of the " fragment of !galactosidase. Complementation between the " fragment and the # fragment (expressed in certain E. coli strains) of !galactosidase would not occur. In the presence of a chromogenic substrate such as 5-bromo-4-chloro-3-indolyl-!D-galactoside (X-gal), !-galactosidase would catalyze the formation of 5-bromo-4-chloro-indigo, a blue-coloured product. If the enzyme is absent, the substrate would not break down and remain colourless. Thus, an effective ribozyme should lead to the formation of white, in contrast to blue, colonies on agar plates containing X-gal and isopropylthio-!D-galactoside (IPTG), an inducer of the lac operon.

FIG. 1A. Indicator cell system for monitoring ribozyme cleavage activity in vivo. "-complementation between the "-peptide produced from plasmids containing the N-terminal portion of the lacZ gene and bacteria which express the #-peptide leads to formation of blue colonies in agar plates with X-gal and IPTG.

110

Gene Therapy and Molecular Biology Vol 4, page 111

Fig 1 (Cont.) (1B) An oligonucleotide was designed which contained the ribozyme and its target sequence downstream. A 7-nt loop was placed between the ribozyme and its target sequence to allow folding and consequent hybridization of the two sequences. This loop (UUCGAAU) was designed so that it closely resembles a naturally occurring loop such as the tRNA anticodon loop (U/CUNNNG/AN; 31). Hind III and EcoR I sites were added on either side of the oligonucleotides to allow cloning between the Hind III-EcoR I sites of the plasmid pGEM4Z located at the lacZ gene. The oligonucleotide was thus Hind III-RzEnv -loop-Env-EcoR I. Additional nucleotides were added such that insertion by itself of the oligonucleotide would not affect the reading frame of the lacZ gene present in pGEM4Z. Upon in vitro transcription, the ribozyme cleaves its target site, thereby inactivating the lacZ mRNA. Because the "-peptide is not produced, "-complementation does not occur, which leads to formation of white colonies in agar plates with X-gal and IPTG. Ribozyme catalytic domain and 7-nt loop are shown in large case. Ribozyme flanking sequences and the target sequences to which they bind are shown in small case. ! denotes cleavage site.

111

Medina and Joshi: Screening HIV-1 RNA-specific ribozymes A yeast splicing protein was found to interact in vivo with a ribozyme and block its intracellular activity (Castanotto et al, 1998), whereas the nucleocapsid protein of HIV-1 (Tsuchihashi et al, 1993; Bertrand and Rossi, 1994; Herschlag et al, 1994; Moelling et al, 1994; Muller et al, 1994; Mahieu et al, 1995; Hertel et al, 1996), the heterogeneous nuclear ribonucleoprotein A1 (Bertrand and Rossi, 1994; Herschlag et al, 1994) and glyceraldehyde-3phosphate dehydrogenase (Sioud and Jespersen, 1996) were found to enhance ribozyme activity. Since hammerhead ribozymes are found in plant pathogens (viroid and satellite RNAs of viruses), plant proteins may also be found which could enhance ribozyme cleavage. Lack of complete cleavage both in vitro and in vivo in bacterial and mammalian cells may reflect the absence of proteins which enhance ribozyme activity. Thus, aside from the assessment of ribozyme cleavage activity in vivo, a bacterial system may also be used for cloning protein cofactors which could affect ribozyme activity in vivo.

II. Results A. Bacterial indicator system for identification of ribozymes capable of in vivo cleavage E. coli DH5" cells contain a portion of the lacZ gene which encodes for the # fragment of !-galactosidase. Transformation of these cells with plasmids expressing the " fragment of !-galactosidase leads to complementation between the " and # fragments and the consequent assembly of an active enzyme, whose activity can be detected by chromogenic substrates (FIG. 1A). A ribozyme (RzEnv ) was therefore designed to cleave the lacZ mRNA coding for the " fragment of !-galactosidase. This was achieved by cloning RzEnv and the env target sequence in frame within the lacZ gene of plasmid pGEM4Z. Upon transcription of this modified lacZ gene, lacZ mRNA would be produced which contains Rz Env and its env target sequence. If this mRNA remains intact, then the ribozyme must have been incapable of in vivo cleavage. This should lead to the formation of blue colonies on agar plates containing X-gal and IPTG. In contrast, if the conditions in vivo are suitable for cleavage, then the ribozyme should hybridize to its target located downstream and cleave it, effectively cutting the lacZ mRNA into two. Bacteria harbouring ribozymes capable of in vivo cleavage would not produce the " fragment of !-galactosidase and, as a result, would give rise to white colonies on plates containing X-gal and IPTG (FIG. 1B). The colour of the colonies should thus correlate with in vivo cleavage of the ribozyme target site present in the lacZ mRNA. A ribozyme's ability to cleave in vivo may therefore be easily and quickly assessed by monitoring the colour of the colonies which result after transformation in E. coli cells.

B. Ribozyme cloning, in vivo screening and characterization Oligonucleotides containing the ribozyme and its target

sequence were synthesized. A 7-nt loop was placed between the ribozyme and its target sequence to allow folding and consequent hybridization of the two sequences. This loop (UUCGAAU) was designed to closely resemble a naturally occurring loop, such as the tRNA anticodon loop (U/CUNNNG/AN; Stryer, 1988). Additional nucleotides were added so that insertion by itself of the oligonucleotide would not affect the reading frame of the lacZ gene present in pGEM4Z. After cloning, ligated plasmids were used to transform E.coli cells. Cells were then plated on X-gal/IPTG plates. Twenty-four colonies which ranged in colour from white to light blue were screened by restriction enzyme analysis and quickly assayed for !-galactosidase activity. Clones #4, #18 and #21 demonstrated correct restriction enzyme patterns and lower !-galactosidase activities compared to cells expressing the plasmid pGEM4Z. Colonies #4 and #21 were light blue, while #18 was white on LB agar plates containing X-gal and IPTG. !-galactosidase activities of extracts from all three colonies were consistently lower, compared to extracts from cells expressing pGEM4Z (Table 1). Table 1. !-galactosidase activity of pGEM4Z clones* clone # 4 18 21 pGEM4Z

colour of colony light blue white light blue dark blue

!-gal activity* 1.42 9.60 <0 23.14

*The values listed are the average of two experiments. Unit of !-galactosidase activity = 1000 x [A420 -(1.75 x A 550)] / (t x 0.1 x A 600), where t = time in minutes N.A., not applicable

Upon sequencing, all three clones were found to contain mutations (FIG. 3). Clone #4 contained an insertion (G) in the ribozyme flanking sequence. Clone #18 contained a substitution (G $ T) in stem loop II of the ribozyme catalytic domain and a deletion (C) in the ribozyme flanking sequence. Clone #21 contained an insertion (C) in the 7-nt loop connecting the ribozyme and the target sequence. Three additional clones that were picked and sequenced also contained mutations (data not shown). In a second set of experiment, twenty-four colonies picked from the ligation using the partially overlapping oligonucleotides were also characterized. Three colonies from this set were sequenced. Instead of single point mutations, tracts of mutated sequences were observed (data not shown). These could have resulted from mis-alignment of the partial overlap during the extension reaction performed prior to cloning.

C. Cis and trans cleavage activity in vitro of a cloned ribozyme Of the three clones selected, clone #21 contained a mutation in the loop region between the ribozyme and the

112

Gene Therapy and Molecular Biology Vol 4, page 113

FIG. 2. Overview of the selection procedure for colonies with reduced !-galactosidase activity. Selection of clones able to cleave in vivo was mediated by the chromogenic substrate X-gal which was added to agar plates. Lighter coloured clones expressing ribozymes were picked and grown in liquid culture, and used for an initial assay. Plasmid DNA isolated was subjected to restriction enzyme analysis. The cultures were re-streaked on agar plates. Isolated colonies were used in a !-galactosidase assay to confirm lack of lacZ mRNA expression.

FIG. 3. Sequences of the RzEnv clones. Mutations in the sequences of clones 4, 18 and 21 are indicated as $ for substitution, ! for insertion and " for deletion. The numbers in parentheses correspond to the clone # in which the mutation was found. Ribozyme catalytic domain and 7-nt loop are shown in large case. Ribozyme flanking sequences and the target sequences to which they bind are shown in small case. ! denotes cleavage site. The sequence of the 75-nt insert is shown at the bottom. Locations of the different mutations within the flanking sequences, ribozyme catalytic domain and the loop region are indicated by arrows.

then used in an in vitro transcription and cleavage reaction (FIGS. 4A, 4B). Cis cleavage occurred during the in vitro transcription reaction itself. This demonstrates that the ribozyme cloned in pGEM-RzEnv -Env #21 was functional in vitro.

target site. This mutation was not expected to affect ribozyme cleavage per se. The ribozyme and target site from pGEM-Rz Env -Env #21 were PCR amplified and the PCR products transcribed in vitro. The PCR product was

113

Medina and Joshi: Screening HIV-1 RNA-specific ribozymes Relative occurrence of cis and trans cleavage in vitro of RNA containing Rz Env -Env sequences was determined as follows. The RNA containing the RzEnv target site was transcribed separately and added to the in vitro transcription mixture of pGEM-RzEnv -Env #21, and the cis and trans cleavage products were analyzed by PAGE (FIG. 4C). Trans cleavage did not occur for up to 2 h incubation. Thus, only cis cleavage occurred under the conditions used for in vitro transcription. To determine whether the ribozyme possesses trans cleavage ability, ribozyme (without the cis target site) was PCR amplified from clone #21 and the PCR product transcribed in vitro. This RNA was then used in an in vitro trans cleavage reaction using a target RNA which was PCR amplified and transcribed separately. The ribozyme was able to cleave the target RNA in trans (FIG. 5). Thus, lack of trans cleavage in the presence of a cis target site (FIG. 4C) is due to the higher efficiency of cis cleavage.

X-gal and IPTG were identified. Lack of !-galactosidase activity within the bacterial cell extracts was confirmed by performing an assay using ONPG as a substrate (Table 1). Plasmid DNA from the clones was isolated and analyzed by restriction enzyme analysis. However, sequencing results revealed that mutations were present in the insert (FIG. 3). The mutations present in the clones may have caused formation of white colonies by disruption of the lacZ ORF. In addition, Rz Env could have cleaved its target site in vivo which could have further decreased the number of lacZ mRNAs available for translation of the " fragment of !-galactosidase. Therefore, the observed reduction in !galactosidase activity in these clones could be due to an additive effect between the mutations and ribozyme activity. However, because clones containing both an active ribozyme and a frameshift mutation were the only ones which reduced !-galactosidase activity to a detectable level, only these clones were selected. Clones containing the correct ribozyme and target sequence may have been missed, as these may have appeared blue on agar plates with X-gal/IPTG and therefore not selected for further analysis. As seen during the in vitro transcription and cleavage reaction using pGEM-RzEnv -Env #21, some of the RNA may have remained uncleaved in E. coli, which could then be used in translation. Using a similar blue/white colour selection, Chuah & Galibert (1989) could successfully demonstrate the activity of a cis cleaving ribozyme but not of a trans cleaving ribozyme. In this study, a ribozyme targeted to lacZ mRNA was cloned within the lacZ coding region of plasmid M 13mp8 to allow co-expression of the ribozyme and its target site within the same RNA molecule in vivo. Upon transcription, the ribozyme was expected to cleave the lacZ mRNA in cis. Out of 18 white plaques tested, 15 contained the correct ribozyme sequence, while 3 were due to cloning of aberrant sequences leading to the loss of the ORF. When the ribozyme was designed and expressed to trans cleave the lacZ RNA encoding # fragment of !galactosidase transcribed in E. coli from the episome, all of the isolated white plaques were due to the presence of incorrect sequences (Chuah & Galibert, 1989). In our study, all of the isolated white colonies were due to mutations. The discrepancy between our results and those by Chuah and Galibert (1989) could be due to a number of reasons. The ribozyme that we designed cleaved the 5' end of the lacZ mRNA coding for the " fragment. This may have been less effective in reducing the amount of protein produced than if the target chosen was further downstream as is the case in Chuah and Galibert's study (1989). The ribozyme used in our study may have been less active than the ribozyme used by Chuah and Galibert (1989). However, RzEnv was shown to cleave the lacZ mRNA in vitro (FIG. 4A, 4B); the majority of the RNA was cleaved in cis, suggesting that the design of the construct was appropriate. Since the ribozyme was in very close proximity to its target site, it is also unlikely that the ribozyme was bound to sequences other than its downstream target, forming an inactive complex.

III. Discussion Although in vitro selection techniques may allow the identification of ribozymes with improved catalytic activity, the in vivo performance of these ribozymes may not correlate with their activity in vivo. In vivo ribozyme activity may be rapidly assessed using a bacterial indicator system, provided that a strategy is designed which allows correlation of in vivo ribozyme activity with a bacterial phenotype. We attempted to test activity of the enzyme !galactosidase produced by lacZ mRNA to monitor ribozyme activity in vivo (FIG. 1). Sequences encoding RzEnv and its target site were cloned in cis within the Nterminal region of the lacZ gene in pGEM-4Z. Bacterial cells were then transformed with pGEM-RzEnv -Env plasmids. Upon transcription, RzEnv should have bound to and cleaved its target site, thereby inactivating the lacZ transcript coding for the " fragment of !-galactosidase. Absence of the " fragment should have prevented formation of a functional enzyme via complementation. White colonies likely to contain active RzEnv were identified on agar plates containing X-gal and IPTG. Ribozyme's ability to cleave in cis was demonstrated during in vitro transcription (FIGS. 4A and B). Upon addition of a target RNA containing RzEnv target site, only the products corresponding to cis cleavage were detected (FIG. 4C). However, this does not rule out the ability of RzEnv to cleave in trans. The ribozyme was indeed able to cleave the target RNA under trans cleavage conditions (FIG. 5). Thus, cis cleavage occurs with higher efficiency than trans cleavage. The use of a cis cleaving ribozyme is therefore a logical choice in establishing a bacterial indicator cell system. pGEM-RzEnv -Env plasmid designed to contain the ribozyme and its target sequence was used to transform E. coli cells. Colonies which had reduced !-galactosidase activity based on their colour on LB agar plates containing

114

Gene Therapy and Molecular Biology Vol 4, page 115

FIG. 4A,B. Cis and trans cleavage activity of RzEnv-Env. (4A, Left): Increasing amounts of PCR DNA (2, 5, 10, 25 or 50 µl) were used for in vitro transcription for 2 h. (4B, Right): PCR DNA (30 µl) was used for in vitro transcription and incubated at increasing time intervals (0.5, 1, 1.5 or 2 h). The full-length transcript (135 nts) along with the 5' (73 nts) and 3' (62 nts) cis cleavage products were detected.

Fig. 4C: Same as B, except that RNA (333 nts) containing the ribozyme target site was added to each transcription mixture. Products (170 nts and 163 nts) which would result from trans cleavage were not detected. Only the cis cleavage products (73 nts and 62 nts) and the full-length transcript (135 nts) were detected. T, target RNA alone. The uncleaved target RNA in lane “0.5” and 5’ cleavage product in lane “2” have been excised from the gel and used for subsequent experiments.

FIG. 5. Trans cleavage activity of pGEM-RzEnv. RzEnv and ["-32P]-labelled target RNA were used in a trans cleavage reaction. Aliquots were taken at the indicated time intervals and analyzed by 8 M - 8% polyacrylamide gel electrophoresis followed by exposure to a phosphor screen and scanning by Storm phosphorimager (Molecular Dynamics; Sunnyvale, USA).

115

Medina and Joshi: Screening HIV-1 RNA-specific ribozymes 5 units of Klenow (Life Technologies; Burlington, Canada). The fill-in products were ethanol-precipitated and resuspended in water and digested with Hind III and EcoR I. In a second set of experiment, complementary oligonucleotides with 5' overhangs to allow cloning (5'-AGC-TTG-GAT-CCaatc-gca-aCT-GAT-GAG-TCC-GTG-AGG-ACG-AAa-cca-gccgtt-cga-atc-ggc-tgg-ttt-tgc-gat-tCG-3' and 5’-AAT-TCG-aatcgc-aaa-acc-agc-cga-ttc-gaa-cgg-ctg-gtT-TCG-TCC-TCACGG-ACT-CAT-CAG-ttg-cga-ttG-GAT-CCA-3’) were synthesized. Ribozyme catalytic domain is in uppercase bold, 8-nt flanking sequences complementary to either side of the cleavage site are in lowercase, target sequence is in lowercase bold, loop sequence is in lowercase italics, restriction enzyme sites are in uppercase italics, and 5' overhangs are underlined. The full-length oligonucleotides (75 nts) are of comparable length to the Hind III-EcoR I fragment (54 nts) being removed from pGEM4Z. Plasmid pGEM4Z (Promega Corp.; Madison, USA) was transformed into E. coli strain DH5", isolated by a miniprep procedure and digested with Hind III and EcoR I. The DNA band corresponding to the EcoR I-Hind III fragment was eluted using the Geneclean kit (BIO 101; Vista, USA) following 1% agarose gel electrophoresis. Full-length oligonucleotides were then cloned as described in (Sambrook et al, 1989) at the Hind III and EcoR I sites within the lacZ gene of pGEM4Z. Ligation reactions (10 µl) containing 50 mM Tris-Cl (pH 7.6), 10 mM MgCl2, 1 mM ATP, 1 mM DTT, 5% (w/v) polyethylene glycol-8000, double stranded insert (300-5000 ng), vector (10 ng) and 1 unit T4 DNA ligase (Life Technologies; Burlington, Canada) were performed at 23 °C for 1h. DH5" competent cells were transformed with the ligation mix and plated on Luria-Bertani (LB) agar plates containing ampicillin (50 µg/ml), X-gal (800 µg) and IPTG (0.4 µmol). A positive transformation control consisting of pGEM4Z DNA yielded over 300 colonies, a negative transformation control without DNA yielded no colonies, and the ligation mixtures each yielded ~50 colonies. Colonies which ranged in size and colour from white to light shades of blue were picked and screened by Csp 45 I, Dra I, Sma I, BamH I, EcoR I and Hind III restriction enzyme analyses. DNA sequencing was performed using the T7 Sequencing Kit (Pharmacia Biotech Inc.; Baie d’Urfé, Canada) using instructions provided by the supplier.

Another possibility is that the white plaques obtained by Chuah and Galibert (1989) may have been due to mutations which occurred elsewhere in the cloning vector and not in the insert. Also, the substrate (X-gal) concentration we have used in our system (800 µg Xgal/plate) was higher than the amount used by Chuah and Galibert (4 µg X-gal/plate). Thus, it is conceivable that our system is too sensitive, allowing small amounts of !galactosidase to produce a detectable blue-coloured product. Ribozymes tested against HIV-1 pro (Ramezani et al, 1997), RT (Ramezani et al, 1997) and IN (Sioud and Drlica, 1991) coding regions were shown to be active in E. coli. However, in these studies ribozyme activities were demonstrated by assays that relied on the presence of cleaved RNA and their translation products. On the other hand, the system we and Chuah and Galibert (1989) have utilized detected the presence of uncleaved products. Thus, although the majority of the lacZ RNA may have been cleaved in vivo, protein translated from the remaining uncleaved transcripts catalyzed the breakdown of the substrate to a blue coloured product, which could still be detected by the assays used. As such, the blue/white colour selection may not accurately report the in vivo cleavage activity of a ribozyme, since colonies containing mutations were the only ones that could be isolated in our study. For successful development of a ribozyme screening system, the amount of substrate used may have to be titrated for each ribozyme. However, this may be time-consuming. Alternatively, substrates may have to be used which have a higher cut-off limit of detection, requiring a higher amount of !-galactosidase before a colour change is observed. Thus, only those cells which are producing high amounts of !-galactosidase may turn blue.

IV. Materials and methods A. Oligonucleotide design and cloning of RzEnvEnv into pGEM4Z Cloning of sequences encoding RzEnv (Medina and Joshi, 1999) and its HIV-1 env target site was performed using two sets of oligonucleotides. The first set consisted of partially overlapping oligonucleotides (53-54-nt) which were first extended in vitro and then cloned. The second set consisted of two complementary oligonucleotides that contained ribozyme and target sequences flanked by restriction sites, which were cloned directly into the plasmid pGEM4Z. Both ligations yielded >100 colonies upon transformation into E. coli. RzEnv was designed to cleave after a highly conserved GUU (nt 665 to nt 667) sequence within the env coding region of HIV-1 HXB2 RNA (Myers et al, 1995). Partially overlapping oligonucleotides (5'-CCC-CCC-AAG-CTT-GGA-TCC-aat-cgcaaC-TGA-TGA-GTC-CGT-GAG-GAC-GAA-acc-agc-3' and 5'GGG-GAA-TTC-Caa-tcg-caa-aac-cag-ccg-att-cga-acg-gctggt-TTC-GTC-CTC-AC-3') were synthesized using the Expedite% Nucleic Acid Synthesis System (Millipore; Etobicoke, Canada). Before cloning, these oligonucleotides were extended to full-length complementary oligonucleotides for 1 h at 37°C in a 40 µl reaction containing 50 mM Tris-Cl (pH 8.0), 10 mM MgCl2, 50 mM NaCl, 8µM of each dNTP, and

&. !-galactosidase activity of individual pGEMRzEnv-Env clones Individual colonies were picked and grown overnight in LB medium containing ampicillin (50 µg/ml). The next day, !galactosidase activity was assayed using cultures at an optical density at wavelength of 600 (OD 600) equivalent to 1.00. The cultures were incubated for another 4 h after adding IPTG (0.1 µmol) and X-gal (200 µg). Cells were pelleted by spinning for 3 min at 8000g. OD 550 and OD 420 of the supernatants were measured to quickly assess !-galactosidase activity using culture medium containing IPTG and X-gal as a blank. For clones selected for further characterization, !galactosidase activity was assayed using o-Nitrophenyl-!-Dgalactopyranoside (ONPG) as substrate (30) and LB cultures at the logarithmic growth phase with OD 600 values between 0.280.70. After cooling on ice for 20 minutes, cell cultures (100 µl) were mixed with 50 µl 0.1% SDS, 100 µl chloroform and 900 µl Z buffer (60 mM Na 2HPO4, 40 mM NaH 2PO4, 10 mM KCl, 1 mM MgSO4, 50 mM !-mercaptoethanol), vortexed for 10 seconds, and incubated at 28°C for 5 minutes. ONPG (4

116

Gene Therapy and Molecular Biology Vol 4, page 117 mg/ml, 200 µl) was added to each tube, and the incubation continued for 80 to 220 min at 37°C. Reactions were stopped by adding 500 µl of 1M Na 2CO3. OD 550 and OD 420 were then measured.

References Bertrand,E.L. and Rossi,J.J. (1994) Facilitation of hammerhead ribozyme catalysis by the nucleocapsid protein of HIV-1 ribozyme catalysis by the nucleocapsid protein of HIV-1 and the heterogeneous nuclear ribonucleoprotein A1. EMBO J. 13, 2904-2912 Birikh,K.R., Heaton,P.A. and Eckstein,F. (1997) The structure, function and application of the hammerhead ribozyme. Eur. J. Biochem. 245, 1-16 Castanato,D., Li,H., Chow,W., Rossi,J.,J. and Deshler,J.,O. (1998) Structural similarities between hammerhead ribozymes and the spliceosomal RNAs could be responsible for lack of ribozyme cleavage in yeast. Antisense Nucl Acid Drug Dev. 8, 1-13 Chuat,J. and Galibert,F. (1989) Can ribozymes be used to regulate prokaryote gene expression? Biochem Biophys Res Comm. 162, 1025-1029 Crisell,P., Thompson ,S. and James,W. (1993) Inhibition of HIV-1 replication by ribozyme that show poor activity in vitro. Nucl. Acid Res. 21, 5251-5255 Domi,A., Beaud,G. and Favre,A. (1996) Transcripts containing a small anti-HIV hammerhead ribozyme that are active in the cell cytoplasm but inactive in vitro as free RNAs. Biochimie 78, 654-662 Haseloff,J. and Gerlach,W.L. (1988) Simple RNA enzymes with new and highly specific endoribonuclease activities. Nature 334, 585-591 Hershlag,D., Khosla,M., Tsuchihaba,Z. and Karpel,R.L. (1994) An RNA chaperone activity of non-specific RNA binding proteins in hammerhead ribozyme catalysis. EMBO J. 13, 2913-2924 Hertel,K.J., Hershlag,D. and Ulenbeck,O.,C. (1996) Specificity of hammerhead ribozyme cleavage. EMBO J. 15, 37513757 Inokuchi,Y., Yuyama,N., Hirashima,A., Nishikawa,S., Ohkawa,J. and Taira,K. (1994) A hammerhead ribozyme inhibits the proliferation of an RNA coliphage SP in E. coli. J Biol Chem. 269, 11361-11366 Joshi,S. and Joshi,R.L. (1996) Molecular biology of human immunodeficiency virus type-1. Transfus Sci. 17, 351378 Koseki,S., Tanabe,T., Tani,K., Asano,S., Shioda,T., Nagai,Y., Shimada,T., Ohkawa,J. and Taira,K. (1999) Factors governing the activity in vivo of ribozymes transcribed by RNA polymerase III. J. Virol. 73, 1868-1877 Kuwabara,T., Warashina,M., Nakayama,A., Ohkawa,J. and Taira,K. (1999) tRNAVal-heterodimeric maxizymes with high potential as gene inactivating agents: simultaneous cleavage at two sites in HIV-1 tat mRNA in cultured cells. Proc. Natl. Acad. Sci. USA 96, 1886-1891 Macpherson,J.L., Ely,J.A., Sun,L.Q. and Symonds,G.P. (1999) Ribozymes in gene therapy of HIV-1. Front Biosci. 1, D497-505 Mahieu,M., Hendrix,C., Ooms,J., Herdewijn,P. and Conent,J. (1995) In vitro NCp7 enhancement of ribozyme-mediated cleavage of full-length human IL-6 mRNA. Biochem Biophys Res Comm. 214, 36-43 Medina,M. and Joshi,S. (1999) Design and Characterization of tRNA3Lys-based hammerhead ribozymes. Nucl Acids Res. 27, 1698-1708

C. Cis and trans cleavage activity of RzEnv-Env To detect cis cleavage activity, ribozyme and HIV-1 env target site were PCR amplified from pGEM-RzEnv -Env #21 using a forward primer (5-CGA-AAT-TAA-TAC-GAC-TCACTA-TA-3') which binds to the T7 promoter and a reverse primer (5'-GTA-AAA-CGA-CGG-CCA-GT-3') which binds downstream of the ribozyme target site. PCRs were performed for 30 cycles (1 min, 95 oC; 1 min, 56 oC; 1 min, 72 oC each). PCR DNA (2-50 µl) containing the T7 promoter sequence was transcribed in vitro at 37°C in a reaction mixture (100 µl) containing 40 mM Tris-Cl (pH 8.0), 25 mM NaCl, 8 mM MgCl 2, 2 mM spermidine, 5 mM DTT, 1 mM of each NTP, and 200 units of T7 RNA polymerase (Life Technologies; Burlington, Canada). The reaction was stopped after 0.5-2 h by digesting the template DNA with 5 units of RQI RNase-free DNase (Promega Corp.; Madison, USA) for 10 min. Cis cleavage at the HIV-1 env target site by Rz Env occurred under the condition used for transcription, without further incubation or addition of reagents. To compare cis and trans cleavage activities, the env target sequence was PCR amplified from the plasmid pHEnv using a primer (5’-ATA-TCA-TATGTA-ATA-CGA-CTC-ACT-ATA-GGG-CGA-GTG-CAG-AAAGAA-TAT-GC-3’) which binds upstream of the ribozyme target site and contains the T7 promoter sequence and a primer (5’GTC-CGT-GAA-ATT-GAC-AG-3’) which binds downstream of the ribozyme target site. PCR DNA was transcribed in vitro for 2 h at 37°C as described above. After phenol extraction and ethanol precipitation, the target RNA was resuspended in water and then added to the in vitro transcription mixture. The cleavage products were analyzed by 8 M urea-8 % polyacrylamide gel electrophoresis (PAGE) followed by methylene blue staining (Sambrook et al, 1989).

D. Trans cleavage activity of RzEnv Target RNA was transcribed in the presence of ["-32P] UTP (3000 Ci/mmol; Amersham Canada Ltd.; Oakville, Canada). RzEnv was transcribed from PCR DNA which was amplified from pGEM-RzEnv -Env #21 using a forward primer (5-CGA-AAT-TAA-TAC-GAC-TCA-CTA-TA-3') which binds to the T7 promoter and a reverse primer (5'-ATA-TAT-ATC-GATAAA-AAA-CGG-CTG-GTT-TCG-TCC-TC-3') which binds near the 3' end of the ribozyme. It was then used in a trans cleavage reaction with ["-32P]-labelled target RNA. Essentially, Rz Env and target RNA were combined in a reaction mixture containing 40 mM Tris-Cl (pH 8.0) and 10 mM NaCl. The sample was heated to 65°C for 5 min, cooled to 37°C, and the reaction initiated by adding 20 mM MgCl 2. Aliquots were taken after 7, 15, 30, 120, 300, 600 and 900 min incubation at 37 °C, and the reaction stopped by addition of loading buffer containing 5 mM EDTA. Cleavage products were analyzed by 8 M urea-8 % PAGE followed by exposure to a phosphor screen and scanning by Storm phosphorimager (Molecular Dynamics; Sunnyvale, USA).

Acknowledgements This work was supported by grants from the National Health and Research Development Program and Medical Research Council of Canada. 117

Medina and Joshi: Screening HIV-1 RNA-specific ribozymes Moelling,K., Mueller,G., Dannull,J., Reuss,C., Beimlimg,P., Bartz,C., Wiedenmann,B., Yoon,K., Surovoy,A. and Jung,G. (1994) Stimulation of Ki-ras ribozyme activity by RNA binding protein, NCp7, in vitro and in pancreatic tumor cell line, capan 1. Ann NY Acad Sci. 733, 113-121 Muller,G., Strack,B., Dannull,J., Sproat,B.,S., Surovoy,A., Jung,G. and Moelling,K. (1994) Amino acid requirements of the nucleocapsid protein of HIV-1 for increasing catalytic activity of a Ki-ras ribozyme in vitro. J Mol Biol. 242, 422429 Myers,G., Hahn,B.,H., Mellors,J.,W., Henderson,L.,E., Korber,B., Jeang,K., McCutchan,F.,E. and Pavlakis,G.,N. (1995) Human Retroviruses and AIDS. Los Alamos National Laboratory, Los Alamos, New Mexico. Pan,T. (1997) Novel and variant ribozymes obtained through in vitro selection. Curr Opin Chem Biol. 1, 17-25 Ramezani,A. and Joshi,S. (1996) Comparative analysis of five highly conserved target sites within the HIV-1 RNA for their susceptibility to hammerhead ribozyme-mediated cleavage in vitro and in vivo. Antisense Nucl Acid Drug Dev. 6, 229235 Ramezani,A. and Joshi,S. (1999) Development of hammerhead ribozymes for HIV-1 gene therapy: Principles and progress. Gene Ther Mol Biol. 3, 1-10 Ramezani,A., Marhin,W., Weerasinghe,M. and Joshi,S. (1997) A rapid and efficient system for rapid screening of HIV-1 Pol mRNA-specific ribozymes. Can J Microbiol. 43, 93-96 Ruffner,D.,E., Stormo,G.,D. and Uhlenbeck,O.,C. (1990) Sequence requirement of the hammerhead RNA selfcleavage reaction. Biochemistry 29, 10695-10702 Sambrook,J., Fritsch,E.,F. and Maniatis,T. (1989) Molecular Cloning, A Laboratory Manual. 2nd Ed., Cold Spring Harbor university Press, Cold Spring Harbor, NY Sioud,M. and Drlica,K. (1991) Prevention of human immunodeficiency virus type 1 integrase expression in Escherichia coli by a ribozyme. Proc Natl Acad Sci USA 88, 7303-7307 Sioud,M. and Jespersen,L. (1996) Enhancement of hammerhead ribozyme catalysis by glyceraldhyde-3-phosphate dehydrogenase. J Mol Biol. 257, 775-789 Stryer,L. (1988) Biochemistry, 3rd Ed., W.,H.,Freeman and Company, New York, NY Tsuchihashi,Z., Khosla,M. and Herschlag,D. (1993) Protein enhancement of hammerhead ribozyme catalysis. Science 262, 99-102. Uhlenbeck,O.,C. (1 9 8 7 ) A small catalytic oligoribonucleotide. Nature 328, 596-600 Vaish,N.,K., Kore,A.,R. and Eckstein,F. (1998) Recent developments in the hammerhead ribozyme field. Nucl Acids Res. 26, 5237-5242 Ventura,M., Wang,P., Franck,N. and Saragosti,S. (1994) Ribozyme targeting of HIV-1 LTR. Biochem Biophys Res Comm. 203, 889-898

Sadhna Joshi 118

Gene Therapy and Molecular Biology Vol 4, page 119 Gene Ther Mol Biol Vol 4, 119-132. December 1999.

Direct redox modulation of p53 protein: potential sources of redox control and potential outcomes Review Article

Hsiao-Huei Wu1 , Mark Sherman2 , Yate-Ching Yuan3 , and Jamil Momand1, * 1

Dept. of Cell and Tumor Biology, Beckman Research Institute of the City of Hope, 1450 E. Duarte Road, Duarte CA 910103000 2 Dept. of Biology, Beckman Research Institute of the City of Hope, 1450 E. Duarte Road, Duarte CA 91010-3000 3 Dept. of Biomedical Informatics, Beckman Research Institute of the City of Hope, 1450 E. Duarte Road, Duarte CA 910103000 ______________________________________________________________________________________________________ *

Correspondence: (Present address) Jamil Momand, Ph.D., Department of Chemistry, California State University at Los Angeles, 5151 State University Drive, Los Angeles CA 90032. Tel: 323-343-2361; Fax: 323-343-6490; E-mail: Jmomand@calstatela.edu Abbreviations: ROI, reactive oxygen intermediates; Ref-1, Redox factor-1; DTT, dithiothreitol; ESR, electron spin resonance; PDTC, pyrrolidine dithiocarbamate; GSH, glutathione; SOD, superoxide dismutase; TPEN, N,N,N',N'-tetrakis(2-pyridylmethyl)-ethylenediamine Key Words: oxidation, cysteine, pyrrolidine dithiocarbamate, stressor, sulfhydryl Received: 29 April 1999; accepted: 17 May 1999

Summary Appropriate response to environmental stressors is essential for life. Many stressors, such as UV light, ionizing radiation, reactive oxygen intermediates (ROI), heat shock and hypoxia alter the redox potential of the cell. Recently, it has been shown that some of these stressors promote direct oxidation of specific protein cysteine residues resulting in either up-regulation or down-regulation of protein activity in the cytosol. In higher eukaryotes, the p53 tumor suppressor gene is a central component of stress response and its activation results in either cell cycle arrest or apoptosis. In cultured cells, p53 appears to become activated by some stressors (hydrogen peroxide, heat) predicted to directly increase cellular redox potential. However, in vitro studies indicate that p53 protein oxidation inhibits its ability to bind its consensus sequence DNA. If p53 is unable to bind consensus sequence DNA, p53 is predicted to be incapable of activating the p21WAF1/CIP1 gene, responsible for mediating G1 cell cycle arrest. Two proteins previously shown to reduce oxidized cytoplasmic proteins, Redox factor-1 and thioredoxin reductase, have been shown to play important roles in maintaining p53 activity, suggesting that they may be responsible for keeping p53 in the reduced state inside the cell. Analysis of the p53 crystal structure revealed several well-conserved cysteine residues exposed on the protein surface that may be susceptible to oxidation. Based on this analysis we predict that cysteine residues 124, 176, 182, 242 and 277 are primary candidates for redox regulation. In this communication, we review the data demonstrating p53 regulation by direct alteration of p53 cysteine residue oxidation, propose a testable mechanism by which p53 oxidation may occur, and discuss the possible implications of p53 oxidation on cell growth control and DNA repair.

Once activated, the intracellular p53 protein level increases and p53 binds, in sequence-specific fashion, to certain DNA promoters which, in turn, leads to activation of genes that mediate cell cycle arrest (El-Deiry et al., 1993; Chin et al., 1997; Hermeking et al., 1997; Bunz et al., 1998) or apoptosis (Miyashta and Reed, 1995). One of the p53 responsive genes that appears necessary for mediating G1 arrest in several cell types is p21WAF1/CIP1 , a cyclin-dependent kinase inhibitor (El-Deiry et al., 1993; Harper et al., 1993). Activation of p53 is complex and inhibition of this process can lead to loss of cell growth control.

I. Introduction The p53 tumor suppressor gene is one of the most frequently mutated genes in human cancers (Baker et al., 1989; Nigro et al., 1989; Hainaut et al., 1998). It is a cell cycle checkpoint gene responsible for committing mammalian cells to a growth arrest phenotype or apoptosis in response to genotoxic and non-genotoxic stressors (Levine, 1997; Giaccia and Kastan, 1998). The p53 gene encodes a transcription factor that is synthesized in a latent form and can be activated by a wide range cell stressors.

119

Wu et al: Redox modulation of p53 be oxidized at cysteine residues in cells and the effect of oxidation on their activities. For illustrative purposes, a few of these will be discussed. Three types of chemical oxidation have been identified on protein cysteine sulfhydryl groups in cells. The first type of oxidation is intramolecular disulfide bond formation. This was shown to occur on two bacterial proteins, OxyR and Hsp33 (Zheng et al., 1998; Jakob et al., 1999). Treatment with hydrogen peroxide or heat leads to the formation of intramolecular disulfide bonds and results in the activation of the protein as a transcription factor, in the case of OxyR, or a chaperone protein, in the case of Hsp33. Oxidized OxyR transactivates a panel of genes responsible for protecting the organism from hydrogen peroxide poisoning including hydroperoxidase I, alkyl hydroperoxidase reductase and glutathione reductase (Jamieson and Storz, 1997). Oxidized Hsp33 protects enzymes from denaturing during heat stress or hydrogen peroxide treatment (Jakob et al., 1999).

Activation of p53 is thought to take place at both the translational and post-translational level. Most recent work has concentrated on understanding the post-translational events that lead to p53 activation. The exact activation pathway is highly dependent on the type of stressor applied. Each class of stressors appears to result in a unique pattern of p53 protein phosphorylation and acetylation to achieve p53-mediated transcription of appropriate downstream targets (Giaccia and Kastan, 1998). Concomitant with these modifications p53 protein levels increase. Another potential post-translational modification system less extensively explored is direct p53 redox regulation. Protein redox alterations in response to environmental agents, was proposed to occur more than four decades ago (Barron, 1951), but it is only within the past 5 years that, with the advent of new techniques, solid evidence has accumulated indicating that redox changes can occur on cytoplasmic proteins in vivo (Ă&#x2026;slund and Beckwith, 1999). Table 1 lists a partial set of cytoplasmic proteins that have been shown to

120

Gene Therapy and Molecular Biology Vol 4, page 121

Figure 1. Schematic diagram of three basic p53 protein functional domains in human p53. Primary protein sequence of p53. The numbered serine (S) and lysine (K) residues are sites of phosphorylation and acetylation respectively. All cysteine residue positions are shown.

modifications, thus far, have been shown to occur in the Nand C-terminal domains. The N-terminal 42 amino acid residues is required for p53-mediated transactivation. For p53 to mediate transcription, this domain must bind hTAFII31 and hTAFII70 accessory transcription factors that form part of the TATA box binding protein complex TFIID. Six serine residues within the N-terminal domain can be phosphorylated. Phosphorylation at Ser 15 is required, in some instances, to upregulate p53 protein levels (Shieh et al., 1997). The C-terminal domain (301-393) includes the tetramerization domain, the nuclear localization sequence and a nuclear export sequence. There are four serine residues regulated by phosphorylation and two lysine residues regulated by acetylation in the C-terminal domain. The central domain of p53 (residues 100-300) can bind DNA in a sequence-dependent manner at two palindromic halfsites with the sequence 5'-PuPuPuC(A/T)(T/A)GPyPyPy-3'. Mutations in p53 observed in human tumors and malignancies almost always map to this central DNA binding domain. Proteins with missense mutations in this domain are usually incapable of binding p53 consensus sequence containing DNA elements. All conserved cysteine residues are observed in this domain and, in some cases, Cys residues are mutated in cancers. The frequency of the mutations is fairly low and therefore, cysteine residues are not considered as mutation hot spots (Sun and Oberley, 1996). Direct redox changes on p53 itself would likely affect one or more of these cysteine residues. Given their close proximity to the DNA binding domain it is likely that oxidation of some critical cysteine residues will affect sequence-specific binding activity of p53.

The second type of chemical oxidation involves the oxidation of a metal-sulfur center active site. The bacterial transcription factor SoxR falls into this category (Ding et al., 1996). Like OxyR, SoxR activates genes responsible for bacterial oxidation defense. Target genes upregulated by SoxR include Mn-containing superoxide dismutase, glucose6-phosphate dehydrogenase and the DNA repair enzyme endonuclease IV. Aside from intramolecular disulfide bond formation and oxidation of metal sulfur centers, a third type of chemical oxidation that occurs in proteins is a disulfide bond formed between a protein cysteine residue and a small molecular weight thiol molecule. Some enzymes have been shown to form a disulfide bond with glutathione (Thomas et al., 1994; Cabiscol and Levine, 1996) and cysteine (Sato et al., 1996). Recently, it was shown that p53 is oxidized in cultured cells in the presence of a metal chelator/oxidant, called pyrrolidine dithiocarbamate (PDTC) (Wu and Momand, 1998). Oxidation of p53 correlated with inhibition of its transactivation activity. The chemical nature of the p53 cysteine residue oxidation is not known nor, aside from PDTC, are the types of other stressors that lead to p53 oxidation. In this review, we hope to shed light on the possible reactive cysteine residues within p53 and the role direct redox regulation plays in modulating this protein's function. To understand how oxidation might affect p53 function it is important to review, briefly, the location of the different functional domains of p53 in relation to the p53 cysteine residues (for extensive reviews on this subject see Gottlieb, 1996; Greenblatt, 1994). As shown in Fig. 1, the p53 protein can be divided roughly into three distinct domains based on function. All regulatory post-translational 121

Wu et al: Redox modulation of p53 increase intracellular hydrogen peroxide or hydroxyl radical concentration can lead to elevation in p53 protein levels, p53 nuclear accumulation, p53-dependent cell growth arrest, and p53-dependent cell death in some cell types.

II. p53 activity is upregulated in response to agents that increase intracellular reactive oxygen intermediates Protein disulfide formation can occur in the cytosol when intracellular reactive oxygen intermediates (ROI) are created (Halliwell and Gutteridge, 1989). ROI have long been thought to be part of the multitude of small intracellular molecules that signal specific transduction pathways within the cell. The three common types of intracellular ROI postulated to be important for modulating protein redox levels are hydrogen peroxide, hydroxyl radicals and superoxide. Specific enzymes have evolved to rid the cell of hydrogen peroxide and superoxide and these enzymes are, in fact, regulated in response to these molecules. Intracellular peroxides are generated at sites of inflammation by secretion of hydrogen peroxide by neutrophils (Vile et al., 1998). Hydroxyl radicals are generated in cells in response to UV light, ionizing radiation and free metals. Interestingly, doxorubicin, a common chemotherapeutic agent used in the treatment of tumors, also appears to generate hydroxyl radicals, that may contribute to its anti-neoplastic properties (Doroshaw, 1986). Both genetic and cell biology studies suggest that p53 can be activated by stimuli that also result in intracellular ROI production. Solar UV radiation leads to cellular p53 protein elevation and agents that scavenge hydroxyl radicals prevent p53 protein elevation (Vile, 1997). In cultured normal human fibroblasts, treatment with Cd, a metal known to catalyze the formation of hydroxyl radicals, or hydrogen peroxide leads to accumulation of p53 in the nucleus, a sign of p53 activation in some cells (Sugano et al., 1995; Uberti et al., 1999). In one study, normal human fibroblasts treated with a sublethal dose of hydrogen peroxide underwent long term growth arrest, suggestive of p53 activation (Chen et al., 1998). In IMR-90 fetal lung cells, p53 and p21WAF1/CIP1 protein levels were transiently elevated in response to hydrogen peroxide (Chen et al., 1998). Hydrogen peroxidemediated upregulation of p53 protein was inhibited by the iron chelator deferoxamine, suggesting that intracellular hydroxyl radical formation, perhaps generated by Fentontype chemical reactions, is an important component of this signaling pathway. When the viral oncoprotein E6, a p53inhibiting protein, was expressed in IMR-90 cells, H2O2 treatment failed to upregulate p53 levels or to induce G1 arrest and there was a diminution in the level of p21WAF1/CIP1 increase. These studies suggest that H2O2 and most likely hydroxyl radicals are important intracellular molecules that lead to p53 activation and cell cycle arrest. Increases in intracellular ROI levels can also lead to p53-dependent programmed cell death. Using murine embryo fibroblasts cells derived from p53 -/- mice, it was shown that hydrogen peroxide leads to p53-dependent cell death (Yin et al., 1998). Similarly, normal human fibroblasts engineered to express E6 failed to undergo programmed cell death while normal human fibroblasts without E6 underwent programmed cell death in response to hydrogen peroxide suggesting that p53 is required for cell death (Yin et al., 1999). In summary, agents known to

III. Studies on redox regulation of p53 in vitro One of the possible mechanisms of p53 activation by ROI is direct oxidation of the p53 protein. Oxidation may activate p53 for cell cycle arrest and apoptosis. However, to date, all evidence from p53 oxidation studies conducted in vitro indicates that sequence specific DNA binding is inhibited by p53 oxidation (Hupp et al., 1992; Hainaut and Milner, 1993; Delphin et al., 1994; Sun and Oberley, 1996). Evidence that p53 cysteine residue oxidation can prevent p53 from properly binding its DNA consensus sequence comes from the fact that: (i) high concentrations of dithiothreitol (DTT) are required (0.1-10 mM) to allow recombinant p53 or p53 in nuclear extracts to bind DNA; (ii) treatment of purified recombinant p53 with the thiol alkylating agent N-ethyl maleimide (Rainwater et al., 1995) or in vitro translated p53 with diamine (Hainaut and Milner, 1993) prevents p53 from binding to its DNA consensus sequence. Thus, it appears that maintenance of p53 cysteine residues in the reduced state is necessary for optimal p53 consensus sequence-dependent DNA binding. Another assay to test whether ROI may modulate p53 activity is the p53-transactivation assay. In this assay, a plasmid expressing p53 is cotransfected with a plasmid encoding a p53-responsive element placed upstream of a gene that codes for a transcription reporter. In one report it was demonstrated, using this assay, that H2O2 inhibited p53mediated transactivation (Parks et al., 1997) consistent with the data demonstrating that oxidized p53 fails to bind DNA in vitro. This result, at the outset, appears inconsistent with data demonstrating that H2O2 treatment correlates with an increase in p53 protein and transactivation of the p21WAF1/CIP1 gene (Chen et al., 1998). Indeed, nuclear extracts from cells treated with H2O2 contain higher levels of p53 sequence specific DNA binding activity than untreated cells (Verhaegh et al., 1997). A possible explanation for this apparent contradiction is that H2O2 treatment of cells may directly oxidize p53 and, in addition, may lead to higher levels of p53 protein. The oxidized p53 is expected to be incapable of binding p53dependent effector genes in vivo. However, it must be kept in mind that nuclear extracts derived from H 2O2-treated cells often include DTT in the DNA-binding buffer. In this case, the high levels of oxidized p53 may be rapidly converted to a reduced form that can bind consensus sequence containing DNA. Rapid reduction of p53 by DTT may explain why the p53 DNA binding capacity appears higher in H2O2-treated cells. Notwithstanding this argument, one must still explain the apparent p53-dependent upregulation of p21WAF1/CIP1 observed in cultured cells after H 2O2 treatment. It is possible that H 2O2 treatment immediately leads to high levels of transcriptionally inactive p53. After initial oxidation of p53, the oxidized cysteine residues on p53 may be reduced by 122

Gene Therapy and Molecular Biology Vol 4, page 123 specific enzymes that are also activated by H2O2. Upon reduction, p53 may then upregulate p21WAF1/CIP1 . If this prediction is correct, one would expect that H 2O2 treatment would lead to delayed activation of p21WAF1/CIP1 . Such a prediction is consistent with the fact that H 2O2 treatment of cells leads to an increase in p53 protein at 1.5 h posttreatment and to p21WAF1/CIP1 increase at 18 h post-treatment (Chen et al., 1998). If this scenario is correct then the molecules required to maintain p53 in a reduced state may, in some instances, be limiting within the cell. This conjecture is supported, to some extent, by the fact that the DNA-binding activity of recombinant mouse p53 in freshly prepared nuclear extracts from baculovirus-infected insect cells is stimulated by treatment with DTT (Delphin et al., 1994). It is possible that p53 protein is overexpressed in insect cells relative to the reducing molecules needed to keep the p53 in the reduced state. This model raises the question of whether enzymes involved in reducing oxidized protein cysteine residues affect p53 activity.

p53 (p53 lacking its 30 C-terminal amino acid residues) is used in the DNA-binding assay. This is somewhat unexpected because there are no cysteine residues within this C-terminal region of p53. However, the C-terminus does appear to normally negatively regulate the sequence specific DNA binding function of p53 (Hupp et al., 1992, 1993; Hupp and Lane, 1994). Ref-1 and p53 do not form a stable complex regardless of whether DNA is present. It is possible, then, that Ref-1 transiently associates within the terminal 30 amino acid residues of p53 and reduces oxidized p53 cysteine residues within the central-DNA binding domain of p53. Importantly, Ref-1 was observed to increase p53 transactivation activity in transient expression assays. When the Ref-1 endonuclease domain was removed, Ref-1's ability to stimulate p53 DNA binding activity was severely inhibited but not completely abolished. The data indicate that Ref-1 may stimulate p53 DNA binding activity through both, a non-redox and a redox mechanism. Genetic studies using REF1 +/- and REF1 +/+ mice in appropriate genetic backgrounds suggest that p53 activation in response to UV irradiation is dependent on Ref-1 (Meira et al., 1997).

IV. Enzymes that may be responsible for maintaining p53 in the reduced state

VI. Thioredoxin reductase Thioredoxin reductase is another enzyme that may be responsible for reducing p53 cysteine residues, either directly or indirectly. Thioredoxin reductase is a protein disulfide reductase that catalyzes NADPH-dependent reduction of the active site disulfide in oxidized thioredoxin, a small protein (12-14 kD), to a vicinal dithiol (Arner et al., 1999). The requirement of thioredoxin reductase for human p53 activity was identified in a genetic complementation study in the yeast strain Schizosaccharomyces pombe (Casso and Beach, 1996). In this yeast strain, ectopically expressed human p53 causes growth arrest (Bischoff et al., 1992). Casso and Beach (1996) found that a mutation in a yeast homologue of the human thioredoxin reductase gene (trr1) rescued p53-dependent growth arrest. p53-mediated transcription was also downregulated by this mutant allele of trr1. Mutant trr1 required O 2 for its inhibitory effect on p53-mediated growth arrest. A strain of S. pombe lacking trr1 acted in a similar manner to the strain expressing the mutant trr1, suggesting that the original mutant trr1 acted as a dominant negative allele. The requirement for thioredoxin reductase in order to maintain the transcriptional activity of p53 was also demonstrated in the evolutionarily distant to S. pombe yeast strain Saccharomyces cerevisiae (Pearson and Merrill, 1998). These results suggested that either p53 itself, or a protein required for p53 function, is susceptible to disulfide bond formation. Once the disulfide bond is formed p53-mediated transactivation is abrogated. Thioredoxin reductase is required to reduce the disulfide bond and restore p53 function. It is possible that some redox reactions are controlled by subcellular localization of redox-sensitive factors. Interestingly, translocation of cytoplasmic thioredoxin to the nuclear compartment of mammalian cells was recently demonstrated (Hirota et al., 1997). Furthermore, thioredoxin and Ref-1 can form a complex in vitro and in vivo . The

Redox control of p53 may be a chemical or enzymatic process. No consistent data has emerged to indicate the presence of a protein oxidizing enzyme in the cytoplasm. However, several enzymes appear to participate in reducing cytoplasmic protein disulfide linkages (Thomas et al., 1995; Rietsch and Beckwith, 1998). Thus, it is possible that redox regulation of p53 occurs by chemical oxidation and enzymatic reduction. Although no redox enzyme can be excluded from involvement, evidence to date indicates that there are two candidate enzymes responsible for maintaining p53 in a reduced state in eukaryotic cells: Ref-1 and thioredoxin reductase.

V. Ref-1 Redox factor-1, or Ref-1, was characterized as an activity from HeLa cell nuclear extracts that increased recombinant p53 binding to a p53 consensus sequence (Jayaraman et al., 1997). Ref-1 was previously shown to increase the activity of Fos-Jun heterodimer binding to DNA in a manner that depended on DTT (Xanthoudakis and Curran, 1992). Similar to p53, Fos and Jun are redoxsensitive transcription factors that directly bind DNA and upregulate transcription of genes involved in cell cycle progression (Abate et al., 1990). Ref-1 also possesses class II hydrolytic apurinic/apyrimidinic (A/P) endonuclease activity (Demple et al., 1991; Robson and Hickson, 1991). Because of this latter function, Ref-1 has also been assigned other names such as APE, APEX, and HAP-1. The redox regulation portion of Ref-1 and the endonuclease activity of Ref-1 lie within separate domains of the protein. In the presence of DTT, Ref-1 stimulates consensus DNA binding of full-length p53 but this stimulatory activity is severely inhibited when a C-terminally truncated form of

123

Wu et al: Redox modulation of p53 precise mechanism and the role of thioredoxin, Ref-1 and thioredoxin reductase in the regulation of p53 activity in mammalian cells is unknown but offers a potentially fertile environment for further experimental exploration.

If p53 forms an intramolecular disulfide bond it is important to consider the relative orientation of the two cysteine residues in question. One measure of the degree of cysteine residue side chain movement required for disulfide bond formation is to compare the !1 dihedral angles of p53 Cys176 and Cys242 to the !1 dihedral angles of disulfides in known protein structures. In early work, Richardson (1981) measured the !1 dihedral angles of 70 protein disulfide bonds and found that the !1 angles tended to, but did not exclusively, cluster at -60° (±20°), +180 (±20°) and +60° (±20°). The relative frequency of these angles in protein disulfides was -60°>180°>+60°. In p53, the measured !1 angle of Cys176 is +74° while that of Cys242 is -90°. Thus, the !1 angle of Cys176 falls within the least frequent cluster of observed disulfide dihedral angles (+60° (±20°)) while the !1 of Cys242 does not fall into any particular class. If the zinc atom is removed, the cysteine residue side chains may undergo slight changes to form more favorable dihedral angles for disulfide bond formation. A third factor influencing the likelihood of disulfide bond formation is the polarity of the amino acid residues near the cysteine residue (Snyder et al., 1981). Residues that are positively charged and in close proximity to a cysteine residue could attract negatively charged disulfides, such as glutathione disulfide, to the reactive cysteine residue on p53. We note that many of the conserved cysteine residues on the surface of p53, Cys176, Cys242, Cys275 and Cys277 lie within a region of p53 that is electrostatically positively charged (Fig. 3 ). This is also the region of p53 that interacts with DNA. Based on our analysis of the crystal structure it is most plausible that Cys176 and Cys242 could form a disulfide bond. However, involvement of other cysteine residues in either intra- or inter-molecular disulfide bond formation can not be overlooked. In fact, it was demonstrated that recombinant p53 purified in the presence of chelex-treated solution (presumably metal ion-free) was unable to bind consensus sequence-containing DNA unless DTT was added (Rainwater et al., 1995). Apparently, zinc is tightly bound to p53 during this purification procedure. Interestingly, adding more zinc to the p53 did not enhance its DNA binding indicating that DTT may increase p53 binding to DNA by generating reduced free sulfhydryl groups on cysteine residues not responsible for zinc binding. In this regard, one group has reported that a small percentage of recombinant p53 could form a p53-p53 dimer that could be converted to a monomer after addition of DTT (Delphin et al., 1994). If p53 redox control plays an important role in its regulation one would expect that the cysteine residues at the surface of the protein would be well conserved. An alignment of p53 amino acid sequences from 23 species shows that solvent accessible residues cysteine 176, 242 and 277 share nearly 100% identity (Soussi and May, 1996) but not other solvent accessible cysteine residues. Cysteine residues 242 and 277 are conserved throughout all 23 species reported while residue 176 is conserved in 22 species.

VII. Potential sites of p53 cysteine residue oxidation-a structural analysis In order for p53 cysteine residues to be redox regulated the sulfur atoms must be accessible to the oxidant. Structural studies can be used to rule out many potential cysteine oxidation sites based on solvent accessibility. Very few cysteine oxidation reactions on cytoplasmic proteins have been mapped, which might explain why no consensus sequence has emerged. To analyze the structural requirements for protein cysteine oxidation investigators have treated purified cytoplasmic proteins of known structures with glutathione disulfide, a compound that forms a mixed glutathione disulfide with protein cysteine residues (Thomas et al., 1995). In the case of rat liver carbonic anhydrase III, the crystal structure of the protein with conjugated glutathione was solved. For other proteins a twopart process was used to determine the structural requirements for glutathione conjugation. First, the glutathione-cysteine disulfide was mapped. Second, the mapped cysteine was assessed for solvent accessibility through analysis of the crystal structure of the unconjugated protein. In these studies it was determined that the residues surrounding the susceptible cysteine residues showed no consensus sequence. The only consistent feature found was that cysteine residues were located on the surface of the protein. Structure analysis of Fos and Jun bound to DNA also indicated that their oxidation-susceptible cysteine residues are exposed to the surface as well (Chen et al., 1998). The fact that solvent accessible cysteine residues are in close proximity to DNA binding residues may explain why Fos and Jun oxidation prevents them from binding to DNA. We analyzed the crystal structure of residues 94-289 of p53 bound to DNA (Cho et al., 1994) in order to determine which of the 10 cysteine residues in this domain may be exposed to solvent (Connolly, 1983). As shown in Fig. 2, the sulfhydryl groups of Cys124, Cys176, Cys182, Cys229, Cys242 and Cys277 can theoretically react with small molecules on the surface of p53 (indicated with an asterisk). We conducted our analysis assuming that zinc was absent in this measurement. It is possible that part of the oxidation reaction involves removal of zinc as was recently shown for Hsp33 (Jakob et al., 1999). To determine which p53 cysteine residues may form intramolecular disulfide linkages the distances between cysteine sulfur atoms were measured. Among the cysteine residues exposed to the solvent only Cys176 and Cys242 can theoretically form a disulfide bond (the interatomic distance of sulfur atoms is 3.66 Å). Once zinc is removed and Cys 176 and Cys242 become exposed and/or oxidized, either of them could attack Cys238, which also appears to be a reasonable candidate to participate in disulfide bond formation. 124

Gene Therapy and Molecular Biology Vol 4, page 125

Figure 2. Analysis of p53 (residues 94-289) for potential cysteine residue oxidation (Cho et al., 1994). Backbone protein chain is in blue and cysteine side chains are in orange (only the bonds connecting the C" and C# and S atoms are represented ). The white sphere represents a zinc atom. A thin line between the three cysteine S atoms and zinc are shown representing metal-sulfur bonds. Numbers refer to cysteine residues; the amino terminus and carboxyl terminus are represented by the 'N' and the 'C', respectively. The solvent exposed residues are Cys124 , Cys176, Cys182, Cys229, Cys242 and Cys277 (denoted by *). Cys176 and Cys242 can potentially form a disulfide bond (denoted by **). The structure was displayed and analyzed using the Insight II software, version 98.0 (Molecular Simulations Incorporated).

predicted to be similar to p53, p73 and p51/p63 (Kaghad et al., 1997; Osada et al., 1998; Yang et al., 1998). The nearby double positive charge on residues in close proximity to Cys176 may render this residue more susceptible to reaction with a negatively charged oxidizing molecule such as glutathione disulfide (Snyder et al., 1981). If one or more of these four cysteine residues is oxidized in p53 it is possible that oxidation may also alter the activity of p73 and p51/p63.

Only in one species (Ovis aries, sheep) does the p53 sequence reveal a change in Cys176 to Ser176 (Dequiedt et al., 1995). This is extremely interesting because a serine at this site is expected to prevent consensus DNA binding (Rainwater et al., 1995). Other conserved cysteine residues are maintained in sheep p53. Fig. 4 shows a sequence alignment between human p53 and squid p53 (the most divergent of the p53 genes overall). We also show an alignment of the p53 cysteine coding region with two genes 125

Wu et al: Redox modulation of p53

Figure 3. Surface electrostatic potential of p53. The crystal structure shown in Fig. 2 was analyzed for electrostatic potential using the MolMol program, version 2.6 (Koradi et al., 1996). Blue color represents atoms with low electron density, red color represents residues with high electron density and white represents neutral and charged residues. The arrows point to positions of the surface exposed sulfur atoms of the cysteine residues on the surface in this orientation. Note that the orientation of the p53 protein is identical to the p53 structure shown in Fig. 2. The sulfur atom of the Cys residue 176 was not exposed on this face of the protein.

Figure 4. Sequence comparison of well-conserved solvent exposed p53 cysteine residues. Cys residues expected to be solvent accessible are in boldface. Conserved charged residues within close proximity to the cysteine residues are also shown at the appropriate polarity and location.

126

Gene Therapy and Molecular Biology Vol 4, page 127 Our structure analysis is based on the crystal structure of the p53 DNA binding domain bound to a p53-consensus sequence within an oligonucleotide (the B monomer, Cho et al., 1994). Although this structure is helpful in providing predictions of possible oxidation sites it may not provide the true picture of the conformations p53 exhibits in the cell. Several reports have shown that p53 is conformationally flexible and, depending on temperature and oligomeric status, it can also exist in different conformations in vitro (Ponchel and Milner, 1998; McLure and Lee, 1999). Within the crystal structure of p53 bound to DNA there are actually three monomers of p53 in the repeating subunit (named A, B and C). When the images of the monomers were superimposed differences in some of the cysteine residue side chain orientations were observed (data not shown). Thus, it is possible that the conformational changes of p53 observed in the cytosol could expose different cysteine sulfhydryl groups at the protein surface.

IX. Metal binding agents may alter p53 redox level A complicating factor in the field of redox chemistry is the fact that metal ions can catalyze the production of hydroxyl radicals. Thus, when studying redox changes, free transition metals (elements in groups IB, IIB, VIB, VIIB and VII in the periodic table) must always be taken into consideration. The biological systems have evolved molecules to bind free metals but the mechanisms by which metals are released by these systems are not clear. Metal ions, in combination with hydrogen peroxide, can form hydroxyl radicals through the Fenton reaction, also known as the metal-catalyzed Haber-Weiss reaction: . Fe(II) + H2O2 $ Fe(III) + OH + OH (1) .-

Fe(III) + O 2

Fe(II) + O 2

______________________________ .. H2O2 + O2 $ OH + OH + O2

VIII. Mutational analysis of p53 cysteine residues

(2) (3)

The hydroxyl radical is an extremely reactive species. It reacts with most substances with diffusion-limiting rate constants (109- 1010 M-1 s -1). Such reactivity implies a very short half-life and the molecule will likely be unable to travel at great distances. Hainaut has investigated the possibility that copper can modulate p53 DNA binding properties (Hainaut and Milner, 1993; Hainaut et al., 1995). Recombinant p53 was translated in rabbit reticulocyte lysate and exposed to Cu(II) sulfate at 30 ÂľM. Copper exposure induced wild-type p53 to adopt a denatured conformation as detected by conformation-dependent antibodies. When Cu(II) was added to purified recombinant p53 there was no change in the electron spin resonance (ESR) spectrum indicating that Cu(II) may not bind to p53. However, when H2O2 was added to the Cu(II)/p53 mixture, a novel ESR signal, assigned to Cu(II), was observed indicating binding of Cu(II) to p53 under these conditions. Hainaut interpreted these data as follows: the added Cu(II) is initially reduced to Cu(I) by p53 cysteine residues thereby remaining ESR silent. After reduction, the Cu(I) is oxidized by H2O2 to Cu(II) but remains bound to p53. In support of Hainaut's model, the Cu(I) chelating agent, bathocuproinedisulfonic acid (BCS) protected the p53-DNA binding activity, an activity usually inhibited by Cu(II). DMSO, a hydroxyl radical scavenger, and sodium azide, a singlet oxygen scavenger, failed to prevent Cu(II)'s ability to inhibit p53's DNA binding activity ruling out the possibility that these two oxygen species are potential mediators of p53 oxidation. Can copper mediate p53 oxidation in cultured cells? To test the idea that copper may mediate changes in p53 protein properties in vivo the cell permeable copper chelating agent, pyrrolidine dithiocarbamate (PDTC), was employed (Nobel et al., 1995; Verhaegh et al., 1997). This agent is similar in structure to dithiocarbamate-based herbicides, insecticides and fungicides commonly used in the pesticide industry (WHO, 1988). Treatment of cells with PDTC led to inhibition of p53 DNA binding activity and inhibition of stressor-mediated activation of p21 (presumably via p53

Based on the fact that oxidizing agents prevent p53 binding to consensus-sequence containing DNA and that some cysteine residues are well-conserved it is predicted that site-directed mutagenesis of these cysteine residues would alter p53 activities. Mann and coworkers have conducted a mutational study examining the importance of p53 cysteine residues in sequence-specific DNA binding, suppression of cell transformation and p53-mediated transactivation (Rainwater et al., 1995). A Cys to Ser substitution at each cysteine residue of murine p53 was created and the biochemical and biological activities of each individual cysteine mutant was compared to wild-type p53. For the sake of convenience, we will adopt the convention of enumerating the cysteine residues based on the human sequence. From the relative activities in DNA binding, transactivation and transformation suppression, p53 cysteine residues were categorized into three groups. One group of cysteines (those at sites 176, 238 and 242) directly interact with zinc and are essential for DNA binding, transactivation and transformation activity of p53. A second group of cysteine residues (those at sites 124, 135, 141 and 275) is required for transactivation and suppression function. DNA binding activity of p53 is maintained when cysteine residues 124, 135, 141 and 275 are changed to serine. The third group (cysteine residues at sites 182 and 277) did not exhibit any alterations of the measured activities of p53 when the residues were changed to serine. A cysteine to serine substitution is a very conservative change. An oxidation reaction resulting in a disulfide linkage may have a more dramatic consequences than a serine residue substitution. In this regard, it was shown that Cys to Ser substitution at a redox-sensitive Cys residue in cFos resulted in DTT-independent DNA binding (Okuno et al., 1993). Thus, oxidation at any of the three classes of cysteine residues may alter p53 activities.

127

Wu et al: Redox modulation of p53 activation). Interestingly, PDTC-treated nuclear extracts were incapable of binding to the p53 consensus sequence DNA. The investigators went on to demonstrate that intracellular copper levels significantly increased in the presence of PDTC (Verhaegh et al., 1997). Wu and Momand (1998) demonstrated that PDTC treatment of cultured cells led to increased oxidation of p53 protein. Oxidation of p53 correlated with inhibition of p53-mediated transactivation, inhibition of p53 nuclear accumulation and inhibition of UV-induced p53 protein level elevation in fibroblasts. Interestingly, PDTC treatment also prevented E6-mediated degradation of p53. These results suggested that PDTC-mediated copper loading into cells may directly oxidize p53 or generate hydroxyl radicals near the surface of p53, resulting in cysteine residue oxidation. We have incorporated the experimental data into a coherent testable mechanism for p53 oxidation. Because a high level of reduced glutathione is present in the cytosol we have included glutathione as a transporter of electron radicals. This reaction mechanism is based on a combination of previous studies conducted, primarily, on oxidation of cysteine (Gerweck et al., 1984; Winterbourn, 1993; Thomas et al., 1995): PDTC-Cu 2+ transport across cell membrane p53-SH + Cu2+ $ p53-S. + Cu1+ + H+

(1)

p53-S. + GSH

(3)

(p53-S-S-G).- + H

is required for p53-mediated consensus-sequence binding (Pavletich et al., 1993; Rainwater et al., 1995; Verhaegh et al., 1998) Evidence in support of chelation of zinc leading to p53 conformational changes in cultured cells has recently been demonstrated with a membrane permeable zinc-specific chelator N,N,N',N'-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN) (Verhaegh et al., 1998). Nuclear extracts from TPEN-treated cells do not bind consensus DNA. However, if zinc is supplemented into the TPENtreated media prior to cell harvesting p53 retains the ability to bind DNA. TPEN alters the conformation of p53 protein. Importantly, TPEN does not affect the DNA binding activity of Oct-1, another transcription factor that does not require zinc. The data suggests that metal ions can modulate p53 transactivation activity, prevent consensus DNA binding by p53, and can alter the p53 protein conformation. Whether zinc chelation by TPEN leads to p53 oxidation is unclear at the moment.

X. Other functions of p53 that may be regulated by redox levels Several studies indicate that p53 oxidation prevents its ability to bind consensus sequence containing DNA. This leads to the possibility that oxidation may be a mechanism of p53 down-regulation during oxidative stress. Oxidation of p53 may constitute a mechanism to turn off p53-mediated transactivation after its checkpoint function has been fulfilled and DNA has been repaired. It is also possible that oxidation regulates p53 function in a positive aspect. Aside from its role as a transcription factor, biochemical and genetic studies have demonstrated that p53 is responsible for the faithful execution of other activities directly related to maintenance of genome stability. Such activities include global genomic nucleotide excision repair (Ford and Hanawalt, 1995; Ford and Hanawalt, 1997; Ford et al., 1998) and inhibition of homologous recombination (Mekeel et al., 1997; Dudenhoffer et al., 1998). The p53 activities related to repair and recombination may result, in part, from direct interaction of p53 protein with DNA. Some DNA binding studies suggest that p53 has a direct role in DNA repair. For example, p53 binds to insertion/deletion mismatches (Lee et al., 1995; Szak and Pietenpol, 1999), single stranded DNA molecules, and can mediate DNA strand exchange reactions (Bakalkin et al., 1994, 1995; Reed et al., 1995; Wu et al., 1995). The idea that oxidized p53 may bind non-consensus DNA and mediate a function has not been extensively explored. Insertion/deletion mismatch DNA mutations may result from polymerase-induced errors during replication. Mann and coworkers have shown that recombinant p53 can bind to insertion/deletion mismatch DNA in the presence or absence of DTT (Parks et al., 1997). This may indicate that this particular DNA binding function of p53 may be preserved under oxidizing conditions. Furthermore, p53 has recently been demonstrated to possess double-strand DNA exonuclease activity (Mummenbrauer et al., 1996; Janus et al., 1999), which may also be related to a repair function. The p53 protein has been shown to bind substrates that

(2)

(p53-S-S-G) .- + O2 p53-S-S-G + O2.(4) O2.- + 1/2H2O $ 1/2O2 + H2O2 (catalyzed by SOD) (5) H2O2

H2O + 1/2O2

(catalyzed by catalase)

(6)

__________________________________________ p53-SH + Cu2+ + GSH $ p53-S-S-G + Cu 1+ + 2H+ According to the reaction mechanism, PDTC transports Cu 2+ into the cell (Reaction 1). The p53 cysteine residues are directly oxidized by p53 bound cupric ion (Reaction 2). After one-electron oxidation of p53 the p53 cysteine residue carries a thiyl radical that rapidly reacts with free glutathione (GSH) to form the disulfide radical (Reaction 3). This radical is very unstable and requires oxidation before it reforms the reactants. Molecular oxygen oxidizes the radical to form the S-glutathiolated form of the p53 cysteine residue (Reaction 4). The superoxide formed by one-electron reduction of oxygen reacts with water to form hydrogen peroxide by superoxide dismutase (Reaction 5). Hydrogen peroxide is then dissociated into water and molecular oxygen by catalase (Reaction 6). In the future, it will be important to test this reaction mechanism in vitro. While free metal ions may lead to p53 oxidation it is also possible that PDTC indirectly leads to p53 oxidation by chelating the zinc atom that is bound to Cys176, Cys238 and Cys242. After zinc removal the p53 cysteine residues may be oxidized by another molecule, possibly hydroxyl radical. Both, in vitro and in vivo evidence support the idea that zinc 128

Gene Therapy and Molecular Biology Vol 4, page 129 DNA segments via the middle domain. Nucleic Acids Res. 23, 362-369. Bakalkin, G., Yakovleva, T., Selivanova, G., Magnusson, K. P., Szekely, L., Kiseleva, E., Klein, G., Terenius, L., and Wiman, K. G. (1994). p53 binds single-stranded DNA ends and catalyzes DNA renaturation and strand transfer. Proc. Natl. Acad. Sci. USA 91, 413-417. Baker, S. J., Fearon, E. R., Nigro, J. M., Hamilton, S. R., Preisinger, A. C., Jessup, J. M., vanTuinen, P., Ledbetter, D. H., Barker, D. F., Nakamura, Y., White, R., and Vogelstein, B. (1989). Chromosome 17 deletions and p53 gene mutations in colorectal carcinomas. Science 244, 217-221. Barron, E. S. (1951). Thiol groups of biological importance. In Adv. Enzymol ., F. F. Nord, ed. (New York: Interscience), pp. 201-266. Bischoff, J. R., Casso, D., and Beach, D. (1992). Human p53 inhibits growth in Schizosaccharomyces pombe. Mol. Cell. Biol. 12, 1405-1411. Bunz, F., Dutriaux, A., Lengauer, C., Waldman, T., Zhou, S., Brown, J. P., Sedivy, J. M., Kinzler, K. W., and Vogelstein, B. (1998). Requirement for p53 and p21 to sustain G2 arrest after DNA damage. Science 282, 1497-1501. Cabiscol, E., and Levine, R. L. (1996). The phosphatase activity of carbonic anhydrase III is reversibly regulated by glutathiolation. Proc. Natl. Acad. Sci. USA 93, 4170-4174. Casso, D., and Beach, D. ( 1996). A mutation in a thioredoxin reductase homologue suppresses p53 induced growth inhibition in the fission yeast Schizosaccharomyces pombe . Mol. Gen. Genet. 252, 518-529. Chen, L., Glover, J. N., Hogan, P. G., Rao, A., and Harrison, S. C. (1998). Structure of the DNA-binding domains from NFAT, Fos and Jun bound specifically to DNA. Nature 392, 42-48. Chen, Q. M., Bartholomew, J. C., Campisi, J., Acosta, M., Reagan, J. D., and Ames, B. N. (1998). Molecular analysis of H2O2induced senescent-like growth arrest in normal human fibroblasts: p53 and Rb control G1 arrest but not cell replication. Biochem. J. 332, 43-50. Chin, P. L., Momand, J., and Pfeifer, G. P. (1997). In vivo evidence for binding of p53 to consensus binding sites in the p21 and GADD45 genes in response to ionizing radiation. Oncogene 15, 87-99. Cho, Y., Gorina, S., Jeffrey, P. D., and Pavletich, N. P. (1994). Crystal structure of a p53 tumor suppressor-DNA complex: understanding tumorigenic mutations. Science 265, 346-355. Collison, M. W., and Thomas, J. A. (1987). S-thiolation of cytoplasmic cardiac creatine kinase in heart cells treated with diamide. Biochem. Biophys. Acta 928, 121-129. Connolly, M. L. ( 1983). Solvent-accessible surfaces of proteins and nucleic acids. Science 221, 709-713. Delphin, C., Cahen, P., Lawrence, J. J., and J., B. (1994). Characterization of baculovirus recombinant wild-type p53. Dimerization of p53 is required for high-affinity DNA binding and cysteine oxidation inhibits p53 DNA binding. Eur. J. Biochem. 223, 683-692. Demple, B. (1999). Radical ideas: genetic responses to oxidative stress. Clinical Exper. Pharm. Phys. 26, 64-68. Demple, B., Herman, T., and Chen, D. S. (1991). Cloning and expression of APE, the cDNA encoding the major human apurinic endonuclease: definition of a family of DNA repair enzymes. Proc. Natl. Acad. Sci. USA 88, 11450-11454.

mimic recombination intermediates including three stranded DNA molecules (Dudenhoffer et al., 1998) and Holliday junctions (Lee et al., 1997). Some p53 mutations found in human tumors produce mutant p53 proteins that disrupt DNA repair and recombination activities. Whether p53 oxidation modulates these different p53 activities is unclear at the moment.

XI. Conclusion Redox regulation of p53 activity was proposed in 1993 based on the observation that sequence specific DNA binding could be inhibited by agents that blocked sulfhydryl groups (Hainaut and Milner, 1993). However, p53 redox regulation is still poorly understood. The fact that p53 activation in cultured cells can be promoted by agents that induce the formation of ROI while the oxidation of p53 protein inhibits its ability to bind consensus sequencecontaining DNA in vitro renders this regulation mechanism even more intriguing. Two disulfide-reducing proteins, Ref1 and thioredoxin reductase may hold the key to understanding this regulation. By analyzing the structure of p53, we found several well-conserved cysteine residues exposed at the protein surface. Another area of investigation into p53 redox control is related to its ability to bind metal ions. Binding of metal ions may directly affect p53 redox potential either at the zinc binding cysteine residues or at other cysteine residues on the protein surface. To date, most evidence suggests that p53 oxidation inhibits p53 activity. However, in vitro studies show that oxidized p53 retains the ability to bind insertion/deletion mismatches in DNA without the addition of DTT. This opens the possibility that redox regulation can be used as a molecular switch inside the cell to promote other functions of p53 that are less wellknown (Parks et al., 1997). Future studies on the prevalence of p53 oxidation in response to stressors and the effects of p53 protein oxidation on its biochemical activities are clearly needed to address these issues.

Acknowledgments J.M. is supported by the City of Hope Cancer Center and H.-H. Wu is supported by the American Cancer SocietyOncology Project Grant (OPG-9-98).

References Abate, C., Patel, L., Rauscher, F. J. I., and Curran, T. (1990). Redox regulation of Fos and Jun DNA-binding activity in vitro. Science 249, 1157-1161. Arner, E. S., Zhong, L., and Holmgren, A. (1999). Preparation and assay of mammalian thioredoxin and thioredoxin reductase. Methods Enzymol. 300 , 226-239. Ă&#x2026;slund, F., and Beckwith, J. (1999). Bridge over troubled waters: sensing stress by disulfide bond formation. Cell 96, 751-753. Bakalkin, G., Selivanova, G., Yakovleva, T., Kiseleva, E., Kashuba, E., Magnusson, K. P., Szekely, L., Klein, G., Terenius, L., and Wiman, K. G. (1995). p53 binds singlestranded DNA ends through the C-terminal domain and internal

129

Wu et al: Redox modulation of p53 Dequiedt, F., Kettmann, R., Burny, A., and Willems, L. (1995). Nucleotide sequence of the ovine p53 tumor-suppressor cDNA and its genomic organization. DNA Seq. 5 , 255-259. Ding, H., Hidalgo, E., and Demple, B. (1996). The redox state of the [2Fe-2S] clusters in SoxR protein regulates its activity as a transcription factor. J. Biol. Chem. 271, 33173-33175. Doroshaw, J. H. (1986). Prevention of doxorubicin-induced killing of MCF-7 human breast cancer cells by oxygen radical scavengers and iron chelating agents. Biochem. Biophys. Res. Commun. 135 , 330-335. Dudenhoffer, C., Rohaly, G., Will, K., Deppert, W., and Wiesmuller, L. (1998). Specific mismatch recognition in heteroduplex intermediates by p53 suggests a role in fidelity control of homologous recombination. Mol. Cell. Biol. 18, 5332-5342. El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., and Vogelstein, B. (1993). WAF1, a potential mediator of p53 tumor suppression. Cell 75, 817-825. Ford, J. M., and Hanawalt, P. C. (1995). Li-Fraumeni syndrome fibroblasts homozygous for p53 mutations are deficient in global DNA repair but exhibit normal transcription-coupled repair and enhanced UV resistance. Proc. Natl. Acad. Sci. USA 92, 8876-8880. Ford, J. M., and Hanawalt, P. C. (1997). Expression of wild-type p53 is required for efficient global genomic nucleotide excision repair in UV-irradiated human fibroblasts. J. Biol. Chem. 272, 28073-28080. Ford, J. M., Baron, E. L., and Hanawalt, P. C. (1998). Human fibroblasts expressing the human papillomavirus E6 gene are deficient in global genomic nucleotide excision repair and sensitive to ultraviolet irradiation. Cancer Res. 58, 599-603. Giaccia, A. J., and Kastan, M. B. (1998). The complexity of p53 modulation: emerging patterns from divergent signals. Genes Dev. 12, 2973-2983. Gottlieb, T. M., and Oren, M. (1996). p53 in growth control and neoplasia. Biochim Biophys Acta 1287, 77-102. Grant, C. M., Quinn, K. A., and Dawes, I. W. (1999). Differential protein S-thiolation of glyceraldehyde-3-phosphate dehydrogenase isoenzymes influences sensitivity to oxidative stress. Mol. Cell. Biol. 19, 2650-2656. Greenblatt, M. S., Bennett, W. P., Hollstein, M., and Harris, C. C. (1994). Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathologies. Cancer Res. 54, 4855-4878. Hainaut, P., and Milner, J. (1993). Redox modulation of p53 conformation and sequence-specific DNA binding in vitro. Cancer Res. 53, 4469-4473. Hainaut, P., Hernandez, T., Robinson, A., Rodriguez-Tome, P., Flores, T., Hollstein, M., Harris, C. C., and Montesano, R. (1998). IARC Database of p53 gene mutations in human tumors and cell lines: updated compilation, revised formats and new visualisation tools. Nucleic Acids Res. 26, 205-213. Hainaut, P., Rolley, N., Davies, M., and Milner, J. (1995). Modulation by copper of p53 conformation and sequencespecific DNA binding: role for Cu(II)/Cu(I) redox mechanism. Oncogene 10, 27-32. Halliwell, B., and Gutteridge, J. M. (1989). Free radicals in biology and medicine., 2nd Edition (Oxford: Oxford University Press).

Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, K., and Elledge, S. J. (1993). The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinase. Cell 75, 805-816. Hermeking, H., Lengauer, C., Polyak, K., He, T.-C., Zhang, L., Thiagalingam, S., Kinzler, K. W., and Vogelstein, B. (1997). 14-3-3% is a p53-regulated inhibitor of G2/M progression. Mol Cell 1, 3-11. Hidalgo, E., Ding, H., and Demple, B. (1997). Redox signal transduction: mutations shifting [2Fe-2S] centers of the SoxR sensor-regulator to the oxidized form. Cell 88, 121-129. Hirota, K., Matsui, M., Iwata, S., Nishiyama, A., Mori, K., and Yodoi, J. (1997). AP-1 transcriptional activity is regulated by a direct association between thioredoxin and Ref-1. Proc. Natl. Acad. Sci. USA 94, 3633-3638. Hupp, T. R., and Lane, D. P. (1994). Allosteric activation of latent p53 tetramers. Curr. Biol. 4, 865-875. Hupp, T. R., Meek, D. W., Midgley, C. A., and Lane, D. P. (1992). Regulation of the specific DNA binding function of p53. Cell 71, 875-886. Hupp, T. R., Meek, D. W., Midgley, C. A., and Lane, D. P. (1993). Activation of the cryptic DNA binding function of mutant forms of p53. Nucleic Acids Res. 21, 3167-3174. Jakob, U., Muse, W., Eser, M., and Bardwell, J. C. (1999). Chaperone activity with a redox switch. Cell 96, 341-352. Jamieson, D. J., and Storz, G. (1997). Transcriptional regulators of oxidative stress response. In Oxidative stress and the molecular biology of antioxidant defenses., J. G. Scandalios, ed. (Cold Spring Harbor: Cold Spring Harbor Laboratory Press), pp. 91-115. Janus, F., Albrechtsen, N., Knippschild, U., Wiesmuller, L., Grosse, F., and Deppert, W. (1999). Different regulation of the p53 core domain activities 3'-to-5' exonuclease and sequencespecific DNA binding. Mol. Cell. Biol. 19, 2155-2168. Jayaraman, L., Murthy, K. G. L., Zhu, C., Curran, T., Xanthoudakis, S., and Prives, C. (1997). Identification of redox/repair protein Ref-1 as a potent activator of p53. Genes Dev. 11, 558-570. Kaghad, M., Bonnet, H., Yang, A., Creancier, L., Biscan, J. C., Valent, A., Minty, A., Chalon, P., Lelias, J. M., Dumont, X., Ferrara, P., McKeon, F., and Caput, D. (1997). Monoallelically expressed gene related to p53 at 1p36, a region frequently deleted in neuroblastoma and other human cancers. Cell 90, 809-819. Koradi, R., Billeter, M., and Wuthrich, K. (1996). MOLMOL: a program for display and analysis of macromolecular structures. J. Mol. Graph. 14, 51-55, 29-32. Lee, S., Cavallo, L., and Griffith, J. (1997). Human p53 binds Holliday junctions strongly and facilitates their cleavage. J. Biol. Chem. 272 , 7532-7539. Lee, S., Elenbaas, B., Levine, A., and Griffith, J. (1995). p53 and its 14 kDa C-terminal domain recognize primary DNA damage in the form of insertion/deletion mismatches. Cell 81, 10131020. Lee, S.-R., Kwon, K.-S., Kim, S.-R., and Rhee, S. G. (1998). Reversible inactivation of protein-tyrosine phosphatase 1B in A431 cells stimulated with epidermal growth factor. J. Biol. Chem. 273, 15366-15372. Levine, A. J. (1997). p53, the cellular gatekeeper for growth and division. Cell 88, 323-331. McLure, K. G., and Lee, P. W. (1999). p53 DNA binding can be modulated by factors that alter the conformational equilibrium. EMBO J. 18, 763-770.

130

Gene Therapy and Molecular Biology Vol 4, page 131 corrects DNA repair and mutagenesis defects in E. coli xth (exonuclease III) mutants. Nucleic Acids Res. 19, 5519-5523. Sato, T., Baba, K., Takahashi, Y., Uchiumi, T., and Odani, S. (1996). Rat liver fatty acid-binding protein: identification of a molecular species having a mixed disulfide with cysteine at cysteine-69 and enhanced protease susceptibility. J. Biochem. 120, 908-914. Shieh, S.-Y., Ikeda, M., Taya, Y., and Prives, C. (1997). DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2. Cell 91, 325-334. Snyder, G. H., Cennerazzo, M. J., Karalis, A. J., and Field, D. (1981). Electrostatic influence of local cysteine environments on disulfide exchange kinetics. Biochemistry 20, 6509-6519. Soussi, T., and May, P. (1996). Structural aspects of the p53 protein in relation to gene evolution: a second look. J. Mol. Biol. 260, 623-637. Sugano, T., Nitta, M., Ohmori, H., and Yamaizumi, M. (1995). Nuclear accumulation of p53 in normal human fibroblasts is induced by various cellular stress which evoke the heat shock response, independently of the cell cycle. Jpn. J. Cancer Res. 86, 415-418. Sun, Y., and Oberley, L. W. (1996). Redox regulation of transcriptional activators. Free Radic. Biol. Med. 21, 335-348. Szak, S. T., and Pietenpol, J. A. (1999). High affinity insertion/deletion lesion binding by p53. J. Biol. Chem. 274, 3904-3909. Thomas, J. A., Chai, Y. C., and Jung, C. H. (1994). Protein Sthiolation and dethiolation. Methods Enzymol. 233, 385-395. Thomas, J. A., Poland, B., and Honzatko, R. (1995). Protein sulfhydryls and their role in the antioxidant function of protein S-thiolation. Arch. Biochem. Biophys. 319 , 1-9. Uberti, D., Yavin, E., Gil, S., Ayasola, K.-R., Goldfinger, N., and Rotter, V. (1999). Hydrogen peroxide induces nuclear translocation of p53 and apoptosis in cells of oligodendroglia origin. Mol. Brain Res. 65, 167-175. Verhaegh, G. W., Parat, M.-O., Richard, M.-J., and Hainaut, P. (1998). Modulation of p53 protein conformation and DNAbinding activity by intracellular chelation of zinc. Mol. Carcinog. 21, 205-214. Verhaegh, G. W., Richard, M.-J., and Hainaut, P. (1997). Regulation of p53 by metal ions and by antioxidants: dithiocarbamate down-regulates p53 DNA-binding activity by increasing the intracellular level of copper. Mol. Cell. Biol. 17, 5699-5706. Vile, G. F. (1997). Active oxygen species mediate the solar ultraviolet radiation-dependent increase in the tumour suppressor protein p53 in human skin fibroblasts. FEBS Lett. 412, 70-74. Vile, G. F., Rothwell, L. A., and Kettle, A. J. (1998). Hypochlorous acid activates the tumor suppressor protein p53 in cultured human skin fibroblasts. Arch. Biochem. Biophys. 359, 51-56. WHO. (1988). . In Environmental Health Criteria (Vammal, Finland: The World Health Organization). Wu, H.-H., and Momand, J. (1998). Pyrrolidine dithiocarbamate prevents p53 activation and promotes p53 cysteine residue oxidation. J. Biol. Chem. 273, 18898-18905. Wu, L., Bayle, J. H., Elenbaas, B., Paveletich, N. P., and Levine, A. J. (1995). Alternatively spliced forms in the carboxy-terminal domain of the p53 protein regulate its ability to promote annealing of complementary single strands of nucleic acids. Mol. Cell. Biol. 15, 497-504.

Meira, L. B., Cheo, D. L., Hammer, R. E., Burns, D. K., Reis, A., and Friedberg, E. C. (1997). Genetic interaction between HAP1/REF-1 and p53. Nat. Genet. 17, 145. Mekeel, K. L., Tang, W., Kachnic, L. A., Luo, C. M., DeFrank, J. S., and Powell, S. N. (1997). Inactivation of p53 results in high rates of homologous recombination. Oncogene 14, 1847-1857. Miyashta, T., and Reed, J. C. (1995). Tumor suppressor p53 is a direct transcriptional activator of the human bax gene. Cell 80, 293-299. Mummenbrauer, T., Janus, F., Muller, B., Wiesmuller, L., Deppert, W., and Grosse, F. (1996). p53 Protein exhibits 3'-to-5' exonuclease activity. Cell 85, 1089-1099. Nigro, J. M., Baker, S. J., Preisinger, A. C., Jessup, J. M., Hostetter, R., Cleary, K., Bigner, S. H., Davidson, N., Baylin, S., Devilee, P., Glover, T., Collins, F. S., Weston, A., Modali, R., Harris, C. C., and Vogelstein, B. (1989). Mutations in the p53 gene occur in diverse human tumour types. Nature 342, 705-708. Nobel, C. S. I., Kimland, M., Lind, B., Orrenius, S., and Slater, A. F. G. ( 1995). Dithiocarbamates induce apoptosis in thymocytes by raising the intracellular level of redox-active copper. J. Biol. Chem. 270, 26202-26208. Okuno, H., Akahori, A., Sato, H., Xanthoudakis, S., Curran, T., and Iba, H. (1993). Escape from redox regulation enhances the transforming activity of Fos. Oncogene 8, 695-701. Osada, M., Ohba, M., Kawahara, C., Ishioka, C., Kanamaru, R., Katoh, I., Ikawa, Y., Nimura, Y., Nakagawara, A., Obinata, M., and Ikawa, S. (1998). Cloning and functional analysis of human p51, which structurally and functionally resembles p53. Nat. Med. 4, 839-843. Parks, D., Bolinger, R., and Mann, K. (1997). Redox state regulates binding of p53 to sequence-specific DNA, but not to nonspecific or mismatched DNA. Nucleic Acids Res. 25, 12891295. Pavletich, N. P., Chambers, N. P., and Pabo, C. O. (1993). The DNA-binding domain of p53 contains the four conserved regions and the major mutation hot spots. Genes Dev. 1993, 2556-2564. Pearson, G. D., and Merrill, G. F. (1998). Deletion of the Saccharomyces cerevisiae TRR1 gene encoding thioredoxin reductase inhibits p53-dependent reporter gene expression. J. Biol. Chem. 273 , 5431-5434. Ponchel, F., and Milner, J. (1998). Temperature sensitivity of human wild-type and mutant p53 proteins expressed in vivo. Br. J. Cancer. 77, 1555-1561. Rainwater, R., Parks, D., Anderson, M. E., Tegtmeyer, P., and Mann, K. (1995). Role of cysteine residues in regulation of p53 function. Mol. Cell. Biol. 15, 3892-3903. Ravichandran, V., Seres, T., Moriguchi, T., Thomas, J. A., and Johnston, R. B. J. (1994). S-thiolation of glyceraldehyde-3phosphate dehydrogenase induced by the phagocytosisassociated respiratory burst in blood monocytes. J. Biol. Chem. 269, 25010-25015. Reed, M., Woelker, B., Wang, P., Wang, Y., Anderson, M. E., and Tegtmeyer, P. (1995). The C-terminal domain of p53 recognizes DNA damaged by ionizing radiation. Proc. Natl. Acad. Sci. USA 92, 9455-9459. Richardson, J. S. (1981). The anatomy and taxonomy of protein structure. Adv. Prot. Chem. 34, 167-339. Rietsch, A., and Beckwith, J. (1998). The genetics of disulfide bond metabolism. Annu. Rev. Genet. 32, 163-184. Robson, C. N., and Hickson, I. D. (1991). Isolation of cDNA clones encoding a human apurinic/apyrimidinic endonuclease that

131

Wu et al: Redox modulation of p53 Xanthoudakis, S., and Curran, T. (1992). Identification and characterization of Ref-1, a nuclear protein that facilitates AP-1 DNA binding activity. EMBO J. 11, 653-665. Yang, A., Kaghad, M., Wang, Y., Gillett, E., Fleming, M. D., Dotsch, V., Andrews, N. C., Caput, D., and McKeon, F. (1998). p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol. Cell. 2, 305-316. Yin, Y., Solomon, G., Deng, C., and Barrett, J. C. (1999). Differential regulation of p21 by p53 and Rb in cellular response to oxidative stress. Mol. Carcinog. 24, 15-24. Yin, Y., Terauchi, Y., Solomon, G. G., Aizawa, S., Rangarajan, P. N., Yazaki, Y., Kadowaki, T., and Barrett, J. C. (1998). Involvement of p85 in p53-dependent apoptotic response to oxidative stress. Nature 391, 707-710. Zheng, M., Ă&#x2026;slund, F., and Storz, G. (1998). Activation of the OxyR transcription factor by reversible disulfide bond formation. Science 279, 1718-1721.

Jamil Momand

132

Gene Therapy and Molecular Biology Vol 4, page 133 Gene Ther Mol Biol Vol 4, 133-141. December 1999.

HIV-1 DNA integration: advancing anti-HIV-1 gene therapy approaches by blocking and modulating the process Review Article

Aaron Geist, Mohamad BouHamdan, Giuseppe Nunnari, Roger J. Pomerantz* and Joseph Kulkosky The Dorrance H. Hamilton Laboratories, Center for Human Virology, Division of Infectious Diseases, Department of Medicine, Jefferson Medical College, Thomas Jefferson University, Philadelphia, Pennsylvania 19107 ______________________________________________________________________________________________________ *C o r r e spo nde nc e : Roger J. Pomerantz, M.D., Professor of Medicine, Biochemistry and Molecular Pharmacology, Chief, Division of Infectious Diseases, Director, Center for Human Virology, Jefferson Medical College, Thomas Jefferson University, Philadelphia, Pennsylvania 19107. Tel: (215) 503-8575; Fax: (215) 923-1956; E-mail: roger.j.pomerantz@mail.tju.edu A b b r e v i a t i o n s : H I V - 1 , human immunodeficiency virus type 1; IN, integrase; RT, reverse transcriptase; PR, protease; PIC, preintegration complex; ASV, avian sarcoma virus; LTRs, long terminal repeats; I g , immunoglobulin; PBMCs, peripheral blood mononuclear cells; S F v , single-chain variable fragment; CAT , chloramphenicol acetyltransferase Key Words: Gene Therapy, Retroviruses, Integration Received: 27 July 1999; accepted: 10 August 1999

Summary The efficient replication of retroviruses requires integration of the double-stranded DNA copy of viral R N A i n t o c h r o m o s o m a l D N A o f t h e i n f e c t e d h o s t c e l l . Integration for all retroviruses, including human immunodeficiency virus type 1 (HIV-1), represents the stable incorporation of viral DNA into that of the host and depends upon the function of the viral encoded enzyme, integrase (IN). Among various classes of animal viruses, integration of viral DNA is a unique and defining step in the life cycle of retroviruses and results in the permanent establishment of the viral genome in the infected host cell. The practical consequences of this activity is the presence of a viral DNA template capable o f d i r e c t i n g t h e p r o d u c t i o n o f p r o g e n y v i r u s f o r t h e l i f e s p a n o f the infected c e l l . This template persists even despite the imposition o f therapeutic regimens available currently that involve the combinatorial use of inhibitors against HIV-1 reverse transcriptase (RT) and protease (PR) that are effective in reducing circulating virus in most HIV-1 seropositive patients. Efficient blockade of integration, or the steps preceding integration, would be far preferable than attempts to suppress the production of viral products from cells that already harbor an HIV-1 provirus. The actual mechanistic details for integration of all retroviruses is remarkably similar and such an u n d e r s t a n d i n g i s v i t a l i n a n e w e r a o f r a t i o n a l d r u g d e s i g n t h a t n o t o n l y r e l i e s upon an intimate knowledge of the structure and function of individual viral products, but alternatively utilizes those details to manipulate these products with the goal of halting viral replication completely or ablating the presence of infected host cells in the seropositive patient. Several small molecule inhibitors of HIV-1 integrase are i n development with the eventual hope o f including such compounds in combination therapy with PR and RT inhibitors. However, there are anti-IN gene therapy strategies being pursued that are highlighted in this review and which can be employed to inhibit function of the protein with a similar goal of blocking HIV-1 infection within host cells.

immune defense functions leading ultimately to death. It is anticipated that shortly beyond the year 2000, cumulative deaths as a consequence of HIV-1 infection, particularly in developing countries, will reach levels approaching those observed previously for the most severe of human plagues.

I. HIV-1 DNA integration HIV-1 is a recently emerged human pathogen (BarrieSinoussi et al., 1983; Gallo et al., 1983). For most individuals, infection by this complex retrovirus results in the gradual loss of host immune surveillance and protective 133

Geist et al: HIV-1 integration and gene therapy As illustrated in Figure 1, the early stages of the lifecycle of HIV-1 bears remarkable similarities with other retroviruses. All retroviruses encode a bank of enzymes that are absolutely required for their efficient replication (Katz and Skalka, 1994). These polypeptides, protease (PR), reverse transcriptase (RT), RT-associated RNase H activity, and integrase (IN) are attractive antiviral targets as they manifest discrete enzymatic functions that are conveniently assayed in vitro and also typically in vivo. All are relatively small proteins that are also amenable to a variety of sophisticated structural analyses. On this basis, it is not surprising that the most effective therapeutic agents currently in use, and developed to inhibit or block HIV-1 replication, are those that effectively neutralize two proteins from this panel of retroviral enzymes, RT and PR (Mellors, 1996). While combination therapy using RT and PR inhibitors can be highly effective in suppressing viral replication in most individuals, the regimen suffers drawbacks. Therapy may not be well-tolerated by some patients, emergence of resistant virus can occur, and perhaps most importantly, HIV-1 proviral DNA remains detectable in certain reservoirs within the body in most patients even after reasonably prolonged treatment (Richman, 1994). These factors prompt the need to develop more effective inhibitors particularly those against additional HIV-1 viral products. The development of novel strategies is also warranted, including judicious gene therapeutic approaches that will either cripple the virus prior to cell infection, prevent HIV-1

F i g u r e 1 . Early events of retroviral replication leading to the establishment to integrated proviral DNA. The viral core enters the cell after attachment of the virus particle to the cell surface. The viral RNA is converted to doublestranded DNA by the process of reverse transcription. This DNA copy enters the cell nucleus where it is integrated into the host genomic DNA.

134

replication during de novo infection of target cells or, finally, eliminate HIV-1 infected cells that are sequestered within tissue reservoirs of patients refractive to anti-HIV-1 compounds in use currently. Integrase (IN) has become an increasingly attractive candidate for the development of antivirals for reasons, which relate to the success in the discovery or design PR and RT inhibitors. For instance, like reverse transcription and PR maturation of viral proteins, integration of retroviral DNA is an obligatory step in the life cycle of retroviruses, including HIV (La Femina et al, 1992), and is required for the efficient replication of all retroviruses in their respective host cells. The process of integration is generally regarded to occur in two important biochemical and temporally discrete steps (Kulkosky and Skalka, 1994). The first step, as illustrated in Figure 2, results in the removal of two nucleotides from the 3â&#x20AC;&#x2122; ends of the double-stranded viral DNA which is generated by reverse transcription of genomic viral RNA. This reaction, referred to as processing, appears to occur primarily, if not exclusively, in the cytoplasmic compartment of the host cell during a natural viral infection (Brown et al, 1989). The second step occurs in the cell nucleus and involves insertion of the processed viral DNA into host chromosomal DNA. As depicted in Figure 2, this reaction is referred to as joining. Both processing and joining activities are mediated by IN and this protein is necessary and sufficient to catalyze these reactions in vitro (Craigie et al., 1990; Katz et al., 1990).

Gene Therapy and Molecular Biology Vol 4, page 135

F i g u r e 2 . The activities of integrase on viral DNA in the cell cytoplasm (t o p ) and nucleus (b o t t o m ). First, linear viral DNA is cleaved to produce recessed ends in the cell cytoplasm resulting in the release of the two 3â&#x20AC;&#x2122; terminal nucleotides. Next, viral DNA is joined to host cell DNA by a concerted cleavage-ligation reaction referred to as joining.

viral DNA to host DNA results in covalent linkage of one DNA to the other (Engleman et al., 1991). IN contains a highly conserved region, referred to as the D, D(35)E domain (see Figure 3), that represents the active site of the protein (Kulkosky et al., 1992). The nomenclature for this domain is based on the presence of a triad of invariant acidic residues in all retroviral INs, the latter two of which are separated by 35 intervening amino acids. Mutations within the D, D(35)E domain have parallel effects on both processing and joining assayed in vitro and these studies provided the first clue that there was likely to be a single catalytic site which mediated both IN functions (Kulkosky et al., 1992). Deletion mutagenesis coupled with specialized activity assays subsequently confirmed this hypothesis (Bushman et al., 1993). The acidic residues of the D, D(35)E domain were originally proposed to coordinate binding of the divalent metal cofactor based on the mechanism of action of enzymes with similar metabolic functions (Kulkosky et al., 1992). A stabilized interaction of the metal cofactor with DNA promoted by these residues serves to facilitate attack on a phosphorous atom in the DNA backbone of the substrate and

The biochemical mechanism involved in processing and joining is a well-understood nucleic acid metabolic activity referred to as a phosphoryltransfer reaction (Mizuuchi, 1992). The removal of nucleotides from the ends of viral DNA by cleavage during processing involves nucleophilic attack by an oxygen atom at a specific phosphorus bond in the viral DNA backbone which is immediately 3â&#x20AC;&#x2122; of a strictly conserved CA dinucleotide present in all retroviral DNAs as illustrated in F i g u r e 2 . In vivo, the oxygen of a water molecule likely serves as the primary nucleophile involved in this attack, although analyses of the processing reaction as performed in vitro, have shown that a variety of nucleophilic molecules can mediate the processing reaction (Vink et al., 1992). In vitro, these nucleophiles can include glycerol since the enzyme is typically stored in this stabilizing solvent, but other alcohols and even free amino acids are able to participate as agents of attack. The mechanism for joining is analogous to that of the processing reaction. In this case the oxygen of the 3â&#x20AC;&#x2122; hydroxyl groups generated at the ends of viral DNA serve as the nucleophiles which attack the phosphorous atoms in the backbone of the enzyme-bound host DNA. Thus, IN-mediated joining represents a DNA transesterification reaction where phophoryltransfer from 135

Geist et al: HIV-1 integration and gene therapy subsequently the target DNA within an IN/DNA/metal complex. High resolution X-ray crystallographic structural data has verified metal binding by at least two of the triad of acidic residues of the D, D(35)E domain (Goldgur et al., 1998). In a natural viral infection, IN is associated with the newly reverse-transcribed DNA initially in the cytoplasm of the infected cell. This association occurs within higher order structures referred to as pre-integration complexes (PICs). PICs can be considered to be a derivative of the viral core, which is released into the cell cytoplasm after fusion of both viral and cellular membranes. These high molecular weight structures, isolated easily from virus infected cells, are comprised of viral DNA, IN as well as other proteins derived from both virus and host (Bowerman et al., 1989; Ellison et al, 1990). Unlike the PICs of oncoretroviruses, lentiviral PICs, including those of HIV-1, appear to have the unique ability of traversing the nuclear membrane of non-dividing cells (Bukrinsky et al., 1992). Several studies have suggested a mechanism of active transport for lentiviral PICs though the nuclear pores of non-cycling cells. This process of active transport may be facilitated by three viral proteins, MA, Vpr and IN (Stevenson and Gendelman, 1995). However the role that MA, Vpr and/or IN play in the nuclear import of PICs, either alone or in combination, is controversial (Reil et al., 1998). Nevertheless, understanding the actual mechanism whereby lentiviruses are able to infect non-dividing cells is important, particularly as it relates to the inability of oncoretroviruses to do so. Retroviral delivery vectors, derived from both simple oncoretroviruses, and more recently from lentiviruses, are in wide-spread use and each possess desirable as well as undesirable characteristics. Merging the attractive features of each vector system and purging those which are

Figure 3. Functional domains of HIV-1 IN and binding sites of anti-IN monoclonal antibodies. IN is comprised of three discrete domains (t o p ), the N-terminal HHCC region, the catalytic core or the D, D(35)E domain in the mid-portion of IN and the C-terminal DNA binding region. The approximate locations of mAb binding are shown at the b o t t o m .

136

less desirable is the focus of investigation by several laboratories, in particular, identifying a means to permit infection of non-dividing cells using vectors based upon simple retroviral templates in order to facilitate gene therapeutic approaches.

II. Gene therapy approaches to block integration in order to ameliorate retroviral-associated diseases Currently, anti-viral therapies based on RT inhibitors, protease inhibitors, cytokines, and receptor blocking agents are not completely successful. This is due to the ability of HIV-1 to mutate rapidly, which results in development of numerous viral variants, some of which can evade host defenses or become resistant to various anti-viral agents (Erice and Balfour, 1994; St Clair et al., 1991). However, an alternative approach of "anti-viral gene therapy" where host cells may be genetically altered or engineered to confer long-lasting protection against viral infection, or replication, appears to be an attractive and convincing technology (Dropulic and Jeang, 1994; Pomerantz and Trono, 1994). Such strategies directed against several targets are currently being applied towards the inhibition of HIV-1 replication within cells or toward the direct elimination of HIV-1 infected cells. They include exploitation of trans-dominant negative mutant HIV-1 protein expression, viral antisense oligonucleotide sequences, specific ribozymes, and HIV-1 trans-activated "suicide" genes (Chen et al., 1994; Lisziewicz et al., 1995; Liu et al., 1994; Sun et al., 1995; Yu et al., 1993).

Gene Therapy and Molecular Biology Vol 4, page 137 Figure 4. Schematic of intracellular expression of anti-IN SFv molecules and their putative binding to IN within the viral core or pre-integration complex (PIC). IN associates with the ends of viral DNA in the cell cytoplasm. The anti-IN SFvs block de novo infection of target cells presumably by binding to IN within the viral core structure as shown or its derivative referred to as the pre-integration complex.

1992; Engelman and Craigie, 1992; Bushman et al., 1993). The recombinant SFv sequences were first inserted into a murine retroviral delivery vector in order to transduce cells susceptible to HIV-1 infection. Intracellular expression of certain anti-IN SFvs within cells susceptible to HIV-1, resulted in cells resistant to highly cytopathic strains of HIV1 and these experiments complement well our additional studies using anti-HIV-1 RT SFvs (Shaheen et al., 1996). SupT1 cells alone, or those expressing the CAT gene product intracellularly as well as the anti-IN SFvs #17 and #21 do not block HIV-1 replication after infection as noted by the release of HIV-1 p24 antigen in the culture supernatants beginning at day 6 and peaking at day 18. In contrast, intracellular expression of anti-IN SFv #12 and #4 provide moderate protection, while SFv #33 results in complete blockade of virus replication even at day 18 following infection by HIV-1 (Levy-Mintz et al, 1996; Levy-Mintz et al., 1998). Our attempts to propagate HIV-1 following direct infection of anti-IN SFv expressing cells suggested there was a block at early stages of viral replication, after the entry of virions into the cellular cytoplasm (see F i g u r e 4 ) rather than at later stages of the HIV-1 life-cycle during which the assembly and production of virions occur. This is based upon a significant early increase in 2-LTR-containing viral DNA within four to eight hours of infection in cells expressing anti-IN SFv#33 at amounts similar to that observed with INmutant virus (Levy-Mintz et al., 1996). Interestingly, the 2LTR DNA appeared consistently at least two hours earlier in anti-IN SFv#33/NU-transduced cells, in which the SFv is located in the nucleus, than in IN-SFv#33-transduced cells, where the SFv is located in the cytoplasmic compartment (Levy-Mintz et al., 1996; Levy-Mintz et al., 1998). The time difference suggests that HIV-1 IN may be initially

Here we present an overview of the efficacy of two independent, but related, gene therapeutic approaches to inhibit HIV-1 replication that target IN function and involve the use of anti-HIV-1 IN SFvs, which are cloned and constructed from their parental murine IgG templates (BizubBender et al., 1994). In the first approach, the inhibitory SFv species are over-expressed within T cell lines or primary PBMCs. Those that specifically target the HIV-1 IN catalytic or carboxy-terminal DNA binding domains, and are localized to either nuclear or cytoplasmic compartments, markedly decrease HIV-1 replication in both types of cells. The second approach involves delivery of the SFv moieties into whole HIV-1 virus particles by fusion to the HIV-1 accessory protein, Vpr, or alternatively to a 23 amino acid Vpr interactor, referred to as the dWF motif (BouHamdan et al 1998). The presence of the SFvs within virions also results in the inhibition of HIV-1 infection of host target cells. The significance of these collective findings is that they provide an opportunity to block HIV-1 replication before integration of viral DNA into the host genome occurs.

A. Inhibiting integrase (IN) function by the over-expression of anti-IN SFvs in target cells As multiple HIV-1 proteins have been targeted via specific SFv strategies, there is now some indication that successful inhibition of HIV-1 replication depends upon neutralizing relevant functional domains within the target protein (Mhashilkar et al., 1995; Shaheen et al., 1996). The isolation of a panel of murine hybridomas synthesizing antiHIV-1 IN IgGs (see F i g u r e 3 ), permitted the subsequent construction of anti-IN SFvs which target a variety of epitopes within functional domains of IN (Kulkosky et al., 137

Geist et al: HIV-1 integration and gene therapy sequestered by the cytoplasmic SFv, delaying nuclear import of the pre-integration complex, where viral DNA ends are joined via nuclear ligase activity. Alternatively, SFv#33/NU may facilitate its transport into the nuclear compartment by virtue of binding to the pre-integration complex. Although the anti-HIV-1 IN IgGs which were used in these studies are not finely mapped to precise peptide domains, the general binding pattern of the SFvs to IN may provide some clues to how anti-HIV-1 IN SFv moieties might function. Binding of SFvs to regions in the highly conserved central catalytic domain (IN-SFv#4) and to the carboxy-terminal domain may indicate that the relevant portions of those domains may be exposed on the surface of the pre-integration complex. Data from high resolution crystal structure analysis of the core domains of the HIV-1 IN (Goldgur et al., 1998) and ASV IN (Bujacz et al., 1995) proteins indicate that the conserved region, to which INSFv#4 binds, is flexible in solution. The binding of a SFv to this domain may interfere with the flexibility required for this domain's enzymatic activity. The carboxy-terminus of IN, to which SFv#33 binds, is a region of least sequence homology among different retroviruses although it has been shown to promote nonspecific DNA binding which may be involved in target DNA interactions (Engelman and Cragie, 1992). When a specific residue in the carboxy-terminus of IN, Trp-235 was substituted with Ala the ability of the HIV1 provirus to replicate was abolished (Cannon et al., 1996). It has been speculated, therefore, that this region may be required for correct positioning of processed retroviral LTRs to interact with the target host cell DNA, either by virtue of an inherent affinity for DNA or by interactions with cellular proteins that may associate with chromatin. IN mutations, which eliminate in vitro enzymatic activity, frequently inhibit viral replication and block the cytopathic effects of HIV-1 in cell culture. Blockade of the integration process in anti-IN SFv-expressing cells suggests that the HIV-1 IN enzyme, is functionally neutralized in the pre-integration complex upon binding of certain SFvs, prior to completing the integration process (F i g 4 ). Hence, this binding must interrupt some steps before DNA integration into the host genome and establishment of a provirus. In summary, the results of our experiments extend previous findings that single-chain antibodies can be stably expressed in cells, function in the cell cytoplasm as well as the nucleus, and are non-toxic to human cells. Choosing carefully from a relevant panel of anti-HIV-1 IN murine hybridomas also indicates that the SFvs, derived from these hybridomas, can be precisely manipulated for appropriate intracellular binding to specific epitopes on the target molecules. Thus, these murine SFvs provide the means to: (i ) control intracellular infections and other diseases; (i i ) understand the biological mechanisms in cells leading to a disease state and, in particular; (i i i ) aid in choosing the most effective human anti-IN SFvs for development and their ultimate applied use in gene therapeutic protocols to benefit human patients. Moreover, the multiple targets for anti-HIV1 SFvs also permits the combination of different SFvs, which may result in longer term and more effective protection against HIV-1 replication. This is important

since the emergence of mutations conferring resistance to antivirals in HIV-1 proteins, especially RT, may also require such combinations of various SFvs against multiple HIV-1 targets.

B. Use of Vpr and the dWF motif to deliver anti-IN SFvs directly into HIV-1 virus particles to block function The highly conserved HIV-1 Vpr protein, expressed in the late-stage of viral production, and incorporated into virions, is an ideal target for testing the "intravirion protein" delivery model for anti-HIV-1 gene therapy. Vpr can be packaged into virions in reasonable quantities, and many investigators have shown that this accessory protein can be fused to proteins for their delivery into progeny virus despite their in trans expression relative to other virion components (Wu et al., 1997; Liu et al., 1997; Fletcher et al 1997). For instance, a chimeric protein, utilizing the conserved protease cleavage site sequences from Gag and Gag-Pol precursor polyproteins as fusion partners to Vpr have been shown to inhibit HIV-1 infectivity (Serio et al., 1997). Other investigators have generated chimeric proteins based on fusion of HIV-1 Vpr and human immunodeficiency virus type II (HIV-2) Vpx, to CAT, staphylococcal nuclease (SN), wild-type and mutated HIV-1 PR, IN, and RT and these have been shown to reduce virion infectivity (Wu et al., 1996; Wu et al., 1995). As has been discussed previously, intracellular expression of anti-IN SFv#33 in the host target cell has a significant inhibitory effect upon HIV-1 replication following virus challenge of these cells. Our most current studies indicate that the Vpr protein can be used to deliver the anti-IN SFv#33 directly into HIV-1 virions in an attempt to bind IN resident within virus particles. Such a “pre-binding” strategy of IN within virus particles could result in interference with the enzyme’s subsequent functions within the host cell after infection. In order to deliver SFvs into HIV-1 virions, expression plasmids bearing the SFv as a fusion to Vpr were constructed. First, immunoprecipitation indicated that the SFvs fused to Vpr were imported into HIV1 virions. Additional studies further demonstrated that intravirion encapsidation of an anti-HIV-1 IN-Vpr-SFv decreases HIV-1 infectivity/replication in T-lymphocytic cells about 10 fold, relative to infection by HIV-1 virions not containing the imported anti-IN SFv. This decrease in virion infectivity was monitored by infection of MAGI reporter cells that stain deeply blue if infection were successful. The significance of this finding is that the delivery of SFvs directly into HIV-1 virions provides a new and novel opportunity to neutralize HIV-1 IN within the virus particle prior to binding of the virion to cell surface receptors and clearly before integration of viral DNA into the host genome. Our laboratory has identified an additional strategy to shuttle polypeptides into virus particles that may have advantages over the import of Vpr chimeras. The putative interactions that occur to mediate such import are outlined in Figure 5. Based on these details, the approach employs an 138

Gene Therapy and Molecular Biology Vol 4, page 139 in-frame fusion of a 23 amino acid peptide, referred to as the dWxxF or dWF domain, which was discovered to be a Vpr interactor using phage display technology (BouHamdan et al., 1998). In previous reports we have shown that both the bacterial CAT gene product, as well as HIV-1 IN, are incorporated into virions, as fusions to the dWF motif, likely by docking to the viral accessory protein, Vpr. In fact, the efficiency of import Vpr versus dWF can be comparable depending upon the virus into which these proteins are delivered (BouHamdan et al. 1998; Kulkosky et al., 1999). Furthermore, the import of SFv inhibitors into HIV-1 virions by dWF may have certain advantages over import as a fusion to Vpr. First, the dWF motif is smaller than Vpr, which is 96 amino acids in length. As has been reported previously, direct fusion to Vpr may hamper the function the fusion partner. This can be overcome by including an intervening HIV-1 protease cleavage site between Vpr and the

fusion partner (Liu et al., 1997; Fletcher et al., 1997) though there may be situations where inclusion of such a site may be undesirable. Finally, Vpr is known to alter cellular functions therefore the use of dWF would presumably not result in the incorporation of additional Vpr molecules into HIV-1 particles. It is important to note that most, if not all anti-HIV-1 gene therapeutics, can be overwhelmed by utilizing very high levels of incoming virus with challenge experiments in vitro. As such, this may be problematic for inhibiting HIV1 replication in the interstices of lymphatic tissues of infected-individuals (Embretson et al., 1993). Therefore, "intravirion" gene therapeutics may address this concern by "immunizing" virions which "escape" from the initially infected cells.

F i g u r e 5 . Model for intravirion incorporation of proteins via the WxxF motif mediated by Vpr. Proteins containing the WxxF motif presumably first dock onto Vpr. As shown in the schematic, the carboxyl terminus or p6 region of p55, the Gag polyprotein precursor, may also play a role in packaging the complex into HIV-1 virus particles. The model further demonstrates a potential interaction of Gag with Ncp7 that may facilitate this virion import process.

139

Geist et al: HIV-1 integration and gene therapy Craigie R, Fujiwara T, and Bushman F ( 1 9 9 0 ) The IN protein of Molone murine leukemia virus processes the viral DNA ends and accomplishes their integration in vitro . C e l l 62, 829-837. Dropulic B, and Jeang KT ( 1 9 9 4 ) Gene therapy for human immunodeficiency virus infection: Genetic antiviral strategies and targets for intervention. Hum Gene Ther 5, 927-939. Ellison V, Abrams H, Roe T, Lifson J and Brown P ( 1 9 9 0 ) Human immuno-deficiency virus integration in a cell free system. J Virol 64, 2711-2715. Embretson J M, Zupancic JL, Ribas A, Burke P, Racz K, TennerRacz B and Haase A ( 1 9 9 3 ) Massive cohort infection of helper T lymphocytes and macrophages by HIV during the incubation period of AIDS. Nature 362, 359-362. Engelman A. and Craigie R ( 1 9 9 2 ) Identification of conserved amino acid residues critical for human immunodeficiency virus type 1 integrase function in vitro. J V i r o l 66, 63616369. Engleman A, Mizuuchi K, and Craigie R ( 1 9 9 1 ) HIV-1 DNA integration: Mechanism of viral DNA cleavage and DNA strand transfer. C e l l 67, 1211-1221 Erice A, and Balfour HH ( 1 9 9 4 ) Resistance of HIV-1 to antiretroviral agents: A review. C l i n I n f e c t D i s 18, 149156. Fletcher III TM, Soares MA, McPhearson S, Hui M, Wiskerson M, Muesing MA, Shaw GM, Leavitt AD, Boek JD and Hahn BM ( 1 9 9 7 ) Complementation of integrase function in HIV-1 virions. EMBO J 16, 5123-5138. Gallo RC, Sarin PS, Gelman EP, Robert-Guroff M, Richardson E., Kalyanaraman VS, Mann D, Sidhu GD, Stahl RE, ZollaPazner S, Leibowitch J, and Popovic M ( 1 9 8 3 ) Isolation of human T-cell leukemia virus in acquired immune deficiency syndrome (AIDS). S c i e n c e 220, 865-867. Goldgur Y, Dyda F, Hickman AB, Jenkins T.M, Craigie R, and Davies DR ( 1 9 9 8 ) Three new structures of the core domain of HIV-1 integrase: an active site that binds magnesium. Proc Natl Acad Sci 95, 9150-9154. Katz R A, Merkel G, Kulkosky J, Leis J. and Skalka AM ( 1 9 9 0 ) The avian retroviral IN protein is both necessary and sufficient for integrative recombination in vitro. C e l l 63, 87-95. Katz RA and Skalka AM ( 1 9 9 4 ) The retroviral enzymes. A n n R e v B i o c h e m 63, 133-173. Kulkosky J, BouHamdan M, Geist A, and RJ Pomerantz ( 1 9 9 9 ) A novel Vpr peptide interactor fused to integrase (IN) restores integration activity to IN-defective HIV-1 virions. V i r o l o g y 255, 77-85. Kulkosky J, Jones KS, Katz RA, Mack JPG, and Skalka AM ( 1 9 9 2 ) Residues critical for retroviral integrative recombination in a region that is highly conserved among retroviral/retrotransposon integrases and bacterial insertion sequence transposases. M o l C e l l B i o l 12, 2331-2338. Kulkosky J. and Skalka AM ( 1 9 9 4 ) Molecular mechanism of retroviral DNA integration. Pharmac Ther 61, 185-203. La Femina RL, Schneider CL, Robins HL, Callahan PL, Le Grow K, Roth E, Schleif WA and Emini EA ( 1 9 9 2 ) Requirement of active human immunodeficiency virus type 1 integrase enzyme for productive infection of human T-lymphoid cells. J Virol 66, 7414-7419. Levy-Mintz P, Duan L-X, Zhang H, Hu B, Dornadula G, Zhu M, KulkoskyJ, Bizub-Bender D, Skalka AM and Pomerantz RJ

In summary, the ability to package functional antiHIV-1 SFv-IN, via Vpr or the dWF motif, represents a novel and potentially useful technology that could contribute towards the efficient inhibition of HIV-1 replication. This new class of anti-retroviral agents could be propagated, in the patients, from cell to cell through HIV-1 virions that carry the molecular anti-retroviral. This concept could further assist in the development of efficient gene therapy-based antiretroviral strategies, as well as in dissection of basic molecular processes within the lentiviral life-cycle.

Acknowledgements The authors wish to thank Ms. Rita M. Victor and Ms. Brenda O. Gordon for their excellent secretarial assistance.

References Barrie-Sinoussi F, Chermann JC, Rey F, Nugeyre MT, Chamaret S, Gruest J, Dauguet C, Axler-Blin C, Vezinet-Brun F, Rouzioux C, Rozenbaum W, and Montagnier L ( 1 9 8 3 ) Isolation of a T-lymphotropic retrovirus from a patient at risk for acquired immune deficiency syndrome (AIDS). S c i e n c e 220, 868-871. Bizub-Bender D, Kulkosky J, and Skalka AM (1994) Monoclonal antibodies against HIV type 1 integrase: clues to molecular structure. AIDS Res and Hum R e t r o v i r u s e s 10, 1105-1115. BouHamdan M, Xue Y, Baudat Y, Hu B, Sire J, Pomerantz RJ and Duan L-X. ( 1 9 9 8 ) Diversity of HIV-1 Vpr interactions involves usage of the Wxxf motif of host cell proteins. J B i o l C h e m 273, 8009-8016. Bowerman B, Brown P, Bishop JM and Varmus H ( 1 9 8 9 ) A nucleoprotein complex mediates the integration of retroviral DNA. G e n e s D e v 2, 469-478. Brown P, Bowerman B, Varmus H, and Bishop JM ( 1 9 8 9 ) Retroviral integration: structure of the initial covalent product and its precursor, and a role for the viral IN protein. Proc Natl Acad Sci USA 86, 2525-2529. Bujacz G M, Jaskolski, Alexandrotas J, Wlodawer A, Merkel G, Katz RA, and Skalka AM ( 1 9 9 5 ) High resolution structure of the catalytic domain of avian sarcoma virus integrase. J M o l B i o l 253, 333-346. Bukrinsky MI, Sharova N, Dempsey MP, Stanwick TL, Bukrinskaya AG, Haggerty S, and Stevenson M ( 1 9 9 2 ) Active nuclear import of human immunodeficiency virus type 1 preintegration complexes. P r o c N a t l Acad S c i USA 89, 6580-6584. Bushman FD, Palmer A, Wingfield P, and Craigie R ( 1 9 9 3 ) Domains of the integrase protein of human immunodeficiency virus type 1 responsible for polynucleotidyl transfer and zinc binding. Proc Natl Acad Sci USA 90, 3428-3432. Cannon PM, ED Byles, SM Kingsman, and AJ Kingsman ( 1 9 9 6 ) . Conserved sequences in the carboxyl terminus of integrase that are essential for human immunodeficiency virus type 1 replication. J . V i r o l 70, 651-657. Chen SY, Khouri Y, Bagley J, and Marasco WA ( 1 9 9 4 ) Combined intra-and extracellular immunization against human immunodeficiency virus type 1 infection with a human anti-gp120. P r o c . N a t l . A c a d . S c i . U S A 91, 5932-5936.

140

Gene Therapy and Molecular Biology Vol 4, page 141 ( 1 9 9 6 ) Intracellular expression of single-chain variable fragments (SFv) to inhibit early stages of the viral life-cycle by targeting HIV-1 integrase. J Virol 70, 8821-8832. Levy-Mintz P, Duan L-X, Zhang H, Hu B, Dornadula G, Zhu M, Kulkosky J, Bizub-Bender D, Skalka AM and Pomerantz RJ. ( 1 9 9 8 ) Intracellular expression of single-chain variable fragments (SFv) to inhibit early stages of the viral life-cycle by targeting HIV-1 integrase. errata. J Virol 72: 3506. Lisziewicz J, Sund D, Lisziewicz A, and Gallo RC. ( 1 9 9 5 ) Anti-tat gene therapy: a candidate for large stage AIDS patients. Gene Therapy 2, 218-222. Liu H, Wu X, Xiao H, Conway JA and Kappes JC ( 1 9 9 7 ) Incorporation of functional human immunodeficiency virus type 1 integrase into virions independent of the Gag-Pol precursor protein. J Virol 71, ,7704-7710. Liu J, Woffendin C, Yang Z-Y, and Nabel GJ ( 1 9 9 4 ) Regulated expression of a dominant negative form of Rev improves resistance to HIV replication in T cells. Gene Therapy 1, 32-37. Mellors JW ( 1 9 9 6 ) Closing in on human immunodeficiency virus-1. Nat Med 3, 274-275. Mhashilkar AB, Bagley J, Chen S-Y, Szilvay AM, Helland DG, and Marasco WA. ( 1 9 9 5 ) Inhibition of HIV-1 Tat-mediated LTR transactivation and HIV-1 infection by anti-tat single chain antibodies. EMBO J 14, 1542-1551. Mizuuchi K ( 1 9 9 2 ) Polynucleotidyl transfer reactions in transpositional DNA recombination. J B i o l C h e m 267, 21273-21276. Pomerantz RJ, and D Trono ( 1 9 9 5 ) Genetic therapies for HIV infections: promise for the future. AIDS 9, 985-993. Reil H, Bukovsky AA, Gelderblom HR and Gottlinger HG ( 1 9 9 8 ) Efficient HIV-1 replication can occur in the absence of the viral matrix protein. EMBO J 17, 2699-2708. Richman DD ( 1 9 9 4 ) Viral resistance to antiretoviral therapy. In: S. Broder, T.C. Merigan and D. Bolognesi (Eds.) T e x t b o o k o f A I D S M e d i c i n e . Baltimore. Williams Serio D, Rizvi TA, Cartas M, Kalyanaraman VS, Weber IT, Koprowski H and Srinivasan A. ( 1 9 9 7 ) Development of a novel anti-HIV-1 agent from within:effect of chimeic Vprcontaining protease cleavage site residues on virus replication. Proc Natl Acad Sci USA 94: 3346-3351. Shaheen F, Duan L-X, Zhu MH, Bagasra O, and Pomerantz RJ. (1 9 9 6 ) Targeting human immunodeficiency virus type 1 reverse transcriptase by intracellular expression of single chain variable fragments to inhibit early stages of the viral life-cycle. J Virol 70, 3392-3400. St. Clair MH, Martin JL, Tudor-Williams G, Bach MC, Vavro CL, King DM, Kellam P, Kemp SD, and Larder BA. ( 1 9 9 1 ) . Resistance to ddI and sensitivity to AZT induced by mutation in HIV-1 reverse transcriptase. S c i e n c e 253, 1557-1559 Stevenson M and Gendelman H (1 9 9 5 ) Cellular and viral determinants that regulate HIV-1 infection in macrophages. J Leuk Biol 56, 278-288. Sun LQ, Pyaty J, Smythe J, Wang L, MacPherson J, Gerlach W, and Symonds G. ( 1 9 9 5 ) Resistance to human immunodeficiency virus type 1 infection conferred by transduction of human peripheral blood lymphocytes with ribozyme, antisense, or polymeric trans-activation response element constructs. Proc Natl Acad S c i USA 92, 7272-7276. Vink C, Yeheskiely E, Van Der Marel GA, Van Bloom J and Plasterk RH ( 1 9 9 2 ) Site-specific hydrolysis and alcoholysis of human immmunodeficiency virus DNA

terminal sequences mediated by the viral integrase protein. Nucl Acids Res 19, 6691-6698. Wu X, Liu H, Xiao H and Kappes JC ( 1 9 9 6 ) Proteolytic activity of human immunodeficiency virus Vpr- and Vpxprotease fusion proteins. V i r o l o g y 219, 307-313. Wu X , Liu H, Xiao H, Kim J, Seshaiah P, Natsoulis G, Boeke JD, Hahn BH and Kappes JC ( 1 9 9 5 ) Targeting foreign proteins to human immunodeficiency virus particles via fusion with Vpr and Vpx. J Virol 69, 3389-3398. Wu X, Xiao H, Conway JA, Hunter E and Kappes JC. ( 1 9 9 7 ) Functional RT and IN incorporated into HIV-1 particles independently of the Gag/Pol precursor protein. EMBO J 16, 5113-5122. Yu M, Ojwang J, Yamada O, Hampel A, Rappaport J, Looney D, and Wong-Staal FW ( 1 9 9 3 ) A hairpin ribozyme inhibits expression of diverse strains of human immunodeficiencY virus type 1. Proc Natl Acad Sci USA 90, 6340-6344.

Roger J. Pomerantz

141

Gene Therapy and Molecular Biology Vol 4, page 143 Gene Ther Mol Biol Vol 4, 143-158. December 1999.

Functional organization of the nuclear lamina Research Article

Michal Goldberg1, Einav Nili2, Gady Cojocaru2, Yonatan B. Tzur1, Raanan Berger2, Michael Brandies1, Gideon Rechavi2, Yosef Gruenbaum1 and Amos J. Simon2 1

Department of Genetics, The Institute of Life Sciences, The Hebrew University of Jerusalem, Jerusalem, 91904 Institute of Hematology, Chaim Sheba Medical Center, Tel-Hashomer and the Sackler School of Medicine, Tel-Aviv University, 52621 Israel. _________________________________________________________________________________________________ 2

Correspondence: Amos J. Simon, Ph.D., Institute of Hematology, Chaim Sheba Medical Center, Tel-Hashomer 52621, Israel. Tel: 9723-5302397; Fax: 9723-5302377; E-mail: wsamos@dapsas1.weizmann.ac.il Abbreviations: NE, nuclear envelope; IF, intermediate filaments; IMPs, integral nuclear membrane proteins; LBR, lamin B receptor; LAP, Lamina Associated Protein; NPC, nuclear pore complex; INM, inner nuclear membrane; ONM, outer nuclear membrane Key words: Apoptosis, nuclear assembly, histone, thymopoietin Received: 2 June 1999, accepted: 12 June 1999

Summary In eukaryotic cells, DNA replication, RNA processing and ribosome assembly all occur in the nucleus, while protein synthesis occurs in the cytoplasm. These activities are physically separated by the nuclear envelope. The inner nuclear membrane and the nuclear lamina are involved in organizing nuclear structure and regulating nuclear events including: nuclear organization, DNA replication, nuclear assembly and disassembly, apoptosis, and correct spacing of nuclear pores. In order to perform these activities the inner nuclear membrane and the nuclear lamina contain a unique set of proteins, including lamin type A and B, otefin, Young Arrest, LAP1, TMPO/LAP2, emerin, LBR and MAN. The proteins of the inner membrane form a complex network of interactions between themselves and with chromatin. Mutations in several of these proteins also result in known diseases. In this paper we review these interactions and discuss their possible roles in normal cell activity and in apoptosis. In addition, we demonstrate that specific regions in lamins can interact with the core histones H2A and H2B. We also show that Thymopoietin (TMPO)/ Lamina Associated Polypeptide 2 (LAP2) gene is alternatively spliced to form a large family of proteins with different size and different distribution within the nucleus. These results are discussed in relationship to the biological roles of the nuclear lamina. meshwork of intermediate filaments (IF) that is located underneath the inner nuclear membrane and abuts the peripheral chromatin (reviewed in Harel et al, 1998; Moir et al, 1995; Stuurman et al, 1998). The inner nuclear membrane and nuclear lamina contain unique proteins. Among them are the integral nuclear membrane proteins (IMPs) p58/lamin B receptor (LBR) (Worman et al, 1990; Worman et al, 1988), LAP1 (Martin et al, 1995), TMPO/LAP2 (Berger et al, 1996; Furukawa et al, 1995; Harris et al, 1994), emerin (Manilal et al, 1996), p34 (Simos and Georgatos, 1994) and p18 (Simos et al, 1996). The peripheral nuclear membrane proteins include type A and type B lamins (Fisher et al, 1986; McKeon et al, 1986), otefin (Harel et al, 1989; Padan et al, 1990), and Young Arrest (YA) (Lin et al, 1991; Lin and Wolfner, 1991). It is yet to be determined whether MAN protein(s) are integral or a peripheral proteins (Paulin Levasseur et al, 1996).

I. Introduction The main feature of eukaryotic cells is the presence of nuclear and cytoplasmic compartments that are separated by the nuclear envelope (NE). The NE is a large complex structure composed of outer and inner lipid bilayer membranes, nuclear pores, nuclear lamina and perhaps chromatin (Figure 1). The outer nuclear membrane is continuous with the endoplasmatic reticulum and is covered with ribosomes. The two membranes are separated by a 2040 nm perinuclear space and are connected at the nuclear pore complexes (reviewed in Burke, 1990a). The nuclear pore complexes are large protein structures that mediate the controlled transport of macromolecules between the cytoplasm and the nucleus (reviewed in Corbett and Silver, 1997; Davis, 1995; Fabre and Hurt, 1997; Gorlich and Mattaj, 1996). The nuclear lamina is a proteinaceous

143

Goldberg et al: Functional Organization of the Nuclear Lamina

Figure 1. Schematic view of the molecular organisation of the nuclear envelope. ONM, outer nuclear membrane; INM, inner nuclear membrane; NPC, nuclear pore complexes; LAP, Lamina Associated Protein; YA, Young Arrest.

gene revealed that lamins have also a central role in spacing the nuclear pores (Lenz-Bohme et al, 1997).

The proteins of the NE form a complex network of interactions (reviewed in Qian et al, 1998). Lamin seems to interact with LBR (Worman et al, 1988), LAP1 (Foisner and Gerace, 1993), TMPO/LAP2 (Foisner and Gerace, 1993), p18 (Simos et al, 1996), otefin and YA (Goldberg et al, 1998). p18 and p34 are associated with LBR (Simos and Georgatos, 1992). LBR (Foisner and Gerace, 1993; Pyrpasopoulou et al, 1996), YA (Lopez and Wolfner, 1997), and lamin (Glass et al, 1993; Glass and Gerace, 1990; Goldberg et al, 1999; Hoger et al, 1991; Taniura et al, 1995; Yuan et al, 1991) are all found to interact with chromatin. The inner nuclear membrane and the nuclear lamina are involved in several biological functions, including: (a) regulation of the NE shape and structure (Furukawa and Hotta, 1993; Furukawa et al, 1994; Newport et al, 1990); (b) disassembly, at the beginning of mitosis, and reassembly, at the end of mitosis of the NE (Ashery-Padan et al, 1997b; Burke and Gerace, 1986; Gant et al, 1999; Gant and Wilson, 1997; Heald and McKeon, 1990; Ulitzur et al, 1992, 1997; (c) providing anchoring sites for chromatin and affecting higher order chromatin organization (Belmont et al, 1993; Gerace and Foisner, 1994; Hutchison et al, 1994; Liu et al, 1997); (d) lamins are required for the elongation phase of DNA replication (Ellis et al, 1997; Gant et al, 1999; Goldberg et al, 1995; Moir et al, 1995; Newport et al, 1990); (e) lamins are very early targets for caspases in apoptosis, and inhibition of lamins apoptotic degradation delays apoptosis (Neamati et al, 1995; Oberhammer et al, 1994; Rao et al, 1997; Shimizu et al, 1998); and (f) the abnormal arrangement of nuclear pores in cells mutated in their lamin

A. Lamins: the main components of the nuclear lamina Lamins are the major and best studied proteins of the nuclear lamina. They are type V intermediate filament (IF) proteins, and, like all IF proteins, contain an ! helical rod domain flanked by a nonhelical amino terminal “head” and carboxy terminal “tail” domains. The lamin rod domain contains three !"helices, each composed of heptad repeats. These helices form coiled-coil interactions between lamin monomers, resulting in lamin dimers. The lamin dimers form longitudinal polar head-to-tail interaction, leading the formation of higher order structures. The outcome polymers further associate laterally to form 10 nm thick filaments. These 10 nm thick filaments can further associate to form the 50-200 nm thick lamin fibers (reviewed in Krohne, 1998; Stuurman et al, 1998). Lamins are divided into type A and type B lamins, according to their sequence and expression patterns. Type A lamins are essentially expressed in differentiated cells, have a neutral isoelectric point and are soluble during mitosis. Type B lamins are expressed in all somatic cells, have an acidic isoelectric point and remain associated with membrane vesicles during mitosis (reviewed in Moir et al, 1995; Stuurman et al, 1998). Metazoan cells contain between one to six lamin proteins. Mammalian lamin A, C (Fisher et al, 1986; McKeon et al, 1986), A delta10 (Machiels et al, 1996) and C2 (Furukawa et al, 1994) 144

Gene Therapy and Molecular Biology Vol 4, page 145 are all products of one alternatively spliced gene. Lamin B1 and B2 are encoded by two separate genes, and lamin B3 is a germ cell-specific spliced variant of lamin B 2 (Furukawa and Hotta, 1993). In Xenopus laevis, five lamin proteins were identified, lamin B1, B2, B3, B 4 and A (Benavente et al, 1985; Benavente, 1985; Stick, 1992, 1994). In chicken, there are three known lamin genes, termed lamin A, B2 and B1 (Peter et al, 1989; Vorburger et al, 1989). Drosophila melanogaster has two lamin genes, lamin Dm0 (Gruenbaum et al, 1988) and C (Bossie and Sanders, 1993). One lamin gene is present in the completely sequenced Caenorhobditis elegans, termed lamin CeLam-1 (Riemer et al, 1993). Interestingly, lamins are not present in the yeast Saccharomyces cerevisiae genome. All nuclear lamins, except lamin C, contain a CaaX box (C is cysteine, a is an aliphatic amino acid residue and X is any amino acid) at their carboxy terminus. The CaaX box undergoes specific post-translation modifications, which include farnesylation of the cysteine residue, endoproteolysis of the last three amino acids and carboxymethylation of the cysteine residue (reviewed in Moir et al, 1995). These modifications are essential, but not sufficient, for lamin association with the NE (Firmbach Kraft and Stick, 1995; Firmbach-Kraft and Stick, 1993; Krohne et al, 1989; Mical and Monteiro, 1998). Lamins are phosphorylated by several protein kinases in a cell-cycle regulated manner (Ottaviano and Gerace, 1985). Phosphorylation and dephosphorylation of lamins regulate their polymerization disassembly (Heald and McKeon, 1990; Ottaviano and Gerace, 1985; Peter et al, 1991), and their import to the nucleus (Haas and Jost, 1993; Hennekes et al, 1993). In mitotic cells, lamins are probably phosphorylated by cdc2 kinase (Dessev et al, 1991; Eggert et al, 1993; Heald and McKeon, 1990; Peter et al, 1991; Peter et al, 1990; Ward and Kirschner, 1990), the # isoform of protein kinase C (Fields et al, 1988; Goss et al, 1994; Hocevar et al, 1993; Hocevar and Fields, 1991), and mitogen-activated protein (MAP) kinase (Peter et al, 1992). In interphase cells, lamins are also phosphorylated by protein kinases A and C (Collas and Alestrom, 1997; Eggert et al, 1993; Eggert et al, 1991; Hennekes et al, 1993; Peter et al, 1990), indicating a separate role for lamin phosphorylation during interphase. The rod domain of lamins can bind otefin (Goldberg et al, 1998), LAP2 (Furukawa et al, 1998) and DNA (Baricheva et al, 1996; Christova et al, 1989; Galcheva-Gargova and Dessev, 1987; Luderus et al, 1992; Luderus et al, 1994; Rzepecki et al, 1998; Shoeman and Traub, 1990; Wedrychowski et al, 1989; Zhao et al, 1996), while the tail domain can bind actin (Sasseville and Langelier, 1998) and histones (Goldberg et al, 1999; Taniura et al, 1995). Lamins are located mainly at the nuclear periphery. However they are also localized internally in the nucleoplasm (Bridger et al, 1993; Goldman et al, 1992; Hozak et al, 1995; Lutz et al, 1992; Moir et al, 1994; Sasseville and Raymond, 1995). The lamin nucleoplasmic foci may represent sites for lamin assembly or modification, prior to their incorporation into the peripheral lamina (Goldman et al, 1992; Lutz et al, 1992; Sasseville and Raymond, 1995) or may be involved in DNA replication (Moir et al, 1994).

B. Lamin-chromatin interaction: implication for NE assembly and disassembly In higher eukaryotes, the NE disassembles during the prophase stage of mitosis and starts to reassemble around daughter chromosomes at late anaphase. Both lamins and lamin-associated proteins seem to be involved in this process. The NE disassembly is driven by phosphorylation of its major structural components, i.e. lamins (reviewed in Marshall and Wilson, 1997; Moir et al, 1995), proteins of the nuclear pore complex (Favreau et al, 1996; Macaulay et al, 1995), and probably LBR, LAP1 and TMPO/LAP2 (reviewed in Chu et al, 1998; Qian et al, 1998). In Drosophila embryos and Xenopus oocytes, the NE is fragmented into small vesicles, while the nuclear pore complexes disperse as soluble complexes and the nuclear lamina is depolymerized. NE assembly includes nuclear vesicles binding to chromatin, fusion of these vesicles, flattening onto the chromatin surface and assembly of the nuclear pore complexes (reviewed in Marshall and Wilson, 1997; Wilson and Wiese, 1996). In mammalian cells, the NE loses its identity as a subcompartment of the endoplasmic reticulum (ER) during mitosis. IMPs are probably resorted to the NE during late anaphase by diffusion through the ER and subsequently association with binding sites on the chromosomes. This process requires cooperative assembly of lamins and IMPs of the inner membrane (Gant et al, 1999; Ulitzur et al, 1997; Yang et al, 1997), LAP2! (Dechat et al, 1998) and chromatin. Both in vivo and in vitro studies proved that lamins have a major role in the assembly and disassembly of the NE, probably via their interaction with chromatin. Injection of anti-lamin antibodies to PtK2 cells resulted in cells that were not able to form normal daughter nuclei. The chromatin in the daughter nuclei remained arrested in a telophase-like configuration, and the telophase-like chromatin remained inactive as judged from its condensed state and by the absence of nucleoli (Benavente and Krohne, 1986). Three-dimensional in vivo studies in Drosophila and mammalian cells revealed that lamin filaments are closely associated with chromatin (Belmont et al, 1993). Genetic analysis also revealed that lamins are important for an intact NE and that lamins interact with chromatin. Mutations in the Drosophila lamin Dm0 gene revealed that it is an essential gene. Flies homozygous for a strong mutation in the Drosophila lamin Dm0 gene died following 9-16 hours of development. These flies demonstrate aberrant nuclear structure, dissociation of chromatin from the nuclear membranes and accumulation of annulate lamellae-like structures (Harel et al, 1998). Flies that carry a weak mutation in the lamin Dm0 gene (<20% lamin expression) show retardation in development, reduced viability, sterility, and impaired locomotion. These flies demonstrate enrichment in nuclear pore complexes in cytoplasmic annulate lamellae and in NE clusters and defects in the nuclear envelopes of several types of cells (Lenz-Bohme et al, 1997). The role of lamins in NE assembly was studied in vitro in several ways. Incubation of CHO cell extracts with antilamin A/C or anti-lamin B antibodies inhibited NE assembly (Burke and Gerace, 1986). Similarly, addition of anti-lamin 145

Goldberg et al: Functional Organization of the Nuclear Lamina Dm0 to a Drosophila extracts, in a cell-free nuclear assembly system, blocked the binding of nuclear vesicles to chromatin. The NE reformed upon supplementation of purified interphase lamin to the assembly extract (Ulitzur et al, 1992; Ulitzur et al, 1997). In the Xenopus nuclear assembly system, antibodies directed against both Xenopus lamins B2 and B3 inhibited the assembly of the nuclear membranes around chromatin (Dabauvalle et al, 1991). Interestingly, addition of antibodies directed only against Xenopus lamin B3 resulted in nuclei that were smaller and fragile, contained high density of nuclear pore complexes and failed to replicate their DNA (Goldberg et al, 1995; Jenkins et al, 1993; Meier et al, 1991; Newport et al, 1990; Spann et al, 1997). The latter experiments strongly suggest a role for lamins in nuclear organization. Lamins can bind specific DNA sequences in vitro. These include matrix/scaffold attachment regions (MAR/SARs, Luderus et al, 1992; Luderus et al, 1994), which lamins bind through their rod domain in a way that depends on lamins polymerization (Zhao et al, 1996). Lamins can also bind telomeric sequences (Shoeman and Traub, 1990), as well as, an A+T-rich DNA fragment that is located in the centromeric regions of Drosophila chromosomes and is organized similar to a MAR/SAR (Baricheva et al, 1996). Photo-crosslinking of Drosophila Kc cells that were grown in the presence of bromodeoxyuridine revealed that interphase lamins, but not mitotic lamins, are associated with both DNA and RNA in vivo (Rzepecki et al, 1998). Lamins also bind chromatin fragments and interphase chromatin (Goldberg et al, 1999; Hoger et al, 1991; Taniura et al, 1995; Yuan et al, 1991), in vitro assembled chromatin (Ulitzur et al, 1992), mitotic chromosomes (Glass et al, 1993; Glass and Gerace, 1990; Goldberg et al, 1999) and specific chromosomal proteins (Burke, 1990b; Goldberg et al, 1999; Taniura et al, 1995; Yuan et al, 1991). Avian type A lamin binds in vitro to polynucleosomes with a dissociation constant of about 1 nM (Yuan et al, 1991). Mammalian type A and B lamins (Glass et al, 1993; Glass and Gerace, 1990) and Drosophila lamin Dm0 can bind mitotic chromosomes in vitro (Goldberg et al, 1999). The chromatin binding activity of the mammalian lamins is localized at their tail domain, is mediated by core histones and possesses a dissociation constant in the range of 0.12-0.3 µM (Taniura et al, 1995). The binding activity of Drosophila lamin Dm0 is also localized at its tail domain, has a dissociation constant in the range of 1 µM and is mediated by histones H2A and H2B (see below). The actual association of lamins to histones may be stronger since lamins form large polymers in vivo. The Drosophila otefin is also involved in the process of NE assembly (Ashery Padan et al, 1997a). Trypsin treatment of membrane fractions of Xenopus and Drosophila extracts abolishes the ability of nuclear vesicles to bind demembranated sperm chromatin (Ulitzur et al, 1997; Wilson and Newport, 1988), indicating that NE assembly also depends on the activity of IMPs. IMPs, such as LBR, LAP1 and TMPO/LAP2, are possible targets for this trypsinization. It is likely that protein complexes containing lamin(s), otefin and IMPs mediate this NE assembly activity.

C. The family of Thymopoietin (TMPO)/ Lamina Associated Polypeptide 2 (LAP2) proteins Thymopoietins (TMPOs) are a family of proteins that are highly conserved in mammals and Xenopus and are putatively involved in functional nuclear architecture and cell cycle control. In mammalian cells, six alternatively spliced TMPO isoforms, designated !, #, #’, $, %, & and ' were isolated and characterized (Berger et al, 1996; Harris et al, 1994). All of them share an identical N-terminal 186 amino acid domain. One member of this family, TMPO# is the homolgue of the rat Lamina Associated Polypeptide 2 (LAP2) (Berger et al, 1996; Furukawa et al, 1995; Harris et al, 1995). TMPO#/LAP2is a type II integral protein of the inner NE, which binds lamin B1 and chromosomes in a phosphorylation dependent manner (Foisner and Gerace, 1993). It can be divided roughly into four domains: a hydrophilic C-terminus domain, a hydrophobic transmembrane domain, a NE targeting and lamina-binding domain and a chromatin-binding domain (Figure 2). As NE disassembly begins at prophase TMPO#/LAP2undergoes phosphorylation by mitotic factors, possibly by p34cdc2 kinase. This has been found to abolish its interphase binding to both lamin B1 and chromosomes. Consistently, during late anaphase, TMPO#/LAP2, as well as lamin-associated and non-associated vesicles, associates independently around chromatin to bind the surface of the decondensing chromosomes in order to complete NE reassembly (Foisner and Gerace, 1993; Yang et al, 1997). Thus, it is suggested that TMPO#/LAP2plays a key role in NE disassembly and reassembly during mitosis and linking chromatin to the NE in interphase. The specific binding of TMPO#/LAP2to lamin was confirmed using the two-hybrid system (Furukawa et al, 1998; Furukawa and Kondo, 1998). Amino acids 78-258 of lamin B 1 rod domain are sufficient for the direct binding of lamin B1 to TMPO#/LAP2 (Furukawa and Kondo, 1998). Interestingly, the lamin binding region of TMPO#/LAP2 coincides with TMPO#/LAP2 NE targeting domain and includes residues 298-373. The chromosomes binding site of TMPO#/LAP2resides within residues 1-85 (Furukawa et al, 1998; Gant et al, 1999; Yang et al, 1997), a region which is common to all TMPO isoforms. By transfecting truncated clones of TMPO!/LAP2, Yang et al (1997), showed that TMPO!/LAP2 has a role in regulating the dynamics of the nuclear lamina and the interaction between TMPO#/LAP2 and the nuclear lamina is required for nuclear growth after mitosis (Yang et al, 1997). Three other isoforms of the TMPO family, TMPOs &, % and $, are splicing variants of TMPO#/LAP2. All four isoforms share a putative hydrophobic transmembrane domain near their C-terminus via which TMPO#/LAP2 was suggested to bind to the NE (Berger et al, 1996; Furukawa et al, 1995). However, whereas TMPO#/LAP2 contains two putative p34cdc2 kinase phosphorylation sites, TMPOs % and & contain only one such site and TMPO$ lacks both sites (Berger et al, 1996), suggesting alternative regulating roles to these isoforms at the NE. 146

Gene Therapy and Molecular Biology Vol 4, page 147

Figure 2. The molecular organisation of TMPO#/LAP2 protein. The positions of the different domains are marked.

gene indicates an additional interaction between lamins and nuclear pore proteins, possible gp210.

Another isoform, TMPO! has a completely different Cterminus, lacking a trans-membrane domain but containing a putative nuclear localization signal (NLS) and a putative tyrosine and cdc2-specific phosphorylation sites. It is localized throughout the nuclear interior in interphase cells. During metaphase it dissociates from chromosomes and becomes concentrated around the spindle poles, probably due to specific phosphorylation. The protein is relocated to chromosomes at early stages of nuclear reassembly during telophase (Dechat et al, 1998; this study-see results). It is therefore suggested that, similar to the TMPO#/LAP2-type isoforms, TMPO! has a major role in the process of NE reassembly but in a different manner. For example, it will be interesting to reveal whether TMPO! is a true LAP that binds one of the lamins either in interphase or during mitosis like its TMPO#/LAP2 family member. TMPO', the shortest isoform, lacks both the trans-membrane domain of the TMPO#/LAP2-type isoforms and the NLS of TMPO!. The expression of the various TMPO isoforms is ubiquitous, with higher levels in proliferative tissues (Berger et al, 1996; Harris et al, 1994; Ishijima et al, 1996; Theodor et al, 1997; Zevin-Sonkin et al, 1992).

a. Lamin-LBR/p58 interaction: LBR/p58 is an IMP that binds lamin B in a saturable and specific fashion, with a dissociation constant of about 200 nM (Worman et al, 1988). LBR is composed of a nucleoplasmic amino terminal domain followed by a hydrophobic domain that contains eight putative transmembrane segments (Worman et al, 1990). LBRâ&#x20AC;&#x2122;s hydrophbic region has high sequence homology to the yeast sterol c-14 reductase (Gerace and Foisner, 1994; Shimanuki et al, 1992), and it was shown to posses c-14 reductase activity in Saccharomyces cerevisiae (Silve et al, 1998). The amino terminal domain of LBR has three DNA-binding motifs (Worman et al, 1990) and was suggested to bind DNA (Ye and Worman, 1994). LBR is phosphorylated in a cell cycle dependent manner, on serine residues in interphase and on serine and threonine residues in mitosis (Courvalin et al, 1992; Nikolakaki et al, 1997; Nikolakaki et al, 1996; Papoutsopoulou et al, 1999; Simos and Georgatos, 1992; Worman et al, 1990). It is degraded at late stages of apoptosis, interacts with chromatin (Pyrpasopoulou et al, 1996), and with several nuclear proteins including lamin B (Worman et al, 1988), p34, p18 (Simos and Georgatos, 1992) and the human protein homologous to the Drosophila heterochromatin associated protein, HP1 (Ye and Worman, 1996; Ye et al, 1997). Though its name, there is a recent work, utilizing biochemical fractionation analysis, that suggests that LBR and lamin B do not bind directly to each other in vivo (Mical and Monteiro, 1998). By using in vitro reconstituted vesicles assay it was shown that LBR can provide chromatin docking sites for nuclear vesicles (Pyrpasopoulou et al, 1996). An LBR-related protein from sea urchin can be

D. Interactions between lamins and other nuclear lamina and inner nuclear membrane proteins. 1. Lamin interaction with IMPs The interactions between lamins and IMPs are discussed below. These interactions are required for lamins to exert their biological roles. The interaction between lamins and TMPO#/LAP2 is discussed above. The abnormal arrangement of nuclear pores in cells mutated in their lamin 147

Goldberg et al: Functional Organization of the Nuclear Lamina involved in providing chromatin docking sites for nuclear vesicles and can bind to the nuclear lamina (Collas et al, 1996). b. Lamin-LAP1A-C interaction: LAP1A-C are related IMPs, which are presumably alternatively spliced transcripts of the same gene (Foisner and Gerace, 1993; Martin et al, 1995). LAP1A and LAP1B, but not LAP1C, can bind to both lamin A and lamin B in vitro. However, LAP1C associates with type B lamin in vivo as a component of a protein complex, albeit a direct LAP1C-lamin interaction has not been demonstrated (Maison et al, 1997). Cloning of LAP1C revealed that it is a type II IMP with a single membrane spanning region and a hydrophilic amino terminal domain that is exposed to the nucleoplasm. LAP1 isotypes are differentially expressed during development. LAP1C protein is present both in non-differentiated and differentiated cells whereas LAP1A and LAP1B proteins are abundant only in differentiate cells (Martin et al, 1995).

development, except for sperm cells. In eggs and young embryos, otefin is also associated with the maternal fraction of membrane vesicles (Ashery Padan et al, 1997a). Otefin is a phosphoprotein in vivo and is phosphorylated by cdc2 protein kinase and cAMP-dependent protein kinase in vitro (Ashery Padan et al, 1997a). The otefin protein contains a large hydrophilic domain and a C-terminal hydrophobic domain of 17 amino acids. The C-terminal domain of otefin is essential but not sufficient for the targeting of otefin to the nuclear periphery. Sequences at the hydrophilic domain of otefin are required for both the localization and stabilization of otefin in the NE (Ashery Padan et al, 1997b), and for interaction with lamins. Otefin and lamin interact with each other in the yeast two-hybrid system and are present in vivo in the same protein complex (Goldberg et al, 1998). This interaction requires the rod domain of lamin Dm0 and two hydrophilic regions in otefin (Goldberg et al, 1998). Otefin was shown to be involved in NE assembly as the inhibition of otefin activity in Drosophila cell-free system, with anti-otefin antibodies, blocked the attachment of the nuclear vesicles to chromatin (Ashery Padan et al, 1997a). The similar phenotype observed in the inhibition studies of both lamin Dm0 (see above) and otefin is probably due to their co-localization to nuclear vesicles in the same protein complex (Goldberg et al, 1998).

c. Lamin-emerin interaction: Mutations in the emerin gene result in the X-linked Emery-Dreifuss muscular dystrophy (Bione et al, 1994). Emerin is a ubiquitous IMP of the inner nuclear membrane (Manilal et al, 1996; Nagano et al, 1996) with sequence and structural homology to TMPO#/LAP2 (Manilal et al, 1998). It is reasonable to speculate that emerin and lamin A/C interact with each other, directly or indirectly, since mutations in the gene encoding the human lamin A/C cause the autosomal dominant Emery-Dreifuss muscular dystrophy (Bonne et al, 1999) and both proteins are associated with the inner nuclear membrane.

E. Lamin and apoptosis During apoptosis, lamins are major targets for the caspase family of proteases, which trigger nuclear lamina breakdown. Lamin degradation has been reported in different cell types and in response to different apoptosisinducing stimuli (Anjum and Khar, 1997; Antoku et al, 1997; Fraser et al, 1997; Kaufmann, 1989; Kawahara et al, 1998; Kluck et al, 1997; Lazebnik et al, 1993; Neamati et al, 1995; Oberhammer et al, 1994; Orth et al, 1996; Rao et al, 1996; Shimizu et al, 1998; Shimizu and Pommier, 1997; Takahashi et al, 1996a; Takahashi et al, 1996b; Ucker et al, 1992; Zhivotovsky et al, 1997). Lamin B is phosphorylated by PKC!during apoptosis before its degradation (Shimizu et al, 1998). The degradation of lamin B1 precedes DNA fragmentation in apoptotic thymocytes (Neamati et al, 1995), HeLa cells (Mandal et al, 1996) and HL60 cells (Shimizu et al, 1998). Lamins may be involved directly in the regulation of apoptosis since ectopic expression of uncleavable mutant lamin A or B caused a 12 hours delay in the onset of apoptosis. The cells expressing the mutant lamins failed to undergo chromatin condensation and nuclear shrinking, which is a hallmark of apoptotic cell death. However, in these cells the NE collapsed and the nuclear lamina remained intact. However, the late stages of apoptosis were not affected (Rao et al, 1996). Thus, the proteolysis of lamin facilitates the nuclear events of apoptosis, perhaps by facilitating nuclear breakdown. A possible role for lamin cleavage during apoptosis is to enable the dissociation of the chromatin from the nuclear lamina and thereby to affect nuclear condensation.

2. Lamin interaction with peripheral proteins a. Lamin-YA interaction: YA is a Drosophila NE protein that is present in the nuclear lamina during the first two hours of zygotic divisions and whose function is essential for initial embryonic development (Lin et al, 1991; Lin and Wolfner, 1991; Lopez et al, 1994). YA molecules are present in complexes with each other (Goldberg et al, 1998; Liu and Wolfner, 1998), and embryonic YA can associate with decondensed chromatin in vitro (Lopez and Wolfner, 1997). The carboxy terminal 179 amino acids of YA are necessary to target and retain YA in the NE. The condensation state of the chromosomes in YA-deficient eggs and embryos is abnormal (Liu et al, 1995) and ectopically expressed YA associates with polytene chromosomes in vivo (Lopez and Wolfner, 1997). YA interacts with Drosophila lamin Dm 0 in the yeast two-hybrid system. This interaction requires the rod and tail domains and part of the head domain of lamin Dm0 and the C-terminal domain of YA (Goldberg et al, 1998). b. Lamin-otefin interaction: Otefin is a peripheral nuclear membrane protein, characterized in Drosophila, which has no apparent homology to other known proteins (Padan et al, 1990). In Drosophila, otefin is present in the NE of essentially all cells examined during Drosophila 148

Gene Therapy and Molecular Biology Vol 4, page 149

II. Results and discussion

The binding of lamin Dm0 to chromosomes did not require lamin polymerization, since lamin Dm0 mutated in Arginine 64 (R64>H), which is unable to polymerize (Zhao et al, 1996), still bound chromosomes in a similar way to that of wild type lamin Dm 0 (Figure 4). Further mapping of the regions that are involved in lamin Dm0 binding to mitotic chromosomes, using deletion mutants of lamin Dm0 protein, revealed that the lamin Dm0 tail domain (amino acids 425-622) mediates this binding (Figures 3, 4). The isolated lamin Dm 0 tail domain bound to the periphery of mitotic chromosomes with similar intensity to that of wild type lamin Dm0 (Figure 4). The dissociation constant of lamin tail domain binding to chromatin is approximately 1 µM (Goldberg et al, 1999). The existence of a chromatin binding site in the lamin tail domain was also reported for Xenopus type A and B2 lamins (Hoger et al, 1991) and for mammalian type A and B lamins (Taniura et al, 1995). Deletion mutants of the lamin Dm0 tail domain were designed in order to map the region in this domain that mediates the binding of lamin Dm0 to chromosomes. Each deletion mutant was expressed in bacteria, purified to near homogeneity and subjected to the binding assay. The ability of the different protein constructs to bind mitotic chromosomes is summarized in Figure 4. The T425-522 protein bound chromosomes with lower immunofluorescence intensity as compared to the complete lamin tail domain, but with significantly higher intensity as compared to the T473-622, T523-622 and T473-572 proteins. In addition, the intensity of T473-572 binding to chromosomes was close to background levels (Figure 4). Taken together, these data indicate that there are two regions that are required for efficient binding of lamin Dm0 to chromosomes. The major binding region includes amino acids 425-473 of lamin Dm0 and the minor binding region includes amino acids 572-622 of lamin Dm0. T425-572 protein that contains the first binding region binds to chromosomes with similar efficiency to the complete tail domain, while T523-622 and T473-622 proteins that contain the second binding region bind to chromosomes with significantly weaker intensity. The lower signal obtained with T425-522, as compared to T425-572, can be explained by a difference in protein folding or requirement of additional residues.

A. Lamin Dm0 binds chromosomes through two separate regions within its tail domain. Lamin Dm0 can bind specifically to decondensed sperm chromatin in Drosophila in vitro nuclei assembly assay, as well as to mitotic chromosomes (Goldberg et al, 1999; Ulitzur et al, 1992). The binding assay of lamin Dm0 to mitotic chromosomes includes incubation of purified lamin Dm0 protein with isolated CHO mitotic chromosomes for 30-60 min at 220C in the presence of excess amount of nonrelevant proteins. Lamin binding to the surface of the chromosomes was observed following immunofluorescence analysis with either monoclonal anti-lamin Dm0 antibody or monoclonal anti-RGSHis antibody (Qiagen, Germany) that recognises a RGS(H)4 epitope, as primary antibodies, and Cy3-conjugated goat-anti-mouse antibodies, as secondary antibodies. Alternatively, it was detected directly by a fusion between lamin Dm0 and the green fluorescence protein (GFP) (Figure 3).

B. Lamin Dm0 tail domain binding to chromosomes can be displaced with histones H2A and H2B Figure 3. The binding of the tail domain of lamin Dm0 to mitotic chromosomes can be observed with different detection systems. Upper panel – Detection was with monoclonal anti-lamin Dm0 antibody, as primary antibody, and Cy3-conjugated goat-antimouse antibodies, as secondary antibodies. Middle panel Detection was with monoclonal anti-RGSHis antibody (Qiagen, Germany), as primary antibody, and Cy3-conjugated goat-antimouse antibodies, as secondary antibodies. Lower panel – The tail domain of lamin Dm0 was fused in frame to the Green Fluorescent Protein (GFP). Chromatin was stained with the DNA-specific dye DAPI. In all three cases, viewing was with a Leitz microscope equipped with epifluorescence. (Bar = 5µ).

In order to detect the molecules that are required for the binding of lamin Dm0 to chromosomes, we analyzed the ability of various chromosomal components to displace the binding of the lamin Dm 0 tail domain to mitotic chromosomes. MAR/SAR DNA sequences were previously found to bind lamin Dm0 in vitro (Baricheva et al, 1996; Luderus et al, 1992; Luderus et al, 1994), through the rod domain of lamin Dm0 (Zhao et al, 1996). MAR/SAR DNA sequences, as well as yeast ARS sequences and plasmid DNA, are not the target of lamin tail binding, since they could not displace the binding of the lamin Dm0 tail domain 149

Goldberg et al: Functional Organization of the Nuclear Lamina to chromosomes (Goldberg et al, 1999). This is expected since it is the tail domain that mediates the binding of lamin Dm0 to chromosomes. In contrast, isolated nucleosome core particles and commercially crude preparation of core histones and histone H1 displaced lamin tail domain binding to chromosomes (Goldberg et al, 1999). To identify specific histone(s) that interact with the lamin Dm0 tail domain, the individual core histones (Figure 5) or histone H1 (Goldberg et al, 1999) were tested for their ability to displace the binding of the lamin Dm0 tail domain to chromosomes. A 30-120-fold molar excess of purified histones H2A or H2B could each displace the binding of lamin Dm0 tail domain to chromosomes (Figure 5). In contrast, a 30-120-fold molar excess of individual histones H3 or H4 did not displace the lamin Dm0 tail domain binding to chromosomes (Figure 5). Similarly, a 32-fold molar excess of purified histone H1 (amino acids 1-142) did not displace the binding of lamin Dm0 tail domain to chromosomes (Goldberg et al, 1999). It is, therefore, likely that the interaction between lamin Dm0 and chromosomes is mediated by histones H2A and H2B. The biological significance of this lamin-histones interaction is probably in the processes of nuclear envelope assembly and nuclear organization. The significance of this interaction in vivo is currently under investigation..

C. Analysis of the mouse TMPO !, #, #â&#x20AC;&#x2122;, &, %, $ and ' isoforms. The first mouse TMPO isoforms that was cloned was TMPO' (Berger et al, 1996; Theodor et al, 1997). Its cDNA clone was isolated from thymus cDNA library and characterized, using a 126 bp fragment, encoding thymopoietin amino acids 1-42, from the bovine TMPO cDNA (Zevin-Sonkin et al, 1992), as a probe. The same library was subsequently screened with a 790 bp probe, derived from the N-terminus of the TMPO' cDNA. Repeated screenings and restriction enzyme analysis revealed at least six distinct TMPO transcripts (Figure 6). Examination of the mouse TMPO! sequence revealed a short region of basic amino acids (amino acid 188-194) suggestive of a NLS (Kalderon et al, 1984), and a possible tyrosine phosphorylation site (amino acids 618-625) (Patschinsky et al, 1982). The sequence S/T-P-X-X, a potential recognition sequence for cdc2-related kinases (Nigg, 1993), is found 10 times throughout the TMPO! sequence. Hydropathy analysis revealed that similar to the human isoforms, all mouse TMPOs lack an N-terminal hydrophobic signal sequence typical of secreted polypeptides (not shown). However, this analysis revealed that TMPOs #, &, % and $ contain, like the human TMPO# and $, a hydrophobic transmembrane domain near their Cterminus (amino acid 409-432 of TMPO#), suggesting their association with cellular membranes such as the nuclear envelope.

Figure 4. Binding of different lamin Dm0 constructs to mitotic chromosomes derived from Chinese Hamster Ovaries (CHO) requires specific sequences in their tail domain. R64>H is a point mutation in lamin Dm0 in Arginine 64 that inhibits its ability to form filaments. The numbers in the name of the construct represent the amino acids of lamin Dm0 that are included in the construct and their position is shown below the map of the construct. ((( indicates strong binding; ((_indicates medium binding, ( indicates weak binding and X indicates lack of detectable binding. ) indicates detection with the monoclonal anti-RGSHis antibody or with GFP.

150

Gene Therapy and Molecular Biology Vol 4, page 151

Figure 6. TMPO/LAP2 is a growing family of nuclear envelope proteins derived from a single alternatively spliced gene. The numbers above each isoform represent the corresponding amino acid. The numbers below the ! and # isoforms represent the corresponding exons in the TMPO gene. The exons that are common to all TMPOs are colored in blue and the alternatively spliced exons have different colors. NLS-nuclear localization signal; TM-transmembrane domain; 5’- the last 5 amino acids of the ' isoform that are included in intron 5 of the TMPO gene. Asterisk in #, & and % isoforms show the location of putative sites for the cdc2-kinase.

Figure 5. The binding of lamin Dm0 to mitotic chromosomes can be displaced with histones H2A and H2B. The tail domain of lamin Dm0 was used to bind the mitotic chromosomes in the presence or absent of histones. DAPI staining – left panwels; MAb anti-lamin Dm0 antibody as primary antibody and Cy3-conjugated goat-antimouse antibodies as secondary antibodies – right panels. (Bar=5µ).

Mouse TMPO !, #, %, &, $ and ' share an identical Nterminal 186 amino acids domain (Figure 6). Like the human and the bovine TMPOs, amino acids 1-49 are highly homologous to the originally purified 49-amino acid bovine TP (Schlesinger and Goldstein, 1975). After Glu186, TMPO! diverges from the other TMPOs. TMPO& differs from TMPO# only in lacking the #-specific residues 220259. TMPO% lacks both the # and the &/#-specific domains, which is contained within residues 220-291 of TMPO#. TMPO$ is missing the #, &/# and %/&/#-specific domains contained within amino acids 220-328 of TMPO#. Interestingly, two TMPO# 3’-UTR versions were isolated in mouse and designated TMPO# and TMPO#’. The two clones are identical in their ORF sequence, but differ in an additional 3’- UTR sequence starting at A1715 of the #’ clone, probably due to an alternative polyadenylation signal.

D. Monoclonal antibodies identify alternatively spliced forms of TMPOs Three monoclonal antibodies (mAbs) were generated in order to characterize the expression and localization of the different TMPO isoforms (Dr. Gideon Goldstein, generous gift). The 6E10 mAb was raised against a synthetic peptide common to the N-terminus of all TMPOs. Indeed, Western blot analysis revealed that the 6E10 antibody recognized both recombinant and native TMPOs !, #, $ and ' as well as native TMPOs & and % (Figure 7a). Although all TMPOs are present in both the hematopoietic K562 and the epithelial HeLa cell lines, TMPO# is clearly more abundant than all

151

Goldberg et al: Functional Organization of the Nuclear Lamina other TMPOs and the relative amounts of the different TMPOs vary between these cell lines. Thus, it would be interesting to investigate if the different levels of expression of the various TMPOs in specific cells have a role in the biological functions that these cells have to exert. The 4G10 mAb was raised against recombinant TMPO! and the 6G11 mAb was raised against a synthetic peptide, derived from TMPO#/LAP2 specific region (amino acids 313-330). These antibodies recognized both recombinant and native TMPO! and TMPO# , respectively (Figure 7b and c). Interestingly, the 6G11 antibody cross-reacted with two close bands in both cell lines, suggesting two forms of TMPO# that differ in their phosphorylation pattern (Figure 7c).

TMPO#/LAP2-type isoforms. The localization of TMPO! was studied by immunofluorescence microscopy during the different cell cycle stages of HeLa cells, using the 4G10 anti TMPO! mAb. In interphase, TMPO! was diffusely distributed throughout the nucleoplasm but was excluded from the nucleoli (Figure 8a). In addition, it was found in the centrosomal region (the bright spots in Figure 8a). In early prophase, prior to its complete disintegration of the nuclear envelope, a large fraction of TMPO! was observed at the nuclear periphery (Figure 8a). In metaphase and early anaphase, TMPO! was disseminated throughout the cell In addition, the two centrosomes at the spindle poles were stained brightly (Figure 8b). In telophase a bulk of the TMPO! protein re-localized to the two newly forming nuclear envelopes. At this stage, the nucleus itself was negatively stained (Figure 8c). Upon completion of mitosis and reentry of the cells to G1, TMPO! regained its typical interphase intranuclear and centrosomal localization, while the nuclear envelope was not stained. This staining pattern of TMPO! throughout the cell cycle was exemplified in other cell lines such as PTK, K562 and HL-60 (not shown). The presence of TMPO! at the nuclear periphery during prophase indicates a possible association of TMPO! with NE components during this stage. The association of TMPO! with the chromatin periphery during telophase strongly suggests a role for this isoform in nuclear assembly.

E. TMPO! is localized both to the nucleus and centrosomes, while the TMPO/ LAP2 type isoforms are localized to the nuclear envelope. TMPO! is the largest member of the TMPO family. It shares the first 187 amino acids of all TMPOs but has a unique long C-terminus. This region lacks a transmembrane domain but contains a NLS, a putative tyrosine phosphorylation site and two perfect p34cdc2 kinase phosphorylation sites (Nigg, 1993). This unique region suggests a different intra-cellular localization than that of the

Figure 7. Different monoclonal anti-TMPO antibodies detect the various of TMPO isoforms and show variability in their relative expression in human cells. (a) Detection was with the 6E10 monoclonal anti-all TMPO antibody. The left lane was loaded with recombinant proteins. Although all TMPO proteins are expressed in both Hela and K562 cells, their relative abundance is different. (b) Detection was with 4G10 monoclonal anti-TMPO! antibody. (c) Detection was with 6G11 monoclonal antiTMPO# antibody.

152

Gene Therapy and Molecular Biology Vol 4, page 153

Figure 8. TMPO! and TMPO# have different localisation in human cells. TMPO! localisation was detected with the 4G10 monoclonal anti-TMPO! antibody (a-c) and TMPO# localisation was detected with the 6G11 monoclonal antiTMPO# antibody (d-f). P, a cell in prophase; M, metaphase; T, telophase. Panel d shows immnogold electron microscope detection of TMPO#; all other panels show indirect immunofluorescence analysis.

References

Previous studies have shown that TMPO#/LAP2 is an integral protein of the inner nuclear membrane (Foisner and Gerace, 1993). Its NE targeting and binding are probably mediated by its carboxy terminus-located lamin binding region and trans-membrane domain (Furukawa et al, 1995). Since TMPOs &, % and $ also share this carboxyterminus and parts of the lamin binding domain, and transfected TMPO$ was localized to the NE (data not shown), we suggest that these isoforms are also nuclear membrane proteins. The localization of TMPO#/LAP2 to the nuclear envelope during the cell cycle was studied in Hela cells, by immunofluorescence and immuno-electron microscopy. In interphase, immuno-gold analysis demonstrated the localization of TMPO#/LAP2 to the NE (Figure 8d). The immuno-fluorescence staining was confined to the NE, as well as to intra nuclear speckles (Figure 8f). These speckles possibly represent replication origins (Gant et al, 1999; Moir et al, 1994). In metaphase, TMPO# was dispersed in the cytoplasm (Figure 8e). In telophase, TMPO# re-localized to the two newly forming nuclear envelopes. However, the staining of TMPO# during telophase is more continuous than that of TMPO!.

Anjum R and Khar A (1997). Caspase-mediated apoptosis in AK5 tumor cells: a cell-free study using peptide inhibitors and antisense strategy. Exp Cell Res 236, 371-377. Antoku K, Liu Z, and Johnson DE (1997). Inhibition of caspase proteases by CrmA enhances the resistance of human leukemic cells to multiple chemotherapeutic agents. Leukemia 11, 1665-1672. Ashery Padan R, Ulitzur N, Arbel A, Goldberg M, Weiss AM, Maus N, Fisher PA, and Gruenbaum Y (1997). Localization and posttranslational modifications of otefin, a protein required for vesicle attachment to chromatin, during Drosophila melanogaster development. Mol Cell Biol 17, 4114-4123. Ashery Padan R, Weiss AM, Feinstein N, and Gruenbaum Y (1997b). Distinct regions specify the targeting of otefin to the nucleoplasmic side of the nuclear envelope. J Biol Chem 272, 2493-2499. Baricheva EA, Berrrios M, Bogachev SS, Borisevich IV, Lapik ER, Sharakhov IV, Stuurman N, and Fisher PA (1996). DNA from Drosophila _-heterochromatin binds specifically to nuclear lamins in vitro and the nuclear envelope in situ. Gene 171, 171-176. Belmont AS, Zhai Y, and Thilenius A (1993). Lamin B

153

Goldberg et al: Functional Organization of the Nuclear Lamina distribution and association with peripheral chromatin revealed by optical sectioning and electron microscopy tomography. J Cell Biol 123, 1671-1685.

mitosis-specific phosphorylation and is a substrate for p34cdc2-type protein kinase. J Biol Chem 267, 1903519038.

Benavente R and Krohne G (1986). Involvement of nuclear lamins in postmitotic reorganization of chromatin as demonstrated by microinjection of lamin antibodies. J Cell Biol 103, 1847-1854.

Dabauvalle MC, Loos K, Merkert H, and Scheer U (1991). Spontaneous assembly of pore complex-containing membranes ("annulate lamellae") in Xenopus egg extract in the absence of chromatin. J Cell Biol 112, 1073-1082.

Benavente R, Krohne G, and Franke WW (1985). Cell typespecific expression of nuclear lamina proteins during development of Xenopus laevis. Cell 41, 177-190.

Davis, L. I. (1995). The nuclear pore complex. Annu Rev Biochem 64, 865-896. Dechat T, Gotzmann J, Stockinger A, Harris CA, Talle MA, Siekierka JJ, and Foisner R (1998). Detergent-salt resistance of LAP2alpha in interphase nuclei and phosphorylationdependent association with chromosomes early in nuclear assembly implies functions in nuclear structure dynamics. EMBO J 17, 4887-4902.

Benavente RK (1985). Change of karyoskeleton during spermatogenesis of Xenopus: expression of lamin LIV, a nuclear lamina protein specific for the male germ line. Proc Natl Acad Sci USA 82, 6176-6180. Berger R, Theodor L, Shoham J, Gokkel E, Brok-Simoni F, Avraham KB, Copeland NG, Jenkins NA, Rechavi G, and Simon AJ (1996). The characterization and localization of the mouse thymopoietin/lamina- associated polypeptide 2 gene and its alternatively spliced products. Genome Res 6, 361370.

Dessev G, Iovcheva DC, Bischoff JR, Beach D, and Goldman R (1991). A complex containing p34cdc2 and cyclin B phosphorylates the nuclear lamin and disassembles nuclei of clam oocytes in vitro. J Cell Biol 112, 523-533. Eggert M, Radomski N, Linder D, Tripier D, Traub P, and Jost E (1993). Identification of novel phosphorylation sites in murine A-type lamins. Eur J Biochem 213, 659-666.

Bione S, Maestrini E, Rivella S, Mancini M, Regis S, Romeo G, and Toniolo D (1994). Identification of a novel X-linked gene responsible for Emery-Dreifuss muscular dystrophy. Nat Genet 8, 323-327.

Eggert M, Radomski N, Tripier D, Traub P, and Jost E (1991). Identification of phosphorylation sites on murine nuclear lamin C by RP-HPLC and microsequencing. FEBS Lett 292, 205-209.

Bonne G, Di Barletta MR, Varnous S, Becane HM, Hammouda EH, Merlini L, Muntoni F, Greenberg CR, Gary F, Urtizberea JA, Duboc D, Fardeau M, Toniolo D, and Schwartz K (1999). Mutations in the gene encoding lamin A/C cause autosomal dominant Emery- Dreifuss muscular dystrophy. Nat Genet 21, 285-288.

Ellis DJ, Jenkins H, Whitfield WG, and Hutchison CJ (1997). GST-lamin fusion proteins act as dominant negative mutants in Xenopus egg extract and reveal the function of the lamina in DNA replication. J Cell Sci 110, 2507-2518.

Bossie CA and Sanders MM (1993). A cDNA from Drosophila melanogaster encodes a lamin C-like intermediate filament protein. J Cell Sci 104, 1263-1272.

Fabre E and Hurt E (1997). Yeast genetics to dissect the nuclear pore complex and nucleocytoplasmic trafficking. Annu Rev Genet 31, 277-313.

Bridger JM, Kill IR, O'Farrell M, and Hutchison CJ (1993). Internal lamin structures within G1 nuclei of human dermal fibroblasts. J Cell Sci 104, 297-306

Favreau C, Worman HJ, Wozniak RW, Frappier T, and Courvalin JC (1996). Cell Cycle Dependent Phosphorylation Of Nucleoporins and Nuclear Pore Membrane Protein Gp210. Biochemistry 35, 8035-8044.

Burke B (1990a). The nuclear envelope and nuclear transport. Curr Opin Cell Biol 2, 514-20.

Fields AP, Pettit GR, and May WS (1988). Phosphorylation of lamin B at the nuclear membrane by activated protein kinase C. J Biol Chem 263, 8253-8260.

Burke B (1990b). On the cell-free association of lamins A and C with metaphase chromosomes. Exp Cell Res 186, 169-176. Burke B and Gerace L (1986). A cell free system to study reassembly of the nuclear envelope at the end of mitosis. Cell 44, 639-652.

Firmbach-Kraft I and Stick R (1995). Analysis of nuclear lamin isoprenylation in Xenopus oocytes: isoprenylation of lamin B3 precedes its uptake into the nucleus. J Cell Biol 129, 1724.

Christova R, Yaneva J, and Galcheva-Gargova Z (1989). Some features of DNA fragments associated in vivo with the nuclear lamina. Biosci Rep 9, 587-592.

Firmbach-Kraft I and Stick R (1993). The role of CaaXdependent modifications in membrane association of Xenopus nuclear lamin B3 during meiosis and the fate of B3 in transfected mitotic cells. J Cell Biol 123, 1661-1670.

Chu A, Rassadi R, and Stochaj U (1998). Velcro in the nuclear envelope: LBR and LAPs. FEBS Lett 441, 165-169.

Fisher DZ, Chaudhary N, and Blobe, G (1986). cDNA sequencing of nuclear lamins A and C reveals primary and secondary structural homology to intermediate filament proteins. Proc Natl Acad Sci USA 83, 6450-6454.

Collas P and Alestrom P (1997). Nuclear localization signals: a driving force for nuclear transport of plasmid DNA in zebrafish. Biochem Cell Biol 75, 633-640. Collas P, Courvalin JC, and Poccia D (1996). Targeting of membranes to sea urchin sperm chromatin is mediated by a lamin B receptor-like integral membrane protein. J Cell Biol 135, 1715-1725.

Foisner R and Gerace L (1993). Integral membrane proteins of the nuclear envelope interact with lamins and chromosomes, and binding is modulated by mitotic phosphorylation. Cell 73, 1267-1279.

Corbett AH and Silver PA (1997). Nucleocytoplasmic transport of macromolecules. Microbiol Mol Biol Rev 61, 193-211.

Fraser AG, McCarthy NJ, and Evan GI (1997). drICE is an essential caspase required for apoptotic activity in Drosophila cells. EMBO J 16, 6192-6199.

Courvalin JC, Segil N, Blobel G, and Worman HJ (1992). The lamin B receptor of the inner nuclear membrane undergoes

154

Gene Therapy and Molecular Biology Vol 4, page 155 Furukawa K, Fritze CE, and Gerace L (1998). The major nuclear envelope targeting domain of LAP2 coincides with its lamin binding region but is distinct from its chromatin interaction domain. J Biol Chem 273, 4213-4219.

19080. Gruenbaum Y, Landesman Y, Drees B, Bare JW, Saumweber H, Paddy MR, Sedat JW, Smith DE, Benton BM, and Fisher PA (1988). Drosophila nuclear lamin precursor Dm0 is translated from either of two developmentally regulated mRNA species apparently encoded by a single gene. J Cell Biol 106, 585596.

Furukawa K and Hotta Y (1993). cDNA cloning of a germ cell specific lamin B3 from mouse spermatocytes and analysis of its function by ectopic expression in somatic cells. EMBO J 12, 97-106.

Haas M and Jost E (1993). Functional analysis of phosphorylation sites in human lamin A controlling lamin disassembly, nuclear transport and assembly. Eur J Cell Biol 62, 237-247.

Furukawa K, Inagaki H, and Hotta Y (1994). Identification and cloning of an mRNA coding for a germ cell-specific A-type lamin in mice. Exp Cell Res 212, 426-430. Furukawa K, and Kondo T (1998). Identification of the laminaassociated-polypeptide-2-binding domain of B-type lamin. Eur J Biochem 251, 729-733.

Harel A, Goldberg M, Ulitzur N, and Gruenbaum Y (1998). Structural organization and biological roles of the nuclear lamina. In Textbook of Gene Therapy and Molecular Biology: "From basic mechanism to clinical applications". Ed. T. Boulikas. 1, 529-542.

Furukawa K, Pante N, Aebi U, and Gerace L (1995). Cloning of a cDNA for lamina-associated polypeptide 2 (LAP2) and identification of regions that specify targeting to the nuclear envelope. EMBO J 14, 1626-1636.

Harel A, Zlotkin E, Nainude, ES, Feinstein N, Fishe, PA, and Gruenbaum Y (1989). Persistence of major nuclear envelope antigens in an envelope-like structure during mitosis in Drosophila melanogaster embryos. J Cell Sci 94, 463-470.

Galcheva-Gargova Z and Dessev GN (1987). Crosslinking of DNA to nuclear lamina proteins by UV irradiation in vivo. J Cell Biochem 34, 163-168.

Harris CA, Andryuk PJ, Cline S, Chan HK, Natarajan A, Siekierka JJ, and Goldstein G (1994). Three distinct human thymopoietins are derived from alternatively spliced mRNAs. Proc Natl Acad Sci USA 91, 6283-6287.

Gant TM, Harris CA, and Wilson KL (1999). Roles of LAP2 Proteins in Nuclear Assembly and DNA Replication: Truncated LAP2beta Proteins Alter Lamina Assembly, Envelope Formation, Nuclear Size, and DNA Replication Efficiency in Xenopus laevis Extracts. J Cell Biol 144, 10831096. Gant TM and Wilson KL (1997). Nuclear assembly. Ann Rev Cell Dev Biol 13, 669-695.

Harris CA, Andryuk PJ, Cline SW, Mathew S, Siekierka JJ, and Goldstein G (1995). Structure and mapping of the human thymopoietin (TMPO) gene and relationship of human TMPO beta to rat lamin-associated polypeptide 2. Genomics 28, 198-205.

Gerace L and Foisner R (1994). Integral membrane proteins and dynamic organization of the nuclear envelope. Trends Cell Biol 4, 127-131.

Heald R and McKeon F (1990). Mutations of phosphorylation sites in lamin A that prevent nuclear lamina disassembly in mitosis. Cell 61, 579-589.

Glass CA, Glass JR, Taniura H, Hasel KW, Blevitt JM, and Gerace L (1993). The alpha-helical rod domain of human lamins A and C contains a chromatin binding site. EMBO J 12, 4413-4424.

Hennekes H, Peter M, Weber K, and Nigg EA (1993). Phosphorylation on protein kinase C sites inhibits nuclear import of lamin B2. J Cell Biol 120, 1293-1304.

Glass JR and Gerace L (1990). Lamins A and C bind and assemble at the surface of mitotic chromosomes. J Cell Biol 111, 1047-1057.

Hocevar BA, Burns DJ, and Fields AP (1993). Identification of protein kinase C (PKC) phosphorylation sites on human lamin B. Potential role of PKC in nuclear lamina structural dynamics. J Biol Chem 268, 7545-7552.

Goldberg M, Harel A, Brandeis M, Rechsteiner T, Richmond TJ, Weiss AM, and Gruenbaum Y (1999). The tail domain of lamin Dm0 binds histones H2A and H2B. Proc Natl Acad Sci USA 96, 2852-2857.

Hocevar BA and Fields AP (1991). Selective translocation of beta II-protein kinase C to the nucleus of human promyelocytic (HL60) leukemia cells. J Biol Chem 266, 2833.

Goldberg M, Jenkins H, Allen T, Whitfield WG, and Hutchison CJ (1995). Xenopus lamin B3 has a direct role in the assembly of a replication competent nucleus: evidence from cell-free egg extracts. J Cell Sci 108, 3451-3461.

Hoger TH, Krohne G, and Kleinschmidt JA (1991). Interaction of Xenopus lamins A and LII with chromatin in vitro mediated by a sequence element in the carboxyterminal domain. Exp Cell Res.197, 280-289.

Goldberg M, Lu H, Stuurman N, Ashery Padan R, Weiss AM, Yu J, Bhattacharyya D, Fisher PA, Gruenbaum Y, and Wolfner MF (1998). Interactions among Drosophila nuclear envelope proteins lamin, otefin, and YA. Mol Cell Biol 18, 4315-4323.

Hozak P, Sasseville AM, Raymond Y, and Cook PR (1995). Lamin proteins form an internal nucleoskeleton as well as a peripheral lamina in human cells. J Cell Sci 108, 635-644. Hutchison CJ, Bridger JM, Cox LS, and Kill IR (1994). Weaving a pattern from disparate threads: lamin function in nuclear assembly and DNA replication. J Cell Sci 107, 3259-3269.

Goldman AE, Moir RD, Montag LM, Stewart M, and Goldman RD (1992). Pathway of incorporation of microinjected lamin A into the nuclear envelope. J Cell Biol 119, 725-735.

Ishijima Y, Toda T, Matsushita H, Yoshida M, and Kimura N (1996). Expression of thymopoietin beta/lamina-associated polypeptide 2 (TP beta/LAP2) and its family proteins as revealed by specific antibody induced against recombinant human thymopoietin. Biochem Biophys Res Commun 226, 431-438.

Gorlich D and Mattaj IW (1996). Nucleocytoplasmic transport. Science. 271, 1513-1518. Goss VL, Hocevar BA, Thompson LJ, Stratton CA, Burns DJ, and Fields AP (1994). Identification of nuclear beta II protein kinase C as a mitotic lamin kinase. J Biol Chem 269, 19074-

Jenkins H, Holman T, Lyon C, Lane B, Stick R, and Hutchison C

155

Goldberg et al: Functional Organization of the Nuclear Lamina (1993). Nuclei that lack a lamina accumulate karyophilic proteins and assemble a nuclear matrix. J Cell Sci 106, 275285

Luderus ME, de Graaf A, Mattia E, den BJ, Grande MA, de Jong L, and van Driel R (1992). Binding of matrix attachment regions to lamin B1. Cell 70, 949-959.

Kalderon D, Roberts BL, Richardson WD, and Smith AE (1984). A short amino acid sequence able to specify nuclear location. Cell 39, 499-509.

Luderus ME, den Blaauwen J, de Smit O, Compton DA, and van Driel R (1994). Binding of matrix attachment regions to lamin polymers involves single-stranded regions and the minor groove. Mol Cell Biol 14, 6297-6305.

Kaufmann SH (1989). Induction of endonucleolytic DNA cleavage in human acute myelogenous leukemia cells by etoposide, camptothecin, and other cytotoxic anticancer drugs: a cautionary note. Cancer Res 49, 5870-5878.

Lutz RJ, Trujillo MA, Denham KS, Wenger L, and Sinensky M (1992). Nucleoplasmic localization of prelamin A: implications for prenylation-dependent lamin A assembly into the nuclear. Proc Natl Acad Sci USA 89, 3000-3004.

Kawahara A, Enari M, Talanian RV, Wong WW, and Nagata S (1998). Fas-induced DNA fragmentation and proteolysis of nuclear proteins. Genes Cells 3, 297-306.

Macaulay C, Meier E, and Forbes DJ (1995). Differential mitotic phosphorylation of proteins of the nuclear pore complex. J Biol Chem 270, 254-262.

Kluck RM, Martin SJ, Hoffman BM, Zhou JS, Green DR, and Newmeyer DD (1997). Cytochrome c activation of CPP32like proteolysis plays a critical role in a Xenopus cell-free apoptosis system. EMBO J 16, 4639-4649.

Machiels BM, Zorenc AH, Endert JM, Kuijpers HJ, van Eys GJ, Ramaekers FC, and Broers JL (1996). An alternative splicing product of the lamin A/C gene lacks exon 10. J Biol Chem 271, 9249-9253.

Krohne G. Lamin assembly in vivo. Vol. 31. Subcellular Biochemistry ed. Herrmann and Harris. New York: Plenum Press, 1998.

Maison C, Pyrpasopoulos A, Theodoropoulos PA, and Georgatos SD (1997). The inner nuclear membrane protein LAP1 forms a native complex with B-type lamins and partitions with spindle-associated mitotic vesicles. EMBO J 16, 4839-4850.

Krohne G, Waizenegger I, and Hoger TH (1989). The conserved carboxy-terminal cysteine of nuclear lamins is essential for lamin association with the nuclear envelope. J Cell Biol 109, 2003-2011.

Mandal M, Maggirwa, SB, Sharma N, Kaufmann SH, Sun SC, and Kumar R (1996). Bcl-2 prevents CD95 (Fas/APO-1)induced degradation of lamin B and poly(ADP-ribose) polymerase and restores the NF-kappaB signaling pathway. J Biol Chem 271, 30354-30359.

Lazebnik YA, Cole S, Cooke CA, Nelson WG, and Earnshaw WC (1993). Nuclear events of apoptosis in vitro in cell-free mitotic extracts: a model system for analysis of the active phase of apoptosis. J Cell Biol 123, 7-22.

Manilal S, Nguyen TM, and Morris GE (1998). Colocalization of emerin and lamins in interphase nuclei and changes during mitosis. Biochem Biophys Res Commun 249, 643-647.

Lenz-Bohme B, Wismar J, Fuchs S, Reifegerste R, Buchner E, Betz H, and Schmitt B (1997). Insertional mutation of the Drosophila nuclear lamin dm(0) gene results in defective nuclear envelopes, clustering of nuclear pore complexes, and accumulation of annulate lamellae. J Cell Biol 137, 10011016.

Manilal S, Nguyen TM, Sewry CA, and Morris GE (1996). The Emery-Dreifuss muscular dystrophy protein, emerin, is a nuclear membrane protein. Hum Mol Genet 5, 801-808. Marshall ICB. and Wilson KL (1997). Nuclear envelope assembly after mitosis. Trends Cell Biol 7, 69-74.

Lin H, Song K, Hutcheson GA, Goutte CE, and Wolfner MF (1991). A maternally encoded nuclear envelope protein required for embryonic mitosis in Drosophila. Cold Spring Harb, Symp Quant Biol 56, 719-727.

Martin L, Crimaud, C, and Gerace L (1995). cDNA cloning and characterization of lamina-associated polypeptide 1C (LAP1C), an integral protein of the inner nuclear membrane. J Biol Chem 270, 8822-8828.

Lin HF and Wolfner MF (1991). The Drosophila maternal-effect gene fs(1)Ya encodes a cell cycle-dependent nuclear envelope component required for embryonic mitosis. Cell 64, 49-62.

McKeon FD, Kirschner MW, and Caput D (1986). Homologies in both primary and secondary structure between nuclear envelope and intermediate filament proteins. Nature 319, 463-468.

Liu J, Lopez JM, and Wolfner MF (1997). Developmental modulation of the nuclear envelope. Curr Top Dev Biol 35, 47-70.

Meier J, Campbell KH, Ford CC, Stick R, and Hutchison CJ (1991). The role of lamin LIII in nuclear assembly and DNA replication, in cell-free extracts of Xenopus eggs. J Cell Sci 98, 271-279.

Liu J, Song K, and Wolfner MF (1995). Mutational analyses of fs(1)Ya, an essential, developmentally regulated, nuclear envelope protein in Drosophila. Genetics 141, 1473-1481.

Mical TI and Monteiro MJ (1998). The role of sequences unique to nuclear intermediate filaments in the targeting and assembly of human lamin B: evidence for lack of interaction of lamin B with its putative receptor. J Cell Sci 111, 34713485.

Liu J and Wolfner MF (1998). Functional dissection of YA, an essential, developmentally regulated nuclear lamina protein in Drosophila melanogaster. Mol Cell Biol 18, 188-197. Lopez JM, Son, K, Hirshfeld AB, Lin H, and Wolfner MF (1994). The Drosophila fs(1)Ya protein, which is needed for the first mitotic division, is in the nuclear lamina and in the envelopes of cleavage nuclei, pronuclei, and nonmitotic nuclei. Dev Biol 163, 202-211.

Moir RD, Montag LM, and Goldman RD (1994). Dynamic properties of nuclear lamins: lamin B is associated with sites of DNA replication. J Cell Biol 125, 1201-1212. Moir RD, Spann TP, and Goldman RD (1995). The dynamic properties and possible functions of nuclear lamins. Int Rev Cytol 162b, 141-182.

Lopez JM and Wolfner MF (1997). The developmentally regulated Drosophila embryonic nuclear lamina protein young arrest (fs(1)ya) is capable of associating with chromatin. J Cell Sci 5, 643-651.

Nagano A, Koga R, Ogaw, M, Kurano Y, Kawada J, Okada R,

156

Gene Therapy and Molecular Biology Vol 4, page 157 Hayashi YK, Tsukahara T, and Arahata K (1996). Emerin deficiency at the nuclear membrane in patients with EmeryDreifuss muscular dystrophy. Nat Genet 12, 254-259.

activated protein kinases phosphorylate nuclear lamins and display sequence specificity overlapping that of mitotic protein kinase p34cdc2. Eur J Biochem 205, 287-294.

Neamati N, Fernandez A, Wright S, Kiefer J, and McConkey DJ (1995). Degradation of lamin B1 precedes oligonucleosomal DNA fragmentation in apoptotic thymocytes and isolated thymocyte nuclei. J. Immunol. 154, 3788-3795.

Pyrpasopoulou A, Meier J, Maison C, Simos G, and Georgatos SD (1996). The lamin B receptor (LBR) provides essential chromatin docking sites at the nuclear envelope. EMBO J 15, 7108-7119.

Newport, J. W., Wilson, K. L., and Dunphy, W. G. (1990). A lamin-independent pathway for nuclear envelope assembly. J Cell Biol 111, 2247-2259.

Qian Y, Barton R, and Worman HJ Nuclear lamin-binding proteins. Vol. 31. Subcellular Biochemistry ed. Herrmann and Harris. New York: Plenum Press, 1998.

Nigg EA (1993). Targets of cyclin-dependent protein kinases. Curr Opin Cell Biol 5, 187-193.

Rao L, Modha D, and White E (1997). The E1B 19K protein associates with lamins in vivo and its proper localization is required for inhibition of apoptosis. Oncogene 15, 15871597.

Nikolakaki E, Meier J, Simos G, Georgatos SD, and Giannakouros T (1997). Mitotic phosphorylation of the lamin B receptor by a serine/arginine kinase and p34(cdc2). J Biol Chem 272, 6208-6213.

Rao L, Perez D, and White E (1996). Lamin proteolysis facilitates nuclear events during apoptosis. J Cell Biol 135, 1441-1455.

Nikolakaki E, Simos G, Georgatos SD, and Giannakouros T (1996). A nuclear envelope-associated kinase phosphorylates arginine-serine motifs and modulates interactions between the lamin B receptor and other nuclear proteins. J Biol Chem 271, 8365-8372.

Riemer D, Dodemont H, and Weber K (1993). A nuclear lamin of the nematode Caenorhabditis elegans with unusual structural features; cDNA cloning and gene organization. Eur J Cell Biol 62, 214-223.

Oberhammer FA, Hochegger K, Froschl G, Tiefenbacher R, and Pavelka M (1994). Chromatin condensation during apoptosis is accompanied by degradation of lamin A+B, without enhanced activation of cdc2 kinase. J Cell Biol 126, 827-837.

Rzepecki R, Bogachev SS, Kokoza E, Stuurman N, and Fisher PA (1998). In vivo association of lamins with nucleic acids in Drosophila melanogaster. J Cell Sci 111, 121-129. Sasseville AM and Langelier Y (1998). In vitro interaction of the carboxy-terminal domain of lamin A with actin. FEBS Lett 425, 485-489.

Orth K, Chinnaiyan AM, Garg M, Froelich CJ, and Dixit VM (1996). The CED-3/ICE-like protease Mch2 is activated during apoptosis and cleaves the death substrate lamin A. J Biol Chem 271, 16443-16446.

Sasseville AM and Raymond Y (1995). Lamin A precursor is localized to intranuclear foci. J Cell Sci 108, 273-285.

Ottaviano Y and Gerace L (1985). Phosphorylation of the nuclear lamins during interphase and mitosis. J Biol Chem 260, 624632.

Schlesinger DH and Goldstein G (1975). The amino acid sequence of thymopoietin II. Cell 5, 361-365. Shimanuki M, Goebl M, Yanagida M, and Toda T (1992). Fission yeast sts1+ gene encodes a protein similar to the chicken lamin B receptor and is implicated in pleiotropic drug-sensitivity, divalent cation-sensitivity, and osmoregulation. Mol Biol Cell 3, 263-273.

Padan R, Nainudel ES, Goitein R, Fainsod A, and Gruenbaum Y (1990). Isolation and characterization of the Drosophila nuclear envelope otefin cDNA. J Biol Chem 265, 78087813. Papoutsopoulou S, Nikolakaki E, and Giannakouros T (1999). SRPK1 and LBR protein kinases show identical substrate specificities. Biochem Biophys Res Commun 255, 602-607.

Shimizu, T, Cao CX, Shao RG, and Pommier Y (1998). Lamin B phosphorylation by protein kinase calpha and proteolysis during apoptosis in human leukemia HL60 cells. J Biol Chem 273, 8669-8674.

Patschinsky T, Hunter T, Esch FS, Cooper JA, and Sefton BM (1982). Analysis of the sequence of amino acids surrounding sites of tyrosine phosphorylation. Proc Natl Acad Sci USA 79, 973-977.

Shimizu T and Pommier Y (1997). Camptothecin-induced apoptosis in p53-null human leukemia HL60 cells and their isolated nuclei: effects of the protease inhibitors Z-VAD-fmk and dichloroisocoumarin suggest an involvement of both caspases and serine proteases. Leukemia 11, 1238-1244.

Paulin Levasseur M, Blake DL, Julien M, and Rouleau L (1996). The MAN antigens are non-lamin constituents of the nuclear lamina in vertebrate cells. Chromosoma 104, 367-379.

Shoeman RL and Traub P (1990). The in vitro DNA-binding properties of purified nuclear lamin proteins and vimentin. J Biol Chem 265, 9055-9061.

Peter M, Heitlinger E, Haner M, Aebi U, and Nigg EA (1991). Disassembly of in vitro formed lamin head-to-tail polymers by CDC2 kinase. EMBO J 10, 1535-1544.

Silve S, Dupuy PH, Ferrara P, and Loison G (1998). Human lamin B receptor exhibits sterol C14-reductase activity in Saccharomyces cerevisiae. Biochim Biophys Acta 1392, 233-244.

Peter M, Kitten GT, Lehner CF, Vorburger K, Bailer SM, Maridor G, and Nigg EA (3 1989 ). Cloning and sequencing of cDNA clones encoding chicken lamins A and B1 and comparison of the primary structures of vertebrate A- and Btype lamins. J Mol Biol 208, 393-404.

Simos G and Georgatos SD (1992). The inner nuclear membrane protein p58 associates in vivo with a p58 kinase and the nuclear lamins. EMBO J 11, 4027-4036.

Peter M, Nakagawa J, Doree M, Labbe JC, and Nigg EA (1990). In vitro disassembly of the nuclear lamina and M phasespecific phosphorylation of lamins by cdc2 kinase. Cell 61, 591-602.

Simos, G. and Georgatos, S. D. (2-3 1994). The lamin B receptor-associated protein p34 shares sequence homology and antigenic determinants with the splicing factor 2associated protein p32. FEBS Lett 346, 225-228.

Peter M, Sanghera JS, Pelech SL, and Nigg EA (1992). Mitogen-

157

Goldberg et al: Functional Organization of the Nuclear Lamina Simos G, Maison C, and Georgato, SD (1996). Characterization of p18, a component of the lamin B receptor complex and a new integral membrane protein of the avian erythrocyte nuclear envelope. J Biol Chem 271, 12617-12625.

membrane vesicles is required for nuclear envelope formation in vitro. J Cell Biol 107, 57-68. Wilson KL and Wiese C (1996). Reconstituting the Nuclear Envelope and Endoplasmic Reticulum In Vitro. Seminars In Cell and Developmental Biology 7, 487-496.

Spann TP, Moir RD, Goldman AE, Stick R, and Goldman RD (1997). Disruption of nuclear lamin organization alters the distribution of replication factors and inhibits DNA synthesis. J Cell Biol 136, 1201-1212.

Worman HJ, Evans CD, and Blobel G (1990). The lamin B receptor of the nuclear envelope inner membrane: a polytopic protein with eight potential transmembrane domains. J Cell Biol 111, 1535-1542.

Stick R (1992). The gene structure of Xenopus nuclear lamin A: a model for the evolution of A-type from B-type lamins by exon shuffling. Chromosoma 101, 566-574.

Worman HJ, Yuan J, Blobel G, and Georgatos SD (1988). A lamin B receptor in the nuclear envelope. Proc Natl Acad Sci USA 85, 8531-8534.

Stick R (5 1994). The gene structure of B-type nuclear lamins of Xenopus laevis: implications for the evolution of the vertebrate lamin family. Chromosome Res 2, 376-382.

Yang L, Guan T, and Gerace L (1997). Lamin-binding fragment of LAP2 inhibits increase in nuclear volume during the cell cycle and progression into S phase. J Cell Biol 139, 10771087.

Stuurman N, Heins S, and Aebi U (1998). Nuclear lamins: their structure, assembly, and interactions. J Struct Biol 122, 4266.

Ye Q and Worman HJ (1994). Primary structure analysis and lamin B and DNA binding of human LBR, an integral protein of the nuclear envelope inner membrane. J Biol Chem 269, 11306-11311.

Takahashi A, Alnemri ES, Lazebnik YA, Fernandes Alnemri T, Litwack G, Moir RD, Goldman RD, Poirier GG, Kaufmann SH, and Earnshaw WC (1996a). Cleavage of lamin A by Mch2 alpha but not CPP32: multiple interleukin 1 betaconverting enzyme-related proteases with distinct substrate recognition properties are active in apoptosis. Proc Natl Acad Sci USA 93, 8395-8400.

Ye Q and Worman HJ (1996). Interaction between an integral protein of the nuclear envelope inner membrane and human chromodomain proteins homologous to Drosophila Hp1. J Biol Chem 271, 14653-14656.

Takahashi A, Musy PY, Martins LM, Poirier GG, Moyer RW, and Earnshaw WC (1996b). CrmA/SPI-2 inhibition of an endogenous ICE-related protease responsible for lamin A cleavage and apoptotic nuclear fragmentation. J Biol Chem 271, 32487-32490.

Ye QA, Callebaut I, Pezhman A, Courvalin JC, and Worman HJ (1997). Domain specific interactions of human hp1 type chromodomain proteins and inner nuclear membrane protein LBR. J Biol Chem 272, 14983-14989. Yuan J, Simos G, Blobel G, and Georgatos SD (1991). Binding of lamin A to polynucleosomes. J Biol Chem 266, 92119215.

Taniura H, Glass C, and Gerace L (1995). A chromatin binding site in the tail domain of nuclear lamins that interacts with core histones. J Cell Biol 131, 33-44.

Zevin-Sonkin D, Ilan E, Guthmann D, Riss J, Theodor L, Shoham J, Berger R, Gokkel E, Brok-Simoni F, Avraham KB, Copeland NG, Jenkins NA, Rechavi G, and Simon AJ (1992). Molecular cloning of the bovine thymopoietin gene and its expression in different calf tissues: evidence for a predominant expression in thymocytes. Immunol Lett 31, 301-309.

Theodor L, Shoham J, Berger R, Gokkel E, Trachtenbrot L, Simon AJ, Brok Simon F, Nir U, Ilan E, Zevin Sonkin D, Friedman E, and Rechavi G ( 1997). Ubiquitous expression of a cloned murine thymopoietin cDNA. Acta Haematol 97, 153-163. Ucker D S, Obermiller PS, Eckhart W, Apgar JR, Berger NA, and Meyers J (1992). Genome digestion is a dispensable consequence of physiological cell death mediated by cytotoxic T lymphocytes. Mol Cell Biol 12, 3060-3069.

Zhao K, Harel A, Stuurman N, Guedalia D, and Gruenbaum Y (1996). Binding of matrix attachment regions to nuclear lamin is mediated by the rod domain and depends on the lamin polymerization state. FEBS Lett 380, 161-164.

Ulitzur N, Harel A, Feinstein N, and Gruenbaum Y (1992). Lamin activity is essential for nuclear envelope assembly in a Drosophila embryo cell-free extract. J Cell Biol 119, 17-25.

Zhivotovsky B, Gahm A, and Orrenius S (1997). Two different proteases are involved in the proteolysis of lamin during apoptosis. Biochem Biophys Res Communi 233, 96-101.

Ulitzur N, Harel A, Goldberg M, Feinstein N, and Gruenbaum Y (1997). Nuclear Membrane Vesicle Targeting to Chromatin In a Drosophila Embryo Cell Free System. Mol Biol Cell 8, 1439-1448. Vorburger K, Lehner CF, Kitten GT, Eppenberger HM, and Nigg EA (1989). A second higher vertebrate B-type lamin. cDNA sequence determination and in vitro processing of chicken lamin B2. J Mol Biol 208, 405-415. Ward GE and Kirschner MW (1990). Identification of cell cycleregulated phosphorylation sites on nuclear lamin C. Cell 61, 561-577. Wedrychowski A, Bhorjee JS, and Briggs RC (1989). In vivo crosslinking of nuclear proteins to DNA by cisdiamminedichloroplatinum (II) in differentiating rat myoblasts. Exp Cell Res.183, 376-387. Wilson KL and Newport J (1988). A trypsin-sensitive receptor on

158

Gene Therapy and Molecular Biology Vol 4, page 159 Gene Ther Mol Biol Vol 4, 159-170. December 1999.

Models of cationic liposome mediated transfection Research Article

Aaron Ahearn1 and Robert Malone1,2 Molecular and Cellular Biology Program1 and Department of Pathology2, School of Medicine, University of Maryland, Baltimore _________________________________________________________________________________________________ Correspondence: Robert Malone, M.D., Molecular and Cellular Biology Program and Department of Pathology, School of Medicine, University of Maryland, 10 South Pine Street, 8-58 MSTF, Baltimore, MD 21201-1192. Phone: (410) 706-5602; Fax: (410) 706-8414; Email: rmalo001@umabnet.ab.umd.edu Key words: cationic lipid, liposome, gene delivery, DOTAP, transfection Received: 3 December 1999; accepted: 21 December 1999

Summary Synthetic gene delivery vectors are being developed for in vivo gene transfer applications, and these systems may circumvent the risks inherent in the use of recombinant virus vectors. The majority of synthetic delivery systems are based on the use of cationic amphiphiles to coat and condense polynucleotides, and to facilitate the uptake and release of the polynucleotide payload into somatic cells. Cationic amphiphile-based vector technology has benefited from over a decade of mechanistic and structure-function analyses, but many aspects of the pharmacology, toxicology, and cell biology of these systems remain unresolved. Important outstanding issues include the structure and characteristics of active and inactive particles, in vivo distribution and clearance of particles, and the cellular events involved in binding, cytoplasmic delivery, and nuclear uptake. These processes are interdependent, and therefore difficult to isolate and examine experimentally. The wide range of compounds, methodologies, and polynucleotides used to study these phenomena further complicate development of general principles. This article focuses on the molecular and cellular processes involved in cationic amphiphile-mediated transfection. Both primary data and current literature will be used to illuminate the complexity that impacts on the development and application of this class of synthetic gene delivery vectors. aspects of the system, and this interrelationship complicates analysis of each process component. However, just as viral agents have evolved solutions to each of these steps, man-made preparations must also optimize each of these process components to enable genebased therapies. The essential difference is that viral biology is optimized to maximize the reproductive fitness of the virus, whereas synthetic gene delivery systems must be engineered to maximize therapeutic benefit while minimizing toxicity. In general, the preparation of synthetic vector formulations begins with the deceptively simple step of mixing a cationic compound with a polyanionic genetic molecule. The resulting interactions neutralize the anionic charge associated with the polynucleotide phosphate backbone and thereby facilitate condensation of the extended molecular structure typical of most polynucleotides. Although the genetic molecule is most often a double stranded covalently closed DNA circle (plasmid), non-viral systems are not restricted to delivering a single type of polynucleotide. mRNA, anti-sense

I. Introduction Cationic lipid/DNA particles are not as active for transfection as their viral counterparts are for transduction. In a typical experiment, about 1 in 100 recombinant Adenovirus particles are successfully expressed, while only about 1 in 10,000 cationic lipid:DNA particles are effective. Transfection efficiency must be improved before cationic lipid:DNA complexes will become useful for creating effective genetic medicines. Many steps in the transfection process strongly influence overall efficiency and efficacy. Therefore analyzing, understanding and optimizing process components may identify new opportunities for developing or improving synthetic molecular medicines. From this reductionistic perspective, successful in vivo transfection may be considered as requiring: (1) formulation of the active particle; (2) distribution to the target cell and avoidance of clearance; 3) Binding to the cell surface and cellular entry; (4) polynucleotide release from the synthetic complex; (5) entry into the nucleus; and (6) transgene expression. Modification of any of these steps often influences other 159

Ahearn and Malone: Cationic liposome mediated transfection oligonucleotides, and other DNA structures can all be delivered by similar methods. This flexibility allows for a wide range of therapeutic strategies, and bypasses the constraints and risks associated with adapting highly evolved viral agents for therapeutic gene transfer applications. While polynucleotides share many structural and chemical features, differences in size, sugar (ribose and deoxyribose), and base pairing (single strand/intramolecular and antiparallel double stranded) significantly influence the structural and functional characteristics of the particles formed from polynucleotide/cationic amphiphile mixtures. Most non-viral gene therapy research has focused on the delivery of plasmids. This class of polynucleotide is attractive for gene transfer applications due to flexibility and manufacturing considerations. Plasmids may be engineered to incorporate very large DNA segments, though practical aspects of cellular uptake and recombination during amplification often limit transgene length to roughly 3 KB. This capacity allows transfection of large open reading frames or combinations of genes. Additionally, molecular cloning techniques involved in synthesizing recombinant plasmids are well understood and easy to reproduce. Finally, once the recombinant molecule has been prepared and isolated, proven bacterial amplification methodology may be employed to prepare milligram to gram quantities of transgene-coding plasmid using pharmaceutically compatible processes. Unfortunately, plasmids are also associated with undesirable bacterial-associated modifications and contaminants such as unmethylated CpG sequences and endotoxin. Thus, although bacterially produced plasmids

are the most frequently used polynucleotide, other polynucleotides may be preferred for some applications. The most developed pharmaceuticals for non-viral genetic transfer are cationic lipids. The chemical structure of these compounds varies greatly, but all include a positively charged hydrophilic head group linked to a lipophilic body. Although the exact nature of the cationic lipid/polynucleotide complex is controversial (see below), it is documented that these particles successfully transfect cells in vitro (Felgner et al., 1987) and in vivo. The molecular structures of some of the important prototype cationic transfection lipids are summarized in Figure 1. The transfection activity of these compounds have led to the synthesis and testing of hundreds of chemical analogs. In some cases, analogs have been prepared, tested, and published in an attempt to correlate chemical structure with function. In other instances, development and testing have been driven by commerce. In these cases, information from the developmental process leading to disclosed or marketed compounds is usually not available to other scientists, making it difficult to infer which chemical modifications correlate with changes in biological activity. Although this diversity has lead to progress in the field, it also complicates the process of drawing broad conclusions about mechanism or structure/function relationships. Despite the lack of definitive mechanistic studies, the proven transfection activity associated with various cationic lipid formulations is likely to continue to attract scientific and commercial interest for the foreseeable future.

Figure 1: The molecular structures of some of the prototypical cationic lipid compounds.

160

Gene Therapy and Molecular Biology Vol 4, page 161 involved in the transfection process. In such experiments, the active particles may be a very small percentage of the lipid particle population. Modifications that change the overall distribution of the entire population may alter important characteristics of the active particles, or may only alter an extraneous population. Therefore, the polydispersity of the population of lipid/polynucleotide complexes complicates all studies involving cationic lipidmediated transfection. In light of these issues, it also becomes clear that even comparative gene expression studies involving transfection of different polynucleotides require rigorously standardized formulation protocols. Although virtually all populations of cationic lipid/polynucleotide complexes are capable of transfection under some conditions, differences in the experimental conditions used for formulation lead to variation in both the population distribution and the resulting transfection activity. Evidence shows that the nature of the complex formed will vary based on: (1) the cationic amphiphile; (2) the helper lipids involved; (3) the ratio of lipid to DNA; (4) the time allowed for formation; (5) the temperature at which it is performed; (6) the concentration of lipid and plasmid during formulation; (7) the nature of the solvent; and (8) whether or not energy is imparted to the system through sonication or extrusion (Bennett et al., 1996; Ahearn, 1999). Analysis of any transfection study must consider the interactions between these various factors, as well as the complexity of the particle population mixture. An added level of complexity derives from the wide range of compounds, formulation conditions, and cellular models employed by different researchers. This fact makes it difficult to integrate the data derived from different laboratories into a cohesive model. We propose that a simple thermodynamic model may greatly facilitate analysis and understanding of the process of particle formation, and of how this process affects the resulting populationâ&#x20AC;&#x2122;s transfection activity. Such a model may also account for how subtle changes in laboratory formulation conditions may result in huge variations in the populationâ&#x20AC;&#x2122;s final transfection activity. In this model, separate cationic liposomes and anionic plasmids join to form a metastable intermediate, then mature to become fully neutralized actively transfecting particles. Two antagonizing forces acting on the cationic lipid molecules, electrostatic interactions and hydrophobic bonding, are proposed to drive this process. The first step in forming the transfection complex begins with initial association between MLV and polynucleotide. In the absence of polynucleotide, the cationic charges of lipids within MLV are neutralized by counterions. The nature of these counterions significantly influences the transfection activity associated with any one compound (Aberle et al., 1996). In this microenvironment, associative hydrophobic forces between lipidic side chains predominate, although the particle/fluid water interface remains highly positive (Figure 2, upper left panel). Upon mixing with polynucleotide, electrostatic interactions draw the positive surface of the MLV together with the negative

This review will employ current literature and primary data from our lab to illustrate models that describe how cationic lipid:DNA complexes facilitate polynucleotide transfection, and will explore some of the molecular and cellular barriers to the overall utility of these particles. Our intention is to integrate the work of different research laboratories that address questions of transfection mechanism from different frames of reference. In particular, we wish to emphasize the difficulties associated with interpreting and comparing structural and mechanistic studies that are based on polymorphic formulations of different polynucleotides and cationic compounds.

II. Complex formation, the first step in transfection Cationic lipid mediated plasmid transfection begins with the formulation of a lipid:DNA complex. The mechanism of interaction between DNA and the cationic liposomes will determine the nature of the resulting transfection particle population, and in turn the characteristics of this population will strongly influence the transfection activity even before the solution is applied to the cells. Understanding these dynamic interactions is crucial to understanding the entire transfection process. Standard laboratory practice involves simple mixing of a cationic lipid emulsion (80-400 nm multilamellar vesicles (MLV)) together with a bacterial plasmid that encodes the transgene of interest. Typically, these MLV are prepared by mixing the lipid components in an organic solvent, evaporating the solvent to yield a lipid film, and hydrating the film in a buffer. In many cases, the resulting emulsion is either sonicated or extruded to reduce the size and polydispersity of the MLV, although subsequent particle aggregation and fusion may reduce the long-term effects of such treatments. Such preparations are often purchased directly from commercial vendors, and the bench researcher may be unaware of the formulation process employed and structure of the particles that are mixed with the plasmid. Incubation of cationic lipid MLVs and DNA will lead to a heterogeneous population of particles (Zabner et al., 1995). It is likely that only subsets of these particles are capable of transfection. The characteristics of transfection active particles have not been conclusively identified, and different subpopulations may interact to influence the usual outcome parameter, transgene expression. These ambiguities complicate the design and interpretation of experiments that focus on transfection mechanism and structure/function relationships. One research strategy employed to address such questions involves modifying formulation parameters, characterizing physical alterations in the resulting particle population, and then correlating changes in these physical characteristics to increases in transfection activity. In other cases, reagents may be employed which alter cellular processes posited to be

161

Ahearn and Malone: Cationic liposome mediated transfection phosphate backbone of the plasmid. Cationic headgroups are then attracted to distribute along the backbone, but hydrophobic inter-lipid interactions slows this process of redistribution and prevents immediate coating and neutralization of the entire plasmid. Thus, the liposome binds to one region of the plasmid supercoil, forming a meta-stable complex with a positive pole –the liposomeand a negative pole – the uncoated plasmid. We will refer to this association as the macro-dipole intermediate. The "map pin" structures described by Sternberg (Sternberg et al., 1998) (see below) may be an example of such an intermediate. It is hypothesized that macro-dipole intermediates may be resolved by one of three pathways; 1) Binding additional plasmids to the cationic lipid portion will yield a larger and more complex anionic particle, 2) Binding additional cationic MLV to the partially neutralized plasmid will facilitate liposome aggregation, and 3) Redistribution of lipids along the polynucleotide axis will

form a cationic lipid-coated particle. Once the surface of a lipid/polynucleotide particle becomes more uniformly cationic, it will repel other cationic particles (such as unbound MLV). We propose that the first two processes will lead to aggregation of particles and the formation of large charged complexes ( Figure ), while the third process may yield smaller, less strongly cationic particles. Since the products of the third process are predicted to be smaller and less charged, they may be more likely to facilitate transfection. By this line of reasoning, either small, fully coated plasmid complexes, and/or small transition state intermediates (map-pins) comprise the majority of the transfection active particles. Under most conditions, such particles are typically a minor subpopulation of the overall mixture. Thus, the model proposes that the final transfection activity of the population will depend on the balance between aggregation (yielding large, inactive complexes) and the formation of smaller (transfectionally active) particles.

Figure 2: A hypothetical depiction of the thermodynamic energy barriers that must be overcome in order for mature particles to form. The inlays are actual measurements of the populations’ zeta potentials measured at different angles (represented by different color tracings: Red: 8.9o; blue: 17.6o; green: 26.3o; black: 35.2o) using the Coulter Delsa 440. The formulations were stored at 4°C for 2 minutes, 1hour, and 3 hours after the plasmid was added. The inlays are placed on the curve to approximate the thermodynamic energy state of the majority of the particles in these populations.

162

Gene Therapy and Molecular Biology Vol 4, page 163 plasmid, macro-dipole intermediates, and mature particles within the solvent. It has been clearly documented that precipitated DNA-lipid complexes are not active for transfection in vitro or in vivo (Sternberg, 1998; Ahearn, 1999). Atomic force microscopy has also shown that vesicles larger than 1.4 Âľm are not good for transfection (Kawaura et al., 1998). Thus, in order to optimize the transfection activity of cationic lipid/DNA formulations, one must consider aggregation as well as polynucleotide coating and condensation. This model, based on ionically driven association, subsequent coating, and non-productive aggregation, provides an explanation of why parameters such as time, temperature, and the concentration of components effect transfection activity of the resulting particle population. These parameters determine the amount of energy available to overcome the activation energy required for mature particle formation, and also determine the frequency of aggregation. For example, as temperature increases, more nascent particles are predicted to be converted to coated particles over any given time interval. These considerations may also explain the enhanced activity associated with different functional groups such as the lipopolyamines. Cooperative binding between polyamine headgroups and multiple backbone phosphates may enhance the free energy changes associated with a shift from MLV hydrophobic interactions to coated polynucleotide. Increased temperature will also impart increased kinetic energy, which will facilitate initial interaction between free plasmid and MLV. However, predicting the exact results of changes in these parameters is difficult because increased time, temperature, and concentration will increase aggregation, shifting the resulting particle distribution toward large, inactive complexes. Concentration cuts both ways, because lower concentrations lead to less aggregation, but in vivo protocols usually require high concentrations of complexes. Only empiric multi-parameter analysis will enable each of these variables to be simultaneously optimized. Based on these considerations, we predict that formulation parameters that facilitate complete neutralization of polynucleotide charge and smaller particle size will provide the highest level of active particle uptake and thus the highest transfection activity. This is distinguished from total particle uptake, which includes non-productive binding and/or endocytosis of large, inactive aggregates. In order to test this hypothesis, we attempted to correlate measurable parameters of the particle population, namely size and surface charge, with the transfection activity of the population. An expression plasmid encoding the P. pyralis Luciferase reporter gene was mixed with DOTAP liposomes at a 3:1 positive to negative charge ratio. These formulations were then stored at 4Â°C for various periods of time resulting in different particle populations that could be physically analyzed and used for transfection.

In order to promote the process of lipid redistribution and charge neutralization, energetic barriers must be overcome to reach an equilibrium state of lipid-coated polynucleotide. Examples of such barriers include the free energy change associated with disruption of hydrophobic lipid-lipid interactions within MLV, and the disruption of lipid headgroup-counterion interactions. The sum of the free energy changes associated with disruption of the MLV structure may be the dominant component of the transition state activation energy. Forces favoring MLV disruption and lipid redistribution include the energy of association between the cationic headgroup and the polynucleotide phosphates. In an ideal system where aggregation does not occur, the balance of these different forces will determine the final equilibrium between macro-dipole intermediates and lipid-coated polynucleotides. If there is sufficient thermal energy available to each nascent complex (macro-dipole) to overcome the activation energy, the lipid molecules will flow from the liposome structure to coat the plasmid. We suggest that the completely coated polynucleotide particle is the thermodynamically favored product, and that the coated particle formation rate is limited by the amount of energy available in the system. Figure 2 shows a hypothetical plot of the free energy of the system and a simplified representation of proposed intermediates. The energy boundary between the intermediate state and the final product is crucial, because it will determine how long it will take for the active particle to form at any one temperature. This energy boundary will be partly determined by the strength of the hydrophobic interactions of the chosen lipids. Manipulation of cohesive forces within the MLV membranes is therefore predicted to influence the frequency of active particle formation. For example, addition of the helper lipid dioleoylphosphatidylethanolamine (DOPE) has been shown to destabilize the cationic liposomes and to increase transfection activity (Farhood et al., 1995). Likewise, selection of a cationic lipid that has better surface hydration leads to increased transfection (Bennett, 1996). It is logical that these changes would lower the energy barrier required for the lipids to dissociate from one another and move to a new position on the plasmid, thus increasing the efficiency of the formation of the final product. Although the above description model purports to account for interactions between a single MLV and plasmid in excess solvent, it does not account for the interactions that occur between these complexes. As discussed above, two aggregation pathways complicate the outcome of the formulation process. We propose that the charged ends of the macro-dipole intermediates tend to facilitate aggregation by binding to other immature particles, unbound MLVs and unbound plasmids (Figure 3). In this way, immature particles may bind to each other forming larger and larger complexes that eventually fall out of solution. This aggregation process is influenced by the concentration and kinetic energy associated with MLV, 163

Ahearn and Malone: Cationic liposome mediated transfection

Figure 3: This Figure is a schematic representation of the process of aggregation. Particles that have not undergone maturation easily stimulate aggregation because of their dipole nature. Aggregated particles can become quite large (> 1500 nm) and are probably inactive for transfection.

Figure 4: This Figure shows the SDP analysis of the population size for particles incubated for 2 min (Black), 1 hour (Red), and 4.5 hours at 4째C. Note that at 2 minutes two populations of particles exist: one at about 250 nm and one at roughly 1000 nm. It is believed that the smaller population represents free liposomes and that the larger population represents small complexes- either initial particles or mature ones. At 1 hour almost all the particles belong to the second category. However, after 4.5 hours the particles have begun to aggregate and therefore appear in larger compartments. Note that SDP analysis calculates all particles larger than 2000 nm as 2000 nm.

164

Gene Therapy and Molecular Biology Vol 4, page 165 populations that have a measured mean size of around 1000 nm, but not larger, and have a positive surface potential. This data is consistent with that obtained in other laboratories. For example, it has been demonstrated that a positive zeta potential directly correlates with the transfection activity of the particles (Takeuchi et al., 1996). The thermodynamic model described above is consistent with the findings of other formulation studies. For example, Yang and Huang have reported studies focused on avoiding the decrease in transfection activity that occurs when complexes enter human serum (Yang and Huang, 1998). This study demonstrated that allowing lipid:DNA complex to incubate without serum for sufficient time nullifies this drop in activity, even if the particles are subsequently placed in serum. To relate this finding to our model, we suggest that the charged components of the serum may bind to the metastable particles, and either facilitate aggregation or interfere with polynucleotide coating. If sufficient time is allowed for mature particles to form before the serum is added, there will be no decrease in the distribution of active particles and thus no decrease in transfection activity. Higher temperatures, higher concentrations, and higher ratios of lipid to DNA all increased the speed at which serum resistance was obtained (Yang and Huang, 1998). The finding that this serum- resistance was obtained faster under conditions that promote thermodynamic conversion further supports this hypothesis. Sternberg et al have employed electron microscopy to identify structures that may represent different stages of maturation. When image analysis is performed less than 20 minutes after the initial mixing of lipid and DNA, structures described as “map pins” become apparent. These “map pins” consist of a large lipid head and DNA tail. Larger aggregations of lipid and DNA are also seen in these preparations.

Particle size analysis was performed after incubating the mixture for 2 minutes, 1 hour, and 4 hours using a Coulter N4 plus submicron particle sizer, which employs multi angle laser refraction with photon correlation spectroscopy to estimate the distribution of particle sizes within a suspension. The results of this analysis are summarized in Figure 4. At 2 minutes, two separate populations are seen, one with a mean of about 250 nm and one with a mean of about 1000 nm. 250 nm is the approximate size of an average free liposome (MLV), representing particles that have not yet encountered plasmid, and the population around 1000 nm represents complexes of plasmid and liposomes. After an hour of incubation, the majority of the resulting particles cluster around the 1000 nm population mean. However, after 4.5 hours more large particles are seen and fewer were measured at 1000 nm, indicating that aggregation has progressed and shifted the population toward the very large particles that are predicted to be inactive for transfection. The zeta potential is a measurement of charge on the structured water surrounding the particles, and may be measured using the Coulter DELSA 440SX Doppler electrophoretic light scattering analyzer. This machine measures zeta potential by propelling the particles in an oscillating electric field and using the Doppler shift of laser-generated photons to measure the speed of the particle movement. This speed is directly proportional to the surface charge of the particle and is used to calculate the population estimates seen in the inlays of Figure 2. As expected, the cationic DOTAP liposomes alone have a strong positive surface charge. Two minutes after the addition of the plasmid, the particles become quite negative. We interpret this data as indicating that very little of the phosphate backbone has been neutralized at this time. After one to three hours of incubation, the surface charge becomes progressively more positive, and this appears to indicate that more and more of the negative charges are neutralized as the lipid coats the DNA. The inlays of Figure 2 are placed to correspond to relative energy states that we predict to exist using this model. To correlate temporal alterations in size and surface charge with transfection activity, particle populations were incubated for different times and then used to transfect cultured fibroblasts. Particle populations formed by mixing DOTAP and DNA for 2 minutes, 2 hours, 8hours, 24 hours, and 30 hours were used to transfect NIH 3T3 cells plated into 24 well plates one day prior to all of the transfection treatments. The formulation conditions were identical to those used to generate the data summarized in Figures 2 and 4. As shown in Figure 5, activity dramatically improved when incubation was extended from 2 minutes to 2 hours, but declined at incubations of 8 hours and continued to decline at incubation periods up to 30 hours. Transfections of DNA without added lipid were used as a negative control, and resulted in less than 1000 total luciferase counts (data not shown). As predicted, the peak transfection activity of the particles correlate with

Figure 5: This shows the relative light activity of several populations of particles with the same DNA to lipid ratio, but incubated for different periods of time at 4°C. Transfection activity is highest at 2 hours and declines at 8 hours and beyond. This activity corresponds with the populations that have positive zeta potentials but have not yet formed large aggregates.

165

Ahearn and Malone: Cationic liposome mediated transfection Formulations with a greater proportion of â&#x20AC;&#x153;map pinsâ&#x20AC;? and fewer large aggregates show higher transfection activity (Sternberg, 1998). It is possible that these structures represent metastable particles that have not yet aggregated and still retain the potential to mature to form active particles. Alternatively, the map pin intermediate may be the active particle. The thermodynamic model and the data summarized above predicts that further incubation of one or two hours may shift the proportion of particles to yield an increased fraction of small, fully coated particles. If a progression were found from map pin structures toward evenly coated particles, then fully coated plasmids would be more likely to be the functional particle. On the other hand, an increase in map-pin structures during incubation would support the importance of what may be the macro-dipole intermediate. Based on the assertion that the conditions of formulation are as important as the lipid being utilized, the current search for new lipids to improve activity may be less efficient than refining the thermodynamic conditions used with existing lipids. Focusing on this aspect would allow the choice of lipid to be based on issues such as toxicity. We suggest that further development of these models may provide a more complete understanding of the structure and formation of the active particle components of the overall mixture. Consideration of such interactions during formulation may also aid rational design of experiments aimed at improving the transfection activity of synthetic cationic lipid-based vectors.

these systems. Smith et al used ultracentrifugation and a sucrose gradient to separate the free liposomes from the lipid:DNA complexes that they had prepared (Smith et al., 1998). They predicted that purified lipid:DNA complexes would have better activity than the heterogeneous mixture of both liposomes and complexes. When purified particles were administered by intravenous injection, the expression observed in lung was actually lower than the mixture of both free liposomes and lipid:DNA complexes. However, when directly administered to the lung by tracheal instillation, the purified complexes yielded higher expression levels. This difference in expression levels may be mediated by the reticuloendothelial system (RES). The free liposomes may saturate RES clearance and prevent sequestration of some of the lipid:DNA complexes. This observation predicts that RES blockade wiith free liposomes, followed by administration of lipid:DNA particles, will provide higher levels of transfection at the intended site than is observed after administering the heterogeneous mixture. In general, more research that focuses on maximizing delivery to target tissue must be performed. Avoiding both macrophage clearance and transfection of non-target tissue will increase the effectiveness of these therapies and reduce the side effects associated with treatment.

IV. Cellular binding and uptake Cationic lipid-plasmid complexes must cross cell membranes before transgene expression may occur, but the mechanism involved is not well understood. Cationic lipid:DNA particles clearly associate with the cell surface, and over time are readily detected within endosomal compartments. Based on this observation, as well as studies involving agents that alter endosomal acidification or transport, it has been inferred that endosomal uptake is required for transfection. However, the large endosomal aggregates typically observed in such studies may not be transfectionally active. In general, the highly polymorphic nature of the transfection particle population makes it very difficult to design definitive experiments to address these issues, and as a result most studies establish correlation rather than prove causation. Since formulations of negatively charged plasmids and anionic liposomes are relatively inactive for transfection, it has been proposed that the positive charge associated with the cationic lipids is essential for cellular binding. One explanation for this difference is that positive surface charge may encourage interaction with anionic cell surfaces. If tissue culture cells are pre-treated with Pronase, an enzyme that removes negatively charged glycoproteins form the cell surface, then the level of transfection activity decreases by fifty percent (Hui et al., 1996). However, removing cytoplasmic membrane glycoproteins disrupts cellular signaling and decreases the rate of cell division, factors that play a key role in transfection (Zabner, 1995). Therefore, these studies do not definitively resolve this issue.

III. Distribution to target cells and avoiding clearance Synthetic vectors must diffuse or be transported to target cells before transfection can occur. In vitro, this process is relatively simple, as particles simply diffuse through the medium to cellular membranes. The only potential obstacles are serum-associated proteins and other molecules that bind lipidic particles, and either do not allow them to reach the cellular membrane or reduce their activity. For this reason, transfections are typically performed in serum free media. On the other hand, the process of distribution to the target cell is much more complicated in vivo. For example, complex modification by serum factors and sequestration either by the reticuloendothelial system or by lodging in the microvasculature can reduce transfection of the intended target cells. The importance of particle maturation in preventing this serum inactivation has already been discussed (see above). Any design of the lipid:DNA complexes for systemic administration must overcome these effects in order to be effective. Systemically administered particles are often sequestered by reticuloendothelial cells including macrophages. After injection, high levels of DNA are observed in the phagocytic cells of the liver and spleen (Liu et al., 1995). This phenomenon can lead to unexpected results that again illustrate the complexity of 166

Gene Therapy and Molecular Biology Vol 4, page 167 Fasbender et al studied a wide range of lipid and helper lipid combinations, and demonstrated that different lipid combinations strongly influence the fraction of DNA which becomes associated with the cells (Fasbender et al., 1997). However, the transfection activity of these formulations did not correlate with the amount of DNA bound. Zabner also showed that only 15% of the cellular associated DNA is taken up by cells (Zabner, 1995) in one experimental system. These findings suggest that cellular binding is probably not the main factor determining the rate of transfection. If cellular binding is not the limiting factor in transfection, it becomes increasingly important to understand how the cationic liposome coat facilitates polynucleotide transport across the cytoplasmic membrane. Two theories have been developed to describe this process. The first theory, originally proposed by Felgner (Felgner, 1987), is based on the idea that cationic lipids directly fuse with the cytoplasmic membrane, permitting direct DNA transfer from the extracellular space to the cytoplasmic compartment. In the original paper describing cationic liposome-mediated transfection, it was asserted that this is the predominant mechanism for DNA entry into the cell. The second major theory proposes that transfection is mediated by endocytosis, and that DNA enters the cytoplasm only after undergoing endocytosis (Gao and Huang, 1995). It is important to note that these theories are not mutually exclusive. In general, the second theory (endosome uptake) has become the most widely accepted, and many experiments have focused on increasing the number of particles that are endocytosed. After interaction with the cellular membrane, lipid:DNA particles are clearly taken up into endosomal vesicles. This process has been demonstrated by EM and florescence studies (Huebner et al., 1999). As many viruses enter the cell via endosomes and then escape into the cytoplasm, it is reasonable to hypothesize that cationic liposome/DNA complexes also follow this pathway. Although the overwhelming majority of plasmid DNA inside the cells is in vesicular compartments, a small amount can be seen in the cytoplasmic fraction (Friend et al., 1996). This cytoplasmic fraction may have escaped from the endosomes, or entered directly through the cellular membrane. Though the majority of DNA enters endosomes, this step is not proven to be the pathway of transfection. Wrobel and Collins have attempted to show that DNA entry involves fusion of cationic liposomes with the endocytic compartments (Wrobel and Collins, 1995). The florescence pattern of N-NBD-PE, a helper lipid, changes upon mixing with another lipid present in either free liposomes or cellular membranes. The amount of lipid mixing between cationic lipid complexes and cellular membranes dictates the magnitude of the change in florescence. The study showed high amounts of lipid mixing under conditions favoring endocytosis. When the

cells were subjected to low temperature or the poison Monensin, endocytosis stops and lipid mixing is reduced. Presuming lipid mixing is crucial for cellular entry, Xu and Szoka focused on the differences between the intracellular and extracellular faces of the cellular membrane(Xu and Szoka, 1996). Employing a purely liposomal model, it was demonstrated that the interaction of cationic lipid:DNA complexes and anionic liposomes that mimic the lipid composition of the cytoplasmic face of the cellular membrane will uncoat plasmid DNA. DNA is not released when liposomes mimicking the extracellular face are used. This finding was interpreted to indicate that cationic lipid complexes disrupt the normal structure of endosomal membranes after endocytosis. It was concluded that disruption allows lipid from the cytoplasmic face to flip over to the extracellular face and interact with the endocytosed cationic lipid. In this model, the two lipids neutralize each other, stimulating release of the plasmid leading to an opening in the cellular membrane that allows for plasmid release into the cytoplasm. Other observations are not consistent with this theory. The lipid mixing observed by Collins et al and Xu and Szoka requires the helper lipid DOPE, but many successfully transfecting cationic lipid formulations function without helper lipid. Also, the lipid-mixing studies performed by Stegmann showed absolutely no correlation between lipid mixing and transfection efficiency (Stegmann and Legendre, 1997). These studies imply either that mixing is not the mechanism for cellular entry or that another process determines the final transfection efficiency. To summarize, there is no evidence that conclusively demonstrates that the small amount of DNA in the cytoplasm originates from endosomes. Endosomal and cytoplasmic membranes may be equally permeable, rendering endocytosis unnecessary. As the hostile endosomal environment may damage the plasmid DNA, development of non-endocytosed lipids would be beneficial. Also, if cationic particles open endosomal compartments, then the cytoplasm of the transfected cells is exposed to this harsh environment. Avoiding of endocytosis may reduce the toxicity seen with cationic lipids and increase effective transfection.

V. Dissociation of lipid and DNA Somewhere between particle uptake and transgene expression, the cationic lipid must dissociate from the plasmid DNA. In theory, this disassociation may occur after particle transport into the nucleus, perhaps even during transcription. To examine this possibility, Zabner et al injected oocyte nuclei with lipid:DNA complexes, but found very little expression of the transgene (Zabner, 1995). This study suggests that dissociation generally occurs before the plasmid passes into the nucleus. As discussed above, Xu and Szoka have shown that DNA becomes exposed when cationic lipid:DNA particles

167

Ahearn and Malone: Cationic liposome mediated transfection interact with anionic liposomes resembling the cytoplasmic face of the cellular membranes. These liposomes released the DNA at a charge ratio of 1:1. On the other hand, a 100-fold excess of charged molecules such as ATP, DNA, RNA, spermidine, and histone do not stimulate release (Capecchi, 1980). It is possible that cellular membranes mediate plasmid release. However, it remains unclear if this event is simultaneous with cellular entry. It is also possible that this process mediates release of the DNA without being the mechanism for cellular entry. After all, particles that have traversed the cellular membrane are exposed to the anionic surfaces of both the cytoplasmic membrane and the membranes of cellular organelles. Finally, if the map-pin structures observed by Sternberg are transfectionally active, it may be that the majority of the polynucleotide is not coated prior to uptake, and the process of migration across cellular membranes requires minimal disassociation of lipid from plasmid.

This process requires transcription, mRNA export, translation, and proper protein folding. Fortunately, this process is handled by the normal cellular machinery and is not usually a barrier to transfection. The only key issue is which promoter should be used. Currently, almost all constructs rely on the CMV promoter, because it is known to give the very high levels of expression. When this therapy is actually applied to a therapeutic situation, this promoter may not be the best choice. Much research has been done to discover tissue specific promoters that allow high levels of expression in a target tissue (Peel et al., 1997). Selection of a naturally occurring promoter may improve long-term gene expression since it is less likely to be inactivated by the cell..

VIII. Conclusion Although our understanding of cationic lipid-based transfection continues to improve, there are many parts of the puzzle that must still be pieced together. Resolving these issues will allow us to improve the effectiveness of non-viral gene delivery. When first reported in 1987, cationic lipid-mediated transfection was considered a laboratory curiosity of marginal significance. Since that time, this technology has become the most widely used method for experimentally transfecting cultured cells. The widespread acceptance of the method derives from it's efficiency and apparent simplicity. However, although apparently simple, the process involves a complex set of interactions and particle populations which confound thorough analysis and elude complete understanding. If the efficiency of cationic-lipid:DNA complexes can be improved, then the resulting improvements in transgene expression may create new opportunities to develop genetic medicines. However, we suggest that further development of the system to yield effective, well defined and reproducible pharmaceuticals will require a much more complete understanding of the processes involved in formulating and applying this promising class of synthetic gene delivery vectors.

VI. Nuclear entry An essential step for gene expression is the entry of plasmid DNA into the nucleus (Capecchi, 1980). Surprisingly, nuclear entry may be the most inefficient process involved in lipid-mediated transfection. To examine the role of nuclear transport in the transfection process, Zabner et al designed an expression plasmid with a T7 promoter rather than the usual eurkaryotic transcriptional elements. This vector provides cytoplasmic transgene transcription when cells are coinfected with vaccinia virus expressing the T7 RNA polymerase (Zabner, 1995). Using this system, it was demonstrated that the majority of a sample of cultured cells could be productively transfected using 0.01 microgram of transfected plasmid. In contrast, 100 fold more of the CMV-promoter driven plasmid yielded transgene expression in only 10% of the cells under similar conditions. This data suggests that movement from the cytoplasm to the nucleus is probably the biggest limiting factor in transfection. One approach to overcoming this barrier is using the natural cellular trafficking to carry DNA into the nucleus. Zanta et al. used a single nuclear localizing signal to increase the entry into the nucleus (Zanta et al., 1999). The reporter gene was cleaved out of a plasmid and capped. A targeting peptide sequence that appears in nuclear proteins was then ligated to the DNA, increasing transfection between ten and 1,000 fold (depending on the cell type). Insertion of a signal mutation into the peptide sequence, eliminates this effect. Thus, it seems that nuclear import machinery recognizes the peptide sequence and is responsible for the increase in transfection.

Material and Methods A. Liposome Preparations. Bulk DOTAP, suspended in chloroform, was purchased from Avanti Polar Lipids. In glass vials, one mg of DOTAP was mixed with 500 Âľl of chloroform. The vials were then desiccated overnight, resulting in a lipid thin film, and stored at â&#x20AC;&#x201C;20Â°C until needed. Just before the formulations, the lipid was resuspended in one ml of purified water for injection. The vial was then sonicated in a bath sonicator for two minutes and placed on ice. B. Plasmid Preparations. For analysis of gene transfer and reporter gene expression, a plasmid encoding an enhanced Photinus pyralis luciferase (pND2Lux) was employed. The pUC19 replicon-based plasmid was constructed in a fashion similar to that described previously (Chapman et al., 1991), To prepare milligram quantities of plasmid DNA for transfection, pND2Lux was transformed into E. coli DH-5! (18258-012; Life

VII. Expression Once plasmid DNA successfully enters the nucleus, the final step in transfection is expression of the gene. 168

Gene Therapy and Molecular Biology Vol 4, page 169 manuscript. This work was supported by funding from the Molecular and Cellular Biology Program and the MDPh.D. Program at the University of Maryland, Baltimore, and the National Institutes of Health (USPHS training grant HL07612-14 and NIH-RO1RR12307 for AA, NIHKO2A101370 and NIH-RO1RR12307 for RM).

Technologies, Gaithersburg, MD), isolated by growth on ampicillin plates, amplified, and purified using either the glass powder/sodium iodide method (MonsterPrep; Merlin Core Services, Bio 101, Vista, CA) or base lysis with double CsCl banding. C. Cationic Lipid Formulations. In general, cationic lipid:DNA formulations were prepared by mixing 1 µg of plasmid DNA with enough DOTAP to bring the final charge ratio to 3:1. This charge ratio was previously determined to be optimal for transfection (unpublished data). These formulations were performed in a total volume of 1 ml in serum free DMEM media and incubated for varying lengths of time at 4°C.

References Aberle A.M., Bennett, M.J., Malone, R.W. and Nantz, M.H. (1996). The counterion influence of cationic lipid-mediated transfection of plasmid DNA. Biochim Biophys Acta1299, 281-283. Ahearn A.J. (1999). Unpublished results. Bennett M.J., Aberle, A.M., Balasubramaniam, R.P., Malone, J.G., Nantz, M.H. and Malone, R.W. (1996). Considerations for the design of improved cationic amphilphile-based transfection reagents. J Liposome Res 6, 545-565. Capecchi M.R. (1980). High efficiency transformation by direct microinjection of DNA into cultured mammalian cells. Cell 2, 479-88. Chapman B.S., Thayer, R.M., Vincent, K.A. and Haigwood, N.L. (1991). Effect of intron A from human cytomegalovirus (Towne) immediate-early gene on heterologous expression in mammalian cells. Nucleic Acids Res 19, 3979-86. Farhood H., Serbina, N. and Huang, L. (1995). The role of dioleoyl phosphatidylethanolamine in cationic liposome mediated gene transfer. Biochim Biophys Acta 1235, 28995. Fasbender A., Marshall, J., Moninger, T.O., Grunst, T., Cheng, S. and Welsh, M.J. (1997). Effect of co-lipids in enhancing cationic lipid-mediated gene transfer in vitro and in vivo. Gene Ther 4, 716-25. Felgner P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W., Wenz, M., Northrop, J.P., Ringold, G.M. and Danielsen. M (1987). Lipofection: a highly efficient, lipid-mediated DNAtransfection procedure. Proc Natl Acad Sci USA 84, 7413-7. Friend D.S., Papahadjopoulos, D. and Debs, R.J. (1996). Endocytosis and intracellular processing accompanying transfection mediated by cationic liposomes. Biochim Biophys Acta 1278, 41-50. Gao X. and Huang, L. (1995). Cationic liposome-mediated gene transfer. Gene Ther 2, 710-22. Huebner S., Battersby, B.J., Grimm, R. and Cevc, G. (1999). Lipid-DNA complex formation: reorganization and rupture of lipid vesicles in the presence of DNA as observed by cryoelectron microscopy. Biophys J 76, 3158-66. Hui S.W., Langner, M., Zhao, Y.L., Ross, P., Hurley, E. and Chan, K. (1996). The role of helper lipids in cationic liposome-mediated gene transfer. Biophys J 71, 590-9. Kawaura C., Noguchi, A., Furuno, T. and Nakanishi, M. (1998). Atomic force microscopy for studying gene transfection mediated by cationic liposomes with a cationic cholesterol derivative. FEBS Lett 421, 69-72. Liu Y., Liggitt, D., Zhong, W., Tu, G., Gaensler, K. and Debs, R. (1995). Cationic liposome-mediated intravenous gene delivery. J Biol Chem 270, 24864-70. Peel A.L., Zolotukhin, S., Schrimsher, G.W., Muzyczka, N. and Reier, P.J. (1997). Efficient transduction of green fluorescent protein in spinal cord neurons using adeno-associated virus

D. Transfections. NIH 3T3 cells were plated at 5 X 104 cells in 24 well plates, and allowed to grow in DMEM with calf serum for 24 hours. The medium was aspirated, and 200 µl of the transfection mixtures was added to four wells for each time point. The plates were incubated for 2 hours and then 800 µl of DMEM with calf serum was added to each well. They were then incubated for 24 hours. E. Luciferase assays. Relative luciferase activity was determined with the enhanced luciferase assay kit (556-866; PharMingen, San Diego, CA) and Monolight luminometer (2010; Analytical Luminescence Laboratories, San Diego, CA) as per the manufactures recommendations. The cells were lysed with 200 µl of 1X lysis buffer and incubated on wet ice for 30 minutes. Luciferase light emissions from 20 µl of the lysate were integrated over a 10-sec period, and results were expressed as a function of the total lysis volume. The data shown is a mean result of the 4 wells measured and the error bars represent one standard deviation. F. Characterization of the Size Distributions. Size distributions of transfection particles were measured using the Coulter N4 Plus Submicron Particle Sizer according to the manufactures recommendations. 200 µls of the formulation mixtures were mixed with 3 mls of water, which had been purified using a 0.2 micron filter, in standard plastic cuvettes. Small adjustments in concentration were made in order to fall within the particle counts recommended by the machine. Laser refraction was performed on these mixtures for 425 sec at angles of 30° and 90° at a temperature of 20°C. SDP analysis of these results was performed using 30 bins with a range of 100 nm to 2000 nm. G. Zeta Potential. The distribution of the surface charges of the particles were determined using the Coulter Delsa 440. The same mixture of particles used for the size experiments was used for these measurements. Zeta potential was determined by Doppler shift measurements of the movement inducted by an oscillating electric field. Measurements at 4 different angles are recorded on each graph.

Acknowledgements The authors would like to thank Martin Woodle for initial training in the use of the Delsa and the N4. We also thank Caroline Toll, Stella Somiari, Jim Mixson, and Joe Drabick for patient assistance in proofreading this

169

Ahearn and Malone: Cationic liposome mediated transfection vectors containing cell type-specific promoters. Gene Ther 1, 16-24. Smith J.G., Wedeking, T., Vernachio, J.H., Way, H. and Niven, R.W. (1998). Characterization and in vivo testing of a heterogeneous cationic lipid-DNA formulation. Pharm Res 15, 1356-63. Stegmann T. and Legendre, J.Y. (1997). Gene transfer mediated by cationic lipids: lack of a correlation between lipid mixing and transfection. Biochim Biophys Acta 1325, 71-9. Sternberg B., Hong, K., Zheng, W. and Papahadjopoulos, D. (1998). Ultrastructural characterization of cationic liposomeDNA complexes showing enhanced stability in serum and high transfection activity in vivo. Biochim Biophys Acta 1375, 23-35. Takeuchi K., Ishihara, M., Kawaura, C., Noji, M., Furuno, T. and Nakanishi, M. (1996). Effect of zeta potential of cationic liposomes containing cationic cholesterol derivatives on gene transfection. FEBS Lett 397, 207-9. Wrobel I. and Collins, D. (1995). Fusion of cationic liposomes with mammalian cells occurs after endocytosis. Biochim Biophys Acta 1235, 296-304. Xu Y. and Szoka, F.C.J. (1996). Mechanism of DNA release from cationic liposome/DNA complexes used in cell transfection. Biochemistry 35, 5616-23. Yang J.P. and Huang, L. (1998). Time-dependent maturation of cationic liposome-DNA complex for serum resistance. Gene Ther 5, 380-7. Zabner J., Fasbender, A.J., Moninger, T., Poellinger, K.A. and Welsh, M.J. (1995). Cellular and molecular barriers to gene transfer by a cationic lipid. J Biol Chem 270, 18997-9007. Zanta M.A., Belguise-Valladier, P. and Behr, J.P. (1999). Gene delivery: a single nuclear localization signal peptide is sufficient to carry DNA to the cell nucleus. Proc Natl Acad Sci USA 96, 91-6.

Robert Malone

170

Gene Therapy and Molecular Biology Vol 4, page 171 Gene Ther Mol Biol Vol 4, 171-176. December 1999.

Essentials of radionanotargeting using oligodeoxynucleotides Review Article

Kalevi J. A. Kairemo1,3 , Antti P. Jekunen2 , and Mikko Tenhunen3 1

Department of Clinical Chemistry, Helsinki University Central Hospital, FIN-00029, Finland Department of Clinical Pharmacology, Helsinki University Central Hospital, FIN-00029, Finland 3 Department of Oncology, Helsinki University Central Hospital, FIN-00029, Finland ___________________________________________________________________________________ 2

Correspondence: Kalevi J A Kairemo, MD, PhD, MSc (Eng), Department of Clinical Chemistry, Helsinki University Central Hospital, FIN-00029, Finland. Tel. +358-9-47172565; Fax. +358-9-47176678; E-mail:Kalevi.Kairemo@huch.fi Key Words: Antisense, oligonucleotide therapy, radionanotargeting, radionuclide, phosphorothioates, Auger electron emitter, cancer Received: 9 August 1999; accepted: 29 August 1999

Summary Antisense oligomers may be used for carrying radiation source into a specific location inside a tumour cell. Effects of radioactive labeled oligos may be exerted both via direct antisense inhibition and radiation. This radionanotargeting approach may provide several benefits to conventional treatment modalities, and radiation is minimized in adjacent tissue. In addition, a combination of radiation and antisense activity of oligodeoxynucleotide may result in synergistic interaction, as there are two different treatment modalities hitting a single mechanism of action. We have previously shown that oligonucleotide therapy is effective with internally labeled oligodeoxynucleotide phosphorothioates P-32, P-33 and S-35. Here, we review our results and discuss the role of radionanotargeting. We refer to our previous results of a large selection of radionuclides; we have calculated in vivo subcellular tissue distribution for oligodeoxynucleotide phosphorothioates using decay characteristics of ten !- and Auger-emitting radionuclides. The absorbed nuclear doses of these radiolabelled oligonucleotides were estimated in different cellular dimensions using the subcellular biodistribution data. These results indicate that Auger-emitter isotopes do not give higher absorbed cell nuclear doses than the isotopes suitable for internal labeling of oligo phosphorothioates. The best isotopes for subcellular targeting were P-33 and S-35 giving smallest variation of nuclear dose in the cell dimensions we studied (nuclear diameter 6-16 Âľm, cellular diameter 12-20 Âľm). Therefore, we conclude that radionanotargeting by oligonucleotides may provide synergistic interaction and should be carried on with short range !-emitters suitable for internal labelling of oligonucleotides unless relative biological effectiveness of Auger-emitters could be remarkably improved. Further preclinical evaluation of radionanotargeting based on radio-oligos should be continued.

sequence. The primary target for ionizing radiation is nuclear DNA, and the radiotoxicity of Auger electron emitters is mainly due to low energy electrons with short ranges (1-10 nm). Antisense techniques have also limitations, e.g. limited uptake and unspecific binding (Kairemo et al. 1999a). Antisense oligos have been studied in vivo with regards to their pharmacokinetics, pharmacological and toxicological properties. The pharmacokinetics of various oligos in animal models have already been determined (Agrawal et al.1991, Sands et al.1994). The selection of

I. Introduction Radionanotargeting may provide several benefits over conventional treatment modalities. Theoretically high subcellular concentration is achievable due to an accumulation effect. Radiation is minimized in adjacent tissue. In addition, a combination of radiation and antisense activity of oligodeoxynucleotide may result in synergistic interaction. Antisense oligodeoxynucleotides (oligos) are designed against the most specific target, a DNA/ RNA 171

Kairemo et al: Radionanotargeting using oligodeoxynucleotides an appropriate nucleotide sequence for oligos is of importance in inducing translation arrest. However, the therapeutic possibilities of radiolabeled antisense oligodeoxynucleotides are very much dependent on the characteristics of radiation. The aim of this article is to demonstrate the importance of radionuclides in radionanotargeting. We also discuss a label optimization procedure based on cellular data, and discuss the suitability of several radionuclides for labeling oligos. This review article updates our research programs focused on developing nanotargeting with radiolabelled oligodeoxynucleotides planned up to the first in human trial.

phosphorothioate (Iversen et al. 1994). The pharmakokinetics of the compounds were not expected to change for phosphorus or sulfur labeling. The activity concentrations were calculated using the estimated ratios of relative organ weights to whole animal weight; for liver, kidneys, bone marrow and tumor these were 9, 2, 3 and 5%, respectively (Cossum et al. 1993, Iversen et al. 1994). The cumulative concentration, i.e., the area under the time-concentration curve, was estimated either from the published time-concentration data assuming monoexponential elimination after the maximum uptake, or using published biological half-lives and areas under the time-concentration curve. Table I shows all the oligonucleotides used in our calculations. The results for the estimated organ doses are presented in Table II. There is only a slight difference between the doses of P-33 and S-35 whereas the doses of P-32 are approximately 10-fold higher due to the greater disintegration energy. The absorbed dose distribution does not merely describe the biological response of the cell or tissue to radiation (Kairemo et al.1999a). With an externally delivered radiation dose in radiotherapy of cancer it is known that total dose and dose rate (fraction size and interval) and total treatment time are the most important external factors affecting radiation reaction of healthy tissue and tumor response. Even if the biokinetics and activity distribution of two tracers labeled with different radionuclides are similar, the differences in physical dose distribution and decay time lead to differences in dose-rate distribution which can be biologically significant.

II. State of the art We have previously shown using the biodistribution data of oligonucleotide phosphorothioates in a xenograft model that this type of therapy can theoretically be given with P-32 and P-33 (Kairemo et al. 1996a). We have extended the analysis to several more isotopes, e.g. beta and Auger emitting radionuclides, and we calculated tumor and organ doses, and in vivo subcellular tissue distribution for oligodeoxynucleotide phosphorothioates (Kairemo et al. 1998). We estimated the dosimetric properties of oligonucleotides at the cellular level, that could be predicted from existing data on the characterization of phosphorolabeled oligos. Predetermined situations differed from each other by nuclear and cell sizes. This made it possible to assess relative radiation exposures in these variable cellular dimensions. We used 10 radionuclides, P-32, S-35, Cr-51, Ga-67, In-111, In114m, I-123, I-125, I-131, Tl-201, with different physical properties in our calculations. Four predetermined cellular dimensions were used, even though in reality cell and nuclear sizes may vary in a particular tumor type (Kairemo et al. 1996b). We chose to keep mathematics simple to get more understandable results. Calculations can give a recommendable source for the labeling of an oligo, and thus allow proper selection of the optimal label. This requires estimation of benefits of one radionuclide over others among different isotopes in several cell models with different cell dimensions. The accumulated dose from internally administrated radionuclides has been estimated in our calculations as described earlier (Kairemo et al. 1996a, b, 1998, 1999a). In addition to the tumor dosimetry it is equally important to consider the dosimetry in the normal tissue for the optimization of the radiation therapy. For that reason we have made dose calculations using published biodistribution data of oligonucleotides in the mouse with a 15-mer In-111-labeled oligonucleotide sequence coupled with diethylenetriamine pentaacetateisothiocyanate (Dewanjee et al.1994). All 15-21-mer oligos gave almost identical liver and kidney distributions in mice as well as in rats with the 25-mer oligodeoxynucleotide phosphorothioate GEM91 ( Zhang et al. 1995) and the 27-mer oligonucleotide

III. Discussion The ideal dose in external radiation therapy has traditionally been defined as a dose which gives as many cures as possible before exponential increase in complications. The dose is always depending on the nature of the complications. Of course, the worst complication of radiotherapy is tumor recurrence. Enhanced local control is obtained when radiotherapy is followed by or administered simultaneously with adjuvant chemotherapy in locally advanced cancer. Combination treatment is based on attempting to increase the therapeutic index. External radiation may be replaced by oligonucleotide radiotherapy which is highly specific minimizing the radiation effects to normal tissue and dramatically reducing complications (Kairemo et al. 1999a). Therapeutic index is increased as dose limiting late side-effects are not occurring or occur only minimally. In order to improve antisense oligonucleotide efficiency, chemical modifications have been developed, and improvement of oligonucleotide uptake has been achieved with different systems of vector development including liposomes (neutral, cationic, immunoliposomes), nanoparticles, or covalent attachment to a carrier (Lefbrevre-dÂ´Hellencourt et al. 1995). Polyalkylcyanoacrylate nanoparticles have been introduced as polymeric carriers of oncogene-targeting antisense DNA (Schwab et al. 1994). Our aim was to optimize radiation 172

Gene Therapy and Molecular Biology Vol 4, page 173 exposure and to select the radiation source that provides the highest amount of radiation to the subcellular target (nucleus) and to diminish the radiation in the surroundings. The envisioned therapeutic use of radiolabeled antisense oligos is based on the assumption that an appropriate amount of radiation is delivered to a targeted location of specific sequence of cellular DNA or RNA causing local damage. In addition, the radiotherapeutic effect may be enhanced by antisense mediated inhibition of gene function. Tumor-specific activity will be obtained by hitting appropriate targets, such as anti-oncogenes or tumor-suppressor genes, e.g. p53-mutations. The number of targets, may be increased in many ways, e.g. by inductive manipulation or gene transfer, and thus the efficacy of nucleotide radiotherapy can be improved. Targets with different cellular locations have been described, such as mRNA translation sites, pre-mRNA splicing sites, or the DNA molecules themselves. Use of antisense oligos to inactivate genes has still several difficulties and requires improvements: these include delivery of the oligo into cells and entry to an appropriate intracellular compartment, nonsequence specificity, optimizing pharmacokinetic properties and designing new and better oligo backbones. Radionanotargeting has a limited range from the radiation source resulting in a rapid dose fall-off effect while avoiding damage to surrounding tissues. However, applications remain local and systems for successful systemic administration need to be established. Our approach has focused on optimizing the type of label for oligos using Auger-emitting radionuclides by calculating subcellular dose distribution. We show that, for subcellular targeting, the S-35 and P-32 internal labels

give the lowest variation in estimated absorbed nuclear doses using our cell model of given dimensions (nuclear diameter 6-16 Âľm, cellular diameter 12-20 Âľm). The doses vary considerably using Auger-emitting isotopes depending on cellular dimensions; however, in small cells Auger-emitting isotopes may give a high dose (Kairemo et al. 1998). In tumors, cell dimensions may vary and, therefore, the above mentioned Auger-emitting isotopes should be applied only when nuclear target circumstances are well characterized. The high energy !-emitter P-32 gives a nuclear dose closest to uniform distribution in cell sizes; however, this is due to high energy.

Table I. List of used oligos in our calculations Name of the oligo c-myc antisense Anti-ICAM-1, ISIS 3082 Anti-ICAM-1, ISIS 9045 Anti-HPV, ISIS 2105 Anti-HPV, ISIS 2911 Anti-HIV-1 Anti-HIV-1 Peptide nucleic acid (PNA)

Size, mer 15 20

Reference Dewanjee, 1994 Crooke, 1996

Crooke, 1996

Crooke, 1995

25 27 15

Iversen, 1994 Iversen, 1994 Mardirossian, 1997

Table II . Calculated organ doses for different internally labelled oligomers in mouse models. The percentage organ doses refer to those obtained in tumor models. Oligomer

Peptide nucleic acid, 15-mer c-myc, antisense, 15 mer ISIS 3082, 20-mer ISIS 9045, 20-mer ISIS 2105, 21-mer c-myc, antisense, 15mer

Initial activity (% of injected dose) 0.19, liver 1.45, kidney 6.95, liver 5.15, kidney 18.0 liver 25.0, kidney 45.0, liver 12.0, kidney 18.0, liver 25.0, kidney 11.0, tumor

Biological halflife, *Tb (hrs)

Liver dose (S35) Gy/MBq

Kidney dose (S35) Gy/MBq

Reference

5.1, liver 4.8, kidney 178.2, liver 170.7, kidney 62.0, liver 112.0, kidney >1000, liver >1000, kidney 62.0, liver 112.0, kidney 194.0, tumor

0.003 (0.078%) 0.40 (100%) 0.40 (90%) 30 (7620%) 0.40 (90) 1.0, tumor (100%, tumor)

0.010 (0.79%) 1.30 (100%) 4.0 (320%) 35.0 (2710%) 4.0 (320%)

Mardirossian, 1997

173

Dewanjee, 1994 Crooke, 1996 Crooke, 1996 Crooke, 1996 Dewanjee, 1994

Kairemo et al: Radionanotargeting using oligodeoxynucleotides Currently, RNA expressions in selecting the target gene are being able to detect by cDNA microarrays in a single experiment screening thousands of genes. Thus the likelihood of the selected gene being the most important target is going to be high. This provides an excellent opportunity of modulating tumor to be more sensitive for chemotherapeutic agents and radiation. For example, use of the p53 as an early model for this approach. Transcriptional activation of genes by p53 may coordinately shut down cell cycle progression and induce a battery of genes involved in DNA repair. DNA damage induces p53 accumulation. Cells lacking p53 are resistant to other forms of apoptotic induction, such as that caused by chemotherapeutic agents and radiation. Tumors that have lost p53, are no longer able to respond to adverse growth conditions by initiating apoptosis. Docetaxel and irinotecan are examples of new efficacious drugs in variety of tumor types with new mode of action: prevention of depolymerization of tubulin and specific DNA topoisomerase I inhibition, respectively. These drugs have a favorable interaction with radiation, and presumably highly usable in radionanotargeting approach. When obstacles have been resolved in antisense approach, including administration, delivery, uptake, accumulation in the target, binding to the receptor molecule, effective time, oligonucleotide radionanotargeting may be provide a new option for radiation therapy. The possible cellular targets have been presented in Table III. Successful radionanotargeting with oligos should result in a therapy where metastases are treated while irradiation is low in normal tissue except close surroundings of tumor. On the contrary, when using external radiotherapy, metastases are not located in the field but are left outside, while healthy normal tissue is also radiated. Radionanotargeting with oligos may many applications, e.g. in treatment of metastases in lymph nodes, locally advanced cancer, adjuvant therapy before and after the surgical operation with intention of either making tumors smaller or treating remnant tumors. In addition, true interactions between chemotherapeutic drugs and gene expressions are continuously being defined, which may provide an opportunity to resensitize tumor cells for already acquired resistance. Our project of developing clinical therapy option based on radionanotargeting with oligos has now been partially completed: the first section which consists of an estimation and rationale for the selection of label. The second section is underway consists both of in vitro and in vivo experiments challenging our current conclusions, and the third section will be the first in human trial.

We have previously found (Kairemo et al. 1996a,b) that P32 labeled oligos destroy non-target cells because of their long range. This is not the case when the !-emitters, P-33 and S-35, were used which are optimal when targets are smaller than 300 Âľm in diameter (Kairemo et al. 1996a). Preliminary studies using I-125 labeled oligos for mammalian cell lines suggested that oligonucleotides delivered with liposomes give a lower nuclear dose than DNA-incorporated I-125-UdR (Sedelnikova et al. 1998). However, our calculations indicate that there are several more optimal radionuclides than I-125 (Kairemo et al. 1999b). For example, S-35 and P-32 doses seem to concentrate more efficiently around the nucleus than I-125, which may be of practical importance in delivering the effective doses to the nuclear target. It should be emphasized that the behavior of the radiation at small distances is crucial. It would be important in oligoradiotherapy to achieve the highest possible uptake in the target cell and minimal radiation toxicity to surrounding cells. I-125 and other radiolabels should not be used unless better specificity is achieved. It is possible to use a mixture of radioisotopes to ensure a complete coverage of targets in more than one locations, e.g. targeting nuclear related and cellular RNA at the same time. In addition, modern imaging technique allows visual control over kinetic events. Sometimes the target is dense, e.g. in nucleus or it can be diffusely spread around the cellular area. Dual labeling with P-32 and S-35 may provide therapeutic benefits when treating smaller and larger targets simultaneously (Kairemo et al. 1999b). P-33 and S-35 have some benefits over P-32, since the organ doses remain smaller and thus the therapeutic index may be wider. Critical organ exposure remains 10-fold lower with P-33 and S-35 than with P-32. Moreover, P-33 and S-35 concentrate more efficiently around the target than P32, which could be of practical importance for the delivery of effective doses to the tumor. It can be addressed that the behavior of the radiation at small distances is crucial. Therefore, the radionuclides P-33 or S-35 are more suitable than P-32 for cell destruction at short distances. This was also clearly demonstrated with calculations of tumor doses for 1 g, 1 mg and 1 Âľg tumor masses (Kairemo et al. 1996b). The microscropic tumors cannot be treated using P-32. Instead, with P-33 or S-35 tumor doses up to 7.5-fold for 1 mg tumor mass and up to 50fold for 1 Âľg tumor mass are achieved. Tumors larger than 1 g could be treated with any of these radionuclide labeled oligos. Oligos are carrying the radioactivity source inside the cell and finally achieve close contact with target RNA macromolecules. Nanotargeting means a specific way of targeting small molecules at nanometer scale by antibodies or antisense oligonucleotides. Isotopes in oligonucleotide phosphorotioates are located in S or P atoms. Emitted radiation of isotopes affects structures close to the binding site of oligo. Antisense oligonucleotides serve as vehicles for radioisotopes enhancing targeted efficacy (Figure 1).

References Agrawal S, Temsamani J, Tang JY. (1991) Pharmacokinetics, biodistribution and stability of oligodeoxynucleotide phosphorothioates in mice. Proc Natl Acad Sci U S A 88: 7595-7599.

174

Gene Therapy and Molecular Biology Vol 4, page 175

Figure 1. Schematic representation of radionanotargeting principles using oligos.

Table III. Schematic presentation of main characteristics in different cellular targets. Factor accessibility target induction efficacy exposure for surrounding cells

Outer membrane receptor easy stable possible, slow low high

Cossum P. A., H. Sasmor, D. Dellinger, L. Truong, L. Cummins, S. R. Owens, P. M. Markham, J. P. Shea and S. Crooke. (1993) J. Pharm. Exp. Ther. 267,1181. Crooke RM, Graham MJ, Cooke ME, Crooke ST. (1995) In vitro pharmacokinetics of phosphorothioate antisense oligonucleotides. J Pharmacol Exp Ther 275, 462-473. Grooke ST, Grtaham MJ, Zuckerman JE (1996) Pharmacokinetic properties of several novel oligonucleotide analogs in mice. J Pharmacol Exp Ther 277: 923-937,

RNA moderate transient usual, fast moderate moderate

Nuclear difficult stable no high low

Dewanjee MK, Ghafouripour AK, Kapadvanjwala M, Samy AT. (1994) Kinetics of hybribdization of mRNA of c-myc 111 oncogene with In-labeled antisense oligodeoxynucleotide probes by high-pressure liquid chromatography. Biotechniques 16 , 844-850. Iversen P. L., J. Mata, W. G. Tracewell and G. Zon (1994) Pharmacokinetics of an antisense phosphorothioate oligodeoxynucleotide against rev from human immunodeficiency virus type I in adult male rat

175

Kairemo et al: Radionanotargeting using oligodeoxynucleotides following single in jections and continuos infusion. Antisense Res Dev 4, 43. Lefebvre-d'Hellencourt C, Diaw L, Guenounou M (1995) Immunomodulation by cytokine antisense oligonucleotides. Eur Cytokine Netw; 6, 7-19. Kairemo KJ, Tenhunen M, Jekunen AP (1996a) Dosimetry of radionuclide therapy using radiophosphonated antisense oligodeoxynucleotide phosphorothioates based on animal pharmacokinetic and tissue distribution data. Antisense Nuclei Acid Drug Dev 6: 215-220. Kairemo KJA, Tenhunen M, Jekunen AP (1996b) Oligoradionuclidetherapy using radiolabelled antisense oligodeoxynucleotide phosphorothioates. Anti-Cancer Drug Design 11: 439-449. Kairemo KJA, Tenhunen M, Jekunen AP (1998) Gene therapy using antisense oligodeoxynucleotides labelled with Auger-emitting radionuclided. Cancer Gene Therapy 5:408-412. Kairemo K, Jekunen A, Tenhunen (1999a) Dosimetry and optimization of in vivo targeting with radiolabeled antisense oligonucleotides. In: M.I.Phillips (ed), Antisense Technology A volume of Methods in Enzymology, Academic Press New York, pp. 506-524.. Kairemo Kja, Jekunen AP, Tenhunen M (1999b) Problems associated with oligonucleotideradiotherapy. J Nucl Med, 40, 1582. Mardirossian G, Lei K, Rusckowski M. (1997) In vivo hybridization of technetium-99.m.labelled peptide nucleic acid (PNA). J Nucl Med 38: 907-913. Sands H, L.J. Gorey-Feret, A.J. Cocuzza, F.W. Hobbs, D. Chidester, and G.L. Trainor. (1994) Biodistribution and metabolism of internally 3H-labeled oligonucleotides. I. Comparison of a phosphodiester and a phosphorothioate. Mol Pharmacol 45, 932-943. Schwab G (1994) Antisense oligonucleotides adsorbed to polyalkylcyanoacrylate nanoparticles specifically inhibit mutated Ha-ras-mediated cell proliferation and tumourigenicity in nude mice. Proc Natl Acad Sci U S A 91 , 10460-10464. Sedelnikova, OA., I.G. Panyutin, A.R. Thierry, and R.D. Neumann. (1998) J. Nucl. Med. 39; 1412. Zhang R, J. Yan, H. Shahinian, G. Amin, Z. Lu, T. Liu, M.S. Saag, Z. Jiang, J. Temsamani, B.R. Martin, P.J. Schechter, S. Agrawal, and R.B. Diasio (1995) Pharmacokinetics and tissue distribution in rats of an oligodeoxynucleotide phosphorothioate (GEM 91) developed as a therapeutic agent for human immunodeficiency virus type-1. Biochem Pharmacol 49, 929-939.

Kalevi Kairemo

176

Gene Therapy and Molecular Biology Vol 4, page 177

177

Gene Therapy and Molecular Biology Vol 4, page 177 Gene Ther Mol Biol Vol 4, 177-182. December 1999.

ets-1 mRNA as target for antisense radiooligonucleotide therapy in melanoma cells Research Article

Kalevi J. A. Kairemo1,3 , Ketil Thorstensen1 , Merete Mack 1 , Mikko Tenhunen2 , and Antti P. Jekunen4 1

Dept. of Clinical Chemistry, Norwegian University of Science and Technology, Trondheim, Norway. Dept. of Oncology, Helsinki University Central Hospital, Finland 3 Dept. of Clinical Chemistry, Helsinki University Central Hospital, Finland 4 Dept. of Clinical Pharmacology, Helsinki University Central Hospital, Finland _________________________________________________________________________________________________ 2

Correspondence: Kalevi J A Kairemo, MD, PhD, MSc (Eng), Department of Clinical Chemistry, Helsinki University Central Hospital, FIN-00029, Finland. Tel. +358-9-47172565; Fax. +358-9-47176678; E-mail:Kalevi.Kairemo@huch.fi Key Words: Antisense, phosphorothioate, oligonucleotide therapy, radiation, phosphorothioates, angiogenesis, endothelial cell, ets-1 Abbreviations: AS-PODNs, antisense phosphorothioate oligodeoxynucleotides; VEGF, vascular endothelial growth factor Received: 9 August 1999; accepted: 29 August 1999

Summary Angiogenesis provides a novel target for anticancer therapy, in particular radiochemo-therapy as endothelial cells in the vascular wall are sensitive to radiation. Antisense phosphorothioate oligodeoxynucleotides (AS-PODNs) may serve as vehicles for carrying cytotoxic or radioactive agents into a particular intracellular location. Radiolabelled AS-PODNs have the potential of having both antisense and radiation effects. Recently, vascular endothelial growth factor (VEGF)-induced invasiveness was shown to be specifically inhibited by AS-PODN directed against ets-1. Previous studies have shown that radio-oligonucleotide therapy may be effective with AS-PODNs internally labelled with 32P, 33P or 35S. Theoretically, 35S gave the smallest variation in nuclear dose in the different cell dimensions studied (Kairemo et al., Cancer Gene Ther 1998; 5: 408-12). This means that cell nuclear targets should be treated with the short range !-emitters 35S or 33P for optimal radio-oligonucleotide therapy. Here we explore this possibility using 33P labeled 17-mer AS-PODNs directed against ets-1 in human melanoma cells in vitro. Inhibition of cell growth was observed in the following order: labeled AS-PODN > nonlabeled AS-PODN > labeled sense PODN > nonlabeled sense PODN > transfection agent. Even with a single 33P at the 5Â´-end of the ASPODN melanoma cell uptake of label was approximately 0.5 mBq/cell. The nuclear doses in this experiment varied from 1.9 to 3.7 cGy. Thus, in vitro and in vivo use of radio-oligonucleotide therapy utilizing 33P radionanotargeting, e.g. in angiogenesis through ets-1, are highly recommended.

human to Drosophila. The common feature of ets proteins is a well-conserved 85 amino acid domain that binds specifically to DNA containing a (G/C)(A/C)GGAAGT consensus sequence (Macleod et al. 1992; Wasylyk et al. 1993; Timms and Kola 1994). The ets gene family includes, in addition to ets-1 and ets-2, also erg, elk-1, elk-2, pu-1, fli-1 and E74 (Fisher et al. 1992). ets-1 encodes a set of phosphoproteins ranging in size from 39 to 51 kD (Fisher et al. 1992). However, whereas the chicken ets protein, which contains both the ets-1 and ets-2 domains,

I. Introduction The oncogene v-ets was originally discovered as a component of a chimeric genome, along with a truncated v-myb gene, present in the genome of E26, an avian leukosis virus (LePrince et al. 1983; Nunn et al. 1983). Since then, a family of transcription factors, known as the ets family, involved in a wide variety of biological processes including growth control and development, transformation, and T-cell activation, have been cloned and sequenced from a variety of species ranging from 177

Kairemo et al: Antisense ets-1 therapy for melanoma distributes equally between the cytoplasm and nucleus, in the human and other mammals, the ets-1 protein is cytoplasmic and the ets-2 protein nuclear. This, together with their noncoordinate expression, suggests that ets-1 and ets-2 have different biologic functions (Fujiwara et al. 1988). Ets-1 is preferentially expressed at high levels in B and T cells and is regulated during both thymocyte development and T cell activation (Chen 1985; Bhat et al. 1989). Studies in mice have shown that Ets-1 is essential for normal maintainance, survival and activation of B- and T-lineage cells (Bories et al. 1995; Muthusamy et al. 1995). Bhat et al. (1990) found that, following T-cell activation, ets-2 mRNA and proteins are induced, while ets-1 gene expression decreases to very low levels. Amplification and rearrangement of ets-1 has also been implicated in human leukemia (Goyns et al. 1987; Rovigatti et al. 1986). ets-1 is also expressed in endothelial cells during blood vessel development and in fibroblasts adjacent to tumor cells in various invasive human carcinomas (Wernert et al. 1992; Wernert et al. 1994), suggesting that the ets-1 gene could be involved in angiogenesis associated with tumor growth and normal development. Tumor growth, progression and metastasis are dependent on the formation of new capillary blood vessels from existing vessel, a process termed angiogenesis (Folkman and Shing 1992); rapidly growing tumors are often hypoxic due to insufficient vascularization. Angiogenesis is a cascade of processes involving both soluble angiogenic factors and insoluble extracellular matrix factors (Jekunen and Kairemo 1997). Multiple soluble molecules that stimulate angiogenesis are released by tumor cells as well as host cells such as endothelial, epithelial and mesothelial cells and leukocytes. In SKMEL-2 human melanoma cells cultured under hypoxic conditions the synthesis of Vascular Endothelial Growth Factor (VEGF) is stimulated (Claffey et al. 1996). In endothelial cells VEGF induces the expression of the proto-oncogene ets-1 and stimulates endothelial cell migration. On the other hand, antisense P-ODNs directed against ets-1 mRNA inhibit the ability of endothelial cells to migrate (Chen et al. 1997). In fact, induction of ets-1 expression appears to be a common phenomenon in endothelial cells stimulated by angiogenic growth factors (Iwasaka et al. 1996). Thus, the ets-1 gene apparently plays a direct role in angiogenesis. Antisense oligodeoxynucleotides (ODNs) are short (typically 15 bases) stretches of synthetic DNA that are complementary to specific regions of cellular mRNA or DNA. This complementarity allows them to hybridize to specific parts of cellular mRNA or DNA, forming mRNADNA or DNA-DNA duplexes. The duplex formation disrupts the function of that particular gene either at the translational or transcriptional level. Due to their specificity antisense ODNs have become attractive potential tools for specific therapeutic applications, e.g. as specific inhibitors of malignant cell growth. In order to be effective the antisense ODNs must first enter the cell and

then escape degradation by intracellular nucleases to achieve adequate concentrations in the correct intracellular compartment. As cellular nucleases effectively degrade phosphodiester ODNs, several more nuclease resistant ODNs have been developed. Of these the phosphorothioate ODNs (P-ODNs) in which a non-bridging oxygen atom has been replaced by a sulphur atom, are the most common. Antisense ODNs may also function as specific carriers of cytotoxic drugs into cells, provided that the target sequence is specifically expressed in that particular cell type. Similarly, by attaching an appropriate radionuclide to the ODN, the selective delivery of radiation to a particular cell may be achieved. The primary target for ionizing radiation is the nuclear DNA and if the radiation source is in close proximity to the DNA, molecular damage would ensue (Kairemo et al. 1999). Theoretical studies evaluating Auger and gammaemitting radionuclides as well as beta-emitters (Kairemo et al. 1998) have suggested that short range beta-emitters such as 33P and 35S may be best suited for delivery of radiation confined to the cell nucleus. ODNs are easily labeled with 33P or 35S at their 5Â´- or 3Â´-end. Furthermore, since antisense P-ODNs contains both sulphur and phosphorus atoms that could be exchanged with 33P and/or 35 S as part of their structure, these radionuclides offer benefits over e.g. transition metal nuclides since they do not require any extra coupling techniques for incorporation of the nuclide into the P-ODN. The aim of this study was to investigate the potential of radiolabeled antisense oligodeoxynucleotides to specifically inhibit the growth and to destroy melanoma cells utilizing ets-1 mRNA as target (Figure 1).

II. Results During the first 24 h of incubation with 100 nM 33Plabelled antisense or sense oligonucleotides the cells accumulated 33P to an activity of approximately 0.15-0.25 mBq/cell. No further cellular accumulation of label was observed during the next 24 h. Incubation of cells with 200 nM 33P-labeled oligonucleotides increased cellular accumulation of label to approximately 0.4-0.5 mBq/cell at 24 h incubation, with no further accumulation of label during the next 24 h (Table 1). Thus, cellular accumulation of 33P-labeled oligonucleotides apparently increases linearly with extracellular oligonucleotide concentration. The cellular accumulation of 33P at 24 h with 200 nM oligonucleotides corresponds to a cellular oligonucleotide uptake of approximately 2.5 pmol DNA per million cells. The cumulative activity per cell was calculated for the two different oligonucleotide concentrations at 48 h incubation. At 100 nM oligonucleotide the cumulative activity was somewhat higher for the antisense compared to the sense oligonucleotide. At 200 nM oligonucleotide concentration there was no difference in cumulative activity between the antisense and the sense

178

Gene Therapy and Molecular Biology Vol 4, page 179 oligonucleotide (Table 1). Based on the data shown in Table 1 and assuming a cell diameter of 14 Âľm, radiation doses per cell were calculated. The radiation doses varied between 1.2 and 4.1 cGy. In cells that were allowed to accumulate 33P -labelled antisense oligonucleotide directed against ets-1 mRNA for 24 h cell growth was inhibited approximately 25% (Table 2). With other oligonucleotides (i.e. labeled sense,

unlabeled antisense or sense) the effect on cell growth was less pronounced, irrespective of whether the oligo was labelled or not. The transfection reagent alone inhibited cell growth by approximately 15%. Incubation of cells with 100 or 200 nM oligonucleotide apparently had no effect on cell growth. After 48 h of incubation no effects on cell growth were observed.

Figure 1. Effects of Ets-1 protein and antisense approach for cell killing.

Table 1: Accumulation of 33P during incubation of cells with labelled oligonucleotides. Activities, cumulative activities and doses per cell are shown.

Oligo concentration (nM) 100 100 200 200

Time (hours)

Antisense Activity (mBq/cell)

24 48 24 48

0.26 0.18 0.47 0.43

Cumulated activity (Bqs/cell)

Nuclear dose (cGy)

30.2

1.9

59.1

3.7

179

Sense Activity (mBq/cell) 0.15 0.14 0.50 0.49

Cumulated activity (Bqs/cell)

Nuclear dose (cGy)

19.0

1.2

64.4

4.1

Kairemo et al: Antisense ets-1 therapy for melanoma Table 2. Effect of oligonucleotide treatment on cell growth.

III. Discussion In this first, to our knowledge, experiment designing internal labeling characteristics for ODNs we used single 5´-end labeling with 33P to obtain a sufficient specific activity for subcellular dosimetric and cell killing experiments. In order to achieve high specific activity, many 33P-atoms may be incorporated as part of the ODN backbone during synthesis. We have shown here that doubling the ODN concentration in the cell may increase cellular radiation doses more than three-fold (Table 1). Theoretically, it is possible to introduce 17 more 33Patoms in this ets-1 P-ODN system. Human melanoma G361 cells were utilized to establish the in vitro model system. These cells are well characterized, simple to maintain in culture, and they express ets-1. They are rather resistant to radiation. These cells are well suited for use in xenograft models in mice, because they grow subcutaneously and they may develop new vessels (angiogenesis). In this preliminary experiment, melanoma cells were incubated with two concentrations of 33P-labeled ODNs for 24 and 48 hours. The accumulation and elimination of ODNs by the cells seemed to be low. Total cellular uptake was less than 1% of added ODNs in the cell culture media. Furthermore, very little events between 24 and 48 hours, this is because influx and efflux are in balance (Table 1). It is important to find the optimal time and concentration for antisense treatment. Actually, in our experiments the lower ODN concentrations gave clearly different doses both for antisense ODN and sense ODN, respectively. The significance of this, if any, is unknown. At higher concentration the doses both for antisense ODN and sense ODN were rather similar. This demonstrates that it is important to find the optimal mode of ODN delivery. During these experiments we did not yet find the optimal amount of transfection reagent, because very small differences on cell growth in various conditions were observed (Table 2). However it was clear that the best effect was obtained using labeled antisense-ODN. This finding has not been shown earlier in the literature. In fact, this to our knowledge, the first report of cell killing utilizing endocytotic therapy with 33P-radionuclide. We calculated internal radiation doses as described previously (Kairemo et al. 1998) as D (target"target). This means that cell-to-cell-interactions as well as activities in the cell media were neglected (negligible?). This should be the case in optimal radionanotargeting. Following loading of the cells with radiolabeled ODNs a simple subcellular fractionation into nuclear, membrane and cytosolic fractions can be performed. This will yield data on the nuclear, cytoplasmic, and cell surface distribution of label (data not shown). Using this approach internal radiation doses were calculated as previously described (Kairemo et al. 1999). Using cell diameter of 14 µm, a nuclear dose of 3 cGy was obtained which is in accordance with the cell doses shown in Table 1.

Treatment None 33 P-Antisense Antisense 33 P-Sense Sense SuperFect

100 nM oligo Cell growth (%) 100 77 86 87 86 84

200 nM oligo Cell growth (%) 100 76 79 82 84 86

The human melanoma G361 cell line produces adequate levels of Ets-1 protein and mRNA under standard culture conditions. Utilizing known inducers (e.g. VEGF) or inhibitors (e.g. tissue plasminogen activator) of ets-1 expression the sensitivity of the detection systems may be changed. Following validation of the detection systems the ability of antisense ODNs to down-regulate the cellular expression of ets-1 can be studied. Our preliminary results indicate that radiolabeled antisense ODNs have to be evaluated with respect to any effects on ets-1 expression in addition to or synergistically with the pure antisense effect. Further studies are needed to decipher the molecular mechanisms of cell killing by radioactive antisense oligonucleotides .

Acknowledgments The authors (KJAK, KT) were supported by a reseach grant from Sintef Unimed Foundation, Trondheim, Norway.

IV. Materials and Methods A. Cell culture Human melanoma cells G631 were cultured in RPMI1640 medium containing 2 mM L-glutamine, 50 U penicillin/ml 50 µg streptomycin/ml and 10% (v/v) FCS, in an atmosphere of 5% CO2/95% air in 60 mm dishes. The doubling time of the cells was approximately 24h. B. Phosphorothioate oligonucleotides Antisense and sense phosphorothioate oligonucleotides against ets-1 were obtained from Amersham Pharmacia Biotech. The oligonucleotide sequences were as described by Chen et al. (Chen et al. 1997) as follows: Antisense: 5’TCGACGGCCGCCTTCAT-3’; Sense: 5’ATGAAGGCGGCCGTCGA-3’. The oligonucleotides were purified by FPLC from the manufacturer and were reconstituted in sterile 1xTE buffer.

180

Gene Therapy and Molecular Biology Vol 4, page 181 Bhat NK, Thompson CB, Lindsten T, et al. (1990) Reciprocal expression of human ETS1 and ETS2 genes during T-cell activation. Proc Natl Acad Sci 87, 3723-3727. Bories J-C, Willerford D, Grévin D, et al. (1995) Increased Tcell apoptosis and terminal B-cell differentiation induced by inactivation of the Ets-1 proto-oncogene. Nature 377, 635-638. Chen JH (1985) The proto-oncogene c-ets is preferentially expressed in lymphoid cells. Mol Cell Biol 5, 2993-3000. Chen Z-Q, Fischer RJ, Riggs CW, et al. (1997) Inhibition of vascular endothelial growth factor-induced endothelial cell migration by ETS1 antisense oligonucleotides. Cancer Res 57, 2013-2019. Claffey KP, Brown LF, del Aguila LF, et al. (1996) Expression of vascular permeability factor/ vascular endothelial growth factor by melanoma cells increases tumor growth, angiogenesis, and experimental metastasis. Cancer Res 56, 172-181. Fisher RJ, Koizumi S, Kondoh A, et al. (1992) Human ETS1 oncoprotein. J Biol Chem 267, 17957-17965. Folkman J, and Shing Y (1992) Angiogenesis. J Biol Chem 267, 10931-10934. Fujiwara S, Fisher RJ, Seth A, et al. (1988) Characterization and localization of the products of the human homologs of the v-ets oncogene. Oncogene 2, 99-103. Goddu SM, Howell RW, and Rao RV (1993) Cellular dosimetry: absorbed fractions for monoenergetic electron and alpha particle sources and S-values for radionuclides uniformly distributed in different cell compartments. J Nucl Med 35: 303-311. Goyns MH, Hahn IM, Stewart J, et al. (1987) The c-ets-1 protooncogene is rearranged in some cases of acute lymphoblastic leukaemia. Br J Cancer 56, 611-613. Iwasaka C, Tanaka K, Abe M, Sato Y (1996) Ets-1 regulates angiogenesis by inducing the expression of urokinasetype plasminogen activator and matrix metalloproteinase-1 and migration of vascular endothelial cells. J Cell Physiol 169, 522-531. Jekunen AP, and Kairemo KJA (1997) Inhibition of malignant angiogenesis. Cancer Treatm Rev 263-286. Kairemo KJA, Tenhunen M, and Jekunen AP (1998) Gene therapy using antisense oligodeoxynucleotides labeled with Auger-emitting radionuclides. Cancer Gene Therapy 5, 408-412. Kairemo KJA, Jekunen AP, Tenhunen M (1999) Dosimetry and optimization of in vivo targeting with radiolabeled antisense oligodeoxynucleotides (oligonucleotide radiotherapy). In Antisense Technology (Phillips MI, ed.), Meths Enzymol, (Abelson JN and Simon MI, eds.) Vol. 314, Academic Press, New York, pp506-524. LePrince D, Gégonne A, Coll J, et al. (1983) A putative second cell-derived oncogene of the avian leukaemia retrovirus E26. Nature 306, 395-397. Loevinger R, and Berman MA (1976) A revised schema for calculating the absorbed dose from biologically distributed radionuclides. MIRD pamphlet no. 1, revised, p.3, Society of Nuclear Medicine, New York. Macleod K, LePrince D, and Stehelin D (1992) The Ets gene family. Trends Biochem Sci 17, 252-256. Muthusamy N, Barton K, and Leiden JM (1995) Defective activation and survival of T cells lacking the Ets-1 transcription factor. Nature 377, 639-642.

C. Labeling of oligonucleotides The oligonucleotides were 5’-end labeled with # -33PATP according to the manufacturers instructions, utilizing a kit from Amersham Pharmacia Biotech. Unincorporated # -33P-ATP was removed by passage through an anion exchange column (Qiaquick oligonucleotide removal kit, Qiagen). Incorporation of 33P into the oligonucleotides was in the range of 60-80% and specific activity was approximately 200 Bq/pmol DNA. D. Incubation of cells Cells were seeded in 60 mm dishes at a density of 5x105 cells per dish in culture medium (se above). The next day cells were washed twice in DPBS before fresh culture medium containing labelled or unlabelled oligonucleotides (200 nM) in SuperFect™ (Qiagen) were added. The cells were incubated in the presence of oligonucleotides for a maximum of 48 h. E. Determination of cell growth At the designated time points the cells were washed in DPBS, trypsinated and the resulting cell suspension counted in a Coulter Z1 (Coulter Electronics, Ltd.). The effect of treatment on cell growth is expressed relative to the control cells receiving no treatment. F. Cell dosimetry The nuclear dose of the internalized oligodeoxynucleotide was estimated using the principles of Medical Internal Radiation Dose (MIRD) schema (Loevinger and Berman 1976). The dose D was calculated as a product of cumulated activity Ã and specific absorption fraction, S, where all the radiation sources k are summed up together:

D=$Ãk ·Sk k

The nuclear dose was calculated using the assumption of uniformly distributed activity and subcellular S-factors of Goddu et al. (Goddu et al. 1993) for the cellular and nuclear diameters of 7 and 4 µm. We assumed that subcellular S-factors of phosphorus-33 are similar to those of sulphur-35, which gives the total Sfactor (cell!nucleus) to be 6.3 x 10-4 Gy/Bqs. The dose from the adjacent cells or from the medium was not taken into account. The tracer concentration was assumed to increase from 0 to 24 hrs linearly and remain constant from 24 to 48 hours.

References Bhat NK, Komschlies KL, Fujiwara S, et al. (1989) Expression of ets genes in mouse thymocyte subsets and T cells. J Immunol 142, 672-678.

181

Kairemo et al: Antisense ets-1 therapy for melanoma Nunn MF, Seeburg PM, Moscovici C, and Duesberg PH (1983) Tripartite structure of the avian erythroblastosis virus E26 transforming gene. Nature 306, 391-395. Rovigatti U, Watson DK, and Yunis JJ (1986) Amplification and rearrangement of Hu-ets-1 in leukemia and lymphoma with involvement of 11q23. Science 232, 398-400. Timms MJ, and Kola I (1994) Regulation of gene expression by transcription factors Ets-1 and Ets-2. Mol Reprod Dev 39, 208-214.

Wasylyk B, Hahn SL, and Giovane A. (1993) The ETS family of transcription factors. Eur J Biochem 211, 7-18. Wernert N, Raes MB, Lasalle P, et al. (1992) c-ets1 protooncogene is a transcription factor expressed in endothelial cells during tumor vascularization and other forms of angiogenesis in humans. Am J Pathol 140, 119127. Wernert N, Gilles F, Fafeur V, et al. (1994) Stromal expression of c-Ets1transcription factor correlates with tumor invasion. Cancer Res 54, 5683-5688.

182

Gene Therapy and Molecular Biology Vol 4, page 183 Gene Ther Mol Biol Vol 4, 183-191. December 1999.

Electronic microarray for DNA analysis Review Article

Lana Feng* and Michael Nerenberg Department of Molecular Biology and Genomics, Nanogen Inc., San Diego, CA _________________________________________________________________________________________________ * Correspondence: Lana Feng, Department of Molecular Biology, Nanogen, Inc., 10398 Pacific Center Court San Diego, CA 92121. Tel: (858) 546-7700; Fax: (858) 410-4650; E-mail: lfeng@nanogen.com Abbreviations: AC, alternate current; DC, direct current; STRs, Short Tandem Repeats; SNPs, single nucleotide polymorphisms; CODIS, Combined DNA Information System. Key words: microarray, electronic hybridization, short tandem repeat, single nucleotide polymorphism, forensics Received: 9 August 1999; accepted: 29 August 1999

Summary Nanogen has developed a microelectronic technology that allows fast and accurate transport and hybridization of DNA on a semiconductor microchip. Based on this technology, we have developed electronic assays for genetic analysis of single nucleotide polymorphisms in several different applications. We have also developed electronic hybridization based short tandem repeats analysis for genetic identification and forensics. Our data show that these assays work in multiple formats under different electronic conditions. A beta instrument is capable of performing these tests on 100-site arrays in a semi-automated fashion. We are also working towards developing an integrated system, which could potentially become a portable device.

perform automatic stringency and quality checks; (v) has a very wide applicability to charged molecules including DNA; and (vi) has broad applicability in biomedical research, medical diagnostics, genomics, and genetic testing and drug discovery. Nanogenâ&#x20AC;&#x2122;s microchips are made by standard semiconductor processing techniques (Mahajan, 1993). The microchips are 1cm by 1cm in size with 25-site, 100-site or 400-site arrays (Sosnowski et al, 1997). The test sites are clustered in the center of the array in a 5 x 5, 10 x 10 and 20 x 20 fashion (Fig. 1). A ring of counter electrodes surrounds them. A platinum wire connects each individual site and their respective counter electrode in the 25 or 100 pad chips. This allows the microinsulation of these sites while in the same time provides equal access to the molecules in an overlaying electrolyte solution. On top of the array is an agarose based permeation layer, which provides the attachment chemistry for anchoring probes as well as an interface between the electrodes and the solution. This interface allows ion flow while retarding penetration of target nucleic acid. It also acts as a buffer zone for DNA from damaging electrochemical reactions on top of the active electrode. In the electronic DNA assays, probe DNA is loaded to the desired site(s) by electronically activating that electrode(s) with a postive charge and the counter electrodes

I. Introduction Recently there has been tremendous interest in microarray technologies ( for reviews, see Nature Genet. Supplement, Vol 12, 1999). The combination of microfabrication, chemistry and molecular biology have allowed the generation of microarrays that permit rapid and miniaturized multiplex analysis of DNA samples. These microarray technologies are also good platforms for applications in drug discovery and genomics. Most microarray technologies require successful integration of miniaturization, array format, microfabrication and highly sensitive detection system to make genetic analysis on a chip possible. They mostly rely on passive hybridization or diffusion of large biological molecules such as nucleic acid or proteins. Nanogen has developed an electronic microchip technology that allows fast and accurate delivery of electronically charged biological molecules to test sites (electrodes) on microarrays followed by hybridization and detection. The technology (i) allows significant acceleration of molecular binding, therefore speeding up reaction time; (ii) affords multiplexing and simultaneous analysis of multiple test results from single sample; (iii) has an open architecture design, which allows the microarrays to be customized easily and quickly by end users; (iv) has high accuracy because it can give end users the ability to precisely manipulate molecular movement electronically and to 183

Feng and Nerenberg: Electronic Microarray for DNA Analysis with a negative charge (Fig.2). The electric field strength generated by this configuration interrogates the solution above the test array and drives the negatively charged DNA molecules to migrate and concentrate over the positively biased sites. Because of the streptavidin attachment chemistry in the agarose permeation layer, we are able to electronically address and anchor our biotinylated capture probes to individual or multiple pads on the array. Subsequently, the target DNA is addressed to these pads electronically in a similar fashion. Our electronic format not only allows fast transport and concentration of target sample on individual test sites, but also affords rapid hybridization of the target DNA to capture probes anchored in situ. Conversely, the electronic polarity can be reversed so that the charge on the test site electrodes is negative and counter electrodes positive. This repulses DNA from the test sites (Sosnowski et al, 1997; Gilles et al, 1999). The resulting electronic washing allows the removal of nonbound target DNA, nonspecific DNA and partially hybridized target DNA respectively with increasing electronic field strength.

order to transport nucleic acid efficiently by free solution electrophoresis, low conductivity buffers are preferred. We have utilized zwitterions which have pIs near neutral pH, where they have little net charge. In the case of histidine, conductivity is less than 100uS/cm. This is three magnitudes less than the buffers commonly used in molecular biology electrophoresis (e.g. Tris-Borate). However, at their pI, zwitterions do not permit hybridization under passive conditions because they are unable to support optimal shielding of charges contributed by the nucleic acid phosphodiester backbone. Buffers that support optimal hybridization should have titratable substituents with pKa values at or near pH 7.0. This requirement is particularly important for our electronic system because acid is generated at the test site electrode by electrolysis during DNA transport. In the absence of buffer, a current of 200nA causes a dramatic drop in pH (Edman et al, 1997). We found that some low ionic buffers containing an imidazole ring, such as histidine and imidazole, could maintain a pH above 5 in the region above the anode during electrolysis. The imidazole ring could serve as the primary source of buffering within the pH range. Furthermore, the imidazole ring can be protonated near neutral pH. This provides cations that shield repulsion between the negatively charged phosphodiester backbone of the two DNA strands during hybridization.

II. pH Generation and the zwitterionic buffer system Our buffer system supports rapid electrode specific transport and concentration of DNA molecules while facilitating accelerated hybridization (Edman et al, 1997). In

Figure 1: Layout for the 25-site and 100-site chips. The array (5 x 5 or 10 x 10) is in the center of the microchip. The ring pads surrounding the array are used as counter electrodes. Contact pads are on the edge of the chip, which are used for contacting with pogo pins that apply electronics to the microchip. For each of the test sites and ring pads, there is a platinum wire connecting it to its respective contact pad. There are two wires for each test site on the 25-site chips.

184

Gene Therapy and Molecular Biology Vol 4, page 185

Figure 2: (A) Capture oligos are biotinylated at the 5â&#x20AC;&#x2122; end. When a test site is positively charged, it will allow negatively charged oligos to move to the site which will become bound in the permeation layer. (B) Target DNA is electronically hybridized to the capture oligo on a desired pad by applying positive charge to that pad. (C) A fluorophore dye labeled reporter oligo is passively hybridized to the target DNA in order to detect the presence of the target. By reversing the polarity of that pad, nonhybridized target DNA will be removed from the vicinity of the test site.

Most importantly, the ability of histidine buffer to support DNA hybridization is dependent on electronic activation at the test sites. Since buffer molecules have very little net charge under neutral pH without electroactivation, they serve to keep complementary single stranded target DNA, such as PCR product from reannealing in solution. In summary, our microelectronic array has the advantage of a programmed pH gradient that allows discrete activation of hybridization zones.

III. AC vs. DC current One drawback of electronic addressing is the generation of damaging agents by electrolysis in solution above the permeation layer. Purines in DNA are particularly susceptible to attack by compounds such as radicals. One way to avoid this is to use thicker permeation layer which allows a greater distance over which the buffer may neutralize these radicals. Another approach is to use a higher 185

Feng and Nerenberg: Electronic Microarray for DNA Analysis concentration of buffer to serve as a scavenger. A third approach is to use alternate current (AC) rather than direct current (DC). The polarity on the DNA test sites is alternated between positive and negative, with the duration under positive current longer than that of negative current. This asymmetric biasing ensures the directed movement and concentration of DNA samples on microarray pads while allowing adequate time for inactivation of damaging radicals. In addition, AC current is better than DC at controlling pH gradient generated by electrolysis at the electrode as it facilitates diffusion and maintains the pH as close to neutral as possible for optimal hybridization. However, in exchange for operating closer to equilibrium, transport and hybridization are slowed down.

(Short Tandem Repeats) genotyping for forensics and human identification (Radtkey et al, Submitted).

V. Comparison of array formats: multiple SNPs per chip vs. multiple patients per chip Most of the microarray technologies currently available focus on large arrays (Lipshutz et al, 1999; Chee et al, 1996; Brown et al, 1999). They provide useful formats for monitoring gene expression, gene screening and drug discovery. But they are less optimal for clinical applications such as medical diagnostics, where relevant target genes are more limited in number. The best examples would be most of the SNPs tests for screening for genetic diseases and drug resistance of microorganisms and cancer cells (Evans 1998; Koza et al, 1996; Wang et al, 1998, 1999). With the electronic platform, screening can be performed in two formats: (i) Multiple SNPs from one patient may be analyzed on one chip. (ii) One SNP from multiple patients may be determined on one chip. The former is a useful format for screening an individual for polymorphisms conferring drug resistance, such as the large numbers of polymorphism in HIV (Vahey et al, 1999). Though passive hybridization may be used, electronic hybridization should provide speed and flexibility and permit different hybridization conditions to be used simultaneously on multiple genes. There is little concern of crosscontamination between patients since only one patient is applied per chip. The second format is better used for higher throughput screening for the same genetic defect (Wang et al, 1999). An example would be a survey for hemochromatosis mutations in a population (Barton et al, 1997). In this case, it is almost impossible to use passive hybridization schemes in an open format since multiple patient samples must be applied to different test sites on the same array. Our microelectronic technology, using an amplicon down format, will perform this with ease because we are able to target DNA sample to specific pads without significant crosscontamination to adjacent sites. The design of cartridges allows easy washes between multiple sample applications (Fig. 4). Therefore, our microchip platform could accommodate both scenarios and perform both assays equally well.

IV. Assay formats Although several different formats are possible, the simplest one for our microelectronic assays is the capture down sandwich format (Fig.2). Capture probes are designed to encompass matched and mismatched regions of the alleles of the target and are complementary to the target sequence. First they are electronically addressed to individual or multiple pads on the microarray (Fig. 2A). The 5â&#x20AC;&#x2122; end of these capture probes is biotinylated so that they will be anchored to the permeation layer via biotin/streptavidin interaction. Subsequently, the target DNA, in the form of amplified product such as PCR product or genomic DNA, is electronically concentrated and hybridized to the pads containing appropriate capture probes (Fig. 2B). Because electronic conditions can be readily fine-tuned on our microchip, we can achieve hybridization under stringent conditions within seconds. Furthermore, we can multiplex readily since different electronic conditions can be administered on different pads based on the optimal hybridization conditions for individual genes. After target hybridization, a fluorescence-labeled reporter oligo is then annealed to the single stranded target DNA outside the capture region to detect the presence of the target. The reporter oligo is long enough so that it will not be pulled apart from the target during subsequent stringency steps. Electronic stringency can be applied to these sites to quickly remove the nonspecifically bound target (Fig. 2C). Importantly, a higher stringency current can then be applied to achieve discrimination between matched and mismatched alleles. Such stringency may be applied serially to obtain optimal discrimination for each test site. Alternatively, the stringency can be achieved thermally. This allows more rapid parallel processing when multiple sites have similar characteristics. A fluorescence scan is then conducted to measure the ratio between the match and the mismatch in order to make a call. Alternatively, fluorescence scans can be integrated into the stringency steps to monitor the discrimination process. Using a reporter oligo has certain advantages over labeling the target DNA directly. It decreases the background without sacrificing sensitivity. So far, we have developed assays for a wide range of SNPs (Gilles et al, 1999; Canter et al, Submitted) and STRs

VI. Comparison between capture down and amplicon down formats Although the capture down format for SNP and STR assays is convenient for electronic hybridization, it has its limitations (Fig.3). The two strands of target DNA are kept apart by our histidine buffer in the absence of electronic activation. However, during electronic activation, histidine is protonated and generates a hybridization zone above the pads. This allows not only hybridization between capture probe and complementing target strand, but also the reannealing between the two target strands. This could result in a decrease in capture probe hybridization efficiency. 186

Gene Therapy and Molecular Biology Vol 4, page 187 Secondly, anchored capture probes are much shorter than target DNA. This may create a steric hindrance problem when short captures near the surface are forced to anneal with a longer DNA that is present in very small quantity. Further, in comparison with transport or hybridization alone, this step is more technically demanding because transport of very small amounts of long DNA strands is coupled simultaneously with hybridization to capture probes,. To get around this problem, we have developed an amplicon down scheme (Canter et al, Submitted). In the amplicon down format, the target DNA (usually amplicon) is anchored to the permeation layer. This is achieved by using a biotinylated PCR primer to synthesize one target strand. After heat denaturation, the amplicons are then electronically addressed in histidine buffer to the desired pads. Since there is no hybridization involved, we utilize a condition that affords maximum transport. At this point, most of the bound DNA should be biotinylated single stranded target DNA. A brief alkaline denaturation step is then performed to remove any residual annealed strand. The

two oligo components which comprise the sandwich in the amplicon down format do not contain biotin groups (Fig. 3). They hybridize to the anchored target and form a basestacked complex (Canter et al, Submitted; Radtkey et al, Submitted). One contains the fluorophore and is the reporter oligo. The advantages of the amplicon down format are multifold. First, it is easier to transport very small amounts of amplicon DNA to test sites without having to accommodate hybridization. Second, the additional alkaline denaturation step after putting down target DNA on the permeation layer ensures that the bound DNA is single stranded, thereby enhancing the assay efficiency. Third, there is less potential steric hindrance during hybridization. Since the complementary region to the capture is in the middle of the target molecule, and the anchor is at the 5' end, it gives the capture oligo plenty of room to interact with the target. Most importantly, the capture oligos, which are shorter and in excess, are the free diffusing component in this hybridization reaction. This is a kinetically favored reaction simply due to mass action.

Figure 3: Comparison of two different formats for SNP assays. In the capture down format, capture probes are loaded on the pads first. Then the denatured amplicon molecules are electronically hybridized to the captured probes on site. A mixture of reporter oligos that encompass the polymorphic region is passively hybridized to the target. The two different reporter oligos are labeled with different fluorophore dyes and represent either wild type or SNP alleles of the gene. Finally, thermal denaturation is applied to the array. As a result, the reporter which contains the mismatched allele in the polymorphic region is removed and that of the matched allele stays on. The signal ratio between two colors allows us to make a call. The amplicon down format is similar to the capture down format with several exceptions. First, biotinylated amplicon DNA is initially addressed to the site. Second, oligo probes are hybridized to the anchored amplicon.

187

Feng and Nerenberg: Electronic Microarray for DNA Analysis

Figure 4: The Nanogen microchip cartridge. The cartridge contains a 100-site microchip mounted on a substrate that has contact areas for electrical pogo pins (panel on the right). A flow cell sits on top of the array and serves as a chamber for electronics, washing and optical window (panel on the left). The cartridge also contains multiple valves and interconnected channels leading to the flow cell for sample application and washing.

VII. Amplification on the test site

VIII. STR typing in forensics and the CODIS database

Although the amplicon down format is not as easily multiplexed as the capture down format, its biggest advantage is to provide a platform for a natural extension of our technology. Chip based nucleic acid tests, especially ones for medical diagnostics, have multiple requirements. These include sample preparation, target amplification, hybridization and detection. In addition to speed and miniaturization, the microelectronics provides a platform for integration of some of these elements. The cartridge used in our semi-automated instruments contains a microelectronic chip mounted on a substrate and a flow cell on top of the array (Fig.4 ). The flow cell can serve as a suitable chamber for aspects of sample prep, for example, providing fluidics steps such as washing the array in between steps. It can also serve as an compartment for on site amplification of the target DNA (Westin et al, Submitted). This could include linear or exponential amplification processes. On site amplification could provide badly needed simplicity and streamlined integration of technologies currently available separately on the market. The on site amplification on our microchip could also potentially alleviate the difficulties of multiplexing in existing amplification schemes.

Short Tandem Repeat sequences are dispersed widely throughout the human genome. They are highly polymorphic markers and the number of copies of the repeat sequence differentiates alleles of these loci (Edwards et al, 1991). STR DNA typing is well established in forensics, paternity testing, cancer diagnostics and plant breading. In the area of forensics, a large DNA database, the Combined DNA Information System (CODIS) has been implemented by Department of Justice to collect DNA genotypes of convicted felons (Budowle et al, 1998; Budowle et al, In Press). Small amounts of biological material collected at a crime scene can be typed, and results run against the CODIS database to determine if a match is found. STRs are not easily analyzed by conventional hybridization technology. The identical nature of repeats in STR loci makes accurate discrimination fastidious due to slippage during hybridization. Thus, current typing technologies are mainly based on sizing of amplicons which allow for estimation of the numbers of repeats (Budowle et al, 1998). Such analysis is quite labor intensive and does not meet the needs of high throughput and multiplex

188

Gene Therapy and Molecular Biology Vol 4, page 189 analysis. Furthermore, none of these technologies is streamlined enough to facilitate a portable device used at a crime scene. Nanogen’s microarray technology allows fast and accurate analysis due to its electronic hybridization, miniaturization of otherwise cumbersome techniques, and potential for highly multiplexed testing (Radtkey et al, Submitted).

microvariants, stutter bands and difficulties dealing with mixed samples (Budowle et al, 1998). Even if SNP analysis does not replace the STR system, it could paly a complementary role. A perfect example is the Y chromosome polymorphism. In recent years these Y chromosome SNPs have surfaced in a number of human population genetic studies (Jobling et al, 1997). With the large number of rape and sexual assault cases in this country, they can be extremely useful in determining the male identity in sexual assault specimens. Nanogen’s semiautomated microchip device provides an attractive platform for this potential application, particularly in an integrated system where amplification can be performed on the chip followed by genetic assay.

IX. Utility of SNPs in forensics Recently, utilization of single nucleotide polymorphism in forensic genotyping has become more and more attractive as a complement to STRs. This is partly due to the limitations of the STR system such as

Figure 5: Example of a capture down STR assay on a 25-site chip. Capture probes representing various alleles of TPOX and CSF genes are loaded to the array according to the map on the upper left panel of the figure. A labeled T12 oligo is used for control of capture attachment in the permeation layer. It contains a biotin at the 5’ end and a Bodipy Texas Red (BTR) dye at the 3’ end. The CTT (CSF, THO1 and TPOX loci) triplex PCR product from a patient is then hybridized to the array. During thermal denaturation, multiple images are taken at each elevated temperature. The lower right panel represents one of those images. The fluorescence signal of each pad is measured within each frame and plotted against increased temperatures. A genotype is determined based on the fact that the reporter for matched alleles is more stable than the mismatched alleles under thermal denaturation. In this case, the patient is 8repeat homozygote for TPOX locus, and 9 and 10-repeat heterozygote for CSF locus.

189

Feng and Nerenberg: Electronic Microarray for DNA Analysis hemochromatosis. Blood Cells Mol Dis 23, 135-145; discussion 145a-b. Brown, P. O., Botstein, D., (1999) Exploring the new world of the genome with DNA microarrays. Nature Genet 21, 33-37. Budowle, B., Moretti, T. R., Niezgoda, S. J., and Brown, B. L., (1998) CODIS and PCR-based short tandem repeat loci: Law enforcement tools. In: Second European Symposium on Human Identification, Promega Corporation, Madison, Wisconsin. Budowle, B., Moretti, T. R., Baumstark, A. L., Defenbaugh, D. A., and Keys, K. M., Population data on the thirteen CODIS core short tandem repeat loci in African Americans, U.S. Caucasians, Hispanics, Bahamians, Jamaicans, and Trinidadians, J Forensic Sci (In Press). Canter, D., Radtkey, R., Sosnowski, R., Oâ&#x20AC;&#x2122;Connell, J., Nerenberg, M., Use of a base stacked probe improves single nucleotide polymorphism discrimination on a microelectronic chip. Submitted. Chee, M,, Yang, R., Hubbell, E., Berno, A., Huang, X.C., Stern, D., Winkler, J., Lockhart, D.J., Morris, M.S., Fodor, S.P., (1996) Accessing genetic information with high-density DNA arrays. Science 274, 610-614. Edman, C.F., Raymond, D. E., Wu, D. J., Tu, E., Sosnowski, R. G., Butler, W. F., Nerenberg, M., Heller, M. J., (1997) Electric field directed nucleic acid hybridization on microchips. Nucleic Acids Res. 25, 4907-4914. Edwards, A., Civitello, A., Hammond, H. A., and Caskey, C. T., (1991) DNA typing and genetic mapping with trimeric and tetrameric tandem repeats, American Journal of Human Genetics 49, 746-756. Evans, G.A., (1998) The Human Genome Project: applications in the diagnosis and treatment of neurologic disease. Arch Neurol. 55, 1287-1290. Gilles, P. N., Wu, D. J., Foster, C. B., Dillon, P. J., Chanock, S.J., (1999) Single nucleotide polymorphic discrimination by an electronic dot blot assay on semiconductor microchips. Nat Biotechnol. 17, 365-370. Jobling, M.A., Pandya, A., Tyler-Smith, C. (1997) The Y chromosome in forensic analysis and paternity testing. Int J Legal Med 110, 118-124. Koza, M.J., Shah, N., Shen, N., Yang, R., Fucini, R., Merigan, T.C., Richman, D.D., Morris, D., Hubbell, E., Chee, M., Gingeras, T.R., (1996) Extensive polymorphisms observed in HIV-1 clade B protease gene using high-density oligonucleotide arrays. Nat Med. 2, 753-759. Lipshutz, R. J., Fodor, S. P., Gingeras, T. R., Lockhart, D. J., (1999) Review: High density synthetic oligonucleotide arrays. Nat Genet. 21(1 Suppl), 20-24. Mahajan, S. (Editor) (1993) Handbook on semiconductors 3; Moss, T. Ed.; North Holland, Amsterdam Radtkey, R., Feng, L., Muralhidar, M., Canter, D., Duhon, M., DiPierro, D., Fallon, S., Tu, E., McElfresh, K., Nerenberg, M., Sosnowski, R., Rapid Determination of STR Alleles on an Electronically Active DNA Microchip. Submitted.

X. Prospect of a portable device In addition to the CODIS national database, several states, including Virginia and Florida have set up similar databases to help them solve crimes within state. This kind of initiative not only requires high-throughput technologies for databasing, it also needs testing at the crime scene to minimize potential contamination and to include or exclude suspects at the scene. The National Institute of Justice has supported DNA chip technology development with an emphasis on portable devices that will accommodate samples such as blood, semen and skin cells at the crime scene (Schmalzing et al, 1997). We are developing a small integrated device containing the cartridges in which the microchips are packaged. The small flow cell will serve as a reaction chamber for amplification. Afterwards, the cartridges can be inserted into a portable reader for assay and detection. Future development which may facilitate miniaturization of instruments may include a direct (nonfluorescent) detection systems, perhaps sensing changes in resistance of magnetism. Since no colormetric detection would be required, we could eliminate the fluorescence scanner that composes a large part of the instrument.

XI. Conclusions In conclusion, the electronic microchip technology could have wide applications in biomedical research, medical diagnostics, genomics, genetic testing and drug discovery. We have developed assays for multiple SNP applications and STR analysis for forensic genotyping. These, in conjunction with the development of integrated systems and portable devices, provide an alternative to the current array technologies available.

Acknowledgements We thank Drs. Tina Nova, Mike Heller and Gene Tu at Nanogen for their intellectual contribution to the microelectronic technology and Drs. Don Ackley and Paul Swanson for chip designs, Dr. Christian Valcke for cartridges and the entire engineering team for the research instruments. We would like to acknowledge Dr. Ron Sosnowski for developing the STR assay, Dr. David Canter for developing the SNP assay, and Drs. Ray Radtkey, Lori Westin, Carl Edman, Ling Wang, Geoff Landis and Dana Vollmer for their contribution to the development of various assays. We are also grateful to the National Institute of Justice and our corporate and grant partners Becton Dickinson and Company and Bode Technology Group for their financial support and intellectual input.

Schmalzing, D., Koutny, L., Adourian, A., Belgrader, P., Matsudaira, P., Ehrlich, D., (1997) DNA typing in thirty seconds with a microfabricated device. Proc Natl Acad Sci USA. 94, 10273-10278. Sosnowski, R. G., Tu, E., Butler, W. F., O'Connell, J. P., Heller, M. J., (1997) Rapid determination of single base mismatch

References Barton, J. C., Shih, W.W., Sawada-Hirai, R., Acton, R. T., Harmon, L., Rivers, C., Rothenberg, B. E., (1997) Genetic and clinical description of hemochromatosis probands and heterozygotes: evidence that multiple genes linked to the major histocompatibility complex are responsible for

190

Gene Therapy and Molecular Biology Vol 4, page 191 mutations in DNA hybrids by direct electric field control. Proc Natl Acad Sci USA Feb 94, 1119-1123. Vahey, M., Nau, M.E., Barrick, S., Cooley, J.D., Sawyer, R., Sleeker, A.A., Vickerman, P., Bloor, S., Larder, B., Michael, N.L., Wegner, S.A., (1999) Performance of the affymetrix GeneChip HIV PRT 440 platform for antiretroviral drug resistance genotyping of human immunodeficiency virus type 1 clades and viral isolates with length polymorphisms. J Clin Microbiol. 37, 2533-2537. Wang, S.S., Virmani, A., Gazdar, A.F., Minna, J.D., Evans, G.A., (1999) Refined mapping of two regions of loss of heterozygosity on chromosome band 11q23 in lung cancer. Genes Chromosomes Cancer. 25, 154-159. Wang, S.S., Esplin, E.D., Li, J.L., Huang, L., Gazdar, A., Minna, J., Evans, G.A., (1998) Alterations of the PPP2R1B gene in human lung and colon cancer. Science 282, 284-287. Westin, L., Xu, X., Miller, C., Wang, L., Edman, C. F., Nerenberg, M., Anchored multiplex amplification on a microelectronic chip array. Submitted.

Michael Nerenberg

Lana Feng

191

Gene Therapy and Molecular Biology Vol 4, page 193 Gene Ther Mol Biol Vol 4, 193-202. December 1999.

Rapid generation of recombinant herpes simplex virus vectors expressing the bacterial lacZ gene under the control of neuronal promoters Research Article

Alistair McGregor1, Alexis Roberts2, R. Wayne Davies2, J. Barklie Clements1 and 1 Alasdair R. MacLean 1 2

Division of Virology and Division of Molecular Genetics , IBLS, University of Glasgow, Church Street, Glasgow, G11 5JR. __________________________________________________________________________________________________ Correspondence: Alasdair R. MacLean, Division of Virology, IBLS, Institute of Virology, University of Glasgow, Church Street, Glasgow, G11 5JR, U.K. tel: 0141-330-4023, fax: 0141-337-2236, email: a.maclean@vir.gla.ac.uk Key words: HSV-1, lacZ, neuronal promoter, GABAA receptor, recombination A b b r e v i a t i o n s : H S V - 1 , herpes simplex virus type 1; NSE, neurone specific enolase; PNS, peripheral nervous system; CNS, central nervous system; HCMV, human cytomegalovirus; GABA, !-amino butyric acid; LAT, latency associated; Received: 8 June 1999; accepted: 28 June 1999

Summary We describe the development of herpes simplex virus type 1 (HSV-1) RL1 negative mutants as vectors expressing the l a c Z reporter gene under the control of neuronal specific promoters and their expression in neuronal cell cultures. Two neuronal promoters were used in this study 1) the rat neurone specific enolase (NSE) promoter, which is expressed in all neurones and 2) the mouse GABA A receptor delta ( ") subunit, which i s expressed within specific regions o f the brain (cerebellum, hippocampus and thalmus). The l a c Z gene was also placed under the control of two viral promoters: 1) the HCMV IE promoter, a 'universally' strong promoter, and 2) the HSV-1 gD promoter, a strong delayed early lytic promoter. Initial expression experiments with recombinant viruses were carried out on neuronal and non-neuronal cell lines and primary mouse hippocampal cells. The promoter/reporter cassettes were introduced into one of two sites in the HSV-1 genome: firstly a non essential gene in the unique long region; and secondly at the LAT locus in the long terminal repeats just downstream from the LAT P2. Two high frequency methods of generating recombinant HSV are described. The first step for both involves insertion, into a non-essential genomic region, of a unique restriction enzyme site (Pac I), which is used to digest the HSV DNA into 2 arms. In the first method the insert, flanked by Pac I sites is isolated and i n v i t r o ligated into the digested HSV vector to generate recombinant virus at a frequency of 10-90%. In the second method the DNA to be inserted is flanked by HSV DNA and co-transfected with the digested viral DNA. In vivo recombination across the digested ends of the HSV DNA and through the sequences to be inserted generates recombinants at a frequency of up to 100%. including introns and large promoter elements which might be important in specifying region specific expression in the CNS. Additionally in latently infected neurons of both the peripheral and central nervous systems (PNS and CNS) HSV is stably maintained as an episomal element with transcriptional activity limited to one region of the genome, the latency associated (LAT) region (Glorioso et al, 1995; Stevens, 1989). HSV-1 encodes

I. Introduction Herpes simplex virus type 1 (HSV-1) is a neurotrophic virus that is being increasingly used as a vector for gene delivery to neurones. The virus has a large ds DNA genome (150kbp) that has been fully sequenced and has the capacity to package up to 30kbp of foreign DNA (McGeoch et al, 1988; Glorioso et al, 1995; Longnecker et al, 1988) allowing insertion of entire genes 193

MacLean et al: Rapid generation of recombinant HSV vectors approximately 80 genes of which half are deemed 'nonessential' for replication in tissue culture, but may contribute to viral propagation in animal models (Ward and Roizman, 1994). At least some of these 'non- essential' genes could be deleted to provide additional room for inserted sequences, or to modify neuronal tropism without drastically altering the virus growth in tissue culture. Before use as a neuronal gene delivery vector it is essential that the capacity of the virus to undergo a lytic infection in the CNS is abolished: currently we are developing non pathogenic vectors based on deletion of the virulence factor, ICP34.5 encoded by the RL1 gene (MacLean, A. et al, 1991; Coffin et al, 1996; McKie et al, 1998). Using lac Z as a reporter gene, we are interested in studying specificity and duration of neuronal promoter expression in a HSV-1 background in the central nervous system (CNS) of animal models; the ultimate aim being the generation of potentially useful gene therapy vectors for the treatment of neurological disorders such as Huntingtonâ&#x20AC;&#x2122;s and Alzheimer's disease. Our current studies involve the use of two neuronal promoters: the GABAA receptor delta (d) subunit promoter, which has a specific regionalised expression pattern in the CNS of rodents (cerebellum, hippocampus and thalmus; Laurie et al, 1992); and the neurone specific enolase (NSE) promoter, which is ubiquitously expressed in neurones throughout the CNS (Forss-Peter et al, 1990). In this paper we describe the generation of a number of recombinant viruses and their pattern of expression in both neuronal and nonneuronal tissue culture cells and primary neuronal cells as well as initial expression data in vivo. One major drawback to the use of HSV-1 as a gene delivery vector lies in the labour intensive method of isolating recombinant viruses. The standard method of introducing foreign DNA into HSV-1 is to target insertion into a 'non-essential' gene on the viral genome by homologous recombination (MacLean, C. et al, 1991); firstly by cloning the foreign DNA into a plasmid between flanking HSV-1 sequences of at least 500-1000bp; and secondly by cotransfecting the recombinant plasmid with intact viral DNA onto a cell monolayer to generate progeny virus. The frequency of recombinant viruses within the population varies between 0.1 to 1%, depending upon the efficiency of recombination and the growth characteristics of the mutant virus. If the inserted foreign DNA is a lacZ reporter gene under a ubiquitous promoter, recombinant progeny can easily be detected by blue/white plaque staining in the presence of X-gal in the medium overlay (MacLean, C. et al, 1991). If the inserted foreign DNA cannot be used to select for or detect recombinants (as in our case where the promoter driving is only functional in certain non plaquing cell types) the DNA profile of individual plaque isolates requires to be analysed

to identify recombinant clones, a labour intensive process (MacLean, A. et al, 1991). Although a variety of systems exist which allow the selective enrichment of recombinant progeny virus, each of these systems has individual drawbacks. One method is to either reduce the parental progeny population, and/ or select for the recombinant progeny population. By targeting the DNA insertion to the thymidine kinase (TK) locus of HSV-1 and selecting for progeny virus with a TK negative phenotype both criteria can be met (Efstathiou et al, 1989). However, it is not always desirable to have a TK negative virus since such viruses are resistant to acyclovir (the only clinically effective anti-HSV drug) - an undesirable phenotype for a gene therapy vector - and additionally have altered neurovirulence. This method is also restricted to one site in the genome. Other methods have been developed to maximise the yield of recombinant progeny in the population. The most powerful system is the use of a cosmid system, with cloned overlapping DNA inserts spanning the length of the viral genome (Cunningham and Davison, 1993); this effectively eliminates background parental virus. However, generation of functional overlapping cosmids is a complicated procedure and has the disadvantage that each set is strain/ mutant dependent. Another technique, widely used in the generation of recombinant viruses, is the P1 phage Cre-loxP system. Here foreign DNA is introduced into the viral genome by cloning the foreign DNA into a loxP bearing plasmid to allow targeted recombination to occur in the presence of the Cre protein between the plasmid and an existing loxP site previously introduced into the parental HSV-1 strain (Gage et al, 1992). Although this procedure is highly efficient at generating recombinant progeny virus, it suffers the drawbacks of firstly having to construct an initial virus mutant with a loxP site and secondly, and, potentially more seriously, the inserted DNA will be flanked by loxP sites, which may affect reporter gene expression. By taking a genome whose normal XbaI sites had been deleted (MacLean and Brown, 1987), Rixon and McLauchlan (1990) developed a procedure for introducing foreign DNA into a unique XbaI site which had been introduced back into this genome: this was carried out by in vitro ligation of foreign DNA with flanking XbaI sites and subsequent transfection of the ligation mixture into cells to generate virus progeny. Huang et al. (1994) improved on this procedure by introducing two unique restriction sites (PacI and SwaI) into the LAT locus of the HSV-1 strain HFEM genome (which contains only a single LAT locus) such that the insert is ligated unidirectionally and re-ligation of viral genome DNA without an insert is significantly impaired because the genomic ends are non-compatible thereby lowering 194

Gene Therapy and Molecular Biology Vol 4, page 195 parental virus background. However, both ligation systems are limited in their ability to be used at any location in the genome and since a lot of work was involved in generating these particular HSV-1 viruses with unique restriction sites these methodologies lack flexiblity because they are again strain/ mutant dependent. We have extended this latter method to develop two alternative procedures for the creation and isolation of HSV-1 recombinants which cannot easily be detected.

HSV-1 strain 17 non-virulent mutants, 1716 and 1764. Both mutants are deleted in the RL1 virulence gene (MacLean, A. et al, 1991), while 1764 carries an additional mutation in the Vmw65 immediate early transactivating gene (Ace et al, 1989; Coffin et al, 1996). A polylinker containing PacI sites was introduced into UL43 at an unique NsiI site at n.p. 94711 in the plasmid p35 containing HSV sequences from a BamHI (n.p. 91610) to EcoRI (n.p. 96751) site (MacLean, C. et al, 1991; Figure 1). A HCMVIE promoter/lacZ reporter cassette flanked by PacI restriction sites was inserted into the PacI sites in p35(PacI) and the recombinant plasmid co-transfected separately with HSV-1 1764 and 1716 DNA onto BHK21/C13 cells and recombinant viruses (1780 and 1780R respectively) isolated under X-gal selection. Three subsequent rounds of plaque purification were carried out to ensure the purity of 1780 and 1780R. The structure of both recombinant viruses was confirmed by restriction enzyme analysis and Southern blotting with p35 (data not shown).

II. Results A. Generation of recombinant viruses carrying PacI restriction enzyme sites HSV-1 does not contain a PacI restriction enzyme site and such a site can be introduced at any non-essential site to enable insertion to be specifically targeted to that location on the HSV-1 genome by one of two methods, both relying on digestion at the unique PacI site. Initial studies have targeted insertion to the non-essential UL43 gene (MacLean, C. et al, 1991) in the long unique region of the viral genome. Insertions were carried out into two

Figure 1. a) A diagram of the HSV-1 genome with the region spanning UL43 expanded. The n.p. of the BamHI and EcoRI sites used to construct p35 are marked MacLean, C. et al., 1991). b ) An expansion of the HSV-1 sequence cloned in p35 with the n.p. of the flanking BamHI and EcoRI sites and the NsiI site used for insertion marked. c) The sequence and sites present in the polylinker inserted into the NsiI site to generate p35(Pac). All expression cassettes were inserted into UL43, via the introduced PacI site, at the position of the unique NsiI site.

195

MacLean et al: Rapid generation of recombinant HSV vectors

Figure 1d) The lacZ expression cassettes inserted into the UL43 locus.

The latter procedure is achieved by firstly cloning the foreign DNA into a vector with flanking HSV sequences homologous to the region surrounding the PacI site in the HSV-1 viral genome (Figures 1 and 2A). The recombinant HSV plasmid and PacI digested HSV-1 viral DNA are co-transfected onto BHK21/C13 cells by calcium phosphate precipitation (Stow and Wilkie, 1976) allowing recombination to take place in vivo to generate progeny virus. The PacI recombination protocol, although lower in efficency at generating virus compared to the in vitro ligation strategy, produces a high proportion of recombinant progeny (up to 100%) that have unidirectional insertions, thus enabling rapid purification of HSV-1 recombinants: this method should be especially useful for purifying recombinant viruses with a selective growth disadvantage. A summary diagram illustrating both the PacI ligation and recombination strategies is illustrated in Figures 2A and B.

DNA prepared from 1780 and 1780R was digested with PacI and viral DNA re-ligated to remove the lacZ cassette and produce a virus with a single PacI site in the UL43 locus (mutants 1784 and 1784R respectively; Figure 2). 1784 and 1784R were isolated under X-gal staining by isolating white plaques, which were subjected to three subsequent rounds of plaque purification. The structure of 1784 and 1784R was confirmed by BamHI restriction enzyme analysis and Southern blotting with p35 (data not shown). The presence of two PacI sites flanking the HCMVIElacZ cassette gives a significantly lower parental virus background, and thus a higher percentage of recombinants, after PacI digestion of viral DNA: for this reason 1780/1780R were the parental strains used in most PacI viral selection system experiments. A detailed methodology for each of the PacI systems is given below. As a prerequisite for both methods, to ensure a minimum level of parental virus background, it is essential to confirm complete digestion of the HSV DNA with PacI by checking the transfectibility of the digested compared to undigested DNA. Following digestion there should be a knockdown of at least 1000-fold in plaque numbers.

B. PacI ligation protocol The PacI ligation protocol was carried out as follows. Firstly, viral DNA was digested overnight with an excess of PacI enzyme, the DNA phenol/ chloroform extracted, ethanol precipitated, washed with 70% ethanol, air dried and carefully resuspended at a concentration of 0.1ug/ul in water overnight. Secondly, a plasmid containing a promoter/ reporter cassette flanked by PacI sites was digested with PacI and gel purified. Two ug of viral DNA was ligated with up to 5ug of purified PacI fragment DNA in a total volume of 40ul with 8 units of T4 DNA ligase in ligation buffer and 1mM ATP and incubated overnight at 16oC.

Introduction of foreign DNA at the inserted unique PacI site was either by 1) a modification of the ligation/ insertion strategy developed by Rixon and McLauchlan (1990); or alternatively 2) a homologous recombination virus rescue protocol using a parental virus digested into 2 arms at the site of insertion by PacI. Although novel to HSV-1, this method is similar to one employed in the generation of recombinant baculovirus (Kitts et al, 1993). 196

Gene Therapy and Molecular Biology Vol 4, page 197

Figure 2A. A diagrammatic illustration of the PacI in vivo recombination selection system used to generate recombinant virus at the UL43 locus. a) 1780 with a CMVIE/lacZ insert flanked by 2 PacI sites in UL43. b) 1784 with the CMVIE/lacZ insert removed and containing an unique PacI site in UL43. Digestion of 1784 with PacI separates the genome into two arms. c) Cotransfection of the two arms with a HSV plasmid spanning UL43 and containing an insert allows in vivo recombination to generate recombinant virus at high frequency. The example given is NSElacZ .

ug of PacI digested viral DNA and 2-10 ug XmnI linearised plasmid DNA were mixed together and co-transfected onto BHK21/C13 cells using the standard calcium phosphate/ DMSO boost protocol (Stow and Wilkie, 1976). Infected monolayers were harvested 7-8 days post transfection regardless of whether complete cpe had been obtained. Recombinant virus was isolated by a similar procedure used in the PacI ligation protocol.

The ligation reaction was transfected by the calcium phosphate precipitation/ DMSO boost protocol (Stow and Wilkie, 1976). Once complete cpe was established on the transfected monolayer (approximately 3-4 days), the plates were harvested and recombinant virus isolated by the standard procedures of DNA analysis of individual plaque isolates or limiting dilution (Rixon and McLaughlan, 1990).

Using these two procedures, a number of recombinant 1716 and 1764 based viruses in UL43 was isolated (F i g u r e 3 ; T a b l e 1 ). By the ligation method: NSE lacZ and GABA 1.6kbp lacZ; and by the recombination method, GABA 1.6kbp lacZ and 4.5kbp lacZ.. The percentage of recombinant virus generated by the PacI ligation sytem varied between 10 to 90% of the total viral progeny and recombinant virus generated via the PacI recombination protocol constituted up to 100% of the total viral progeny.

C. PacI Recombination protocol The PacI recombination protocol was carried out as follows. Firstly, viral DNA was digested with PacI, purified and resuspended at 0.1ug/ml as described above. Five - ten ug of recombinant p35(PacI) plasmid, containing a promoter/ reporter cassette, was linearised by digestion with XmnI, which cuts once in the ampicillin resistance gene of the plasmid backbone. The plasmid DNA was phenol/ chloroform extracted, ethanol precipitated, washed with 70% ethanol, air dried and resuspended at a concentration of 0.5ug/ul in water. Five

Equivalent viruses in the LAT locus were isolated by standard cotransfection (Figure 4; Table 1).

197

MacLean et al: Rapid generation of recombinant HSV vectors

F i g u r e 2 B . A diagrammatic illustration of the PacI in vitro ligation selection system used to generate recombinant virus at the UL43 locus. a) 1780 with a CMVIE/lacZ insert flanked by 2 PacI sites in UL43. b) 1784 with the CMVIE/lacZ insert removed and containing an unique PacI site in UL43. Digestion of 1784 with PacI separates the genome into arms. c) Ligation of the two arms in vitro in the presence of an excess of an insert with PacI ends generates a mixture of recombinant virus and 1784. The example given is 4.5kbp GABAA d lacZ.

Figure 3 Diagram of the GABAA d promoter and the fragments, with their sizes, used to generate lacZ expression cassettes. The initiating ATG is also marked. Restriction enzyme sites used in the cloning are marked.

198

Gene Therapy and Molecular Biology Vol 4, page 199

Figure 4 A diagram of the HSV-1 genome with IRL expanded showing the location of the identified LAT transcripts and promoters and other genes (ICP0, ICP34.5, ICP4) encoded in the repeats. Expression cassettes were inserted downstream of LAP2 at the 5' end of the 2kb LAT. Below the line are illustrated the three lacZ expression cassettes inserted into the LAT locus (CMVIE, NSE, 1.6 GABAd). An equivalent arrangement is found in TRL.

standard calcium phosphate/ DMSO boost protocol (Stow and Wilkie, 1976).

UL43 locus

LAT locus

4.5 GABA 1.6 GABA

1.6 GABA

NSE

HSVgD HCMVIE

HCMVIE

E. Expression viruses

lacZ

recombinant

1. In vitro expression Initially lacZ expression from the recombinant viruses was assayed in a range of neuronal (ND7 and PC12) and non-neuronal (BHK21/C13) tissue culture and primary hippocampal cells as outlined in Table 2. As no significant differences in expression levels were observed between either 1716 or 1764 or either genomic locus, only 1 set of results is given. Varying multiplicities of infection (0.01-10 pfu/cell) were used and samples were fixed 48 h post infection: the numbers of positive cells increased with multiplicity, but expression levels were independent of multiplicity. Both the HCMVIE and gD promoters expressed strongly in the neuronal and nonneuronal cell lines, but slightly less in the primary neuronal cells. NSE expressed weakly in both the neuronal cell lines and primary hippocampal cells, whereas expression from the GABA d 1.6kb promoter was only faintly detected in the neuronal cell lines. No expression was detected from the GABA d 4.5kb delta promoter in any cell type used.

T a b l e 1 . 1716 and 1764 recombinant viruses.

D. Construction expression cassettes

GABAd/lacZ

A 10kbp upstream fragment of the GABAd promoter was used as the starting plasmid and a range of fragments deleted from the 5' end constructed to generate 5 lacZ expression cassettes with promoter fragments ranging in size from 1.6-10.1 kbp (Figure 5). All five were inserted into the SpeI site of p35(Pac) and the 1.6 and 4.5 kbp promoter fragment constructs used to generate recombinant viruses. In addition the 1.6kbp NSE promoter was cloned in front of the lacZ gene, inserted into p35(Pac) and used to generate recombinant virus. The 1.6 kbp GABAd and NSE promoter/lacZ constructs were also inserted into a plasmid carrying RL sequences just downstream of LAP2 and recombined into virus by the 199

MacLean et al: Rapid generation of recombinant HSV vectors

F i g u r e 5 . A restriction map of the GABAA " upstream region (5' UTR), showing the sizes of fragments used to generate lacZ expression cassettes.

Cell line/ Promoter

HCMV IE gD

NSE

BHK21/C13 PC12 ND7 hippocampal cells hippocampal neurones

+++ ++ ++ + +

+ + + +

+++ ++ ++ + +

GABA 1.6kbp + + -

GABA 4.5kbp -

Table 2. Tissue culture expression of re c o m bi na nt v ir u s e s. The number of + refers to strength of blue staining in individual cells. Primary hippocampal cells were established from rat embryos with neurone enriched cultures being derived by treatment with ara C to kill the dividing cells (Kennedy et al., 1995).

based viruses, possibly due to the higher particle numbers used to give an equivalent multiplicity (Ace et al, 1989; Coffin et al, 1996). Expression peaked at 7 days post inoculation, was declining rapidly by day 21 post inoculation and was undetectable 35 days post inoculation. To detect expression in the CNS, 3 week old female Balb/C mice were infected with 105 pfu/mouse as described by MacLean, A. et al. (1991). LacZ expression could only be detected with the gD based constructs for the first 7 days post inoculation, presumably due to expression during the abortive lytic infection observed with ICP34.5 negative viruses.

III. Discussion One of the practical difficulties in using HSV as a gene therapy vector is the inability to rapidly isolate recombinants. Standard methods generate a low frequency of recombinants and involve time consuming analysis of a large number of individual plaque isolates (MacLean, A. et al, 1991). Both methods described in this paper generate recombinants at significantly increased frequency, thus reducing the number of isolates which have to be screened. The recombination method, while requiring subcloning of the sequence to be inserted into a HSV flanking plasmid generates recombinants at nearly 100% in the desired orientation. The ligation method has the advantage that

2. In vivo expression All viruses were assayed for expression in the PNS and CNS of Balb/C mice. To detect expression in the PNS, 4 week old Balb/C mice were infected with 107 pfu/mouse in the footpad to allow acute replication/ establishment of latency in the dorsal root ganglia (DRG) of the PNS as previously described (Robertson et al, 1992). In the PNS lacZ expression could only be detected from the gD and HCMVIE promoters. As previously observed a larger number of lacZ positive neurones were detected with 1764 200

Gene Therapy and Molecular Biology Vol 4, page 201 the sequences to be inserted do not require to be first subcloned into a plasmid with flanking HSV sequences, but produces a lower frequency of recombinants (5-90%) in random orientations. The isolation of a parental virus with PacI sites surrounding lacZ under the control of a strong ubiquitous promoter, HCMVIE, is non labour intensive as recombinant viruses can be readily detected under X-gal staining. Construction of one plasmid and recombinant viruses allows insertion into the same site of all mutants of the one HSV-1 strain, as demonstrated by insertion into 1716 and 1764 in this paper.

expression; increasing the sensitivity of our assays; and determination of whether lacZ expression obtained in the CNS under GABAA d promoter control can be region specific.

Acknowledgements Some of this work was first presented at the International Congress of Virology, Israel, August, 1996. We thank Jan Gairns and Prof Peter Kennedy, Dept. of Neurology, University of Glasgow for the supply of primary hippocampal cells. This work was supported by the Medical Research Council, U.K.

Using these techniques a range of recombinant viruses have been constructed and are described inFigure 4. One of the major problems of using HSV as a gene delivery vehicle is obtaining long term expression from the latent genome, preferably in a cell type dependant manner. Towards this aim we are particularly interested in studying neuronal promoter expression in a HSV-1 background in the CNS/PNS of animal models. Our current studies involve the use of two neuronal promoters: the GABAA receptor d subunit promoter, which has a specific regionalised expression pattern (cerebellum, hippocampus and thalmus) in the CNS of rodents; and the neurone specific enolase (NSE) promoter, which is ubiquitously expressed in neurones throughout the CNS. Both the level and specificity of expression from these promoters in a HSV-1 vector is being analysed using betagalactosidase as a reporter gene. As controls we are using 2 ubiquitous strong promoters: the HCMV IE and HSV gD promoters. Our inability to detect expression in vivo from either the NSE or GABAA d constructs suggests any expression is low level and below the limits of detection by Xgal staining and we may require to detect expression by the more sensitive method of RT-PCR, or use of a more sensitive reporter gene such as green fluorescent protein (GFP). This hypothesis is supported by the low level of expression in tissue culture (Table 2).

References Ace AI, McKee TA, Ryan JM, Cameron JM, and Preston CM (1 9 8 9 ) Construction and characterization of a herpes simplex virus type 1 mutant unable to transinduce immediate-early gene expression. J V i r o l 63 , 22602269. Coffin RS, MacLean AR, Latchman DS and Brown SM (1 9 9 6 ) Safe and efficient trans gene expression in the PNS and CNS of mice using ICP34.5 deletion mutants. Gene Ther 3, 886-891. Cunningham C and Davison A (1 9 9 3 ) A cosmid based system for constructing mutants of herpes simplex type 1. V i r o l o g y 197, 116-124. Efstathiou S, Kemp S, Darby G, and Minson AC (1 9 8 9 ) The role of herpes simplex virus type 1 thymidine kinase in pathogenesis. J Gen Virol 70 , 869-879. Forss-Petter S, Danielson PE, Catsicas S, Battenburg E, Price J, Nerenberg M and Sutcliffe JG (1 9 9 0 ) Transgenic mice expressing beta-galactosidase in mature neurons under neron-specific enolase promoter control. Neuron 5, 187197. Glorioso JC, DeLuca NA and Fink CA (1 9 9 5 ) Development and application of herpes simplex vectors for human gene therapy. A n n u R e v M i c r o b i o l 49, 675-710. Gage PJ, Sauer B, Levine M, and Glorioso JC (1 9 9 2 ) A cellfree recombination system for site-specific integration of multigenic shuttle plasmids into the HSV-1 genome. J V i r o l 66 , 5509-5515. Huang QS, Deshmane SL and Fraser NW (1 9 9 4 ) An in vitro ligation and transfection system for inserting DNA sequences into the latency associated transcripts (LATs) gene of HSV-1. Gene Ther 1, 300-306. Kennedy PGE, Gairns J, Kew JNC, Sofroniew MV (1 9 9 5 ) Attempted modulation of herpes simplex virus (HSV) infection of neurons in culture by fibroblast growth factor. J Neurovirol 1,399-404. Kitts PA and Possee RD (1 9 9 3 ) A method for producing recombinant baculovirus expression vectors at high frequency. B i o t e c h n i q u e s 14 , 810-817. Laurie DJ, Seeburg PH, and Wisden, W (1 9 9 2 ) The distribution of 13 GABAA receptor subunit mRNAs in the

In the case of the GABA A d 1.6kbp promoter both transcriptional and translational fusion cassettes were generated in an effort to maximise reporter gene expression: no difference was detected. However, to obtain maximal/ regional specific levels of expression of a reporter gene may require the presence of further upstream sequences (6.5 to 10kbp 5â&#x20AC;&#x2122; UTR): we have constructed such lac Z expressing cassettes and are currently generating recombinant viruses using the PacI recombination system, as we anticipate viruses contaning such large inserts 1014kbp will be at a significant growth disadvantage. Recently, it has been shown with the rat tyrosine hydroxylase promoter a 9kbp flanking sequence is required for regional specific expression of a lacZ reporter gene in transgenic mice (Min et al, 1994). Future work will be concentrated on constructing recombinant viruses using the full length 10kbp GABAA d fragment to drive lacZ 201

MacLean et al: Rapid generation of recombinant HSV vectors rat brain. II. Olfactory Bulb and Cerebellum. J N e u r o s c i e n c e 12 , 1063-1076. Longnecker R, Roizman B and Meigner B (1 9 8 8 ) Herpes simplex viruses as vectors: properties of a prototype vaccine strain suitable for use as a vector in “Viral V e c t o r s ”, pp68. Edited by Gulzman, Y. and Hughes, S.H. Published by Cold Spring Laboratory Press, NY. MacLean AR and Brown SM (1 9 8 7 ) Generation of a herpes simplex virus type 1 (HSV1) variant devoid of Xba1 sites. J Gen Virol 68 , 1165-1171. MacLean AR, Ul-Fareed M, Robertson L, Harland J and Brown SM (1 9 9 1 ) Herpes simplex virus type 1 deletion variants 1714 and 1716 pinpoint neurovirulence-related sequences in Glasgow strain 17+ between immediate early gene 1 and the ‘a’ sequence. J Gen Virol 72 , 631-639. MacLean CA, Efstathiou S. Elliott ML, Jamieson FE, McGeoch DJ (1 9 9 1 ) Investigation of herpes simplex virus type 1 genes encoding multiply inserted membrane proteins. J Gen Virol 72 , 897-906. McGeoch DJ, Dalrymple MA, Davison AJ, Dolan A, Frame MC, McNab D, Perry LJ, Scott JE, Taylor P (1 9 8 8 ) The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J G e n V i r o l 69 , 1531-74. McKie EA, Brown SM, MacLean AR and Graham DI (1 9 9 8 ) Histopathological responses in the CNS following inoculation with a non-neurovirulent mutant (1716) of herpes simplex virus type 1 (HSV-1): relevance for gene and cancer therapy. N e u r o p a t h A p p l N e u r o b i o l 24 , 367-372. Min N, Joh TH, Kim KS, Peng C, and Son JH (1 9 9 4 ) 5’ upstream DNA sequence of the rat tyrosine hydroxylase directs high-level and tissue specific expression to catecholaminergic neurons in the central nervous system of transgenic mice. M o l B r a i n R e s 27 , 281-289. Rixon FJ and McLauchlan J (1 9 9 0 ) Insertion of DNA sequences at a unique restriction enzyme site engineered for vector purposes into the genome of HSV-1. J G e n V i r o l 71 , 2931-2939. Rixon FJ and McLauchlan J (1 9 9 3 ) Chapter 10: 'Herpes simplex virus vectors' in “Molecular V i r o l o g y A Practical Approach”, pp 285-307. Edited by Davison, A.J. and Ellioitt, R.M. Published by Oxford University Press. Robertson L, MacLean AR and Brown SM (1 9 9 2 ) Peripheral replication and latency kinetics of the HSV1 non-neurovirulent variant 1716. J G e n V i r o l 73 , 967970. Sawtell NW and Thompson RL (1 9 9 2 ) Herpes simples virus type 1 latency associated transcription unit promotes anatomical site dependent establishment and reactivation from latency. J Virol 66 , 2157-2169. Stow ND and Wilkie NM (1 9 7 6 ) An improved technique for obtaining enhanced infectivity of herpes simplex virus type 1 DNA. J Gen Virol 33 , 447-458. Stevens JG (1 9 8 9 ) Human herpesviruses: A consideration of the latent state. M i c r o b i o l R e v 53 , 318-332.

Ward PL and Roizman B (1 9 9 4 ) Herpes simplex genes: the blueprint of a successful human pathogen. Trends Genet 10 , 267-274.

202

Gene Therapy and Molecular Biology Vol 4, page 203 Gene Ther Mol Biol Vol 4, 203-208. December 1999.

Recent progress in gene therapy for eye diseases Review Article

Yasushi Ikuno, Andrius Kazlauskas Schepens Eye Research Institute Department of Ophthalmology, Harvard Medical School, 20 Staniford St., Boston MA 02114. __________________________________________________________________________________________________ Correspondence: Andrius Kazlauskas, Tel: (617) 912-2517; Fax (617) 912-0111; E-mail: kazlauskas@vision.eri.harvard.edu Abbreviations: ARMD, age-related macular degeneration; C.F.U., Colony forming unit; ERG, retino-electrogram; PDE, phosphodiesterase; RPE, retinal pigment epithelial; PDR, proliferative diabetic retinopathy; PDGF, platelet-derived growth factor; PVR, proliferative vitreoretinopathy; RCV, replication competent virus; RP, retinitis pigmentosa; VEGF, vascular endothelial growth factor Key Words: Eye diseases, gene therapy, retinitis pigmentosa, proliferative vitreoretinopathy, adenovirus vectors, retroviral vectors, lentivirus Received: 10 October 1999; accepted 20 October 1999

Summary Gene therapy for eye diseases is still developing and requires further effort to bring it to clinical trials. Both a better understanding of the pathological aspect of the diseases and establishment of better viral systems is required. These efforts will facilitate the development of gene therapy-based treatments, which will complement conventional approaches to manage and potentially cure blinding diseases of the eye.

I. Introduction Despite recent advances in the treatment of ocular diseases, conventional approaches are not yet able to cure many retinal diseases. The retina, being a neuronal tissue, regenerates poorly when damaged, and hence poses additional challenges. The retinal diseases which lack satisfactory treatments or cures include retinitis pigmentosa (RP), proliferative vitreoretinopathy (PVR) and diseases arising from unscheduled neovascularization such as proliferative diabetic retinopathy (PDR), and age-related macular degeneration (ARMD). The etiology of a disease is usually the foundation for developing cures and treatments. Fortunately, recent advances in the fields of molecular biology and genetics are providing a better understanding of the pathogenesis of these diseases. For example, vascular endothelial growth factor (VEGF), a potent mitogen for endothelial cells, and enhancer of new vessel growth, plays a cardinal role in most neovascular disease of the retina and choroid including PDR and ARMD. Several kinds of mutations have been identified in the proteins for transduction of the light signal. These proteins include rhodopsin, !-, "-phosphodiesterase (!, " PDE),

peripherin/rds, and are responsible for some forms of RP. The identification of molecular targets within these diseases provides new opportunity for intervention. For instance, these findings make it possible to develop a gene therapy-based treatment for the incurable eye diseases mentioned above. Investigators have begun to develop gene therapeutic treatments by employing the appropriate DNA constructs in animal models of the diseases. In this paper, we will review some of the work done so far regarding gene therapies in the eye with a particular focus on retinal diseases and vector systems. Finally we will consider relevant issues for the realization of gene therapy-based treatment of some of the most problematic retinal diseases.

II. Vector systems for expressing proteins in the eye Because of the unique features of the eye, the following issues should be addressed when considering gene therapy. The relatively small size of the eye limits the injection volume, and hence the gene delivery system must be highly effective. The volume issue becomes especially important when injecting into the subretinal space, which is much smaller than the vitreous cavity and has more 203

Ikuno and Kazlauskas: Gene therapy for eye diseases restrictions. For ARMD and RP the subretinal space is the ideal injection site because the target cells are photoreceptor and the RPE cells. Intravitreal injection is technically easier than subretinal injection, but the target cells are limited to those within the inner layer of the retina. In cases where the retina is detached, intravitreal injections will potentially access cells in the outer retina and RPE cells through the retinal break. In summary, two different types of injections are feasible, and the choice depends on the target cells within the eye, as well as the condition of the eye itself.

III. Adenovirus vector A. Adenovirus has a relatively high titer and infects many cell types, including neuronal cells. Adenovirus is a member of DNA viruses which can enter both post-mitotic and dividing cells. One of the advantages of this system is the high titer that is readily achieved (greater than 1012 C.F.U. (= Colony forming unit) / ml). This increases the likelihood of expressing genes at high levels, and is particularly advantageous for injection into small spaces such as the subretinal compartment mentioned above. The other advantage is that this viral system is suitable for both dividing and postmitotic cells, including neuronal cells. Consequently, the adenoviral approach is appropriate for diseases such as RP, where the target cell type is the postmitotic photoreceptor cells.

B. Adenovirus is easily rejected by the hostâ&#x20AC;&#x2122;s immune system When genes are introduced using adenoviruses, a common consequence is humoral and cellular immunoreaction. A typical observation following an intravitreal adenovirus injection is severe uveitis characterized by infiltration of inflammatory cells into vitreous cavity, retina, and choroid. Retinal thinning is also observed after intravitreal injection of adenovirus, however this change is transient. Retino-electrogram (ERG) examination has shown that the retina functionally recovers within weeks after the injection (Sakamoto et al., 1998). Subretinal injection has been reported to induce a milder immune response than subcutaneous injection. (Bennett et al., 1996) This may be in part due to the immune-privileged status of the subretinal space which would diminish the immune response to the injected virus. As mentioned above, the introduced adenovirus triggers both cellular and humoral responses. The cellular response leads to the appearance of cytotoxic T cells which kill the virally infected cells. Whereas the humoral response is characterized by the production of

antibodies against the adenoviral proteins. These antibodies persist even after the expression of the introduced proteins has ceased, and hence decreases the success of repeat infections. An additional drawback of this system for gene transfer is that most of the human population probably have antibodies to these adenovirus from the previous infection with adenovirus which are the cause of the common cold. Despite the immunological reaction following gene therapy using adenovirus, it is possible to achieve high expression of the introduced protein. Expression tends to be transient, usually restricted to a month especially the retinal tissue (Bennett et al., 1994; Li et al., 1994), and the shutdown in expression is most probably due to the immune response. This hypothesis is supported by the observation that adenovirus-dependent transgene expression lasts much longer in nude mice or immunosuppressive-reagent treated mice (Dai et al., 1995).

C. Adenovirus is not integrated into the host genomic DNA In addition to transient expression of proteins, a second feature of adenovirus-mediated gene transfer is that the introduced DNA is not integrated into the hostâ&#x20AC;&#x2122;s genomic DNA. The introduced DNA is replicated extrachromosomally in the nucleus. Replication-deficient adenovirus can be generated by replacing the E1 region which is a critical part for virus replication with the gene of interest. In this case, recombinant vectors are replicated only in the cells expressing the E1 gene. In summary, the adenoviral gene delivery system enables high level transient expression. This approach may not be appropriate for treating chronic diseases such as RP, PDR or ARMD. Even in PVR, which has a much quicker onset, it is not clear therapeutic genes would be expressed long enough when introduced with an adenovirus.

IV. Adenovirus associated vector (AAV) A. AAV has a high titer and can be integrated into genomic DNA AAV is a simple and non-pathogenic single-stranded DNA virus, and is newer than either adenovirus or retrovirus in the gene therapy area. The advantage of the vector is that it integrates into genomic DNA of the infected cells, which enables permanent expression in the tissue. The other advantage is that AAV has relatively higher titer (greater than 1012 C.F.U./ml), and that both growing and resting cells can be infected. This viral approach has been used to stably express lacZ or green fluorescent protein (GFP) in photoreceptors of mice and rabbits (Ali et al., 1996; Flannery et al., 1997).

204

Gene Therapy and Molecular Biology Vol 4, page 205

B. Production of AAV is technically involved, and accommodates only relatively small DNA inserts The disadvantage of AAV is that preparation of virus is challenging, and requires that the investigator be vigilant to avoid contamination with adenoviruses. The other disadvantage is that AAV cannot accommodate DNA inserts bigger than 4.0 kbp. This characteristic limits the size of the gene which can be expressed by AAV. The advantages include high titer, wide target range, and prolonged expression. As with the other gene transfer systems, the characteristics of the disease and genes that will be expressed should be matched to the features of the approach for gene expression.

V. Retroviral vector A. Retroviral vector is the most commonly used The retrovirus belongs to the family of RNA viruses, which convert genomic RNA into DNA with RNA transcriptase. Of the viral vehicles, the retrovirus vector is the most widely used both in basic research and clinical trials. Retroviruses have been studied extensively, and consequently more is known about this viral system for gene therapy as compared with the other available approaches. Retroviruses are easier to generate and purify than adenovirus or AAV. The retroviral gene transfer system is the most commonly used approach in gene therapy, and many of the past and current clinical trial protocols employ this vector system.

B. Retroviral vector can be integrated into genomic DNA of the target cells An important feature of retroviruses is that they integrate into the genomic DNA of the target cells by passing through the cellular and nuclear membrane. Despite the integration of the introduced gene into the cellâ&#x20AC;&#x2122;s genome, expression of protein is not as long as with the AAV system. This reason is still unclear (Verma and Somia, 1997). A caveat of integrating the gene into the host genome is the potential for insertional mutagenesis.

C. Retroviruses preferentially infect proliferating cells A feature of this system is that retrovirus poorly infect post-mitotic cells such as neuronal cells. This is a potential problem to treat the diseases in which neuronal cells are the target cell type. In fact, some phenotypes of the RP result from point or null mutation of rhodopsin, or peripherin / RDS gene (Dryja et al., 1990; Kajiwara et al., 1994), both of which are specifically expressed in the

photoreceptor cells. So while photoreceptor cells are likely a target for gene therapy to treat RP, and a retroviral vector delivery system is probably not a good choice. Also the viral titer possible with retroviruses is much lower (greater than 108 C.F.U./ml) as compared with adenoviruses or AAV. These unique features of the retrovirus make this delivery system particularly well suited to certain diseases. For instance, since retroviruses selectively infect proliferating cells, administration of virus to an anatomical site that has both resting and dividing cell will result in infection of only the cells that are proliferating. In an individual with a brain tumor, one approach would be to use retroviruses containing a conditional suicide gene such as thymidine kinase. Subsequent administration of gancyclovir will kill only the infected tumor cells. The application of this approach to ocular diseases is discussed below.

D. An ex vivo approach can increase the efficacy of infection To circumvent the relatively low viral titer, an ex vivo method is sometimes employed instead of in vivo infection. This method involves following four steps: (i) remove some of the target cells from the body; (ii) culture them under condition where they proliferate; (iii) infect the target cells in vitro with the retrovirus harboring the gene of interest, expand them; (iv) reintroduce the modified cells into the body. Although a high transfer efficacy can be achieved with this ex vivo method, it can only be applied when the target cells are readily accessible and proliferate in tissue culture. Another problem is that expression of the introduced gene sometimes shuts off after transplantation. While the mechanism involved has not been clearly resolved, one solution for this problem is to use receptors with different / stronger promoters (Verma and Somia, 1997).

VI. Lentivirus A. Lentivirus is the newest approach for gene transfer Lentivirus belongs to the retroviral family, and human immunodeficiency virus (HIV) is the best-known member of this group. The advantages of this vector include: (i) integration into the host genomic DNA; (ii) both dividing and non-dividing cells can be effectively infected; (iii) like AAV, expression of target protein is sustained. Green fluorescent protein (GFP) expression in the retina was still stable even 12 weeks after gene transfer with this vector. (Miyoshi et al., 1997). A relatively minor drawback is that titers of only 108 can be readily achieved for lentivirus. A much more serious disadvantage of this vector is the possibility of generating replication competent virus (RCV). Since this vector system

205

Ikuno and Kazlauskas: Gene therapy for eye diseases is based on the HIV, contamination of these unwanted RCV would lead the host infection of acquired immunodificiency syndrome (AIDS), a deadly disease. The usage of this vector endangers not only the recipient but the medical staff and researchers. Efforts have been made to develop safe lentivirus based on the HIV virus. For instance, the required viral genes have been divided into three independent plasmid in order to minimize recombinations leading to RCV (Naldini et al., 1996). While the possibility of generating RCV has been severely reduced, it is prudent to exercise caution when using this viral system. Furthermore, it is likely that a variety of safety issue will need to be addressed before the clinical use of lentiviruses becomes widespread.

VII. Specific strategies for each ocular diseases A.Retinitis pigmentosa Retinitis pigmentosa (RP) is often an inherited disease characterized by progressive degeneration of photoreceptor cells in the retina. The degeneration leads to apoptotic cell death of the photoreceptor cells (Portera-Cailliau et al., 1994). The degenerated retina is recognized as a pigmented area of the retina, which results from accumulation of RPE cells. Typically, it initially involves the mid-peripheral retina, and then progresses toward the central part of retina, called macula. Since the macula is responsible for central vision, loss of this portion of the retina compromises visual acuity. Even if macular function is preserved, the visual field is sometimes severely impaired. The severity and progression of the disease generally depends on whether both of the patientâ&#x20AC;&#x2122;s alleles are mutated. Homozygous patients usually have an early onset, and have worse prognosis than heterozygous patients. Heterozygous patients have a relatively late onset, and sometimes retain good vision throughout their entire life. These facts suggest that RP includes multiple genotypes. Genetic studies have revealed mis- or null mutations in multiple genes including rhodopsin, peripherin / RDS, or RPE 65 (Dryja, 1997; Morimura et al., 1998). The existence of an animal model for the disease has enabled investigators to test gene therapy as a means of treatment. The rd (rd/rd) mouse harbors a mutation in the beta subunit of cGMP phosphodiesterase gene ("PDE), and this results in progressive retinal degeneration similar to RP. Adenoviruses have been used to successfully transfer normal "PDE into the photoreceptors of these mice, and retinal degeneration is delayed in such animals (Bennett et al., 1996). Thus, this seems to be an appropriate approach to transduce a gene encoding a functional protein into autosomal recessive

patients that suffer from a disease caused by a dysfunctional protein. However, this approach is not suitable for patients with autosomal dominant RP, because the disease arises from an accumulation of mutated protein. The build up of such proteins is believed to be cytotoxic to the photoreceptors, and consequently induces cell death. In this case, therapy requires the elimination of the mutant gene/protein. One way is to use a ribozyme to cleave the mutant mRNA while leaving the wild type RNA. This approach has proven to be effective in slowing retinal degeneration in autosomal dominant retinal degeneration mice (Lewin et al., 1998). A second approach to treat autosomal dominant RP is by expression of a neurotrophic factor which has a protective effect on the photoreceptor cells. So far, ciliary neurotrophic factor (CNTF), and basic fibroblast growth factor (bFGF) have shown to rescue the photoreceptor cell death. (Akimoto et al., 1999; Cayouette and Gravel, 1997) Similar to these, some investigators have shown that brainderived neurotrophic factor (BDNF) prevented ganglion cell death from the optic nerve axotomy (Di Polo et al., 1998). Both the ribozyme and neurotrophic factor approaches hold promise for treating patients with autosomal dominant RP.

B. Proliferative vitreoretinopathy Proliferative vitreoretinopahty (PVR) is the most common reason for failure of retinal reattachment surgery in patients with retinal detachment (RD). It is characterized by pre- and/or subretinal membrane formation and contraction resulting in tractional retinal detachment with or without rhegmatogenous component. Membrane formation is believed to be caused by unscheduled proliferation and collagen synthesis of the RPE cells migrating through the retinal break (Pastor, 1998). Once this occurs, surgery is the only treatment, and only 50% of such patients are cured. In the other patients PVR reoccurs, and often lead to at least a partial loss of vision (Pastor, 1998; Yoshida et al., 1984). The pathogenesis of PVR is still unproven, but growth factors are commonly believed to make an important contribution to disease progression. While many growth factors have been implicated, platelet-derived growth factor (PDGF) is thought to be the strongest candidate for this process. The data to support this idea include: (i) PDGF is a strong mitogen and chemoattractant for RPE cells and the retinal glial cells both of which are present in the membrane (Bressler et al., 1985; Campochiaro and Glaser, 1985), (ii) cultured RPE cell secrete PDGF and express PDGF receptors, thereby establishing a functional autocrine loop (Campochiaro et al., 1994), (iii) the concentration of PDGF is elevated in the vitreous of PVR patients and PDGF can be found in the epiretinal membrane (Cassidy et al., 1998; Robbins et al., 1994) and (iv) in an animal model of PVR,

206

Gene Therapy and Molecular Biology Vol 4, page 207 the ability to respond to PDGF greatly enhances the PVR potential of cells (Andrews et al., 1999). Transforming growth factor-beta (TGF-") and hepatocyte growth factor (HGF) are thought to be other potential inducer of PVR, because both TGF-" and HGF are a mitogen for RPE cells and both are up-regulated in PVR. (Connor et al., 1989; Lashkari et al., 1999) These findings suggest that growth factors contribute to the development of PVR, and identify growth factors and their receptors as a targets for gene therapy-based strategies to prevent PVR. Some investigators have shown that gene therapy can be successfully applied to an animal model of PVR (Sakamoto et al., 1995). They injected fibroblasts with or without a retrovirus harboring herpes simplex virus thymidine kinase (HSVtk) gene which makes the infected cells sensitive for gancyclovir. Gancyclovir was given after the injection, and they found that PVR progression was inhibited by the retrovirally transferred HSVtk gene using this ex vivo approach. Subsequent studies showed that an in vivo approach was also effective for preventing PVR (Kimura et al., 1996).

C. Other disease The lack of effective treatment of neovascular diseases of the choroid and retina requires the development of new approaches, one of which is likely to be gene therapy. Retinal neovascularization occurs in PDR, and in retinopathy of prematurity (ROP), whereas new vessel growth in the choroid is a hallmark of ARMD. In all of these diseases, it is likely that VEGF plays an important role in the pathological angiogenesis. Up-regulation of VEGF due to hypoxia in the retinal glial cells and/or the RPE cells is thought to be the strongest inducer for these new vessels. (Frank, 1997; Kliffen et al., 1997; Lopez et al., 1996; Pe'er et al., 1996). Hence, VEGF, and its receptors are excellent targets for gene therapy-based treatments for these diseases.

References Ali, R. R., Reichel, M. B., Thrasher, A. J., Levinsky, R. J., Kinnon, C., Kanuga, N., Hunt, D. M., and Bhattacharya, S. S. (1996). Gene transfer into the mouse retina mediated by an adeno-associated viral vector. Hum Mol Genet 5, 591-4. Akimoto, M., Miyatake, S., Kogishi, J., Hangai, M., Okazaki, K., Takahashi, J. C., Saiki, M., Iwaki, M., and Honda, Y. (1999). Adenovirally expressed basic fibroblast growth factor rescues photoreceptor cells in RCS rats. Invest Ophthalmol Vis Sci 40, 273-9. Andrews, A., Balciunaite, E., Leong, F. L., Tallquist, M., Soriano, P., Refojo, M., and Kazlauskas, A. (1999). Platelet-derived growth factor plays a key role in

proliferative vitreoretinopathy. Invest Ophthalmol Vis Sci 40, 2683-9. Bennett, J., Wilson, J., Sun, D., Forbes, B., and Maguire, A. (1994). Adenovirus vector-mediated in vivo gene transfer into adult murine retina. Invest Ophthalmol Vis Sci 35, 2535-42. Bennett, J., Pakola, S., Zeng, Y., and Maguire, A. (1996). Humoral response after administration of E1-deleted adenoviruses: immune privilege of the subretinal space. Hum Gene Ther 7, 1763-9. Bennett, J., Tanabe, T., Sun, D., Zeng, Y., Kjeldbye, H., Gouras, P., and Maguire, A. M. (1996). Photoreceptor cell rescue in retinal degeneration (rd) mice by in vivo gene therapy. Nat Med 2, 649-54. Bressler, J. P., Grotendorst, G. R., Levitov, C., and Hjelmeland, L. M. (1985). Chemotaxis of rat brain astrocytes to platelet derived growth factor. Brain Res 344, 249-54. Campochiaro, P. A., and Glaser, B. M. (1985). Platelet-derived growth factor is chemotactic for human retinal pigment epithelial cells. Arch Ophthalmol 103, 576-9. Campochiaro, P. A., Hackett, S. F., Vinores, S. A., Freund, J., Csaky, C., LaRochelle, W., Henderer, J., Johnson, M., Rodriguez, I. R., Friedman, Z., and et al. (1994). Plateletderived growth factor is an autocrine growth stimulator in retinal pigmented epithelial cells. J Cell Sci 107, 2459-69. Cassidy, L., Barry, P., Shaw, C., Duffy, J., and Kennedy, S. (1998). Platelet derived growth factor and fibroblast growth factor basic levels in the vitreous of patients with vitreoretinal disorders . Br J Ophthalmol 82, 181-5. Cayouette, M., and Gravel, C. (1997). Adenovirus-mediated gene transfer of ciliary neurotrophic factor can prevent photoreceptor degeneration in the retinal degeneration (rd) mouse. Hum Gene Ther 8, 423-30. Connor, T. B., Jr., Roberts, A. B., Sporn, M. B., Danielpour, D., Dart, L. L., Michels, R. G., de Bustros, S., Enger, C., Kato, H., Lansing, M., and et al. (1989). Correlation of fibrosis and transforming growth factor-beta type 2 levels in the eye. J Clin Invest 83, 1661-6. Dai, Y., Schwarz, E. M., Gu, D., Zhang, W. W., Sarvetnick, N., and Verma, I. M. (1995). Cellular and humoral immune responses to adenoviral vectors containing factor IX gene: tolerization of factor IX and vector antigens allows for longterm expression. Proc Natl Acad Sci U S A 92, 1401-5. Di Polo, A., Aigner, L. J., Dunn, R. J., Bray, G. M., and Aguayo, A. J. (1998). Prolonged delivery of brain-derived neurotrophic factor by adenovirus- infected Muller cells temporarily rescues injured retinal ganglion cells. Proc Natl Acad Sci U S A 95, 3978-83. Dryja, T. P. (1997). Gene-based approach to human genephenotype correlations. Proc Natl Acad Sci U S A 94, 12117-21. Dryja, T. P., McGee, T. L., Reichel, E., Hahn, L. B., Cowley, G. S., Yandell, D. W., Sandberg, M. A., and Berson, E. L. (1990). A point mutation of the rhodopsin gene in one form of retinitis pigmentosa. Nature 343, 364-6. Frank, R. N. (1997). Growth factors in age-related macular degeneration: pathogenic and therapeutic implications. Ophthalmic Res 29, 341-53. Flannery, J. G., Zolotukhin, S., Vaquero, M. I., LaVail, M. M., Muzyczka, N., and Hauswirth, W. W. (1997). Efficient photoreceptor-targeted gene expression in vivo by recombinant adeno-associated virus. Proc Natl Acad Sci U S A 94, 6916-21.

207

Ikuno and Kazlauskas: Gene therapy for eye diseases Kajiwara, K., Berson, E. L., and Dryja, T. P. (1994). Digenic retinitis pigmentosa due to mutations at the unlinked peripherin/RDS and ROM1 loci. Science 264, 1604-8. Kimura, H., Sakamoto, T., Cardillo, J. A., Spee, C., Hinton, D. R., Gordon, E. M., Anderson, W. F., and Ryan, S. J. (1996). Retrovirus-mediated suicide gene transduction in the vitreous cavity of the eye: feasibility in prevention of proliferative vitreoretinopathy. Hum Gene Ther 7, 799808. Kliffen, M., Sharma, H. S., Mooy, C. M., Kerkvliet, S., and de Jong, P. T. (1997). Increased expression of angiogenic growth factors in age-related maculopathy. Br J Ophthalmol 81, 154-62. Lashkari, K., Rahimi, N., and Kazlauskas, A. (1999). Hepatocyte growth factor receptor in human RPE cells: implications in proliferative vitreoretinopathy. Invest Ophthalmol Vis Sci 40, 149-56. Lewin, A. S., Drenser, K. A., Hauswirth, W. W., Nishikawa, S., Yasumura, D., Flannery, J. G., and LaVail, M. M. (1998). Ribozyme rescue of photoreceptor cells in a transgenic rat model of autosomal dominant retinitis pigmentosa. Nat Med 4, 967-71. Li, T., Adamian, M., Roof, D. J., Berson, E. L., Dryja, T. P., Roessler, B. J., and Davidson, B. L. (1994). In vivo transfer of a reporter gene to the retina mediated by an adenoviral vector. Invest Ophthalmol Vis Sci 35, 25439. Lopez, P. F., Sippy, B. D., Lambert, H. M., Thach, A. B., and Hinton, D. R. (1996). Transdifferentiated retinal pigment epithelial cells are immunoreactive for vascular endothelial growth factor in surgically excised agerelated macular degeneration-related choroidal neovascular membranes. Invest Ophthalmol Vis Sci 37, 855-68. Miyoshi, H., Takahashi, M., Gage, F. H., and Verma, I. M. (1997). Stable and efficient gene transfer into the retina using an HIV-based lentiviral vector. Proc Natl Acad Sci U S A 94, 10319-23. Morimura, H., Fishman, G. A., Grover, S. A., Fulton, A. B., Berson, E. L., and Dryja, T. P. (1998). Mutations in the RPE65 gene in patients with autosomal recessive retinitis pigmentosa or leber congenital amaurosis. Proc Natl Acad Sci U S A 95, 3088-93. Naldini, L., Blomer, U., Gage, F. H., Trono, D., and Verma, I. M. (1996). Efficient transfer, integration, and sustained long-term expression of the transgene in adult rat brains injected with a lentiviral vector. Proc Natl Acad Sci U S A 93, 11382-8. Pastor, J. C. (1998). Proliferative vitreoretinopathy: an overview. Surv Ophthalmol 43, 3-18. Pe'er, J., Folberg, R., Itin, A., Gnessin, H., Hemo, I., and Keshet, E. (1996). Upregulated expression of vascular endothelial growth factor in proliferative diabetic retinopathy. Br J Ophthalmol 80, 241-5. Portera-Cailliau, C., Sung, C. H., Nathans, J., and Adler, R. (1994). Apoptotic photoreceptor cell death in mouse models of retinitis pigmentosa. Proc Natl Acad Sci U S A 91, 974-8. Robbins, S. G., Mixon, R. N., Wilson, D. J., Hart, C. E., Robertson, J. E., Westra, I., Planck, S. R., and Rosenbaum, J. T. (1994). Platelet-derived growth factor ligands and receptors immunolocalized in proliferative

retinal diseases. Invest Ophthalmol Vis Sci 35, 3649-63. Sakamoto, T., Kimura, H., Scuric, Z., Spee, C., Gordon, E. M., Hinton, D. R., Anderson, W. F., and Ryan, S. J. (1995). Inhibition of experimental proliferative vitreoretinopathy by retroviral vector-mediated transfer of suicide gene. Can proliferative vitreoretinopathy be a target of gene therapy?. Ophthalmology 102, 1417-24. Sakamoto, T., Ueno, H., Goto, Y., Oshima, Y., Yamanaka, I., Ishibashi, T., and Inomata, H. (1998). Retinal functional change caused by adenoviral vector-mediated transfection of LacZ gene. Hum Gene Ther 9, 789-99. Verma, I. M., and Somia, N. (1997). Gene therapy -- promises, problems and prospects. Nature 389, 239-42. Yoshida, A., Ho, P. C., Schepens, C. L., McMeel, J. W., and Duncan, J. E. (1984). Severe proliferative vitreoretinopathy and retinal detachment. II. Surgical results with scleral buckling. Ophthalmology 91, 1538-43.

Andrius Kazlauskas

208

Gene Therapy and Molecular Biology Vol 4, Page 209 Gene Ther Mol Biol Vol 4, 209-219. December 1999

The role of HSV amplicon vectors in cancer gene therapy Review Article

Kutubuddin Mahmood, Khaled Tolba, Howard J. Federoff, and Joseph D. Rosenblatt University of Rochester Cancer Center and Division of Molecular Medicine and Gene Therapy

__________________________________________________________________________________ Correspondence: Joseph D. Rosenblatt, M.D. Chief, Hematology-Oncology Unit, University of Rochester Cancer Center and Division of Molecular Medicine and Gene Therapy, 601 Elmwood Avenue, Box 704, Rochester, N.Y. 14642. Tel: (716) 275-9484; Fax: (716) 273-1051; E-mail: joe_rosenblatt@urmc.rochester.edu Key Words: Herpes Simplex Virus vectors, HSV amplicon, cancer gene therapy, cytokine, chemokine, immunotherapy Received: 20 February 1999; accepted: 23 February 1999

Summary Recent progress in tumor biology, virology and immunology has led to new approaches to the gene therapy for cancer. Herpes Simplex Virus (HSV) based vectors are attractive vectors for gene therapy use due to a number of favorable biologic features. Several characteristics render HSV suitable for gene therapy, including high transduction efficiency, ability to transduce non-dividing cells, high packaging capacity, wide cellular tropism and the ability to package multiple copies of the same gene or several genes. Newer HSV vectors differ in replication potential, sensitivity to anti-viral agents, neurotoxicity, tumor-specific cytotoxicity and persistence in the host cell. Socalled â&#x20AC;&#x153;oncolyticâ&#x20AC;? HSV based vectors demonstrate selective replication in tumor cells relative to normal tissue. HSV amplicon based vectors allow genetic transfer of multiple transgene copies in the absence of viral genes. This degree of flexibility, relative to other viral vector systems, has allowed for the use of HSV vectors in a variety of antitumor strategies including oncolytic, as well as immune-based strategies. Successful immune based strategies in animal models have included transfer of cytokines, costimulatory molecules and/or chemokines. Phase I/II clinical trials using HSV based vectors have been initiated.

I. Introduction Current advances in the understanding of the viral structure and function, mechanisms of viral entry, and replication have resulted in the use of engineered viruses for the treatment of cancer. Although effective gene transfer based on the murine retrovirus vectors was developed in the 1980s, the use of retroviral vectors in cancer gene therapy was hampered by low titers, poor transduction efficiency and limited expression of the desired gene products. As a result, other viruses such as adenovirus, adeno-associated virus and herpesviruses have been explored as viral vectors for gene therapy of cancer. Herpes simplex virus type 1 (HSV-1) is one of the most extensively characterized of all the herpes viruses and has several advantages over other viruses used as gene therapy vectors. Among the advantages are a wide host

209

range, high efficiency of gene transfer and ability to replicate in dividing and non-dividing cells. Multiple genes can be simultaneously expressed due to the large genome size that can be replaced with the desired gene(s) up to ~30 kb (Mocarski et al. 1980). The HSV genome consists of a linear double stranded DNA of ~150 kb size coding for over 83 genes (Sears and Roizman.1996). The genome is divided into long and short segments each flanked by their own terminal repeat sequences . Following entry of HSV into a susceptible cell, the viral capsid is transported to the nucleus and the viral genome released in the nucleoplasm where it is transcribed. Transcription of HSV genes is regulated in a tightly controlled manner. Immediate early (IE) or !-genes which are not dependent on the viral protein synthesis are the first to be expressed. The second set of genes, early genes (E) or "-genes are expressed in the presence of IE gene products and encode enzymes

Mahmood et al: HSV amplicon vectors in cancer gene therapy required for DNA synthesis and virus genome replication. Other genes, late or # -genes specify products for virus assembly, tegument and envelope glycoproteins. HSV replicates through a “rolling circle” mechanism and the viral genome is packaged into assembled capsids ( Sears and Roizman.1996). The mature capsids acquire envelope proteins at the nuclear membrane and the virus particles exit from the cell by exocytosis (Locker et al. 1979). During the natural course of infection, HSV may be a lytic virus or can remain latent in the infected cells of neural origin (Ward et al. 1994). During HSV latency, lytic genes are silenced and a set of stable processed intronic RNA transcripts known as latency associated transcripts (LATs) are expressed (Stevens et al. 1987; Stevens. 1989). The ability of HSV to remain latent in the infected host, can be advantageous for vector development if long-term gene expression is desired.

replication and packaging and do not encode any viral proteins (‘ori’ and ‘pac’ are non-coding) , they must be supplied with both viral structural and nonstructural proteins for viral DNA synthesis and exocytosis. HSV amplicons are packaged in the presence of a defective HSV-1 helper virus. Detailed methods for amplicon packaging have been previously described (Geschwind et al. 1994a; Kwong and Frenkel. 1995).The basic components of HSV amplicon packaging systems include: a) helper virus HSV-1 deleted in an essential gene such as (IE3, also known as ICP4), b) a packaging cell line such as RR-1 cells, which is an IE3 expressing BHK derived cell line that complements the deleted function of the helper virus in trans, and c) amplicon plasmid DNA. The packaging procedure involves transfection of the amplicon plasmid into the packaging cell line, followed by infection with HSV-1 helper virus (F i g u r e 1 ). The HSV-1 genes present in the helper virus and the packaging cell line allow for the replication and packaging of a concatemeric form of amplicon DNA (containing the gene of interest) packaged in HSV assembled virions. The resultant virus stock obtained therefore is a mixture of helper and vector virions that can subsequently be concentrated, titrated and stored at -700C for use. The amplicon titers that can be obtained by the use of helper virus range from 108 to 109 pfu/ml. Until recently, production of amplicons required repeated passaging of the amplicon/helper preparation to increase amplicon vector titer.

II. HSV in cancer gene therapy The development of HSV vectors for gene therapy of cancer coincided with efforts to create a prototype HSV recombinant vaccine for HSV-1 and HSV-2. Mutants of HSV-1 were generated by making deletions in certain regions to eliminate virulence and toxicity of the virus (Post et al. 1981; Jenkins et al. 1985). HSV derived deletion mutants can be divided into either replication competent or replication defective vectors and generally fall into four categories:

Due to the HSV tropism for tissues of CNS origin in vivo in the infected host, amplicon vectors have been tested for treatment of diseases of the central and peripheral nervous system both in vitro (Federoff et al. 1992; Battleman et al. 1993; Geller et al. 1993; Geschwind et al. 1994b) and in vivo (During et al. 1994; Kaplitt et al. 1994). Significant progress has been made in the use of HSV amplicons for expressing several neuronal genes including nerve growth factor (NGF) (Federoff et al. 1992; Bahr et al. 1994; Geschwind et al. 1994b), brain-derived neurotrophic factor (BDNF) (Geschwind et al. 1996; Garrido et al. 1998), the NGF receptor (trkA) for converting NGF non-responsive cells to NGF responding cells (Xu et al. 1994), and neurotropic glutamate receptor (Bergold et al. 1993).

A. Amplicons HSV amplicons are plasmids containing only an HSV origin of replication (‘ori’) and packaging sequence (‘pac’). The recognition that such amplicons could be constructed grew out of the observation that defective interfering particles accumulating in HSV stocks passaged at high moi were simple reiteration of ‘ori’ and ‘pac’ sequences (Spaete et al. 1982; Kwong et al. 1995; Sears and Roizman. 1996). Such defective viral particles are comprised of amplified head-to-tail concatemers, hence plasmids designed for propagation in HSV-1 have been called “amplicons”. These amplicon plasmids are engineered to contain the HSV sequences for ‘ori’ and ‘pac’ in cis, which allow them to be packaged as a concatemeric repeated unit into the viral capsid (Frenkel et al. 1994). These sequences represent < 1% of the ~150 kb HSV genome. Since the ‘ori’ and ‘pac’ sequences together measure less than 1 kb, amplicon vectors can be used to clone large fragments of foreign DNA and can serve as excellent vectors for delivery of the desired genes into the host cells. Several genes can be cloned into a single amplicon vector and different amplicon vectors can be mixed during packaging to generate vector stocks containing multiple combinations of different genes. Since the HSV amplicons only include cis signals for

More recently, amplicon vectors have been tested for utility in treatment of cancer in several experimental models in different laboratories including owr own. Amplicons have been used to transduce an immunestimulating gene (e.g. B7.1[CD80], IL-2, IL-12, and GMCSF) or drug-susceptibility genes (e.g. HSV-tk) for cancer therapy. Although amplicon use affords great flexibility, persistence of helper virus in amplicon preparations is 210

Gene Therapy and Molecular Biology Vol 4, Page 211 currently a potential limitation. Prolonged expression of the desired gene is possible in post-mitotic cells; however, in dividing cells, non-replicating amplicon genomes are segregated which results in progressive loss. One attempt to address this issue was by the inclusion of specific EBV sequences required to maintain the amplicon as an episome in the nucleus of transfected cells (Wang et al. 1997). Epstein-Barr virus (EBV) has been shown to contain a unique latent replication origin (ori P) which directs viral self-replication and maintenance in cells without entering the lytic cycle (Yates et al. 1985). EBV nuclear antigen1 (EBNA-1) is a DNA binding transactivator for ori P. HSV amplicon vectors carrying the combination of EBNA-1 and EBV ori P should be more effective for stable segregation following division of eukaryotic cells.

amplicon systems. Amplicon preparations that are free of helper virus have been made using cosmid sets containing the entire HSV genome with deletions in the packaging sequences (Cunningham et al. 1993; Fraefel et al. 1996). These cosmids provide the necessary structural and nonstructural proteins for amplicon replication and packaging into virus capsids but cannot be packaged themselves due to absence of packaging sequences. Cotransfection of cells with this modified cosmid sets and amplicon plasmid DNA results in the production of amplicon stocks free of detectable helper virus. Helper virus free amplicon preparations have been produced by some groups but virus yield produced by this method is relatively low (~107 pfu/ml). More recently, efficient helper free amplicon methods were reported using bacterial artificial chromosomes (BAC) encoding the HSV genome minus the packaging sequences (Stavropoulos et al. 1998). Such helper free amplicon preparations avoid disadvantages related to expression of ICP47 and other HSV encoded genes that may influence immune response to transduced tumor cells.

B. Helper virus free amplicons Helper virus/amplicon mixture preparations are toxic to most cells because of helper virus expressed IE genes and this has stimulated efforts to produce helper-free

Figure 1. Generation of HSV-1 Amplicon Virus : HSV amplicon plasmids are transfected into RR1 cells and 24 hours later IE3 deleted HSV-1 virus is added to the monolayer. Once cytopathic changes are observed, the cells are harvested, freeze-thawed and sonicated. Viral supernatent is repassaged on RR1 cells to increase titer, and then harvested concentrated and titered. The amplicon plasmid is packaged in a concatemeric form.

211

Mahmood et al: HSV amplicon vectors in cancer gene therapy this leads to phosphorylation of the ! subunit of the translation initiation factor eIF-!2 and total shutoff of protein synthesis (Chou et al. 1995). A second function of # 134.5 is to enable the virus to multiply efficiently in the CNS. Viruses constructed with mutations in this gene are highly attenuated in experimental animals (Chou et al. 1990; Bolovan et al. 1994). This function of # 134.5 appears to be independent of its ability to preclude the protein shutoff. Evidence in support of this is based on the observation that viruses carrying GADD34 in place of # 134.5 show no evidence of protein shutoff but still are attenuated - i.e. replicate poorly in neural tissue (Andreansky et al. 1996). A number of HSV-1 mutants for # 134.5 have been tested as a treatment for CNS tumors in experimental animals with a high degree of success and no incidence of encephalitis (Andreansky et al. 1997). These mutants are still sensitive to acyclovir, thus adding a safety feature to eliminate the virus should an unforeseen complication arise. The success of these mutants prompted others to try them in other solid tumors of non-CNS origin. Kucharczuk et.al. (Kucharczuk et al. 1997) showed the effectiveness of an HSV mutant for # 134.5 in an animal model for mesothelioma. G207 had also been tested in a metastatic breast cancer model with growth inhibition and/or improved survival in mice harboring susceptible tumors (Toda et al. 1998b).

C. Replication defective HSV vectors In an attempt to decrease toxicity, recombinant replication defective vectors have been designed with deletions in the ICP4, ICP22, ICP27, or ICP0 genes (Wu et al. 1996; Samaniego et al. 1997; Samaniego et al. 1998). An example of this is the disabled single-cycle virus (DISC-HSV). DISC-HSV lacks the gene for the essential glycoprotein H (gH) and can be grown to high titers in complementing cells expressing the gH gene. DISC-HSV particles have been shown to transduce a variety of human cell types. Murine pre-B cell leukemia cells transduced with DISC-HSV encoding GM-CSF had an improved survival in mice with the tumor cells acting as a potent vaccine (Dilloo et al. 1997). These vectors are less toxic than replication competent vectors since infected cells do not produce infectious virus.

D. Replication restricted HSV vectors A variety of replication restricted HSV mutants have been adapted to selectively grow in tumor cells. Replication competent mutants for gene therapy of cancer include viruses that are deleted or attenuated for the ICP-6 and/or # 134.5 genes, ICP0, thymidine kinase, or the US3 protein kinase genes (Glorioso et al. 1995; Kesari et al. 1995; Mineta et al. 1995; Yazaki et al. 1995; Coffin et al. 1996). Most of these genes code for enzymes involved in nucleic acid metabolism (e.g. thymidine kinase, DNA polymerase, ribonucleotide reductase) which are not expressed in post-mitotic neuronal or other differentiated cells such as hepatocytes, but are induced in tumor cells. Therefore, replication restricted mutants have limited ability to replicate in non-dividing cells and preferentially replicate in tumor cells (McKie et al. 1996). The G207 vector (mutated in ribonucleotide reductase and the # 134.5 genes) or the HSV-tk- mutant replication competent viruses have been successfully tested in brain tumor models (Markert et al. 1993; Kesari et al. 1995; Mineta et al. 1995; Yazaki et al. 1995; Coffin et al. 1996). Use of these vectors is based on the ability of these mutants to replicate in actively growing glioma cells while sparing normal post-mitotic brain cells, thereby reducing the collateral damage. # 134.5 is present in two copies and encodes a 263 amino acid protein in the type 1 herpesvirus. The 70-amino acid carboxyl terminus is highly homologous to the mammalian growth arrest and DNA damage gene GADD34, and the carboxyl terminus of GADD34 restores function to # 134.5-deleted virus (Chou et al. 1994a; Chou et al. 1994b). Viral infection normally triggers a host stress response that shuts off protein synthesis and causes apoptosis aborting viral replication. # 134.5 effectively prevents this protein shutoff and allows viral replication to proceed. Specifically, HSV-1 infection activates the host cell PKR kinase and in the absence of a functional # 134.5,

Although replication restricted HSV vectors are engineered to selectively kill tumor cells while sparing the normal host tissues, these vectors have their limitations including lack of efficacy (DNA polymerase mutants), resistance to the two main HSV therapeutics in clinical use, acyclovir and ganciclovir (thymidine kinase negative mutants), retained capacity to induce encephalitis (thymidine kinase and DNA polymerase mutants); lack of an animal model due to significant differences in human and mouse or rat susceptibility to genetically engineered HSV (ribonucleotide reductase mutants); and potential loss of susceptibility of tumor cells previously treated with alkylating agents (ribonucleotide reductase mutants) (Andreansky et al. 1996).

III. Cancer gene therapy using HSV amplicons Direct activation of cytolytic T cells or immunomodulatory support for their activation has been extensively applied as a cancer gene therapy strategy. Specifically, these efforts are directed at generating a tumor specific CD8+ T-cell response with the goal of generating systemic and lasting immunity. To generate protective immunity, antigen presenting cells and/or tumor cells need to deliver at least two signals to T cells: 1) an antigen specific signal mediated by the T cell receptor (TCR) recognizing tumor specific peptide bound by molecules of

212

Gene Therapy and Molecular Biology Vol 4, Page 213 the major histocompatibility complex (MHC) class I or II proteins and 2) a costimulatory signal mediated by ligand/receptor interaction of CD28 on activated T cells with members of the B7 family (CD80,CD86) of â&#x20AC;&#x153;costimulatoryâ&#x20AC;? proteins (Green et al. 1994; Bluestone. 1995). Presentation of the first without the second signal may lead to the induction of T cell anergy and such anergic T cells lack the ability to mount a tumor specific T cell response (Mueller et al. 1989). B7.1 transduction has been shown to be effective in generating anti-tumor specific CD8+ T cell responses (Ramarathinam et al. 1994), and several tumor cell lines are rendered immunogenic upon expression of B7.1 (Chen et al. 1994). B7.1 expression has been shown to have a synergistic effect in combination with the secretory cytokines such as IL2, and IL12 in several tumor models (Gaken et al. 1997). The immune response may be further potentiated by providing chemotactic signals at the site of tumor designed to elicit active effector cell migration to the tumor site. Chemokine expression such as the use of RANTES, a C-C chemokine have been shown to have anti-tumor effects (Mule et al. 1996). Identification of novel chemokines and their receptors, and their biological functions in recruitment of immune T cells and monocytes will allow increased flexibility in the use of these molecules to augment tumor specific immunity.

(Tung et al. 1996). Such immunized mice were protected upon intra-portal challenge with viable parental HEPA1-6 tumor cells. In contrast, mice immunized with HEPA1-6 cells transduced with the HSV amplicon for lacZ developed liver tumors upon inoculation of parental tumor cells. In another study HSV amplicons encoding IL12 were used for intratumoral delivery in a murine tumor model. Delayed death was observed by the use of HSV-IL12 amplicon (Toda et al. 1998a). More recently the utility of these vectors in treating hematologic malignancies has been explored by our laboratory and others (see below). The characteristics of amplicon vectors, including ease of packaging, high transduction efficiency and ability to achieve high levels of expression in hematopoietic cells render them potentially useful in gene therapy of a variety of malignancies.

IV. Expression of immunogenes using the HSV amplicon in murine and human cells To examine the infectivity of HSV amplicons, murine EL4 lymphoma cells, as well as human leukemic cells (CML or AML) were infected at different multiplicity of infections (MOIs) with the HSVlac amplicon. 24 hours after transduction cells were stained for beta-galactosidase using X-Gal staining. High levels of expression of the cell associated beta-galactosidase (> 95% cells were positive for beta-galactosidase) were observed at an estimated moi of 0.5 with the use of HSVlac amplicon vector in EL4 cells.

HSV amplicons have now been used in several solid tumor models. In a murine hepatoma model, Tung et.al., (Tung et al. 1996) have used the HSV amplicon vector for expressing human IL2 to transduce murine hepatoma cells (HEPA1-6) which were irradiated to immunize naĂŻve mice

Figure 2a. Expression of B7.1 in primary human CML cells. CML cells are isolated from peripheral blood of a patient in blast crisis by Ficoll-hypaque density centrifugation. Cells are infected with HSVB7.1. 24 hours later, control cells (A) or transduced cells (B) are stained with PE-conjugated anti-B7.1 and samples analyzed by flow cytometry.

213

Mahmood et al: HSV amplicon vectors in cancer gene therapy

Figure 2b. Expression of B7.1 in primary human AML cells. AML cells are isolated from peripheral blood of an AML patient by Ficoll-hypaque density centrifugation. Cells are infected with HSVB7.1. 24 hours later, control cells (A) or transduced cells (B) are stained with PE-conjugated anti-B7.1 and samples analyzed by flow cytometry.

Flow cytometric analyses were performed for the cell surface expression of B7.1 on murine and human leukemic cells following transduction with the B7.1 encoding HSVB7.1 amplicon vector. Cells were stained for B7.1 expression (anti-CD80 PE, Becton-Dickinson) and analyzed by flow cytometry. Murine EL4 cells or the human leukemic cells of chronic myelogenous (CML), acute myelogenous (AML), or acute lymphocytic (ALL) origin were negative for B7.1 expression. However, following infection with the HSVB7.1 amplicon vector at an moi of 1 pfu/cell, over 95% of EL4 cells were B7.1 positive (Kutubuddin et al. 1999), and 20-60% of AML, CML or ALL cells stained positively as analyzed by flow cytometry ( F i g u r e 2 a & 2 b ). When dual staining for B7.1 and MHC class I molecules was performed in human leukemic cells infected with HSV amplicons, we observed significant down modulation of the MHC class I expression in the HSVB7.1 or HSVlac amplicon infected cells relative to the uninfected control cells (Figure 3). It has previously been shown that ICP47 expressed by HSV-1 is known to interact with the human TAP-1 protein and thereby prevents the loading of processed peptide onto the MHC class I molecule. Down regulation of MHC-1 may be due to effects of ICP47 carried by helper virus used in HSV amplicon packaging. In addition to costimulation, we wished to use amplicon mediated gene transfer to recruit immune effector cells to tumor. Murine EL4 or human leukemic cell cultures supernatants were tested following transduction with HSV amplicon vectors encoding the CC chemokine RANTES in a sandwich ELISA using anti-RANTES 214

antibody (R & D Systems) for capture and biotinylated anti-RANTES (R & D Systems) followed by alkaline phosphatase conjugated avidin for detection. No RANTES was detected in control culture supernatants. HSVrantes infected cells produced soluble RANTES in the range of 2ng/ml to 60ng/ml depending upon the moi of infection and the cell type. Soluble human RANTES produced by the HSVrantes infected cells was biologically active as demonstrated by the murine T cell migration assay using transwell chambers (Kutubuddin et al. 1999).

V. Intra-tumoral delivery of HSV amplicon in a murine tumor model EL4, a murine T lymphoma cell line was used to develop tumors in syngeneic adult C57BL6 mice (H2b) by subcutaneous inoculation. It has previously been shown that B7.1 expression in EL4 cells leads to tumor rejection in mice due to the establishment of systemic cytolytic T cell responses (Chen et al. 1994). We had been successful in eradicating a pre-established EL4 tumor by the delivery of HSV amplicons for B7.1 (HSVB7.1) or RANTES (HSVrantes). Although approximately 50% of the mice were tumor free with HSVB7.1 (17/26) or HSVrantes (11/22) treatment alone, when both amplicons were used in combination, an increased number of mice (23/26) were tumor free after one month. Control mice injected with untreated EL4 or those whose tumors were treated with the HSVlac (14/14) showed uniform tumor growth. Immune mice consistently rejected rechallenge with parental EL4 cells (Kutubuddin et al. 1999).

Gene Therapy and Molecular Biology Vol 4, Page 215

F i g u r e 3 . Down-regulation of MHC-I on CML cells transduced with HSV virus coding for B7.1 (CD80) and HSV-LacZ. CML cells are stained with a FITC- conjugated anti-human MHC-I 24 hours after transduction with HSVB7.1 or HSVlac and analyzed by flow cytometry.

VI. Generation of systemic cell mediated immunity using HSV amplicon vectors

In a separate experiment, we tested whether the delivery of HSV amplicon into preestablished tumor resulted in induction of systemic immunity against a contralateral tumor inoculated at the same time on the left hind limb. EL4 cells were inoculated subcutaneously (sc) bilaterally in both the hind limb of mice and tumors allowed to grow to a diameter of 5-6 mm (6-7 days). On day 7 and day 14, HSV amplicons were delivered to the tumor established on the right hind limb (10 mice/group). Complete regression was seen in 5/10 mice inoculated with HSVB7.1, 5/10 inoculated with HSVrantes, and 8/10, in combined HSVB7.1 and HSVrantes treated animals. In animals treated with HSVB7.1 alone (5/5, p=0.0325), or with a combination of HSVB7.1 and HSVrantes (8/8, p=0.0007), regression of the contralateral untreated tumor was consistently observed along with the treated tumor (Figure 4). In one HSVrantes treated animal where the treated tumor regressed completely, the contralateral tumor grew at a reduced rate relative to the untreated tumors in control animals. In HSVlac treated animals 10/10 animals demonstrated tumor growth on both the sides, although we noticed that HSVlac treated tumors grew at a slightly reduced rate compared to the control untreated tumors. These experiments demonstrate that systemic immunity generated as a result of intratumoral HSV amplicon injection could prevent progression of the contralateral untreated tumor.

In order to examine the induction of CTL responses in mice transduced intratumorally with HSV amplicon, splenocytes from these mice were harvested and were cocultured in vitro along with irradiated stimulator EL4 cells for 6-7 days. In vitro primed splenocytes were used at different effector : target ratios and 51Cr release from a fixed number of labeled EL4 cells counted as a measure of CTL activity. EL4 specific CTL activity was seen in splenocytes from mice receiving HSVB7.1 or HSVrantes alone or in combination . CTL responses were only seen in mice in which EL4 tumor regressed following HSVB7.1 and/or HSVrantes amplicon treatment. Control animals which were mock treated or treated with HSVlac amplicon did not show any significant CTL responses. The effector population involved in the observed CTL responses were the CD8+ T cells as lysis was markedly inhibited in the presence of either an anti-T cell monoclonal antibody cocktail (CD4, CD8, or Thy-1), or anti-CD8 antibody, but not by anti-CD4 antibody. Low levels of NK activity were detected when the NK sensitive Yac-1 cells were used as target cells in the lytic assay. Therefore, the predominant effector population appeared to be CD8+ CTLs (Kutubuddin et al. 1999).

215

Mahmood et al: HSV amplicon vectors in cancer gene therapy

F i g u r e 4 . Tumor growth following inoculation of HSV amplicon into preestablished EL4 tumor and growth of parental EL4 cells contra-laterally. Viable EL4 cells (106 ) were implanted sc. on both hind limbs of 8 weeks old C57BL6 mice. Tumors were allowed to develop to a size of 5 to 6mm diameter. HSV amplicon virus (2 X 106 amplicon- containing virus particles) was injected into the right tumor on days 7 and 14, and growth of the HSV amplicon- treated and untreated tumors were monitored every 3-5 days.

viruses deleted in the ICP47 gene. Current attempts involve packaging with multiple cosmids encoding HSV virion proteins in trans. More recently, bacterial artificial chromosomes incorporating a non-replicating HSV genome have been used to provide helper free packaging functions (Federoff H, et al., unpublished). Efforts are also needed to improve the yield of the HSV amplicon when using helper free packaging, as the yield of the HSV amplicon titer using the cosmid system is considerably lower than that achievable with standard helper virus. New generations of HSV amplicons also may encode multiple transgenes as well as genes derived from other virus vectors. One example is the recent incorporation of the adeno-associated virus (AAV) inverted terminal repeats (ITRs) and the AAV rep gene, in addition to the HSV replication and packaging elements (Fraefel et al. 1997; Johnston et al. 1997). AAV rep function can, in theory, recognize the ITRs and allow for subsequent integration in a site specific manner in

VII. Future directions HSV amplicon vectors are safe and effective gene therapy vectors for the expression of immune molecules. High titers of the amplicon virus particles can be generated using a helper virus for packaging, and high levels of expression of the transduced genes can be obtained with the use of the HSV amplicon. HSVB7.1 or HSVrantes amplicons were successfully used to eradicate preestablished tumors in mice. Combined use of the HSVB7.1 and HSVrantes amplicons were more effective than either alone. A tumor specific memory T cell response is established which is effective in preventing tumor growth upon rechallenge. Although HSV amplicon vectors readily infect human cells, including primary leukemic cells, helper virus encoding ICP47 presumably causes decreased MHC-1 expression due to effects on TAP-1. This problem may be overcome using helper free packaging or helper

216

Gene Therapy and Molecular Biology Vol 4, Page 217 Bluestone, JA (1 9 9 5 ). New perspectives of CD28-B7 mediated T cell costimulation. Immunity 2(555).

chromosome 19 (Berns. 1996). Such an HSV/AAV hybrid vector may be useful for targeted gene delivery to the nucleus of the host cells. In hereditary liver disorders, gene transfer to the liver is highly desirable, and various viral vectors such as adenoviruses, retrovirus, baculovirus have been evaluated. Since HSV amplicon can infect postmitotic hepatic cells, they are also good candidates for the gene therapy of liver diseases. Further studies are needed to test these vectors in tumor immuno-therapy in the setting of pre-established immunity to HSV. Replication restricted HSV viruses such as the # 134.5 genes deleted G207 virus may preferentially replicate in tumor cells, and can also be used to package amplicons. Whether the lytic function of replication restricted viruses will contribute to tumor antigen uptake and â&#x20AC;&#x153;representationâ&#x20AC;? by antigen presenting cells remains to be tested. HSV vectors have considerable promise as safe, effective vectors for the gene therapy of cancer and other diseases.

Bolovan CA, Sawtell NM and Thompson RL (1 9 9 4 ). ICP34.5 mutants of herpes simplex virus type 1 strain 17syn+ are attenuated for neurovirulence in mice and for replication in confluent primary mouse embryo cell cultures. J V i r o l 68(1), 48-55. Chen I, McGowan P, Ashe S, Johnston J, Li Y, Hellstrom I and Hellstrom KE (1 9 9 4 ). Tumor immunogenicity determines the effect of B7 costimulation on T cell mediated tumor immunity. J Exp Med 179, 523. Chou J, Chen JJ, Gross M and Roizman B (1 9 9 5 ). Association of a M(r) 90,000 phosphoprotein with protein kinase PKR in cells exhibiting enhanced phosphorylation of translation initiation factor eIF-2 alpha and premature shutoff of protein synthesis after infection with gamma 134.5- mutants of herpes simplex virus 1. Proc Natl Acad Sci USA 92(23), 10516-20. Chou J, Kern ER, Whitley RJ and Roizman B (1 9 9 0 ). Mapping of herpes simplex virus-1 neurovirulence to gamma 134.5, a gene nonessential for growth in culture. S c i e n c e 250(4985), 1262-6.

Acknowledgements We gratefully acknowledge Dr. R. Spaete for critical reading of the manuscript and Sue Sullivan, University of Rochester Cancer Center, Rochester, NY for technical assistance.

Chou J, Poon AP, Johnson J and Roizman B (1 9 9 4 a ). Differential response of human cells to deletions and stop codons in the gamma(1)34.5 gene of herpes simplex virus. J Virol 68(12), 8304-11. Chou J and Roizman B (1 9 9 4 b ). Herpes simplex virus 1 gamma(1)34.5 gene function, which blocks the host response to infection, maps in the homologous domain of the genes expressed during growth arrest and DNA damage. Proc Natl Acad Sci USA 91(12), 5247-51.

References Andreansky S, Soroceanu L, Flotte ER, Chou J, Markert JM, Gillespie GY, Roizman B and Whitley RJ (1 9 9 7 ). Evaluation of genetically engineered herpes simplex viruses as oncolytic agents for human malignant brain tumors. Cancer Res 57(8), 1502-9.

Coffin RS, MacLean AR, Latchman DS and Brown SM (1 9 9 6 ). Gene delivery to the central and peripheral nervous systems of mice using HSV1 ICP34.5 deletion mutant vectors. Gene Therapy 3, 886-891.

Andreansky SS, He B, Gillespie GY, Soroceanu L, Markert J, Chou J, Roizman B and Whitley RJ (1 9 9 6 ). The application of genetically engineered herpes simplex viruses to the treatment of experimental brain tumors. Proc Natl Acad Sci USA 93(21), 11313-8.

Cunningham C and Davidson AJ (1 9 9 3 ). A cosmid based system for constructing mutants of herpes simplex virus type 1. V i r o l o g y 197, 116-124.

Bahr B, Neve R, Sharp J, Geller AI and Lynch G (1 9 9 4 ). Rapid and stable gene expression in hippocampal slice cultures from a defective HSV-1 vector. M o l B r a i n R e s 26, 277285.

Dilloo D, Rill D, Entwistle C, Boursnell M, Zhong W, Holden W, Holladay M, Inglis S and Brenner M (1 9 9 7 ). A novel herpes vector for the high-efficiency transduction of normal and malignant human hematopoietic cells. B l o o d 89(1), 119-27.

Battleman DS, Geller AI and Chao MC (1 9 9 3 ). HSV-1 vector mediated gene transfer of the human nerve growth factor receptor p75 huNGFR defines high affinity NGF binding. J Neurosci 13, 941-951.

During, MJ, Naegele JR, O'Malley KL and Geller AI (1 9 9 4 ). Long-term behavioral recovery in parkinsonian rats by an HSV vector expressing tyrosine hydroxylase [see comments]. S c i e n c e 266(5189), 1399-403.

Bergold PJ, Casaccia-Bonnefil P, Zeng XL and Federoff HJ (1 9 9 3 ). Transsynaptic neuronal loss induced in hippocampal slice cultures by a herpes simplex virus vector expressing the GluR6 subunit of the kainate receptor. Proc Natl Acad Sci USA 90(13), 6165-9.

Federoff HJ, Geschwind MD, Geller AI and Kessler JA (1 9 9 2 ). Expression of nerve growth factor in vivo from a defective herpes simplex virus 1 vector prevents effects of axotomy on sympathetic ganglia. Proc Natl Acad S c i USA 89(5), 1636-40.

Berns KI (1 9 9 6 ). Parvoviridae, the viruses and their replication. F i e l d s V i r o l o g y . BN Fields, DM Knipe and PM Howley. Philadelphia, Lippincott-Raven, 2173-2197.

Fraefel C, Jacoby DR, Lage C, Hilderbrand H, Chou JY, Alt FW, Breakefield XO and Majzoub JA (1 9 9 7 ). Gene

217

Mahmood et al: HSV amplicon vectors in cancer gene therapy transfer into hepatocytes mediated by helper virus-free HSV/AAV hybrid vectors. Mol Med 3(12), 813-25.

Kaplitt MG, Kwong AD, Kleopoulos SP, Mobbs CV, Rabkin SD and Pfaff DW (1 9 9 4 ). Preproenkephalin promoter yields region-specific and long-term expression in adult brain after direct in vivo gene transfer via a defective herpes simplex viral vector. P r o c N a t l A c a d S c i U S A 91(19), 8979-83.

Fraefel C, Song S, Lim F, Lang D, Yu L, Wang Y, Wild P and Geller AI (1 9 9 6 ). Helper virus-free transfer of herpes simplex type 1 plasmid vectors into neural cells. J V i r o l 70, 7190-7197.

Kesari S, Randazzo BP, Valyi-Nagy T, Huang QS, Brown M, MacLean AR, Lee VM-L, Trojanowski JQ and Fraser NW (1 9 9 5 ). Therapy of experimental human brain tumors using a neuroattenuated herpes simplex virus mutant. Lab I n v e s t i g a t i o n 73(5), 636-648.

Frenkel N, Singer O and Kwong AD (1 9 9 4 ). Minireview: the herpes simplex virus amplicon--a versatile defective virus vector. Gene Ther 1(Suppl 1), S40-6. Gaken JA, Hollingsworth SJ, Hirst WJ, Buggins AG, GaleaLauri J, Peakman M, Kuiper M, Patel P, Towner P, Patel PM, et al. (1 9 9 7 ). Irradiated NC adenocarcinoma cells transduced with both CD80 and IL-2 induce CD4+ mediated rejection of established tumors. Hum Gene Ther 8, 477.

Kucharczuk JC, Randazzo B, Chang MY, Amin KM, Elshami AA, Sterman DH, Rizk NP, Molnar-Kimber KL, Brown SM, MacLean AR, et al. (1 9 9 7 ). Use of a replication-restricted herpes virus to treat experimental human malignant mesothelioma. Cancer Res 57(3), 466-71.

Garrido JJ, Alonso MT, Lim F, Carcinero E, Giraldez F and Schimmang T (1 9 9 8 ). Defining responsiveness of avian cochlear neurons to brain-derived neurotrophic factor and nerve growth factor by HSV-1 mediated gene transfer. J Neurochem 6, 2336-2346.

Kutubuddin M, Federoff HJ, Challita-Eid PM, Halterman M, Day B, Atkinson M, Planelles V and Rosenblatt JD (1 9 9 9 ). Eradiction of pre-established lymphoma using HSV amplicon vectors. B l o o d 93(2), 643-54.

Geller AI, During MJ, Haycock JW, Freese A and Neve R (1 9 9 3 ). Long-term increases in neurotransmitter release from neuronal cells expressing a constitutively active adenylate cyclase from a herpes simplex virus type 1 vector. Proc Natl Acad Sci USA 90(16), 7603-7.

Kwong A, and Frenkel N (1 9 9 5 ). Biology of herpes simplex virus (HSV) defective viruses and development of the amplicon system. Viral Vectors: Gene Therapy and N e u r o s c i e n c e A p p l i c a t i o n . ed: Kaplitt M. G.& A.D. Loewy, New York, Academic Press, 25-42.

Geschwind MD, Hartnick CJ, Liu W, Amat J, Van De Water TR and Federoff HJ (1 9 9 6 ). Defective HSV-1 vector expressing BDNF in auditory ganglia elicits neurite outgrowth: model for treatment of neuron loss following cochlear degeneration. Hum Gene Ther 7(2), 173-82.

Locker H and Frenkel C (1 9 7 9 ). Structure and origin of defective genomes contained in serially passaged herpes simplex virus type1 (Justin). J Virol 29, 1065-1077. Markert JM, Malick A, Coen DM and Martuza RL (1 9 9 3 ). Reduction and elimination of encephalitis in an experimental glioma therapy model with attenuated herpes simplex mutants that retain susceptibility to acyclovir. Neurosurgery 32(4), 597-603.

Geschwind MD, Kessler JA, Geller AI and Federoff HJ (1 9 9 4 b ). Transfer of the nerve growth factor gene into cell lines and cultured neurons using a defective herpes simplex virus vector. Transfer of the NGF gene into cells by a HSV-1 vector. B r a i n R e s M o l B r a i n R e s 24(14), 327-35.

McKie EA, MacLean AR, Lewis AD, Cruickshank G, Rampling R, Barnett SC, Kennedy PG and Brown SM (1 9 9 6 ). Selective in vitro replication of herpes simplex virus type1 (HSV-1) ICP34.5 null mutants in primary human CNS tumors- evaluation of a potentially effective clinical therapy. Bri J Cancer 74, 745-752.

Geschwind MD, Lu B and Federoff HJ (1 9 9 4 a ). Expression of neurotrophic genes from herpes simplex virus type 1 vectors: modifying neuronal phenotype. M e t h o d s i n N e u r o s c i e n c e s 21, 462-482. Glorioso JC, De Luca NA and Fink DJ (1 9 9 5 ). Development and application of herpes simplex virus vectors for human gene therapy. A n n u a l R e v M i c r o b i o l 49, 675-710.

Mineta T, Rabkin SD, Yazaki T, Hunter W and Martuza RL (1 9 9 5 ). Attenuated multi-mutated herpes simplex virus-1 for the treatment of malignant gliomas. Nature Med 1(9), 938-943.

Green JM, Noel PJ, Sperling AI, Walunas TL, Gray GS, Bluestone JA and Thompson CB (1 9 9 4 ). Absence of B7 dependant resposnes in CD28-defecient mice. Immunity 1, 501.

Mocarski ES, Post LE and Roizman B (1 9 8 0 ). Molecular engineering of the herpes virus genome: insertion of a second L-S junction into the genome causes additional genome inversions. C e l l 22, 243-255.

Jenkins FJ and Roizman B (1 9 8 5 ). Genetic engineering of novel genomes of large DNA viruses. S c i e n c e 229, 12081214.

Mueller DL, Jenkins MK and Schwartz RH (1 9 8 9 ). Clonal expansion versus functional clonal inactivation: a costimualtory signalling pathway determines the outcome of T cell antigen receptor occupancy. Ann R e v Immunol 7, 445.

Johnston KM, Jacoby D, Pechan PA, Fraefel C, Borghesani P, Schuback D, Dunn RJ, Smith FI and Breakfield XO (1 9 9 7 ). HSV/AAV hybrid amplicon vectors extend transgene expression in human glioma cells. Hum Gene Ther 8, 359-370.

Mule, JJ, Custer M, Averbook B, Yang JC, Weber JS, Goeddel DV, Rosenberg SA and Schall TJ (1 9 9 6 ). RANTES secretion by gene modified tumor cells results in loss of

218

Gene Therapy and Molecular Biology Vol 4, Page 219 tumorigenicity in vivo: role of immune subpopulations. Hum Gene Ther 7, 1545.

cell

Wang S, Di S, Young WB, Jacobson C and Link CJ, Jr. (1 9 9 7 ). A novel herpesvirus amplicon system for in vivo gene delivery. Gene Ther 4(11), 1132-41.

Post LE and Roizman B (1 9 8 1 ). A generalized technique for deletion of specific genes in large genomes: alpha gene 22 of herpes simplex virus 1 is not essential for growth. C e l l 25, 2227-2232.

Ward PL and Roizman B (1 9 9 4 ). Herpes simplex genes: the blueprint of a successful human pathogen. Trends Genet 10, 267-274.

Ramarathinam L, Castle M, Wu Y and Liu Y (1 9 9 4 ). T cell costimulation by B7/BB1 induces CD8 T cell dependent tumor rejection: an important role of B7/BB1 in the induction, recruitment, and effector function of antitumor T cells. J Exp Med 179, 1205.

Wu N, Watkins SC, Schaffer PA and DeLuca NA (1 9 9 6 ). Prolonged gene expression and cell survival after infection by a herpes simplex virus mutant defective in the immediate-early genes encoding ICP4, ICP27, and ICP22. J Virol 70(9), 6358-69.

Samaniego LA, Neiderhiser L and DeLuca NA (1 9 9 8 ). Persistence and expression of the herpes simplex virus genome in the absence of immediate-early proteins. J V i r o l 72(4), 3307-20.

Xu H, Federoff H, Maragos J, Parada LF and Kessler JA (1 9 9 4 ). Viral transduction of trkA into cultured nodule and spinal motor neurons conveys NGF responsiveness. D e v B i o l 163(1), 152-61.

Samaniego LA, Wu N and DeLuca NA (1 9 9 7 ). The herpes simplex virus immediate-early protein ICP0 affects transcription from the viral genome and infected-cell survival in the absence of ICP4 and ICP27. J V i r o l 71(6), 4614-25.

Yates JL, Warren N and Sugden B (1 9 8 5 ). Stable replication of plasmids derived from Epstein-Barr virus in various mammalian cells. Nature 313, 812-5. Yazaki T, Manz HJ, Rabkin SD and Martuza RL (1 9 9 5 ). Treatment of human malignant meningiomas by G207, a replication-competent multimutated herpes simplex virus 1. Cancer Res 55 , 4752-4756.

Sears AE and Roizman B (1 9 9 6 ). Herpes simplex viruses and their replication. Fundamental Virology. B. N. Fields Philadelphia, Lippincott-Raven., 1043-1107. Spaete RR and Frenkel N (1 9 8 2 ). The herpes simplex virus amplicon: a new eukaryotic defective-virus cloningamplifying vector. C e l l 30, 295-304. Stavropoulos TA and Strathdee CA (1 9 9 8 ). An enhanced packaging system for helper-dependent herpes simplex virus vectors. J Virol 72(9), 7137-7143. Stevens JG (1 9 8 9 ). Human herpesviruses: A consideration of the latent state. M i c r o b i o l R e v 53, 318-332. Stevens JG, Wagner EK, Devi-Rao GB, Cook ML and Feldman LT (1 9 8 7 ). RNA complementary to a herpesvirus alpha gene mRNA is prominent in latently infected neurons. S c i e n c e 235, 1056-1059. Toda M, Martuza RL, Kojima H and Rabkin SD (1 9 9 8 a ). In situ cancer vaccination: an IL-12 defective vector/replication- competent herpes simplex virus combination induces local and systemic antitumor activity. J Immunol 160(9), 4457-64. Toda M, Rabkin SD and Martuza RL (1 9 9 8 b ). Treatment of human breast cancer in a brain metastatic model by G207, a replication-competent multimutated herpes simplex virus 1 [In Process Citation]. Hum Gene Ther 9(15), 217785. Tung C, Federoff H, Brownlee M, Karpoff H, Weigel T, Brennan MF and Fong Y (1 9 9 6 ). Rapid production of interleukin-2 secreting tumor cells by herpes virus mediated gene transfer: Implications for autologous vaccine production. Hum Gene Ther 7, 2217-2224.

Joseph D. Rosenblatt

219

Gene Therapy and Molecular Biology Vol 4, page 221 Gene Ther. Mol. Biol. Vol 4, 221-232. December 1999

Quantitative detection of CFTR mRNA in gene transfer studies in human, murine and simian respiratory tissues in vitro and in vivo Research Article

Elena Nicolis1, Paola Melotti1, Anna Tamanini 1, Monika Lusky 2, Majid Methali2, Andrea Pavirani2 and Giulio Cabrini1 1

Laboratory of Molecular Pathology, Cystic Fibrosis Center, Verona, Italy Transgene S.A., 11 rue de Molsheim, 67082 Strasbourg, France __________________________________________________________________________________________________ 2

Correspondence: Dr. Giulio Cabrini, M.D., Laboratorio di Patologia Molecolare, Centro Regionale Fibrosi Cistica, Piazzale Stefani, 1, 37126 Verona, Italy. Tel: +39-045-807 2364; Fax +39-045-807 2840; E-mail: cabrini@linus.univr.it Abbreviations: CF, Cystic Fibrosis; CFTR, Cystic Fibrosis Transmembrane conductance Regulator; RT-PCR, reverse transcriptase polymerase chain reaction Key Words: cystic fibrosis, gene therapy, quantitative mRNA Received: 4 January 1999; accepted 24 January 1999

Summary Pre-clinical and clinical studies aimed to correct the basic defect of cystic fibrosis (CF) by transferring the wild type gene into the airway cells have already shown promising results. One of the main unsolved questions of these studies is the quantitation of the amount of gene required to correct the defect. In this respect, suitable technologies able to measure how much gene is actually transferred into the airway cells are under development. In this article we present a series of application of a single-tube competitive RT-PCR assay to quantitate the vector-encoded CFTR mRNA after gene transfer in human, mouse or monkey respiratory cells. These assays are able to measure very low levels of mRNA, with advantages over traditional expression systems for vector-encoded transcript (Northern analysis) or protein (immunocytochemistry). In some instances they may indicate directly the ratio of vector-encoded versus endogenous transcript, being applicable with both viral-derived and synthetic vectors. The assays presented here are instrumental for future application in gene delivery pre-clinical and clinical trials designed to identify corrective doses of the vector, by providing the basic information on the amount of vectorencoded CF gene transcript expressed in the airway cells. glands of the proximal respiratory tract (Engelhardt et al, 1992) and in ~ 2-5 % of the surface cells of the distal airways (Engelhardt et al, 1994 ). Pre-clinical studies in which CF gene transfer vectors have been delivered to airways of mice or monkeys are providing important information to compare different vectors in terms of safety, level and duration of transgene expression. Also, clinical trials aiming at transferring the wild type gene into the airway cells of CF patients have already given evidence of the expression of the transgene in the cells collected from nasal or lung mucosa a few days after vector administration (Bellon et al, 1997; Caplen et al, 1995; Crystal et al, 1994; Gill et al, 1997; Hay et al, 1995; Knowles et al, 1995, 1998; Porteous et al, 1997; Wagner

I. Introduction Cystic Fibrosis (CF) is a common autosomal recessive disease affecting many organs, pulmonary morbidity being the most life-limiting aspect of the disease. The wild type CF gene encodes for a protein termed Cystic Fibrosis Transmembrane conductance Regulator (CFTR), involved in the transport of Cl- and probably of other molecules across several epithelia. The most common mutation of the CF gene is the deletion of 3 bases encoding phenylalanine at position 508 in the aminoacidic chain (!F508); defective CFTR ultimately leads to chronic lung infection and respiratory insufficiency (Welsh et al, 1995). CFTR is expressed in all serous cells of the submucosal 221

Nicolis et at: Quantitative detection of CFTR transcripts et al, 1998; Zabner et al, 1993, 1996, 1997). Among the different unsolved questions arising from these studies, a key issue is the development of the detection system(s) to quantitate the amount of transgene required to correct the defect in vivo. The measurement of the transmucosal potential difference in vivo can be an efficient tool to distinguish the CF phenotype as a function of ion transport impairment (Knowles et al, 1981), but the application of this technique to the human bronchial mucosa is still confined to a small number of experienced laboratories (Alton and Geddes, 1997). The relationship between CFTR functional correction and the extent of wild type CF gene transfer has been studied also in vitro in CF polarized respiratory cells (Johnson et al, 1992; Zabner et al, 1994 ) and in CF bronchial xenografts (Goldman et al, 1995). These studies agree that complete correction of Cltransport can be achieved with as little as 6-10 % of the CF cells transduced. In this respect, a semi-quantitative evaluation to detect the fraction of cells expressing wild type CFTR protein by classic immunocytochemistry or single cell digital imaging (Renier et al, 1995) can provide useful information. An additional possibility could be to quantitate the wild type CFTR mRNA encoded by the gene transfer vectors. The mRNA approach is particularly interesting in view of the recent understanding that the classic CF phenotype is expressed when the wild type CFTR mRNA decreases below 1-3 % of the normal amount (Estivill, 1996), and that approximately 5 % of normal CFTR mRNA is sufficient to correct the electrical abnormality in CF mouse models (Dorin et al, 1996). In order to provide a suitable technological support to this approach, we developed a series of single-tube competitive mRNA detection assays for both endogenous and vector-encoded CFTR transcripts, both in absolute and relative terms, from human, murine and simian respiratory cell and tissue samples. The evaluation of the expression of the vector-encoded versus the endogenous CFTR transcript can be directly applied to pre-clinical and clinical gene therapy studies utilising different types of viral-derived or synthetic vectors.

in the human tumour respiratory cell line A549, which is negative for endogenous CFTR expression (Renier et al, 1995). A representative electrophoretogram in which different amounts of the CFTR competitor pTG6525 are amplified together with reverse transcribed exogenous CFTR template is shown in Figure 1, panel B. The signal ratio of the pTG6525 competitor over the template is plotted against the number of copies of the competitor itself, as shown in Figure 1, panel C. From the equivalence of the ratio signal of the fitted equation the number of copies of reverse transcribed template can be calculated. Since the DNA competitor is double-stranded while the target cDNA is single-stranded, the value obtained at the equivalence point was multiplied by a factor of 2. In the example reported in Figure 1 , panel C, assuming that reverse-transcription efficiency was equal to 1, the exposure of one plaque-forming unit (pfu) of Ad.CFTR every 10 cells for 24 h produced a steady-state level of 3280 copies of Ad.CFTR-encoded mRNA over 50 ng total RNA. The number of CFTR mRNA transcripts was measured as a function of different Ad.CFTR infection doses, from 0.01 to 1. The plot in Figure 1 , panel D indicates that the assay makes it possible to detect CFTR transcripts in a dose-dependent fashion, which reflects the consistency of the method. The detection limit of this assay has been calculated by a dilution assay in the minimum level of 20 transcript copies, which can be obtained in our experimental conditions by infecting A549 cells at the infection dose of 0.01 MOI. As expected, the detection limit of this transcript assay is much lower than that observed by detecting gene expression by protein analysis, e.g. by the classic APAAP immunocytochemistry assay applied to CFTR protein detection (Cordell et al, 1984; Renier et al, 1995). In parallel experiments of infection of A549 cells with Ad.CFTR in vitro, we found that the lowest infection dose allowing a detectable signal with APAAP assay was higher by two orders of magnitude (1-5 MOI, data not shown).

B. Expression of endogenous CFTR transcript in human airway cells ex vivo

II. Results A. Vector-encoded CFTR transcript expression in human airway cells in vitro

Levels of vector-encoded transcript can be compared in parallel to the endogenous transcript for estimation of the extent of gene replacement. Therefore, single-tube competitive RT-PCR was applied to measure the steadystate copies of endogenous CFTR mRNA in respiratory cells. Surface nasal cells from human subjects were brushed and analysed for endogenous transcript expression.

We addressed the issue of mRNA quantitation by using a PCR assay, in which a homologous competitor differing only a few bases from the target transcript is amplified in the same reaction tube. The principle of the assay is summarised in Figure 1, panel A. Competitive PCR was initially applied to quantitate the absolute amount of CFTR transcript encoded by the adenovirusderived CFTR vector, Ad.CFTR (Rosenfeld et al, 1992), 222

Gene Therapy and Molecular Biology Vol 4, page 223

. Figure 1. Expression of vector-encoded CFTR mRNA in human respiratory cells in vitro. (A) Two consecutive PCR amplifications were performed in the presence of reverse transcribed cDNA (upper fragment), homologous competitor (lower fragment) containing a small insertion (shaded box) and appropriate upstream and downstream primers (solid arrows), as specified in each quantitation. Amplified fragments were labelled by one PCR cycle (run-off) using an internal 5-fluorescently-labeled primer. (B) A549 cells were infected for 24 h with Ad.CFTR at MOI 0.1. The number of copies of competitor template pTG6525 added in the first PCR is indicated on the bottom of each panel. Fragment size related to Ad.CFTR mRNA (290 nt) and pTG6525 (303 nt) is indicated on top. Peak areas are reported close to each peak. (C) Competitor over target areas were plotted on the y-axis versus the number of copies of competitor (x-axis), and fitted by least-square regression analysis to quadratic equations. The copies of target mRNA were quantified as twice as much those of the competitor corresponding to the equivalence of competitor-target areas (y-axis value=1). Representative of 5 independent experiments. (D) Dose-response of CFTR mRNA transfer to A549 cells. The number of copies of CFTR mRNA corresponding to the infection doses of 0.01, 0.1 and 1 MOI were 123, 2640 and 31874, respectively. Representative of 3 independent experiments.

C. Relative expression of wild type vs. !F508 CFTR transcript in human airway cells

Preliminary results showed a great sample to sample variability in the steady-state levels of endogenous CFTR mRNA, suggesting that RNA degradation from these samples obtained ex vivo from nasal mucosa is a major problem. This limitation was overcome by quantitating the Human Cytokeratin - 15 (HCK-15) transcript (Leube et al, 1988), which is constitutively expressed in respiratory epithelial cells. Both CFTR and HCK-15 were coamplified in the presence of their respective homologous competitor templates. Table 1 reports the CFTR/HCK-15 transcript ratio, in 10 non-CF individuals (mean 1.45, SD = 0.74). Therefore, the method in question can provide reliable evaluation of endogenous CFTR transcripts, after normalisation with a reporter gene.

Since the most common mutation of the CF gene is the deletion of the three bases encoding for Phe 508 (!F508), the assay can be made much simpler and more rapid by direct comparison between the amounts of wt and mutated transcript. This is possible for two reasons. First, it should be recalled that the DNA automated sequencer utilised in these assays can discriminate fragments differing only a few bases in length. Second, in the simultaneous presence of the wt and mutated alleles, each transcript competes with and normalizes the other, without the need for a homologous competitor and analysis of a 223

Nicolis et at: Quantitative detection of CFTR transcripts Table 1. Quantitation of endogenous expression in respiratory cells.

reporter gene. This approach was first applied to endogenous transcripts of healthy carriers of the !F508 mutation in nasal respiratory cells, obtained ex vivo by brushing. Figure 2, panel A, shows a representative experiment in which very similar levels of transcription were observed for normal and !F508 CFTR alleles. The almost equivalent areas of the peaks at 100 and 103 nucleotides correspond to mutated and normal CFTR fragments, respectively. Extension of the analysis to other healthy carriers of !F508 mutation gave a wt/!F508 CFTR mRNA ratio of 0.90 ± 0.15 (mean ± SD, 4 subjects tested in triplicate; individual values = 0.75 ± 0.17, 0.92 ± 0.05, 1.05 ± 0.07, 0.86 ± 0.10), which is consistent with results obtained by other investigators utilising a different technique (Trapnell et al, 1991). The second application was to test the direct wt/mutated approach in respiratory cells obtained from nasal polyps, excised from CF patients homozygous for the !F508 mutation, cultured and infected in vitro with Ad.CFTR (72 h at MOI 10). In our experimental conditions, the steady state level of the wt transcript encoded by Ad.CFTR was 3 to 4 times higher than in the case of the endogenous !F508 mutation, as judged by the peak areas shown in Figure 2, panel B. We found a wt/mutated ratio of 3.64 ± 1.27 (mean ± SD, 4 separate cultures from 2 individuals tested in duplicate). Though we consider the direct wt/mutated transcript comparison a simplified application of the competitive method, it seems potentially appropriate for gene transfer experiments as a source of rapid information on relative transcript expression in human clinical trials enrolling CF patients homozygous for !F508 mutation.

Subject

CFTR/HCK-15 ratio of number of copies

1.02

1.80

1.72

0.72

1.65

0.43

2.12

1.11

1.00

2.94

mean

1.45

0.74

CFTR

mRNA

After extraction from 4 x 10 5 cells recovered from non-CF individuals (from A to J) by nasal brushing, total RNA (50 ng) was reverse-transcribed into cDNA in the presence of CFTR11R1 and hCK15-6. CFTR9-D1 plus hCK15-5 primers were added in PCR-1 which was performed in the presence of the competitors pCFTR-!C and phCK15-C. PCR-2 was performed in the presence of CFTR9-D1, CFTR10-11R2, hCK15-D2 and hCK15-R2. Run-off reaction was done with CFTR10-D2 FAM and hCK15-D2 FAM. HCK-15 mRNA was utilised as reporter transcript in order to normalise the absolute amount of CFTR mRNA. Normalized data were provided by the ratio of the number of copies of CFTR/HCK-15 mRNA.

D. Relative expression of vector-encoded vs. endogenous CFTR transcript in mouse lung The direct comparison of vector-encoded and endogenous CFTR mRNA can be usefully applied to preclinical studies in which interspecies homologies can be utilised to amplify with the same set of primers the human vector-encoded and the endogenous mouse CFTR sequences. Similarly to wt and !F508 transcripts, exogenous and endogenous CFTR transcripts can compete each other, and the endogenous transcript directly normalizes for RNA degradation. By choosing a portion of the genes in which the two species differ for at least one base in length, it is possible to distinguish the specific signal by running the samples labelled with a fluorescent primer in a 4-6 % denaturing polyacrylamide gel electrophoresis with the automated sequencer.

transcripts by running the ethidium bromide stained fragments in 2 % agarose gel electrophoresis. In this case, the signal can be collected with a videocamera and quantified by image analysis. For instance, mouse CFTR exon 19 differs from the human homolog at an additional GAA codon. We took advantage of this difference by reverse transcribing vector-encoded and endogenous mouse transcripts expressed with a sequence specific homologous primer annealing to exon 20. The cDNA was amplified with a first PCR cycle (PCR-1) from exon 20 to 17b and a second PCR cycle (PCR-2) from exon 18 to 19. As summarised in Figure 3, panel A, the amplified fragments from human and mouse transcripts differ of 3 bases in

Alternatively, by choosing a portion of the genes in which the restriction pattern generated by an endonuclease enzyme is different in the two species, it is possible to perform a selective digestion and distinguish the different

224

Gene Therapy and Molecular Biology Vol 4, page 225

. length (total lengths were 284 and 287 bp, respectively). The amplified fragments contain restriction sites for the endonuclease Rsa I , allowing the distinction of the human and mouse origin. As shown in Figure 3, panel B, by ethidium bromide stained agarose gel a single band is visible from the human amplification (lane 1), two bands from mouse amplification (lane 2), and the fragments can be easily distinguished in lung samples of animals exposed to the CFTR vector (lane 3). The assay was directly applied to the relative quantitation of the CFTR transcript encoded by two adenovirus-derived vectors instilled in the trachea of SCID mice, in which the quantitation of the CFTR mRNA by Northern analysis was unsuccessful because of low expression level (data not shown). In the example reported in Figure 3, panel C, it has been possible to observe that intratracheal instillation of the Ad.TG5643 vector resulted in higher level of vector-encoded transcript in comparison to Ad.CFTR vector, particularly in the lungs of mice sacrificed at day 3 and 14. In our hands, the assay is rapid and reproducible. Moreover, since the reverse transcription and amplification is performed on a sequence within the CFTR gene, the assay can be applied to the relative quantitation of the exogenous transcripts encoded by any CFTR vector, either viral-derived or synthetic.

E. Relative expression of vector-encoded vs. endogenous CFTR transcript in monkey lung Even though the human-mouse CFTR sequence homologies are quite different from those of the humanmonkey (Tata et al, 1991; Wine et al, 1998), the set of primers and the PCR conditions utilised for relative quantitation of vector-encoded versus mouse endogenous CFTR mRNA can be directly applied also to CFTR transcript from Rhesus monkey (Macaca mulatta). The single-tube competitive RT-PCR assay again generates two fragments of 284 and 287 bp, from human and monkey CFTR transcripts, respectively. The Rsa I digestion of exons 18/19 gives the same restriction pattern, since monkey and human CFTR sequences are highly homologous. On the other side, the fragment amplified from exon 18 to 19 of monkey CFTR contains a unique ScrF I restriction site, generating two fragments of 163 and 124 bp. The relative quantitation assay has been applied to evaluate the level of CFTR gene expression in relation to ICAM-1 transcript induction. We have previously found that steady-state levels of ICAM-1 transcript are increased in the lung portions of Rhesus monkeys instilled by bronchoscope with high doses of the E1-E3-deleted vector Ad.CFTR (Nicolis et al, 1998).

Figure 2. Relative expression of human wt and !F508 CFTR mRNA. (A) Electrophoretogram of endogenous wt and !F508 mutated CFTR transcripts from human nasal respiratory cells obtained ex vivo from a !F508 healthy carrier. (B) Electrophoretogram of the relative expression of vectorencoded wt and endogenous !F508 CFTR mRNA from primary cultures of nasal polyp cells from a CF subject homozygous for !F508 mutation. Cells were infected with Ad.CFTR at MOI 10 for 72 h, as previously described (Renier et al, 1995 ). Peak areas at 103 and 100 nt in the electrophoretogram correspond to the amounts of amplification products of wt and ! F508 CFTR mRNA, respectively.

225

Nicolis et at: Quantitative detection of CFTR transcripts Figure 3. Relative expression of vectorencoded versus endogenous CFTR mRNA in mouse. (A) Schematic drawing of the expected restriction pattern with Rsa I endonuclease enzyme of human vectorencoded and mouse endogenous sequences after RT-PCR. The length of the fragments (bp) is indicated. (B) Ethidium bromide-stained agarose gel electrophoresis of the amplified fragments digested with Rsa I. Lane 1: human CFTR cDNA from Ad.CFTR vector. The restriction fragment of 15 bp is usually not visible; lane 2: mouse CFTR mRNA reverse transcribed and amplified from total RNA extracted from the lungs of a mouse not infected with CFTR vectors; lane 3: total RNA reverse transcribed and amplified from lungs of a mouse infected with AdTG5643 in vivo, containing both endogenous and vectorencoded amplified fragments. The relative migration of the molecular weight marker (M) "X174 cleaved with Hae III is shown (apparent sizes are 1353, 1078, 872, 603, 310, 281/271, 234, 194, 118 and 72 bp). (C) Relative expression of CFTR mRNA from lungs of SCID mice infected with CFTR vectors Ad.CFTR (closed circles) and Ad.TG5643 (open circles) and sacrificed 3, 14 or 30 days after intratracheal instillation. Data are mean Âą SD of determinations performed in 3 different days.

Therefore, the present results confirm the dependence of ICAM-1 mRNA induction as a function of the infection efficiency of the E1-E3-deleted adenovirus-derived vector Ad.CFTR.

To test whether the upregulation was directly related to the efficiency of infection of the vector, we studied ICAM-1 mRNA expression as a function of vectorencoded CFTR mRNA by the relative quantitation assay. Figure 4 shows that ICAM-1 transcript increases in direct relation to the levels of Ad.CFTR-encoded CFTR mRNA. In particular, ICAM-1 transcript levels increase above the basal levels (crosshatched area) in the lung portions in which vector-encoded CFTR mRNA is expressed more than 10 % the levels of Rhesus monkey endogenous CFTR mRNA (vector-encoded/endogenous ratio > 0.1).

III. Discussion Northern blot, slot-blot, RNase protection assay, in situ hybridisation and run-on assay can be utilised for detection of RNA expression (Trapnell, 1993). Unfortunately, these techniques suffer different drawbacks

226

Gene Therapy and Molecular Biology Vol 4, page 227

. in effective quantitation of mRNA transcripts expressed at low level. Most of them require relatively large amounts of RNA, involve cumbersome steps or simply are indicated for qualitative but not quantitative detection. To overcome at least some of these limitations, we designed a further development of a single-tube competitive RT-PCR method (Pannetier et al, 1993), with specific application to wt and mutated CFTR gene transcripts, either endogenous or encoded by a gene-transfer vector in three different species. The main advantages of the method here described are based on the possibility of detecting and linearly quantitating few copies of cDNA derived from reverse transcription of more than one target, without the need of radioactive molecules. In fact, the assay is able to detect as few as 10 initial copies of cDNA and the linearity of the signal ranged over 2 orders of magnitude (data not shown), which is a wider range compared to that obtained by densitometric analysis of autoradiograms (Pannetier et al, 1993). On the other side, the calculation of the number of copies of transcript is conditioned by RNA degradation, efficiency of reverse transcription and specific set-up of PCR.

normalise the data of the target mRNA with those of a reporter mRNA, which is extracted and processed in the same tube. Assuming that under the same experimental conditions both target and reporter transcripts are stable to the same extent, normalisation can cope with the presence of RNA degradation, with the only disadvantage of giving the result in terms of relative and not absolute number of copies. It should be recalled that we utilised HCK-15 as reporter gene (Trapnell et al, 1991), which is a marker expressed in respiratory epithelial cells (Hamosh et al, 1991) on levels similar to those of endogenous CFTR mRNA, rendering the co-amplification convenient for further appropriate and simultaneous computerised analysis. An uncompleted efficiency of reverse transcription (RT) of mRNA into cDNA may also result in underestimation of the transcript copies. Efficiency of RT can be calculated by adding known amounts of RNA transcript, that is an external standard RNA, before reverse transcription. Similarly, for homologous competitors it is possible to utilise alternatively cRNA or cDNA sequences (Becker-AndrĂ¨ and Hahlbrock, 1989; Santagati et al, 1993; Wang et al, 1989). However, if the use of RNA standards or cRNA competitors might overcome the variability of efficiency of cDNA synthesis, they do not correct at the same time for losses of RNA. In this respect, we preferred to assume arbitrarily the maximal efficiency of reverse transcription while controlling RNA degradation, both with normalisation and by choosing cDNA competitors, which are also easily synthesised and accurately measured without need of radiolabeled compounds (Santagati et al, 1993).

Some of our RNA samples were intrinsically prone to degradation because of high RNAase content of the mucosa of the airways. Moreover, RNA can be further degraded during the step of digestion of vector-encoded cDNA. These artefacts induce large sample to sample variability, reduce the sensitivity of the assay and may result in underestimating the absolute number of copies of transcript. Even though the risk was reduced by performing RT reaction immediately after RNA extraction, in some cases we have been obliged to

Figure 4. ICAM-1 induction as a function of vector-encoded CFTR mRNA in Rhesus monkeys. Copies of ICAM-1 mRNA are plotted against the relative expression of vector-encoded over endogenous transcript from total RNA extracted from the lungs of Rhesus monkeys instilled with Ad.CFTR (1.5 x 1010 pfu) as previously reported (Bout et al, 1994; Nicolis et al, 1998). Data are mean Âą SD of both transcript quantitation. Crosshatched area is the 99 % confidence interval of the mean of ICAM-1 expression in lungs of Rhesus monkeys exposed to saline solution instead of Ad.CFTR vectors, as previously reported (Nicolis et al, 1998).

227

Nicolis et at: Quantitative detection of CFTR transcripts CF gene transfer in mouse and Rhesus monkey in vivo

In the present assay, the between-run variation in efficiency of PCR of the target (or reporter) is rendered practically irrelevant since quantitation is based on direct comparison with the respective homologous competitor templates processed in the same reaction tube. Generation of competitor sequences diverging only few bases from the targets is relatively easy either by consecutive digestion, filling and ligation, as described in Material and Methods, or by mutagenesis through PCR (Higuchi, 1989). In most quantitative PCR reaction, amplification must be maintained in exponential phase, which requires difficult and long steps for setting conditions. In case of homologous competitors, amplification is not required to run in the exponential phase. It was previously demonstrated that running to saturation, in the plateau phase, did not affect accuracy of quantitation (BeckerAndrè and Hahlbrock, 1989 ; Pannetier et al, 1993 ; Porcher et al, 1992).

Immunodeficient mice C.B17-scid/scid (Iffa Credo, Lyon, France) were instilled intratracheally with 1.5x10 9 pfu/animal of Ad.CFTR or Ad.TG5643 vectors diluted in a solution of 0.9 % NaCl in a volume of 25 µl. Animals were sacrificed 4, 14 or 30 days post-infection. Lungs portions were snap-frozen before total RNA extraction. Lung portions from Rhesus monkeys subjected to Ad.CFTR instillation by bronchoscope (1.5 x 10 10 pfu) (Bout et al, 1994), in which ICAM-1 induction was previously observed (Nicolis et al, 1998), were analysed for relative vectorencoded versus simian endogenous CFTR mRNA expression.

Primers for reverse transcription and PCR Oligonucleotides were synthesised using the 392 Applied Biosystems DNA/RNA synthesiser, following the Supplier's instructions. FAM (5-carboxyfluorescein) and TAMRA (N,N,N’,N’-tetramethyl-6-carboxyrhodamine) (Applied Biosystems-Perkin Elmer) are fluorescent dye-N-hydroxyl succinimide esters (Dye-NHS ester) that have been conjugated to different oligonucleotides. Oligonucleotides utilised in competitive RT-PCR are reported in Table 1. Primers for endogenous CFTR transcript expression and for relative vectorencoded vs. endogenous expression for human !F508 alleles, murine and simian transcripts are designed inside CFTR gene sequence.

Assuming arbitrarily that in our experimental conditions the stability of target and reporter mRNAs are similar and that the efficiency of RT is maximal, this assay seems a convenient method to quantitate steady-state levels of transcripts expressed in low copies. On the other side, we presented data in which the wild type vectorencoded transcript was competing with the endogenous CFTR transcript itself, as for the human !F508 allele, the mouse and monkey CFTR mRNA. In these cases, the limitation of different RNA degradation and efficiency of reverse transcription between target and competitor should be considered solved, since the endogenous transcript is a normalizer itself. Also, since the primers anneal inside the CFTR gene, the relative exogenous/endogenous assay can be utilised irrespectively from the type of vector utilised, either viral or synthetic, in pre-clinical and clinical studies.

Homologous competitors for human CFTR mRNA and Cytokeratin 15 The competitor phCK15-C was obtained as follows. Total RNA from A549 cells was reverse-transcribed using the primer hCK15-6, as described by Others (Trapnell et al, 1991). A 589bp PCR product was generated with primers hCK15-5 and hCK15-6 (Trapnell et al, 1991), then cloned into the pCRII vector using the TA cloning kit (Invitrogen) as specified by the Supplier. Eight bases were then inserted in the cloned hCK-15 sequence by digestion with Nco I, after which the protruding termini thus created were filled with DNA polymerase I Klenow fragment and ligated. The construct obtained was named phCK15-C. The competitor phCFTR-!C was prepared as follows: a fragment of !F508 CFTR cDNA sequence was reverse-transcribed from total RNA, extracted from epithelial respiratory cells of !F508 homozygous individuals, using the primer hCFTR11-R1. PCR reaction was then performed in the presence of upstream and downstream primers specific for human CFTR sequence (hCFTR7-D1 and hCFTR11-R1, respectively). The PCR product was further amplified with the internal upstream primer hCFTR7-D2 and hCFTR11-R1. A 706bp fragment was generated and cloned in the pCRII vector as described above. The construct obtained was named pCFTR-!C and utilised as a homologous competitor for quantitation of human endogenous wt CFTR mRNA.

IV. Materials and Methods CF gene transfer in vitro A549 cells (American Type Culture Collection, CCL 185, Rockville Pike, MD, USA) and primary cultures of nasal polyp cells were grown as previously described (Renier et al, 1995). The replication-defective adenoviral-derived vectors encoding for wt CFTR gene, Ad.CFTR, which is deleted in the E1 and E3 regions (Major Late promoter), and Ad.TG5643, which is deleted in the E1, E3 and E4 regions (CMV promoter), were prepared and titrated by plaque-forming units (iu) assay as described (Lusky et al, 1998). Infection dose was expressed as Multiplicity Of Infection (MOI), where 1 MOI corresponds to 1 pfu per cell. Primary cultures of nasal polyp cells from a CF subject homozygous for ! F508 mutation were infected in vitro with Ad.CFTR (MOI 10, 72 h), as previously described (Renier et al, 1995).

228

Gene Therapy and Molecular Biology Vol 4, page 229

. Table 2 Sequence of primer oligonucleotides Primer name

Oligonucleotide sequence

Localization

Position

Human CFTR gene (Riordan et al, 1989 ) hCFTR7-D1

5' –GAAGGCAGCCTATGTGAGATAC- 3'

exon 7

1023-1044

hCFTR7-D2

5' –TGTGCTTCCCTATGCACTAATC- 3'

exon 7

1095-2016

hCFTR9-D1

5' –GGACAGTTGTTGGCGGTTGC- 3'

exon 9

1483-1502

hCFTR10-D2- FAM

FAM- 5' –GCCTGGCACCATTAAAGAA- 3'

exon 10

1626-1644

hCFTR11-R1

5' –AGAAATTCTTGCTCGTTGACCTCC- 3'

exon 11

1803-1780

hCFTR10/11-R2

5 '-CTTGGAGATGTCCTCTTCTAGTTG- 3'

exon 10-11

1728-1705

Murine CFTR gene (Mus musculus) (Tata et al, 1991 ) mCFTR17b-D1

5’-TCCATTTTAACAACAG-3’

exon 17b

3473-3488

mCFTR20-R1

5’-CCTGATCCAGTTCTTCC-3’

exon 20

3870-3855

mCFTR18-D2

5’-ATGAATATCATGAGTA-3’

exon 18

3530-3545

mCFTR19-R2

5’-GAAATGTTCTCTAATA-3’

exon 19

3816-3801

mCFTR18/19-D3TAMRA

TAMRA-5’-TGATGCGATCTGTGAG-3’

exon 18-19

3588-3603

Simian CFTR gene (Macaca mulatta) (Wine et al, 1998 ) mCFTR17b-D1

5’-TCCATTTTAACAACAG-3’

exon 17b

3424-3439

mCFTR20-R1

5’-CCTGATCCAGTTCTTCC-3’

exon 20

3821-3805

mCFTR18-D2

5’-ATGAATATCATGAGTA-3’

exon 18

3481-3496

mCFTR19-R2

5’-GAAATGTTCTCTAATA-3’

exon 19

3767-3752

mCFTR18/19-D3TAMRA

TAMRA-5’-TGATGCGATCTGTGAG-3’

exon 18-19

3539-3554

Ad.CFTR vector gene (Senè et al, 1995 ) OTG3042

5' –GCAGTTGATGTGCTTGGCTAGAT- 3'

exon 22

4185-4207

OTG5349

5' –TTGTGAAATTTGTGATGCTATTGC- 3'

SV40 polyA

4696-4673

hCFTR24-D2

5' –CATAGAAGAGAACAAAGTGCGGC- 3'

exon 24

4377-4399

OTG4741 FAM

FAM 5' –GTAACCATTATAAGCTGCAATAAAC- 3'

SV40 polyA

4665-4641

Human cytokeratin 15 (HCK-15) gene (Trapnell et al, 1991 ) hCK15-5

5' –TGAAGGAGTTCAGCAGCCAGCTGG- 3’

804-827

hCK15-6

5' –ACTGACTCTTCTACATTGATGTGG- 3'

1392-1369

hCK15-D2-FAM

FAM- 5’-GCAGAGATGAGGGAGCAGTAC- 3'

888-909

hCK15-R2

5' –GCTGGTCTGGATCATTTCTGTTG- 3'

1016-995

The competitor pTG6525 utilised for Ad.CFTR-encoded transcript, which was obtained by adding 13 bases to the Ad.CFTR cDNA between the 3’ end of hCFTR and the SV40 polyA sequence, was a generous gift of Prof. Jean-Luc Imler.

Reverse transcription Total RNA was extracted by RNAzolTM B (Biotecx, Houston, TX) following the Supplier's instructions. When CFTR gene was transferred with vectors, total RNA was treated for 1 h at 37°C with RNAase-free RQ1 DNAase (10 units/µg of total RNA) (Promega, Madison, WI), in the presence of 800 U/ml RNAase inhibitor (Perkin Elmer, Norwalk, CT), in 40 mM Tris-HCl pH 7.9, 10 mM NaCl , 6

mM MgCl2 and 10 mM CaCl2 (final volume 50 µl), after which the total RNA was ethanol-precipitated. 50 ng of total RNA were reverse transcribed by GeneAmp RNA PCR kit (Perkin Elmer) in 50 mM KCl, 10 mM Tris-HCl, 5 mM MgCl2 , 4 mM dNTP, 15 pmol appropriate downstream oligonucleotides (as reported in Tables 2 and 3), 20 units RNAase inhibitor and 50 U Moloney murine leukaemia virus RT, final volume 20 µl, for 90 min at 42°C.

Amplification conditions The total volume of RT reaction was utilised for PCR-1 in 50 mM KCl, 10 mM Tris-HCl, 2 mM MgCl2 , 0.8 mM dNTP, 15 pmol appropriate upstream primer, 2.5 units AmpliTaq DNA Polymerase

229

Nicolis et at: Quantitative detection of CFTR transcripts (Perkin Elmer), and homologous competitor was added to quantitate and normalise the endogenous human and the Ad.CFTR-encoded transcripts (for primers and competitors refer for details to Table 3). Relative expression of the transcripts (from human !F508 allele, mouse and monkey) did not require addition of exogenous competitors. 3 µl of the PCR-1 were subjected to further amplification (PCR-2) in the same conditions as for PCR-1, except that each sample included 30 pmol of upstream and downstream primers, as detailed in Table 3. The amplification conditions for PCR-1 and PCR-2 were: 80°C for 5 min, then 30 cycles of denaturation (94°C for 30 sec), annealing (30 sec at the temperature indicated in Table 3), and extension (72°C for 30 sec). A run-off reaction (Pannetier et al, 1993) of 4 µl of the final amplification reaction was performed in 10 µl containing 0.1 µM fluorescently-labeled primer, 50 mM KCl, 10 mM Tris-HCl, 200 µM dNTP, 3 mM MgCl2 and 20 U/ml AmpliTaq DNA Polymerase (Perkin Elmer). Run-off reaction started with a denaturation step at 95°C for 2 min, followed by 30 sec at the appropriate annealing temperature (see Table 3)

and finally to 15 min at 72°C.

Electrophoresis and gel analysis 2 µl of run-off reaction products were mixed with 0.5 µl of molecular weight standard GS 500 ROX (Applied BiosystemsPerkin Elmer) and 2.5 µl of a 8.3 mM EDTA/formamide solution. After denaturation at 90°C for 2 min, samples were loaded on a 4-6 % polyacrylamide, 8.3 M urea gel, run for 5 h at 1500 volts / 40 watts in an Applied Biosystems 373A DNA Sequencer. Peak area and length of PCR fragments were determined using Genescan 672 Software (Applied Biosystems-Perkin Elmer). For endogenous hCFTR and Ad.CFTR-encoded transcripts, competitor over target areas were plotted on the y-axis versus the number of copies of competitor (x-axis), and fitted by least-square regression analysis to quadratic equations by Sigma Plot software (Jandel Scientific, Erkrath, Germany). The copies of target mRNA were taken to be those of the competitor corresponding to the equivalence of competitor-target areas (y-axis value = 1).

Table 3. RT-PCR assays, The size of the amplified fragments reported refers in the order to the exogenous and endogenous ones, respectively. RT primer

PCR-1primers

PCR-2 primers

annealing T

Run-off primer annealing T

Run-off

Competitor

fragment size

Endogenous human CFTR mRNA normalized with human Cytokeratin 15 mRNA hCFTR11-R1

hCFTR9-D1

55°C

hCFTR10-11-R2

hCFTR10-D2-FAM

100–103 nt

phCFTR-! C

136-128 nt

phCK15-C

303-290 nt

pTG6525

229-232 nt

none

55°C

55°C hCK15-6

hCK15-5 55°C

hCK15-5

hCK15-D2-FAM

hCK15-R2

55°C

55°C Ad.CFTR-encoded transcript OTG5349

OTG3042 55°C

hCFTR24-D2 OTG5349

OTG4741-FAM

55°C

Human vector-encoded versus Murine (mus musculus) and monkey (macaca mulatta) CFTR transcript mCFTR20-R1

mCFTR17b-D1

mCFTR18-D2

47°C

mCFTR19-R2

mCFTR18/19-D3TAMRA

44°C

43°C

and Rossella Rolfini for critical discussion, to Peter Mead for revising the manuscript, to Angela Bozzoli, Nicoletta Zorzi, Huguette Schultz and Dominique Dreyer for excellent technical assistance. The financial support of the French Association for Cystic Fibrosis (AFLM) and of the “Fondo Riservato Centro Fibrosi Cistica” from the “Azienda Ospedaliera Istituti Ospitalieri di Verona” is

Acknowledgements We are grateful to the CF parents and the healthy subjects who agreed to donate nasal epithelial cells by brushing, to Dr. Abraham Bout for animal experimentation, to Jean-Luc Imler for providing pTG6525, to Alberto Bonizzato, Maria Cristina Dechecchi 230

Gene Therapy and Molecular Biology Vol 4, page 231

. Gill DR, Southern KW, Mofford KA, Seddon T, Huang L, Sorgi F, Thomson A, Macvinish LJ, Ratcliff R, Bilton D, Lane DJ, Littlewood JM, Webb AK, Middleton PG, Colledge WH, Cuthbert AW, Evans MJ, Higgins CF, and Hyde SC (1997). A placebo-controlled study of liposome-mediated gene transfer to nasal epithelium of patients with cystic fibrosis. Gene Ther 4, 199 -209.

gratefully acknowledged.

References Alton EWFW and Geddes D (1997) Gene therapy for cystic fibrosis: steady progress, should do well. Eur Resp J 10, 257-259.

Goldman MJ, Yang Y, and Wilson JM (1995) Gene therapy in a xenograft model of cystic fibrosis lung corrects chloride transport more effectively than the sodium defect. Nat Genet 9, 126-131.

Becker-AndrĂ¨ M and Hahlbrock K (1989) Absolute mRNA quantitation using the polymerase chain reaction (PCR). A novel approach by a PCR aided transcript titration assay (PATTY). Nucleic Acids Res 17, 9437-9446.

Hamosh A, Trapnell BC, Zeitlin PL, Montrose-Rafizadesh BJ, Rosenstein BJ, Crystal RG, and Cutting GR (1991) Severe deficiency of cystic fibrosis transmembrane conductance regulator messenger RNA carrying nonsense mutations R553X and W1316X in respiratory cells of patients with cystic fibrosis. J Clin Invest 88, 1880-1885.

Bellon G, Michel-Calemard L, Thouvenot D, Jagneaux V, Poitevi, F, Malcu, C, Accar, N, Layani MP, Aymard M, Bernon H, Bienvenu J, Courtne, M, Doring G, Gilly B, Gilly R, Lamy D, Levre, H, Morel Y, Paulin C, Perrau, F, Rodillon L, SenĂŠ C, So S, Touraine-Mouli, F, Schatz C, and Pavirani A (1997) Aerosol administration of a recombinant adenovirus expressing CFTR to cystic fibrosis patients: a phase I clinical trial. Hum Gene Ther 8, 15-25.

Hay JG, McElvaney NG, Herena J, and Crystal RG (1995) Modification of nasal epithelial potential differences of individuals with cystic fibrosis consequent to local administration of a normal CFTR cDNA adenovirus gene transfer vector. Hum Gene Ther 6, 1487-1496.

Bout A, Imler J-L, Schultz H, Perricaudet M, Zurcher C, Herbrink P, Valerio D, and Pavirani A (1994) In vivo adenovirus-mediated transfer of human CFTR cDNA to Rhesus monkey airway epithelium: efficacy, toxicity and safety. Gene Ther 1, 385 -394.

Higuchi R (1989) Using PCR to engineer DNA. In PCR technology. Herlich, H. A., editors. Stockton Press, New York NY. 61-70.

Caplen NJ, Alton EWFW, Middleton PG, Dorin JR, Stevenson BJ, Gao X, Durham SR, Jeffery PK, Hodson ME, Coutelle C, Huang L, Porteous DJ, Williamson R, and Geddes D (1995) Liposome-mediated CFTR gene transfer to the nasal epithelium of patients with cystic fibrosis. Nat Med 1, 39 46.

Johnson LG, Olsen JC, Sarkadi B, Moore KL, Swanstrom R, and Boucher RC (1992) Efficiency of gene transfer for restoration of normal airway epithelial function in cystic fibrosis. Nat Genet 2, 21-25. Knowles MR, Gatzy JT, and Boucher RC (1981) Increased bioelectrical potential difference across respiratory epithelia in cystic fibrosis. New Engl J Med 305, 1489-1495.

Cordell JL, Falini B, Erber WN, Ghosh AK, Abdulaziz Z, Macdonald S, Pulford KAF, Stein H, and Mason DJ (1984) Immunoenzymatic labelling of monoclonal antibodies using immune complexes of alkaline phosphatase (APAAP). J Histol Cytochem 32, 219-229.

Knowles MR, Hohneker K, Zhou Z, Olsen JC, Noah TL, Hu PC, Leigh MW, Engelhardt JF, Edwards LJ, Jones KR, Grossman M, Wilson JM, Johnson LG, and Boucher RC (1995) A controlled study of adenoviral-vector-mediated gene transfer in the nasal epithelium of patients with cystic fibrosis. New Engl J Med 333, 823-831.

Crystal RG, McElvaney NG, Rosenfeld MA, Chu, C-S, Mastrangeli A, Hay JG, Brody SL, Jaffe HA, Eissa NT, and Danel C (1994) Administration of an adenovirus containing the human CFTR cDNA to the respiratory tract of individuals with cystic fibrosis. Nat Genet 8, 42-51. Dorin JR, Farley R, Webb S, Smith SN, Farini E, Delaney SJ, Wainwright BJ, Alton EWFW, and Porteous DJ (1996) A demonstration using mouse models that successful gene therapy for cystic fibrosis requires only partial gene correction. Gene Ther 3, 797-801.

Knowles MR, Noone PG, Hohneker K, Johnson LG, Boucher RC, Efthimiou J, Crawford C, Brown R, Schwartzbach C, and Pearlman R (1998) A double-blind, placebo controlled, dose ranging study to evaluate the safety and biological efficacy of the lipid-DNA complex GR213487B in the nasal epithelium of adult patients with cystic fibrosis. Hum Gene Ther 9, 249-269.

Engelhardt JF, Yankaskas JR, Ernst SA, Yang Y, Marino CR, Boucher RC, Cohn JA, and Wilson JM (1992) Submucosal glands are the predominant site of CFTR expression in human bronchus. Nat Genet 2, 240-248.

Leube RE, Bader BL, Bosch FX, Zimbelmann R, Achtstaetter T, and Franke WW (1988) Molecular characterization and expression of the stratification-related cytokeratins 4 and 15. J Cell Biol 106, 1249-1261.

Engelhardt JF, Zepeda M, Cohn JA, Yankaskas JR, and Wilson JM (1994) Expression of the cystic fibrosis gene in adult human lung. J Clin Invest 93, 737-749.

Lusky M, Christ M, Rittner K, Dieterle A, Dreyer D, Mourot B, Schultz H, Stoeckel F, Pavirani A, and Methali M (1998) On vitro and in vivo biology of recombinant adenovirus vectors with E1, E1/E2A, or E1/E4 deleted. J Virol 72, 2022-2032.

Estivill X (1996) Complexity in a monogenic disease. Nat Genet 12, 348-350.

Nicolis E, Tamanini A, Melotti P, Rolfini R, Berton G, Cassatella MA, Bout A, Pavirani A, and Cabrini G (1998)

231

Nicolis et at: Quantitative detection of CFTR transcripts ICAM-1 induction in respiratory cells exposed to a replication- deficient recombinant adenovirus in vitro and in vivo. Gene Ther 5, 131-136.

Trapnell BC (1993) Quantitative evaluation of gene expression in freshly isolated human respiratory epithelial cells. Am J Physiol 264, L199-L212.

Pannetier C, Delassus S, Darche S, Saucier C, and Kourilsky P (1993) Quantitative titration of nucleic acids by enzymatic amplification reaction run to saturation. Nucleic Acids Res 21, 577-583.

Trapnell BC, Chu C-S, Paakko PK, Banks TC, Yoshimura K, Ferrans VJ, Chernik MS, and Crystal RG (1991) Expression of the cystic fibrosis transmembrane conductance regulator gene in the respiratory tract of normal individuals and individuals with cystic fibrosis. Proc Natl Acad Sci USA 88, 6565-6569.

Porcher C, Malinge M-C, Picat C, and Grandchamp B (1992) A simplified method for determination of specific DNA or RNA copy number using quantitative PCR and an automated DNA sequencer. Biotechniques 13, 106-113.

Wagner JA, Reynolds T, Moran ML, Moss RB, Wine JJ, Flotte TR, and Gardner P (1998) Efficient and persistent gene transfer of AAV-CFTR in maxillary sinus. Lancet 351, 1702-1703.

Porteous DJ, Drin JR, Mclachlan G, Davidson-Smith H, Davidson H, Stevenson BJ, Carothers AD, Wallace WAH, Moralee S, Hoenes C, Kallmeyer G, Michaelis U, Naujoks K, Ho LP, Samways JM, Imrie M, Greening AP, and Innes JA (1997) Evidence for safety and efficacy of DOTAP cationic liposomes mediated CFTR gene transfer to the nasal epithelium of patients with cystic fibrosis. Gene Ther 4, 210-218.

Wang AM, Doyle MV, and Mark DF (1989) Quantitation of mRNA by the polymerase chain reaction. Proc Natl Acad Sci USA 86, 9717-9721. Welsh MJ, Tsui L-C, Boat TF, and Beaudet AL (1995) Cystic fibrosis. In The metabolic and molecular bases of inherited disease. Scriver, C. R., Beaudet, A. L., Sly, W. S., and Valle, D., editors. McGraw-Hill Inc, New York, NY. 3799-3876.

Renier M, Tamanini A, Nicolis E, Rolfini R, Imler J-L, Pavirani A, and Cabrini G (1995) Use of a membrane potentialsensitive probe to assess biological expression of the cystic fibrosis transmembrane conductance regulator. Hum Gene Ther 6, 1275-1283.

Wine JJ, Glavac D, Hurlock G, Robinson C, Lee M, Potochik U, Ravnik-Glavac M, and Dean M (1998) Genomic DNA sequence of Rhesus (M. mulatta) cystic fibrosis (CFTR) gene. Mamm Genome 9, 301-305.

Riordan JR, Rommens JM, Kerem BS, Alon N, Rozmahel R, Grzelczak Z, Zielenski J, Lok S, Plavsic N, Chou JL, Dumm ML, Iannuzzi MC, Collins FS, and Tsui L-C (1989) Identification of the cystic fibrosis gene: cloning and characterization of complementary DNA. Science 245, 1066 1073.

Zabner J, Cheng SH, Meeker D, Lanspach J, Balfour R, Perricone MA, Morris JE, Marshall J, Fasbender A, Smith AE, and Welsh MJ (1997) Comparison of DNA-lipid complexes and DNA alone for gene transfer to cystic fibrosis airway epithelia in vivo. J Clin Invest 100, 1529-1537.

Rosenfeld MA, Yoshimura K, Trapnell BC, Yoneyama K, Rosenthal ER, Dalemans W, Fukayama M, Bargon J, Stier LE, Stratford-Perricaudet L, Perricaudet M, Guggino WB, Pavirani A, Lecocq J-P, and Crystal RG (1992) In vivo transfer of the human cystic fibrosis transmembrane conductance regulator gene to the airway epithelium. Cell 68, 143-155.

Zabner J, Couture LA, Gregory RJ, Graham SM, Smith AE, and Welsh MJ (1993) Adenovirus-mediated gene transfer transiently corrects the chloride transport defect in nasal epithelia of patients with cystic fibrosis. Cell 75, 207-216. Zabner J, Couture LA, Smith AE, and Welsh MJ (1994) Correction of cAMP-stimulated fluid secretion in cystic fibrosis airway epithelia: efficiency of adenovirus-mediated gene transfer in vitro. HumGene Ther 5, 585-593.

Santagati S, Bettini E, Asdente M, Muramatsu M, and Maggi A (1993) Theoretical considerations for the application of competitive polymerase chain reaction to the quantitation of a low abundance mRNA: estrogen receptor. Biochem Pharmacol 46, 1797-1803.

Zabner J, Ramsey BW, Meeker DP, Aitken ML, Balfour RP, Gibson RL, Launspach J, Moscicki RA, Richards SM, Standaert TA, Williams-Warren J, Wadsworth SC, Smith AE, and Welsh MJ (1996) Repeat administration of an adenovirus vector encoding cystic fibrosis transmembrane conductance regulator to the nasal epithelium of patients with cystic fibrosis. J Clin Invest 97, 1504-1511.

SenĂ¨ C, Bout A, Imler J-L, Schultz H, Willemot JM, Hennebel V, Zurcher C, Valerio D, Lamy D, and Pavirani A (1995) Aerosol-mediated delivery of recombinant adenovirus to the airways of non-human primates. Hum Gene Ther 6, 1587 1593. Tata F, Stanier P, Wicking C, Halford S, Kruyer H, Lench NJ, Scrambler PJ, Hansen C, Braman JC, Williamson R, and Wainwright BJ (1991) Cloning the mouse homolog of the human cystic fibrosis transmembrane conductance regulator gene. Genomics 10, 301-307.

232

Gene Therapy and Molecular Biology Vol 4, Page 233 Gene Ther Mol Biol Vol 4, 233-248. December 1999.

Gene therapy for prostate cancer Review Article

Mitchell S Steiner and Jeffrey R Gingrich Department of Urology, University of Tennessee, Memphis, Tennessee 38163 _________________________________________________________________________________________________ Correspondence: Mitchell S. Steiner, MD, Department of Urology, University of Tennessee Medical Center, 956 Court Avenue, Memphis, Tennessee 38163. Telephone: 901-448-1492; Fax: 901-448-4743; E-mail: MSteiner@utmem.edu Abbreviations: GDEPT, gene directed enzyme prodrug treatment; 6 MPDR, 6-methyl-9 (2 deoxy-!-D erythro-pentofuranosyl) purine; Ad, adenovirus; CMV, cytomegalovirus; MMTV, mouse mammary tumor virus; HSV, herpes simplex virus; tk, thymidine kinase; GCV, ganciclovir; DT, diptheria toxin; EBRT, external beam radiation treatment; PSA, prostate-specific antigen; PSE, prostate specific enhancer; ARE, androgen responsive element; PB, probasin; MHC, major histocompatability complex; TIL; tumor infiltrating lymphocytes; CTL, cytotoxic T cells; IL, interleukin; GM-CSF, granulocyte maturation-colony stimulating factor; bFGF, basic fibroblast growth factor Key Words: Prostate cancer, gene therapy, gene replacement, immunotherapy, oncolytic virus, suicide gene, prostatectomy, prodrug Received: 7 July, 1999; accepted 15 July 1999

Summary The advent of recombinant DNA technology has sparked the age of molecular medicine. The ability to deliberately recombine pieces of DNA and then transfer these specific genes into diseased cells has revolutionized medical research. In fact, the ability to modify these genes in the living person is now possible. Several innovative approaches are being developed to circumvent the limitations of current vectors including more effective delivery routes for gene therapy, the incorporation of tissue specific promoters and other enhancers into vectors, and increasing cell death by a phenomenon known as the bystander effect. Gene therapy strategies are rapidly evolving as new gene targets, better vectors and improved gene expression systems become available. Innovative gene therapy strategies currently being employed for the treatment of prostate cancer include: immunotherapy, gene corrective therapy, exploitation of programmed cell death therapy, gene therapy to target critical biological functions of the cell, suicide gene therapy, oncolytic virus gene therapy, and finally combination gene therapy. At this time, 17 gene therapy trials have been approved by the NIH for the treatment of prostate cancer. Overall, current gene therapy to treat advanced localized prostate cancer has been shown to be safe and feasible. There are many challenges that lie ahead for gene therapy. Nonetheless, it is almost certain that gene therapy will be part of the armamentarium against prostate cancer and other human diseases in the next century.

I. Introduction Prostate cancer is the most frequently diagnosed malignancy and the second leading cause of cancer deaths in American men today with an estimated 179,300 new cases of prostate cancer and 37,000 deaths predicted this year (Landis et al, 1999). The risk of prostate cancer rises steeply with age and will continue to increase by 3-4% each year in older men as fewer men are dying from cardiovascular diseases (Walsh, 1994). Despite concerns that increased detection of early prostate cancer by the wide spread use of serum PSA would lead to many more patients being treated unnecessarily for small indolent cancers, no change in the proportion of such cases has been observed by many large medical centers between 1983-1996 (Soh et al, 1997). In fact, the majority of patients that are carefully selected for treatment of clinically localized disease by radical

233

prostatectomy are found pathologically to have advanced localized disease. Badalament et al (1996) evaluated 4 large prostatectomy series totaling 5,661 patients and found that 55% of patients indeed had extracapsular disease at the time of surgery. Locally advanced prostate cancer, defined as the presence of extracapsular extension, increases the likelihood of positive surgical margins at radical prostatectomy and portends a poor prognosis. Re-examination of the role radical prostatectomy as monotherapy for T3 disease suggests that it is really not curative as the 10 and 15 year survival following radical prostatectomy are 12-60% and 2028%, respectively. The local recurrence rates are as high as 41% by 5 years following radical prostatectomy (Partin et al, 1993). Hence, radical prostatectomy alone is not curative in the majority of patients with significant extracapsular disease. It is also a well established fact that testosterone, or

Steiner and Gingrich: Gene Therapy for Prostate Cancer hormonal deprivation alone does not cure prostate cancer (Schroder, 1995). Neoadjuvant hormone deprivation has been recently used to “down size” or “down stage” locally advanced prostate cancer prior to radical prostatectomy in an attempt to improve the chances of achieving local cancer control. Although earlier studies have shown a reduction in positive surgical margins rate (Wieder and Soloway, 1998), there has been no change in the rate of PSA recurrence either from a retrospective analysis (Wood et al, 1997) or by prospective randomized studies with 2 years follow-up (Goldenberg et al, 1997; Soloway et al, 1997). Thus, neoadjuvant hormone deprivation has not been shown to alter tumor progression or survival rates (Abbas et al, 1996; Cookson and Fair, 1997; Goldenberg et al, 1997; Soloway et al, 1997). External beam radiation treatment (EBRT) alone also has a high local failure rate in advanced prostate cancer. Zagars et al (1991) have reported a rising serum PSA following EBRT in 17% of patients with a PSA less than 40 ng/ml and as high as 60% in patients whose PSA is > 40 ng/ml. Overall, Holzman et al have shown a 53% local recurrence rate by 8 years after EBRT (Holzman et al, 1991). However, pathologically the rate of local control is even more disappointing. The Stanford series found, that after a mean follow-up of 17 months, following definitive EBRT, greater than 60% of patients had a rising serum PSA indicating cancer progression (Stamey and McNeal, 1992) and over 90% of these patients had a positive biopsy for prostate cancer (Kabalin et al, 1989). Even more alarming, the grade of the recurrent prostate cancer has been shown to be higher than the original cancer (Cumming et al, 1990; Wheeler et al, 1993). The reported disease free survival for T3 disease is 64% at 5 years, 10-35% at 10 years, and 1518% by 15 years (Bagshaw, 1993; Schellhammer and Lynch, 1997). If a serum PSA criterion is used then the biochemical failure rate exceeds 90% at 10 years for stage T3 prostate cancer (Schellhammer and Lynch, 1997). Thus, surgery, radiation, or hormone deprivation alone will not be adequate enough to locally control clinical or pathologic stage T3 prostate cancer which will ultimately lead to a higher incidence of morbidity, distant metastasis, and decreased survival (Schellhammer and Lynch, 1997). Clearly, other novel therapies for this devastating and common disease are desperately needed to achieve long term local cancer control. The focus of new therapies should be to intervene at the cellular level as a way to locally directly affect prostate cancer cells in way not possible by current standard therapies. As it is the androgen independent prostate cancer cells that eventually kill the patient (Isaacs, 1995), any strategy that will modify the biologic behavior of these cells may potentially have the most significant clinical impact to achieve local cancer control.

II. Gene therapy and prostate cancer There are several features of prostate cancer that make it a particularly useful model to study gene therapy: 1) Prostate cancer is common and for the majority of patients who are diagnosed with advanced disease there is no cure. 2) The prostate gland produces over 500 unique gene products and each specific prostate antigen, or protein may be exploited for vector targeting or gene vaccine immunization. Moreover, these prostate specific promoters and other enhancers that direct transcription of these prostate unique proteins may also be incorporated into vectors to direct prostate specific expression of therapeutic genes (Simons et al, 1999). 3) The prostate gland does not serve any critical life sustaining functions which obviates the need to distinguish between normal and cancerous prostate tissues for targeted gene therapy. 4) The prostate gland is also easily accessible by transurethral, transperineal, and transrectal approaches for intratumoral administration of gene therapy. 5) The prostate may be easily evaluated by transrectal ultrasound, digital rectal examination, and other standard radiologic imaging (magnetic resonance imaging and computer tomography) following gene therapy. 6) The pattern of prostate cancer spread is also predictable; it spreads to pelvic lymph nodes and then to the axial skeleton. This allows for the development of gene therapy strategies to purge the bone marrow of malignant cells or develop vectors that have tropism for bone (Kim et al, 1997; Ko et al, 1996; Malkowicz and Johnson, 1998). 7) Although controversial, prostate cancer gene therapy may be followed by serum PSA. Even though it should not be an endpoint by which to make decisions, it may serve as a surrogate marker of prostate cancer progression (Cech and Bass, 1986; Partin and Oesterling, 1994). Prostate cancer also poses some interesting challenges for current gene therapy technology as the doubling time of prostate cancer is usually greater than 150 days with less than 5% of cells actively dividing at any one time (Berges et al, 1993). By having such a low proliferative index, vectors are required that are capable of providing high gene transfer independent of cell division. Another concern is that prostate cancer is a heterogeneous cancer, not only from individual to individual, but also within the same individual. For example, the underlying genetic mutations that dictate the phenotype of primary tumor may not be identical to those that are responsible for metastasis (Isaacs, 1995). Thus, corrective gene therapy approaches to correct leading gene mutations for one individual may not necessarily be useful to treat another individual or a different lesion in the same patient.

III. Prostate targeted delivery of gene therapy Current vector technology cannot achieve the ultimate

234

Gene Therapy and Molecular Biology Vol 4, Page 235 goal of in vivo cancer gene therapy; that is, to administer a vector systemically that will result in the expression of therapeutic gene exclusively in 100% of target cells. Several innovative approaches are being developed to circumvent these limitations of current vectors including more effective delivery routes for gene therapy, the incorporation of tissue specific promoters/enhancers into vectors, and increasing cell death by enhancing the phenomenon known as the bystander effect.

A. Delivery approaches Theoretically, metastatic disease may only be treated by the systemic delivery of gene vectors. However, for locally advanced prostate cancer (stages T3 and T4), variations in delivery strategy may help target gene therapy vectors to the prostate. Lu et al (1999) compared the effectiveness of gene transfer by an adenoviral vector containing !-galactosidase reporter gene delivered by three routes of administration: intravenous, intraarterial, and intraprostatic injection in the canine model. The intraarterial approach was performed by cannulating the internal iliac artery and threading the catheter to the inferior vesicle and prostatic arteries. The transduction efficiencies of adenovirus !-gal were assessed by X-gal staining of prostate tissue sections, by colorimetric !-gal assay, and by PCR. The intraprostatic administration of the adenoviral vector resulted by far in the highest gene transfer rate with the least adenoviral systemic dissemination (Lu et al, 1999; Steiner et al, 1999). This study provides direct support for the use of intratumoral prostatic injections of gene therapy vectors for prostate cancer gene therapy trials either by the transrectal route or by the transperineal approach under ultrasound guidance (Steiner et al, 1998).

B. Tissue specific promoters Viral vectors will transfer therapeutic genes to any mammalian cell exposed to that vector. One way for vectors to theoretically target only prostate cells is by incorporating a prostate tissue specific promoter and/or enhancer that will limit the expression of the therapeutic gene to prostate cells. Only prostate cells will have the appropriate complement of transcription factors to activate the prostate specific promoter, and thus, the therapeutic gene will be expressed only in prostate cells. This assumes, however, that this prostate tissue specific promoter/ gene vector expression system is tightly controlled and not active or â&#x20AC;&#x153;leakyâ&#x20AC;? resulting in the inadvertent expression of the therapeutic gene in unintended tissues. Unfortunately, prostate promoters currently employed in gene therapy are leaky to some degree, but what is not known is whether this low level of expression in non-prostate tissues is clinically relevant. Prostate epithelial cell promoters that have been used for gene therapy include prostate specific antigen (PSA)

235

promoter, (Dannull and Belldegrun, 1997; Dannull et al, 1999; Pang et al, 1997; Pang et al, 1995; Steiner et al, 1999) probasin (PB) promoter (Greenberg et al, 1994; Zhang et al, 1998), mouse mammary tumor virus (MMTV) promoter, (Muller et al, 1990; Steiner et al, 1998; Tutrone et al, 1993) and the prostate specific membrane antigen promoter (Israeli et al, 1993). The PSA promoter has been most commonly employed in vector constructs (Dannull and Belldegrun, 1997; Dannull et al, 1999). One theoretical concern is that the PSA promoter requires the presence of both functional androgen receptors and circulating androgens to be active. The majority of patients who have advanced and hormone refractory prostate cancer, however, are also undergoing androgen deprivation therapy which may not optimally activate the PSA promoter. Addressing this potential obstacle, Gotoh et al (1998) have shown that a better promoter is the long PSA promoter (5837 bp) which is more active than the short PSA promoter (631 bp) both in an androgen depleted environment and in androgen insensitive cells in vitro. Another strategy is to be able to preserve the tissue specificity of the promoter like the PB promoter, but be able to activate the promoter with another hormone. Zhang et al (1998) have developed a retroviral construct containing a glucocorticoid responsive element upstream of the androgen responsive element (ARE). This allows the activation of the PB promoter with dexamethasone. Similarly, Rodriguez et al (1999) demonstrated that the androgen sensitive probasin promoter may also be activated by phenylbutyrate in the absence of androgens. Thus, gene therapy using constructs containing modified prostate specific promoters may be used to treat patients who have advanced prostate cancer and are concomitantly receiving androgen deprivation therapy. Another experimental observation is that the activity of prostate specific promoters tend to be less than that of other non specific or viral promoters (Steiner et al, 1999). There appears to be an inverse relationship between promoter activity and tissue specificity. In the canine prostate model, PSA, PB, and MMTV promoters are prostate specific, but have a 10 to 100 fold less activity than the Rous sarcoma virus promoter in vivo (Steiner et al, 1999). To help circumvent this problem, Pang et al (1997) have cloned a mutated PSA promoter (PCPSA) from a prostate cancer patient who had a high serum PSA. The PCPSA promoter has 50-fold greater activity than the wild type PSA promoter. Similarly, upstream regulatory sequences of the PSA promoter referred to as prostate specific enhancer (PSE) sequences were cloned from normal and cancerous prostate tissue. In vitro, PSE sequences increased the PSA promoter activity by 72 fold when isolated from normal prostate compared to 1000 fold when cloned from prostate cancer tissue (Dannull and Belldegrun, 1997). Recently, Dannull et al (1999) incorporated the PSE (822 bp) and PSA promoter (611 bp) into an E1 deleted adenoviral vector.

Steiner and Gingrich: Gene Therapy for Prostate Cancer Intratumoral injection of this vector into a variety of different human prostate xenografts growing subcutaneously in SCID mice resulted in PSA promoter activity that was not as robust as seen in vitro. In fact, the promoter activity was markedly less than the CMV viral promoter in the same system. Another tactic uses gene therapy to treat the supporting bone stromal cells in an effort to eradicate prostate epithelial cells metastatic to bone. Using the osteocalcin promoter which is active in bone stroma including osteoblasts (Ko et al, 1996), prostate cancer gene therapy may be directed to the bony metastatic sites. Ko et al (1996) have shown that osteosarcoma tumors are inhibited following intratumoral injection with adenoviral vector composed of the osteocalcin promoter controlling thymidine kinase (tk) followed by ganciclovir (GCV). Moreover, the osteocalcin promoter is active in spontaneous canine prostate cancer bony metastasis suggesting that the osteocalcin may be useful for targeting bone metastasis in humans (Ou et al, 1999). Combinations of prostate specific antigen enhancers and other prostate specific epithelial and stromal promoters are currently under intense investigation. Whether or not any of these promoter combinations will be ultimately effective in the systemic treatment of prostate cancer with the required level of promoter activity remains to be shown.

C. Bystander effect A unique observation that became apparent only after the application of gene therapy in clinical in vivo models is the bystander effect. The bystander effect occurs when more cells are destroyed or biologically altered following gene therapy than would have been predicted by transduction rates alone. This is fortuitous as no one vector system is currently available that can transfer therapeutic genes into 100% of target cells; thus, the ability to affect more cells than just those that have been transfected makes gene therapy more clinically applicable. There are several hypotheses to explain the mechanism of the bystander effect: 1) There is intratumoral cell to cell transfer of the therapeutic gene, gene product, or gene activated toxic prodrug from cell to cell by cellular vesicles and endocytosis or by diffusion through gap junctions and cell channels. 2) The therapeutic gene or vector antigens may induce an intense immune response which contributes to cell kill. 3) Tissue ischemia resulting from endovascular injury secondary to toxic gene product or non-specific immunologic response (Simons and Marshall, 1998). Although the exact mechanism is not known, the bystander effect is a real clinical phenomenon that may help compensate in part for the inefficient in vivo gene transfer limitations of the currently available vectors.

236

IV.Gene therapy strategies to combat prostate cancer Gene therapy strategies are also rapidly evolving with new gene targets, better vectors and improved gene expression systems. Initially, ex vivo gene therapy, the transfer of genes into cells growing outside the body in tissue culture, was the primary approach. Primarily because of the new viral-based gene therapy technologic advances, in vivo gene therapy, the ability to transfer genes into cells that are still part of a living organism, has also become possible. The innovative gene therapy strategies that are being currently employed for the treatment of prostate cancer include: immunotherapy, corrective gene therapy, exploitation of programmed cell death gene therapy, gene therapy to target critical biological functions of the cell, suicide gene therapy, oncolytic virus gene therapy, and combination gene therapy.

A. Immunotherapy Class I major histocompatability complex (MHC) proteins are critical for appropriate antitumor immunologic responsiveness. Alterations or loss of Class I MHC is one common way that prostate cancer cells may evade the hostâ&#x20AC;&#x2122;s immune system (Blades et al, 1995; Sanda et al, 1995). Several approaches have been used to stimulate or augment the bodyâ&#x20AC;&#x2122;s own antitumor immune response to essentially circumvent the loss of the critical Class I MHC proteins. Four general immunologic approaches have evolved for the immunotherapy of prostate cancer: autologous or nonautologous gene vaccine therapy using ex vivo gene transfer techniques, direct in vivo intratumoral injection of gene therapy vectors containing cytokine genes, adoptive immunotherapy to treat effector immune cells such as dendritic cells, tumor infiltrating lymphocytes (TIL), or cytotoxic T cells (CTL) by ex vivo gene transfer techniques, and lastly, cytokine immunotherapy, which is not truly gene therapy as the patient is treated systemically with purified cytokines such as Interleukin 4 (IL-4), IL-2, granulocyte maturation-colony stimulating factor (GM-CSF), or B7 (Dannull and Belldegrun, 1997).

1. Gene vaccine therapy Autologous tumor or fibroblast cells harvested by patient biopsy or surgical specimen, or alternatively nonautologous cells are modified by ex vivo gene therapy with genes that encode for immunity enhancing proteins (Fisher et al, 1989). The genetically altered cells are irradiated to destroy their capacity to replicate prior to subcutaneous patient re-inoculation. CTL not only recognize tumor specific antigens present on the surface of these inoculated, irradiated cells, but also they are induced by local secretion of the transferred stimulatory cytokine.

Gene Therapy and Molecular Biology Vol 4, Page 237 The activated CTL cells expand, target, and destroy tumor cells that share these same antigens on their cell surface throughout the body (Sikora and Pandha, 1997). Gene therapy using cytokine genes, IL-2 and GM-CSF, theoretically will stimulate antitumor responses independent of Class I MHC levels, alternatively utilizing class II MHC expression and natural killer cell mediated tumor lysis (Fearon et al, 1990; Golumbek et al, 1991). In contrast, TNF-" and IFN-# depend on Class I MHC proteins which are altered in prostate cancer to mediate their immune effects (Catalona et al, 1991; Lange et al, 1989). Preclinical studies have clearly shown that gene vaccines are not efficacious in the presence of a large tumor burden, but theoretically may be more useful against micrometastasis following debulking of the primary tumor (Simons et al, 1998a, 1999a,b). Other concerns about autologous prostate cancer gene vaccines include the fact that prostate cancer tissue may not be resectable, the tumor tissue may not be easily cultured, poor cytokine gene transfer efficiencies, and the inability to produce large scale yields of genetically modified tumor cells (Jaffee et al, 1993). Gene vaccines that have shown efficacy against prostate cancer in animal models include gene transfer of cytokine genes GM-CSF (Sanda et al, 1994), IL-2 (Kawakita et al, 1997; Vieweg et al, 1994), IFN-gamma, TNF-, and B7 (Kawakita et al, 1997). Cytokine producing cancer gene vaccines have been evaluated for IL-2, GM-CSF, and IFN-# in MAT LyLu cells grown in Copenhagen rats (Moody et al, 1994; Sanda et al, 1995; Schmidt et al, 1995; Vieweg et al, 1994; Yoshimura et al, 1996). The retroviral delivery of IL2 or GM-CSF inhibited prostate cancer tumors and increased animal survival with up to 30% of animals cured when treated with GM-CSF gene vaccines (Gansbacher et al, 1990; Vieweg and Gilboa, 1995; Vieweg et al, 1994; Yoshimura et al, 1996). Similar results have been reported for ex vivo IL-2 liposomal gene vaccine therapy (Vieweg et al, 1995). In contrast, Kawakita et al (1997) used canary pox virus for utilizing genes IL-2, IFN-#, TNF-", or B7 and found that only TNF-" or IL-2 delayed RM1 tumors in C57BL/6 mice (Kawakita et al, 1997). Another approach alters prostate cancer cells or immune cells by ex vivo gene transfer of genes encoding tumor specific antigens. When inoculated, these modified prostate cells recruit immune effector cells capable of eliciting a wide array of immunologic antitumor responses and thereby sensitize the hostâ&#x20AC;&#x2122;s immune system against these newly introduced prostate tumor specific antigens (Dannull and Belldegrun, 1997; Naitoh et al, 1998). Tumor specific antigens that appear to have prostate cancer specificity include PSA, PSMA (Israeli et al, 1993), GAGE-7 (Chen et al, 1996), PAGE (Chen et al, 1996), TAG-72 (Chen et al, 1996), prostate mucin antigens, mucin-1 (MUC-1) (Apostolopoulos and McKenzie, 1994, Carrato, et al, 1994), and mucin-2 (MUC-2) (Dannull and Belldegrun, 1997, 237

Gambus, et al, 1993, Naitoh, et al, 1998, Price, et al, 1993). Lubaroff et al (1999) have demonstrated that PSA producing adenoviral vector induced potent antitumor immunity in vivo mediated by both CTL and humoral immune responses. Overall, cancer gene vaccines whether cytokine producing or presenting tumor specific antigens, aim to enhance the induction of T cell immunity to eradicate prostate cancer micrometastases.

2. Direct in vivo intratumoral injection of gene therapy vectors containing cytokine genes This second immunotherapy approach treats the tumor directly by intratumoral injection of vectors containing cytokine genes. Utilizing animal models of prostate cancer, Naitoh et al (1998) have shown that liposome and adenoviral vectors containing the IL-2 gene produced IL-2 following intratumoral injection resulting in the activation of specific T cell antitumor responses. Similarly, Sanford et al (1999) have employed adenoviral vector containing IL-12 (Ad IL12) to intratumorally inject primary prostate cancer tumors in mice which significantly reduced the number of lung metastasis. The molecular mechanism may include stimulation of T cell and NK cells, induction of IFN-#, and upregulation of fas expression (Hyer et al, 1999; Sanford et al, 1999). Phase I clinical trials utilizing IL-2 gene transfer vectors for intratumoral injection of prostate cancer are nearing completion (Naitoh and Belldegrun, 1998).

3. Adoptive immunotherapy Effector immune cells such as dendritic cells, TIL, or CTL cells are genetically modified by ex vivo gene transfer techniques. This form of therapy for prostate cancer is still in its infancy. The difficulty with this approach in prostate cancer has been the ability to selectively obtain specific effector cell types from the patient, ex vivo amplification of the effector cells, and ex vivo gene transfer of specific biological modifiers such as cytokine genes. In summary, immunotherapy may ultimately be effective against micrometastatic disease, but patients with greater metastatic or primary tumor volumes will require some other non-immunologic therapeutic intervention. Another concern is the theoretical possibility that by autologous or nonautologous cells producing low levels of tumor associated antigens they may paradoxically induce immune tolerance which would suppress the hostâ&#x20AC;&#x2122;s immunity against prostate cancer. Finally, tumor antigens in general are also very weak inducers of the immune system. Thus, it is only through well-designed clinical trials employing gene immunotherapy that these critical questions may be answered.

Steiner and Gingrich: Gene Therapy for Prostate Cancer

B. Corrective gene therapy Corrective gene therapy seeks to replace inherited or acquired defective genes which are important for normal growth regulation of the cell cycle. The molecular components of the cell cycle targeted include protooncogenes, tumor suppressor genes, and growth factors and their receptors. Since prostate cancer is estimated to be a consequence of an average of 5 genetically accumulated mutations, it is hard to conceptualize that the correction of any single gene alteration would have any major biological consequence on the cancer cellâ&#x20AC;&#x2122;s phenotype. This is confounded by the fact that current vector technology does not achieve the stable integration of therapeutic genes into 100% of prostate cancer cells comprising the tumor. Unexpectedly, the replacement or correction of one gene alteration has been shown to indeed alter the malignant phenotype and in some cases even completely eradicate prostate tumors in preclinical studies. In fact, this phenomenon has been repeatedly shown for different genes and vectors (Bookstein et al, 1990; Isaacs et al, 1991; Kleinerman et al, 1995; Steiner et al, 1998, 1999). These observations suggest that some genetic mutations are more critical to cell control than others and that the bystander effect may be playing an important role as well (Gotoh et al, 1997; Hall et al, 1997, 1998; Steiner et al, 1998b, c). Most corrective gene therapy strategies have employed either retroviral or adenoviral vectors administered by intratumoral injection. Prostate cancer preclinical studies have been reported for the replacement of an assortment of tumor suppressor genes including AdCMVp53 (Asgari et al, 1998; Eastham et al, 1995; Gotoh et al, 1997; Ko et al, 1996), retroviral LXSN BRCA-1 (Steiner et al, 1998), AdCMVp21 (Eastham et al, 1995; Gotoh et al, 1997), and AdCMV CAM1 (Hsieh et al, 1995). Another critical cell cycle component, cell cycle dependent kinase inhibitor p16, is commonly altered in prostate cancer (Cairns et al, 1994; Cairns et al, 1995). In prostate cancer, p16 inactivation is a common event observed in the majority of human prostate cancer cell lines (Itoh et al, 1997; Jarrard et al, 1997), and alterations of p16 have also been reported in patients who have prostate cancer (Cairns et al, 1995). Controversy arose as to whether p16 inactivation was critical only in rapidly dividing cancer cells in tissue culture rather than in primary human prostate cancer because homozygous deletions or intragenic mutations of p16 were apparently infrequent (Liggett and Sidransky, 1998). This controversy, however, was laid to rest by the recent discovery of microdeletions within the p16 gene. These microdeletions of the p16 gene were difficult to confirm by standard molecular techniques because of the presence of normal cells within the tumor specimen (Liggett and Sidransky, 1998). Microsatellite analysis employing markers close to the p16 gene revealed that a wide range of tumor types including prostate cancer had small (< 200 kb) deletions of both p16 alleles (Liggett 238

and Sidransky, 1998). Unlike other tumor suppressor genes that are commonly inactivated by point mutation, small homozygous deletions represented a major mechanism of p16 inactivation in cancer (Liggett and Sidransky, 1998). In fact, using this technique Cairns et al found that p16 homozygous deletions occurred in 40% of human primary prostate cancers (Cairns et al, 1995; Jarrard et al, 1997). Moreover, with progression 46% of prostate cancer metastatic lesions demonstrate loss of heterozygosity (Jarrard et al, 1997). Even more interesting, patients who have failed androgen deprivation have a 71% loss of 9p allelic loss (Isaacs, 1995). Using an adenoviral RSV vector containing p16, Steiner et al (1999) have shown that p16 replacement suppresses cell growth and induces cell senescence in a variety of prostate cancer cell lines. Moreover, in vivo, a single intratumoral injection of Adp16 resulted in 70% reduction of PPC-1 human prostate xenografts in nude mice and prolonged animal survival (Lu et al, 1999; Lu et al, 1998; Steiner et al, 1999). Using a different prostate cancer animal model, Gotoh et al (1997) have shown similar results employing an AdCMVp16 vector. Interestingly, peptide growth factor receptor FGFR 2 IIIb becomes altered with prostate cancer progression (Feng et al, 1997). Matsubara et al (1998) have shown that following transfection with FGFR2 kinase, AT3 had restoration of KGF response resulting in suppression of AT3 prostate cancer growth. Thus, restoration of a single underlying growth factor pathway may favorably alter the malignant phenotype. Oncogene overexpression is another way that cancer cells commonly lose control of the cell cycle (Steiner et al, 1995). One therapeutic approach utilizes expression of an antisense mRNA to the oncogene. The antisense mRNA anneals to the sense strand and effectively prevents the translation of the protein from that mRNA, thereby suppressing the protein level. Since prostate cancer commonly has overexpression of c-myc, Steiner et al (1998a) constructed a retroviral LXSN vector containing a prostate specific MMTV promoter driving the antisense cmyc gene. A single intratumoral injection of retroviral MMTV antisense c-myc was able to markedly suppress and even eradicate some of the DU145 prostate cancer xenografts growing in nude mice. The molecular mechanism was down regulation of c-myc expression and protein and the induction of apoptosis and downregulation of bcl-2 protein (Steiner et al, 1998a). Using a similar approach, Kim et al (1997) have shown that an adenoviral vector containing the antisense erb-B-2 gene (Ad anti-erb-B2) inhibited the overexpression of growth factor erb-B-2 in prostate cancer cells resulting in their destruction. This tactic was used to selectively purge bone marrow cells of metastatic prostate cancer cells in vitro (Kim et al, 1997). In general, corrective gene therapy holds the promise that when expression of one or more genes is restored, the

Gene Therapy and Molecular Biology Vol 4, Page 239 malignant phenotype of the cancer cell may be restored towards a more normal cell. Other corrective gene therapy approaches like AdCMVp53 have shown that gene replacement induces cell death, while others like retroviral antisense c-myc may incite host responses such as the bystander effect and other immunologic host responses (Gotoh et al, 1997; Hall et al, 1997, 1998; Steiner et al, 1998). Preliminary studies that employ corrective gene therapy also raise important clinical concerns unique to gene therapy. It has always been the dictum in cancer therapy that every cancer cell must be eradicated to effect a longterm cure. Corrective gene therapy, whose goal is to correct or repair alterations of the cell cycle and its components, challenges this concept. It is quite possible that the clinical endpoint of a corrective gene therapy strategy would simply be that the cell behaves more normally and no longer threatens the life of the patient. To this end, leaving a restored to more normal cancer cell in the patient may be an acceptable clinical endpoint. As the human genome project progresses and new prostate cancer genes are identified, corrective gene therapy will play a pivotal role in the treatment of prostate cancer.

C. Exploitation of programmed cell death gene therapy Gene therapy strategies are being developed to activate apoptotic pathways toward the ultimate goal of forcing the cancer cell to irreversibly commit to programmed cell death. Segawa et al (1998) have used an elaborate PSA promoter based system (GAL-4-VP16) to activate GAL-4 responsive elements. One GAL-4 responsive element is placed upstream of the polyglutamine gene. Polyglutamine is a potent apoptotic protein which in this case is selectively expressed in PSA producing cells. Similarly, Hyer et al (1999) have shown that adenovirus mediated transduction of the fas ligand, a component of cell death pathways, induced apoptosis in LNCaP, PC3, and DU145 prostate cancer cell lines in vitro. Marcelli et al (1999) have reported that transduction of prostate cancer cell line LNCaP with adenoviral vector containing caspase-7, a potent and critical modulator of apoptosis, also induced programmed cell death. Another molecular approach has targeted the bcl-2, an oncogene that has anti-apoptotic activity, with an adenoviral vector containing a hammerhead ribozyme directed against bcl-2 (Dorai et al, 1997, 1999). A bcl-2 ribozyme is an RNA molecule which specifically catalyzes or disrupts bcl-2 mRNA making the cell more susceptible to apoptosis. Interestingly, adenoviral hammerhead bcl-2 ribozyme treatment induced apoptosis in androgen sensitive, but not androgen insensitive prostate cancer cells. Although no preclinical in vivo studies have yet been reported, exploitation of programmed cell death gene therapy approaches are attractive and may potentially be quite effective. 239

D. Gene therapy to target critical biological functions of the cell Like classical pharmacology, gene therapy may be used to target critical cellular processes as the basis of rational anticancer gene therapy design. Lee et al (1996) have designed liposomal vectors that contain a PSA promoter upstream of either antisense topoisomerase II or antisense DNA polymerase ". Both topoisomerase II and DNA polymerase " are critical molecular components of DNA replication. The treatment combination of both liposome PSA-antisense topoisomerase II and liposome PSA-antisense DNA polymerase " had the greatest inhibitory effects on prostate cancer cell lines (LNCaP, DU145, and PC3) in vitro. Similarly, a retroviral vector that incorporated antisense eIF4E gene was used to treat prostate cancer cells (Williams et al, 1998). Prostate cancer cells have been previously shown to have overexpression of eIF4E which is a rate limiting factor in the translation initiation of growth controlling genes like cyclin D1, c-fos, c-myc, VEGF, and bFGF. A single intratumoral injection of retroviral antisense eIF4E suppressed prostate cancer xenograft growth for up to 65 days (Williams et al, 1998). Thus, the rational design of gene therapy vectors to disrupt critical molecular events required for cellular function is an enticing strategy against prostate cancer.

E. Suicide gene therapy Suicide gene therapy may have the most promising clinical application. Vectors introduce the therapeutic gene into the cancer cells, and once the gene product is expressed, the cell is destroyed without regard to the underlying genetic mutations responsible for the malignant phenotype. Two types of suicide gene therapy strategies have emerged: gene directed enzyme prodrug treatment (GDEPT) and gene directed production of a cellular toxin.

1. Gene directed enzyme prodrug treatment (GDEPT) GDEPT approach utilizes a system that couples prodrug enzyme gene therapy followed by systemic administration of its specific prodrug. Following gene transfer of the prodrug enzyme gene, the cancer cell produces that prodrug enzyme, and as a consequence, is capable of converting a nontoxic prodrug into an activated, lethal metabolite. This activated drug not only kills the cell that produced the toxic drug, but also its neighboring cancer cells. This bystander effect can be quite impressive with the ability to kill 100 to 1000 times more cells than would be predicted by gene transfer rates alone. Thus, a low gene transfer efficiency may be compensated by the high bystander effect. By having the cancer cell itself manufacture the activated cancer killing drug which acts locally minimizes systemic toxicity as the

Steiner and Gingrich: Gene Therapy for Prostate Cancer toxic drug is greatly diluted in the volume distribution of the blood stream. The most widely used GDEPT system against prostate cancer is the Herpes simplex virus thymidine kinase (HSVtk) and ganciclovir (GCV) system (Eastham et al, 1996; Hall et al, 1997). The nucleoside analogue GCV is converted by HSV-tk into a phosphorylated compound that is then incorporated into DNA during DNA replication. This causes DNA chain termination and selective killing of dividing cells. Eastham et al (1996) have used an adenoviral vector containing HSV-tk (AdHSV-tk) to sensitize both human and murine prostate cancer cells to the toxic effects of GCV both in vitro and in vivo models. AdHSV-tk gene therapy followed by GCV in both suppressed prostate cancer growth and prolonged survival rates in mice bearing prostate tumors (Eastham et al, 1996). Hall et al (Hall et al, 1997; Hall et al, 1998) have also shown that AdHSV-tk followed by GCV in the mouse prostate reconstitution orthotopic model suppressed tumor growth and decreased the rate of spontaneous prostate metastases to the lung. An immune basis for these effects was demonstrated by challenging mice with an injection of prostate cancer cells into the tail vein followed by excision of primary prostate cancer tumors. The animals that had treated primary tumors had a 40% reduction in lung metastases. This bystander effect appeared to be mediated in part by NK cells (Hall et al, 1998). Other GDEPT systems including the prodrug enzyme cytosine deaminase-flucytosine strategy in which cytosine deaminase converts flucytosine to the chemotherapeutic agent 5 fluorouracil have been investigated in prostate cancer (Blackburn et al, 1998; Kim et al, 1999). Kim et al (1999) transferred either the cytosine deaminase gene or the HSV-tk gene into stromal cells of the bone marrow derived murine cell line D1. Co-cultures of D1 cells and human prostate cancer cell lines followed by the appropriate prodrug resulted in prostate cancer cell death with as low as 20% of D1 cells producing the prodrug enzyme in the coculture. Blackburn et al (1998) used an adenoviral vector incorporating a heat shock protein (HSP 70) promoter and either cytosine deaminase or HSV-tk gene to treat PC3 cells. In this system, hyperthermia to 41o C activated the HSP-70 promoter resulting in prodrug enzyme expression. Thus, systemic administration of the prodrug and local heat allowed selective expression of the prodrug enzyme in intended tissues (Blackburn et al, 1998). Another system used an E1a deleted adenovirus containing prodrug enzyme E. coli DeoD gene product purine neocleoside phosphorylase (PNP) under the control of the PSA promoter (Martiniello-Wilks et al, 1998). The prodrug is 6-methyl-9 (2 deoxy-!-D erythro-pentofuranosyl) purine (6 MPDR) is converted into a toxic nonphosphorylated purine capable of killing both quiescent and proliferating cells when incorporated in mRNA or DNA during synthesis (Martiniello-Wilks et al, 1998). The PNP-6 MPDR system 240

had efficacy against human prostate cancer cell line PC3 (Martiniello-Wilks et al, 1998). Other GDEPT systems utilizing assorted prostate specific promoters and vector types are currently under intense investigation.

2. Gene directed production of cell toxin This strategy is similar to GDEPT where the transferred gene kills the cell independent of the underlying cancer gene mutations, but unlike GDEPT, this approach does not require a prodrug. Rodriguez et al (1998) screened numerous direct biological toxins known to kill mammalian cells by cell cycle independent mechanisms to determine which would be the best one against human prostate cancer. Diptheria toxin (DT) was found to be the most toxic. DT kills rapidly, independent of p53 or androgen sensitivity status and it kills both dividing and nondividing cells alike (Rodriguez et al, 1998). This approach, however, has several limitations. First, this toxic gene must be incorporated into vectors that contain promoters that are highly prostate specific and under tight regulatory control; DT is such a toxic biological toxin that even small amounts of â&#x20AC;&#x153;leakyâ&#x20AC;? promoter activity in non-prostatic tissues may be very lethal. In addition, mass production of adenoviral vector- DT gene is very difficult because of the toxic effects of the DT gene on the packaging cell line resulting in low production titers (Simons et al, 1999).

F. Oncolytic virus gene therapy Because of safety reasons, practically all current vectors are engineered to be replication incompetent meaning that the virus cannot express those viral genes that commandeer cells to enter the lytic cycle producing more virus. Consequently, the effectiveness of the viral vector is directly correlated to its transduction efficiency and its ability to be given in repeated doses. Recently, two types of replication competent viral vectors have been developed. One conditionally competent adenoviral vector has been mutated such that the virus cannot express viral protein E1b (Bischoff et al, 1996). The wild type adenovirus uses the E1b protein to stop p53 from preventing the replication of cells that have damaged DNA. Theoretically, the mutant E1b- virus can infect, replicate, and lyse p53 deficient cells, but does not affect normal cells that have functional p53 (Bischoff et al, 1996). Thus, these mutant viruses are oncolytic to cancer cells that harbor p53 mutations. In prostate cancer, however, the p53 mutation rate is lower than perhaps other types of cancer. Only 10-20% of prostate cancers having nonfunctional p53 and most p53 mutations are only found in tumors that have a higher grade and stage (Brooks et al, 1996; Dahiya et al, 1996; Eastham et al, 1996). Another oncolytic virus is CN706 which is a replication

Gene Therapy and Molecular Biology Vol 4, Page 241 competent, attenuated cytotoxic adenovirus type 5 vector with a prostate specific enhancer and promoter coupled to the E1a gene (Rodriguez et al, 1997). The E1a viral product allows the virus to reproduce and to enter the lytic cycle. The PSA promoter theoretically limits E1a production to PSA producing cells (Rodriguez et al, 1997). The level of E1a production has been shown to be several logs higher in PSA producing cells like LNCaP than in cells that produce little or no PSA (Simons et al, 1999). In vivo, CN706 viral vector produced tumor regression of LNCaP tumors and decreased PSA production following a single intratumoral injection (Rodriguez et al, 1997; Simons et al, 1999). Although a clinical Phase I trial of intratumoral injection of CN706 in patients who have prostate cancer is in progress, there are several clinical concerns that are raised about CN706. First, there is no experimental support that systemically distributed CN706 will exclusively lyse prostate cells. In some systems, wide variation in the level of E1a expression has demonstrated little effect on viral replication suggesting that even low level expression may be sufficient to support viral replication and subsequent cell lysis. This is especially worrisome since the PSA promoter has been shown to be leaky as other types of cells in addition to prostate cells produce PSA. For example, cells that line the urethra produce abundant PSA. Theoretically, CN706 will only cease replication and lysis when all PSA producing cells are eradicated. Furthermore, there is questionable utility of any PSA promoter vector for the systemic treatment of PSA negative prostate cancer cells or in patients undergoing androgen deprivation therapy. Nonetheless, studies employing the CN706 or any other vector containing the PSA promoter by using intratumoral injections will be critical in increasing our understanding of this field until newer tissue specific vector technology becomes a reality.

G. Combination of gene therapy approaches and other treatment modalities Use of gene therapy as monotherapy against prostate cancer is currently in its infancy. The use of gene therapy in combination with surgery, radiation, or chemotherapy to improve cancer treatment will most likely be the way that gene therapy will be used clinically. Moreover, clinical trials using this multimodal approach are more justified until the science of gene therapy improves. The most commonly employed strategy is the use of gene therapy in combination with DNA damaging agents. Improvements in prostate cancer tumor suppression and induction of apoptosis have been reported for the combinations of Ad5CMVp53 and paclitaxel (Nielsen et al, 1998), Ad5CMVp53 (Wilson, et al, 1999), Adp16, Ad.Egr TNF-" (Chung, et al, 1998), and ionizing radiation. Other combination approaches that have been employed in prostate cancer include GDEPT Ad cytosine deaminase and flucytosine and radiation (Yin et al, 1998), or GDEPT AdHSV-tk/GCV and AdIL12 cytokine 241

therapy (Hassen et al, 1999). Improved prostate cancer treatment by including gene therapy as part of a multimodal treatment regimen is where gene therapy may have the most immediate clinical application.

VI. Prostate cancer gene therapy clinical trials At this time, 17 gene therapy trials for the treatment of prostate cancer have been approved by the NIH (Table 1). To date, the approved trials have all been either Phase I or Phase I/II in design. Preliminary results of these trials are only recently forthcoming as meeting abstracts and peer reviewed publications. The first approved gene therapy trial for prostate cancer by Simons et al (1998, 1999) (Table 1) was designed for patients who were found to have metastatic prostate cancer in the lymph nodes at the time of radical prostatectomy. Similar to previous studies for the treatment of metastatic renal cell carcinoma (Simons et al, 1997), ex vivo transduction of autologous, irradiated prostate tumor cells with retroviral MFG-GM-CSF was performed to generate vaccines which were administered subcutaneously every 2 weeks until available cells were exhausted. Vaccination site biopsies revealed infiltrating macrophages, dendritic cells, eosinophils, and T cells. No dose limiting toxicities were observed. Seven out of eight patients had progressed by ultrasensitive PSA criteria by an average follow-up of 20 weeks. This study demonstrated the feasibility of autologous GM-CSF transduced prostate cancer vaccines was limited only by the in vitro expansion of vaccine cells. To circumvent this limitation, a follow-up trial by the same investigators powered to estimate efficacy utilizing ex vivo GM-CSF transduced allogeneic prostate cancer cell lines [PC-3 (Kaighn et al, 1979) and LNCaP (Horoszewicz et al, 1983)] as vaccines has been approved and is currently in progress (Table 1). Patients are vaccinated weekly for eight weeks with irradiated, GM-CSF secreting PC-3 and LNCaP prostate cancer cells. One of 21 patients treated to date has had a partial PSA response of >7 months duration, 14/21 have stable disease, and 6/21 progressed (Simons et al, 1999). At three months after treatment, PSA velocity or PSA slope have decreased in 71% of patients. No dose limiting toxicities have been identified. Although the dose and schedule are still undergoing optimization to demonstrate potential therapeutic efficacy, numerous new post-vaccination IgG1 antibodies have been identified demonstrating that immune tolerance to prostate cancer associated antigens may be broken and therefore appears to be clinically feasible for prostate cancer treatment. Preliminary results of the first trial approved utilizing direct transrectal prostatic gene therapy injection has recently been reported by Steiner et al (1998). This is the first study designed using a gene replacement strategy.

Steiner and Gingrich: Gene Therapy for Prostate Cancer Table 1. Approved prostate cancer gene therapy trials Patient Population

Vector

Gene

Modality

Johns Hopkins

Metastatic

Retrovirus

MFG-GMCSF

Memorial SloanKettering Univ. of Tennessee/ Vanderbilt Bethesda, Naval

Prostate cancer

Retrovirus

IL-2 + gamma IFN

Advanced

Retrovirus

Anti-sense myc

Prostate Cancer

Vaccinia virus

PSA cDNA

Paulson

Duke Univ.

AAV/ Liposome

IL-2

Scardino

Adenovirus

RSV-HSV-tk /ganciclovir

9609-160

Kufe & Eder

Baylor College of Medicine DanaFarber

Locally Advanced, Metastatic RadioRecurrent

Ex vivo, autologous PCA vaccine Ex vivo, allogeneic PCA vaccine In vivo, intraprostatic injection In vivo, intradermal injection Ex vivo

Prostate Cancer

Vaccinia virus

PSA cDNA

9702-176

I/II

Sanda

Univ. of Michigan

Therion Biologics Corp

PSA Recurrence after RRP

Vaccinia virus

PSA cDNA

9703-184

Belldegrun

UCLA

Vical, Inc.

Locally Advanced

Liposome

IL-2

9705-187

Hall

T1c, T2b & c

Adenovirus

RSV-HSV-tk /ganciclovir

9706-192

Belldegrun

Memorial SloanKettering UCLA

Adenovirus

p53 wild type cDNA

9708-205

I/II

Simons

Johns Hopkins

Locally Advanced, Recurrent Prostate Cancer

Retrovirus

GM-CSF

9710-217

I/II

Logothetis/ Steiner

No Prior Therapy, high risk

Adenovirus

p53 wild type cDNA

9801-229

I/II

Kadmon

RSV-HSV-tk /ganciclovir

9802-236

Simons

No Prior Therapy, high risk RadioRecurrence

Adenovirus

9805-251

I/II

Figlin

MD Anderson/ Univ. of Tennessee Baylor College of Medicine Johns Hopkins Univ. UCLA

9812-276

Gardner/ Chung

Univ. of Virginia

MUC-1 Positive Metastatic

Vaccinia virus Adenovirus

PSA-E1a (replication competent) MUC-1/IL2

Pending

I/II

Gingrich

Univ. of Tennessee

Locally Advanced

Adenovirus

NIH

Phase

Institution

9408-082

I/II

Simons

9503-102

I/II

Gansbacher

9509-123

Steiner

9509-126

Chen

9510-132

9601-144

Sponsor

ScheringPlough Corp.

Introgen, Inc.

Calydon Transgene SA

Genotherapeutics

Primary Source: http://www.nih.gov/od/orda

242

Osteocalcin HSV-tk/ valacyclovir p16

In vivo, intraprostatic injection In vivo, intradermal injection In vivo, intradermal injection autologous In vivo, Autologous, intratumoral In vivo, intraprostatic injection/RRP In vivo, intratumoral injection Ex vivo, allogeneic PCA vaccine In vivo, intraprostatic injection/RRP In vivo, intraprostatic injection/RRP In vivo, intraprostatic injection Immunotherapy In vivo, intratumoral injection In vivo, intraprostatic injection/RRP

Gene Therapy and Molecular Biology Vol 4, Page 243 Prostate cancer patients (n=21) who had failed standard therapy underwent ultrasound guided injection of retrovirus LXSN containing BRCA1 under the control of the viral promotor (LTR). No patients developed viral symptoms or evidence of viremia. Viral DNA was detected in prostate tissue by PCR at least 2 years after injection (unpublished results). Although average prostate volume was decreased one month after injection (30.9 ± 6.4cc vs. 36.5 ± 6.5cc), serum PSA in those patients who had metastatic disease remained unchanged. This study demonstrated the safety of direct injection of prostate cancer gene replacement therapy. Initial clinical gene therapy results utilizing a similar gene replacement strategy of replication-defective adenovirus containing wild-type p53 (AdCMVp53) driven by the CMV promoter injected into patients prior to radical prostatectomy have recently been reported (Logothetis et al, 1999). To date, 17 patients with locally advanced prostate cancer have received at least one course of AdCMVp53 which consists of 3 administrations 14 days apart. Transperineal injections of 3 ml in 4-6 divided doses are delivered under transrectal ultrasound guidance. In this phase I/II study, the number of viral particles (vp) delivered was escalated from 3x1010 vp per treatment per patient to 3x1012 vp. Three patients completed a second course of therapy for > 25% reduction in tumor size as measured by endorectal coil MRI after the first course of treatment. There were no grade 3 or 4 toxicities in 14 evaluable patients. This study confirms the relative feasibility and safety of intraprostatic gene therapy injection. Further results regarding efficacy due to the apparent radiologic responses, correlations with gene expression, pathologic findings at prostatectomy, and surgical outcomes are pending at this time. The first trial utilizing intraprostatic injection of a locally cytotoxic or “suicide” type of gene therapy has recently been completed by Scardino et al (Herman et al, 1999; Scardino et al, 1998). For the treatment of locally recurrent prostate cancer after definitive radiation therapy, eighteen patients received a single 1.1 cc injection of replication-deficient adenovirus containing the HSV-tk gene (AdHSV-tk) under the RSV promoter followed by 14 days of intravenous ganciclovir. The dose of AdHSV-tk was escalated from 1 x 108 to 1 x 1011 IU. Three patients had > 50% reduction in PSA and 1 had a negative biopsy after treatment. Local and systemic toxicity was mild except for the final patient who developed severe thrombocytopenia and abnormal liver function tests. A second study to investigate multiple sites of injection prior to radical prostatectomy is currently in progress. These follow-up studies will provide important histologic confirmation of tumor response to the intraprostatic cytotoxic gene therapy. In addition, assessment of long term patient outcomes due to beneficial local or possible systemic immunologic “bystander” effects will be interesting. 243

An alternative approach to intraprostatic gene replacement or cytotoxic therapy which may indirectly enhance a host immunologic response against the tumor cells is to administer gene therapy to intentionally stimulate an immune response. Utilizing a liposomal vector, Belldegrun et al have treated 12 patients prior to radical prostatectomy and 9 patients with recurrent prostate cancer after radiation or cryotherapy with two injections of intraprostatic IL-2 (Patel et al, 1999). A total of 40 injections of 300-1500 µg of IL-2 were administered. For patients treated prior to radical prostatectomy, the average PSA declined by 5.3 ng/ml prior to surgery and 75% maintain undetectable PSA levels at 24-56 weeks after surgery. Average PSA declines in patients with recurrent disease were 3.6 ng/ml and 1.3 ng/ml after one or two injections, respectively. Although the significance of these PSA responses is not yet clear, this strategy utilizing liposomal delivery appears to be safe and may be locally and systemically therapeutic. Three trials (Table 1) have been approved and initiated using recombinant vaccinia virus expressing PSA (PROSTVAC) as a tumor associated antigen (TAA) immunotherapy strategy. Chen et al treated 30 patients with hormone refractory prostate cancer who had undergone withdrawal of anti-androgen therapy with a vaccinia inoculation followed by PROSTAVAC immunization (Chen et al, 1998). Patients received 2.65 x 105 or 2.65 x 106 plaque forming units (PFU) by dermal scarification or subcutaneous delivery of 2.65 x 106 or 2.65 x 108 PFU once a month for 3 months followed by restaging. All patients developed local erythema at the vaccination site with all toxicities grade 0-1. Of 14 patients who completed the treatment course, 4 had stable disease (2 subsequently progressed at 5 and 6 months) and 10 had continued disease progression. In a patient population with less advanced disease, Eder et al treated 24 men with rising PSA after radical prostatectomy, radiation therapy, or both with 2.65 x 106-8 PFU as three consecutive monthly doses without significant toxicity (Eder et al, 1998). Patients were removed from the protocol for clinical progression or 3 monthly rises in PSA > 50% of baseline. Twelve out of 23 men maintained a stable disease status for $ 10 months. In a third study, Sanda et al administered PROSTVAC once to six patients with androgen-modulated recurrence of prostate cancer after radical prostatectomy (Sanda et al, 1999). In addition to evaluation for toxicity, time until rise in serum PSA after interruption of androgen deprivation therapy and Western blot analysis for anti-PSA antibody production were determined. Again, no dose limiting toxicity was observed. One patient maintained an undetectable serum PSA for over 8 months after withdrawal of anti-androgen therapy and immunization. This study emphasizes the variability in time to return of serum testosterone levels after withdrawal of anti-androgen therapy as observed by others (Oefelein,

Steiner and Gingrich: Gene Therapy for Prostate Cancer 1998). One patient developed an IgG antibody against PSA after immunization. It is interesting to note that 2/6 patients had anti-PSA antibodies prior to immunization. The significance of this finding and the implications regarding vaccination against tumor associated antigens is unclear at this time. However, it was demonstrated that an immune response may be solicited by this gene therapy strategy.

VI.Conclusions This decade marks the real birth of human gene therapy as a biomedical commodity. Since the first gene therapy trial in 1990, there have been over 125 Phase I, 25 Phase II, and 1 Phase III human clinical trials in the United States. Worldwide, over 363 clinical gene therapy trials have been approved with over 4000 patients enrolled. Among the various diseases being treated malignancies rank first (68%) followed by AIDS (18%) and cystic fibrosis (8%). Wide spread clinical acceptance of this technology and the Human Genome Project have further fueled the rapid expansion of this new technology. Gene therapy to treat advanced localized prostate cancer has been shown to be safe and feasible. The challenges that lie ahead for the widespread use of this technology in the next century include: 1) To find the appropriate genes to use for gene therapy; 2) To systemically deliver gene therapy to target prostate cancer cells and treat distant disease; 3) To identify strong tissue specific promoters and other ways to exclusively target prostate cancer; and 4) To determine the ultimate safety and efficacy of gene therapy in humans.

References Abbas F, Kaplan M, and Soloway MS (1996) Induction androgen deprivation therapy before radical prostatectomy for prostate cancer--initial results. Br J Urol 77, 423-428. Apostolopoulos V, and McKenzie IF (1994) Cellular mucins: targets for immunotherapy. Crit Rev Immunol 14, 293-309. Asgari K, Xu L, Naga S, Sesterhenn I, McLeod D, Seth P, Moul J, and Srivastava S (1998) Evaluation of the mechanisms of anti-tumorigenic effects of the AdWTp53 on DU145 cells derived tumors in nude mice (San Diego, CA) vol. 159, pp. 12. Badalament RA, Bahn DK, and Lee F (1996) Is radical prostatectomy primarily a surgical debulking procedure? Urol Oncol 2, 88-91. Bagshaw MA (1993) In The challenge of radiation treatment of carcinoma of the prostate H Lepor and RK Lawson, eds. (Philadelphia: W.B. Saunders Co.). pp. 326-346. Berges RR., Furuya Y, Remington L, and English HF (1993) Cell proliferation, DNA repair, and p53 function are not required for programmed death of prostatic glandular cells induced by androgen ablation. Proc Nat Acad Sci USA 90, 8910-8914. Bischoff JR, Kirn DH, Williams A, Heise C, Horn S, Muna M, Ng L, Nye JA, Sampson-Johannes A, Fattaey A, and McCormick F (1996) An adenovirus mutant that replicates selectively in p53deficient human tumor cells. Science 274, 373-376. Blackburn RV, Galoforo SS, Corry PM, and Lee YJ (1998)

244

Adenoviral-mediated transfer of a heat-inducible double suicide gene into prostate carcinoma cells. Cancer Res 58, 1358-1362. Blades RA, Keating PJ, McWilliam LJ, George NJ, and Stern PL (1995) Loss of HLA class I expression in prostate cancer: implications for immunotherapy. Urology 46, 681-686; discussion 686-687. Bookstein R, Shew J-Y, Chen P-L, Scully P, and Lee W-H (1990) Suppression of tumorigenicity of human prostate carcinoma cells by replacing a mutated RB gene. Science 247, 712-715. Brooks JD, Bova GS, Ewing CM, Piantadosi S, Carter BS, Robinson JC, Epstein JI, and Isaacs WB (1996) An uncertain role for p53 gene alterations in human prostate cancers. Cancer Res 56, 3814-3822. Cairns P, Mao L, Merlo A, Lee DJ, Schwab D, Eby Y, Tokino K, van der Riet P, Blaugrund JE, and Sidransky D (1994) Rates of p16 (MTS1) mutations in primary tumors with 9p loss. Science 265, 415-417. Cairns P, Polascik TJ, Eby Y, Tokino K, Califano J, Merlo A, Mao L, Herath J, Jenkins R, Westra W, and et al (1995) Frequency of homozygous deletion at p16/CDKN2 in primary human tumours. Nat Genet 11, 210-212. Carrato C, Balague C, de Bolos C, Gonzalez E, Gambus G, Planas J, Perini JM, Andreu D, and Real FX (1994) Differential apomucin expression in normal and neoplastic human gastrointestinal tissues. Gastroenterology 107, 160-172. Catalona W, Smith D, Ratliff T, Dodds K, Coplen D, Yuan J, Petros J, and Andriole G (1991) Measurement of prostatespecific antigen in serum as a screening test for prostate cancer. New Engl J of Med 324. Cech TR, and Bass BL ( 1986) Biological catalysis by RNA. Annu Rev Biochem 55, 599-629. Chen AP, Bastian A, Dahut W, Chung Y, Tsang K, Khleif S, Blake K, Gelmann E, Marshall J, Schlom J, Allegra C, and Hamilton JM (1998) In A phase I study of recombinant vaccinia virus (rV) that expresses prostate specific antigen (PSA) in adult patients (pts) with adenocarcinoma of the prostate. (Los Angeles: Lippincott Williams & Wilkins). vol. 17 (suppl), pp. 314a. Chen M, Sikes R, Troncosco P, Sue-Hwa L, von Eschenbach A, and Chung L (1996) PAGE and GAGE 7 are novel genes expressed in the LNCaP prostatic carcinogenesis model that share homology with melanoma-associated antigens. vol. 155, pp. 624a. Chung TD, Mauceri HJ, Hallahan DE, Yu JJ, Chung S, Grdina WL, Yajnik S, Kufe DW, and Weichselbaum RR (1998) Tumor necrosis factor-alpha-based gene therapy enhances radiation cytotoxicity in human prostate cancer. Cancer Gene Ther 5, 344-349. Cookson MS, and Fair WR (1997) Neoadjuvant androgen deprivation therapy and radical prostatecotmy for clinically localized prostate cancer. AUA Update Series 16, 98. Cumming JA, Ritchie AW, Goodman CM, McIntyre MA, and Chisholm GD (1990) De-differentiation with time in prostate cancer and the influence of treatment on the course of the disease. Br J Urol 65, 271-274. Dahiya R, Deng G, Chen KM, Chui RM, Haughney PC, and Narayan P (1996) P53 tumour-suppressor gene mutations are mainly localised on exon 7 in human primary and metastatic prostate cancer. Br J Cancer 74, 264-268. Dannull J, and Belldegrun AS (1997) Development of gene therapy for prostate cancer using a novel promoter of prostatespecific antigen. Br J Urol. 79 Suppl 1, 97-103.

Gene Therapy and Molecular Biology Vol 4, Page 245 Dannull J, Patel B, Pang S, Kaboo R, and Belldegrun A (1999) Tissue specific expression of a prostate-specific antigen promoter/enhancer in an adenoviral vector used in a human prostate cancer SCID mouse model (Dallas, TX). vol. 161, pp. 127. Dorai T, Olsson CA, Katz AE, and Buttyan R (1997) Development of a hammerhead ribozyme against bcl-2. I. Preliminary evaluation of a potential gene therapeutic agent for hormonerefractory human prostate cancer. Prostate 32, 246-258. Dorai T, Perlman H, Walsh K, Goluboff T, Olsson C, and Buttyan R (1999) Enhanced apoptosis of prostate cancer cells by adenoviral hammerhead ribozyme mediated disruption of bcl-2 (Philadelphia, PA). vol. 40, pp. 482. Eastham JA, Chen SH, Sehgal I, Yang G, Timme TL, Hall SJ, Woo SL, and Thompson TC (1996) Prostate cancer gene therapy: herpes simplex virus thymidine kinase gene transduction followed by ganciclovir in mouse and human prostate cancer models. Hum Gene Ther 7, 515-523. Eastham JA, Hall SJ, Sehgal I, Wang J, Timme TL, Yang G, Connell-Crowley L, Elledge SJ, Zhang WW, Harper JW, and et al (1995) In vivo gene therapy with p53 or p21 adenovirus for prostate cancer. Cancer Res 55, 5151-5155. Eder Jr JP , Kantoff PW, Bubley GJ, O'Connor LA, Rodriguez D, Xu G, Tsang A, Schlom J, Mazzara G, Panicali D, Horzempa ML, and Kufe DW (1998) In A phase I trial of recombinant vaccinia virus, PROSTVAC that expresses prostate specific antigen (rV-PSA) as a vaccine in men with advanced prostate cancer (Los Angeles:Lippincott Williams & Wilkins). vol. 17 (suppl), pp. 434a. Fearon E, Pardoll D, Itaya T, Golumbek P, Levitsky H, Simons J, Karasuyama H, Vogelstein B, and Frost P (1990) Interleukin-2 production by tumor cells engineered to secrete interleukin-4. Cell 60, 397. Feng S, Wang F, Matsubara A, Kan M, and McKeehan WL (1997) Fibroblast growth factor receptor 2 limits and receptor 1 accelerates tumorigenicity of prostate epithelial cells. Cancer Res 57, 5369-5378. Fisher B, Packard BS, Read EJ, Carrasquillo JA, Carter CS, Topalian SL, Yang JC, Yolles P, Larson SM, and Rosenberg SA (1989) Tumor localization of adoptively transferred indium-111 labeled tumor infiltrating lymphocytes in patients with metastatic melanoma. J Clin Oncol. 7, 250-261. Gambus G, de Bolos C, Andreu D, Franci C, Egea G, and Real FX (1993) Detection of the MUC2 apomucin tandem repeat with a mouse monoclonal antibody. Gastroenterology 104, 93-102. Gansbacher B, Bannerji R, Daniels B, Zier K, Cronin K, and Gilboa E (1990) Retroviral vector-mediated gamma-interferon gene transfer into tumor cells generates potent and long lasting antitumor immunity. Cancer Res 50, 7820-7825. Goldenberg SL, Klotz L, Jewett M, Barkin J, Chetner M, Fradet Y, Chin J, and Group, C (1997) A randomized trial of neoadjuvant androgen withdrawal therapy prior to radical prostatectomy: 24 month post-treatment PSA results. J Urol part 2 157, 92. Golumbek P, Lazenby A, Levitsky H, Jaffee E, Karasuyama H, Baker M, and Pardoll D (1991) Treatment of established renal cell cancer by tumor cells engineered to secrete interleukin-4. Science 254, 713. Gotoh A, Kao C, Ko SC, Hamada K, Liu TJ, and Chung LW (1997) Cytotoxic effects of recombinant adenovirus p53 and cell cycle regulator genes (p21 WAF1/CIP1 and p16CDKN4) in human prostate cancers. J Urol 158, 636-641. Gotoh A, Ko SC, Shirakawa T, Cheon J, Kao C, Miyamoto T,

245

Gardner TA, Ho LJ, Cleutjens CB, Trapman J, Graham FL, and Chung LW (1998) Development of prostate-specific antigen promoter-based gene therapy for androgen-independent human prostate cancer. J Urol 160, 220-229. Greenberg NM, DeMayo PC, Sheppard PC, Barrios R, Lebovitz R, Finegold M, Angelopoulou R, Dodd JG, Duckworth ML, Rosen JM, and Matusik RJ (1994) The rat probasin gene promoter directs hormonally and developmentally regulated expression of a heterologous gene specifically to the prostate in transgenic mice. Mol Endocrinol 8(2), 230-239. Hall SJ, Mutchnik SE, Chen SH, Woo SL, and Thompson TC (1997) Adenovirus-mediated herpes simplex virus thymidine kinase gene and ganciclovir therapy leads to systemic activity against spontaneous and induced metastasis in an orthotopic mouse model of prostate cancer. Int J Cancer 70, 183-187. Hall SJ, Sanford MA, Atkinson G, and Chen SH (1998) Induction of potent antitumor natural killer cell activity by herpes simplex virus-thymidine kinase and ganciclovir therapy in an orthotopic mouse model of prostate cancer. Cancer Res 58, 3221-3225. Hassen W, Sanford M, Atkinson G, Chen S-H, and Hall S (1999) Natural killer cells induced by herpes simplex virus thymidine kinase gene transduction and ganciclovir therapy are enhanced by interleukin-12 gene therapy in an othotopic model of mouse prostate cancer (Dallas, Texas). pp. 53. Herman JR, Adler HL, Aguilar-Cordova E, Rojas-Martinez A, Woo S, Timme TL, Wheeler TM, Thompson TC, and Scardino PT (1999) In situ gene therapy for adenocarcinoma of the prostate: A phase I clinical trial. Hum Gene Ther 10, 12391249. Holzman M, Carlton CE Jr, and Scardino PT (1991) The frequency and morbidity of local tumor recurrence after definitive radiotherapy for stage C prostate cancer. J Urol 146, 15781582. Horoszewicz JS, Leong SS, Kawinski E, Karr JP, Rosenthal H, Chu TM, Mirand EA, and Murphy GP (1983) LNCaP model of human prostatic carcinoma. Cancer Res 43, 1809. Hsieh JT, Luo W, Song W, Wang Y, Kleinerman DI, Va NT, and Lin SH (1995) Tumor suppressive role of an androgenregulated epithelial cell adhesion molecule (C-CAM) in prostate carcinoma cell revealed by sense and antisense approaches. Cancer Res 55, 190-197. Hyer M, Rubinchik S, Dong J-Y, Matusik R, and Norris J (1999) Gene therapy of prostate cancer with adenovirus-mediated fas ligand (FasL) expression (Philadelphia, PA). vol. 40, pp. 173. Isaacs WB (1995) In Molecular genetics of prostate cancer, pp. 357-379. Isaacs WB, Carter BS, and Ewing CM (1991) Wild-Type p53 suppresses growth of human prostate cancer cells containing mutant p53 alleles. Cancer Res 51, 4716-4720. Israeli RS, Powell CT, Fair WR, and Heston WD (1993) Molecular cloning of a complementary DNA encoding a prostate-specific membrane antigen. Cancer Res 53, 227-230. Itoh N, Kakehi Y, Akao T, Kinoshita H, Okada Y, and Yoshida O. (1997) Concomitant presence of p16/cyclin-dependent kinase 4 and cyclin D/cyclin-dependent kinase 4 complexes in LNCaP prostatic cancer cell line. Jpn J Cancer Res 88, 229-233. Jaffee EM, Dranoff G, Cohen LK, Hauda KM, Clift S, Marshall FF, Mulligan RC, and Pardoll DM (1993) High efficiency gene transfer into primary human tumor explants without cell selection. Cancer Res 53, 2221-2226. Jarrard DF, Bova GS, Ewing CM, Pin SS, Nguyen SH, Baylin SB,

Steiner and Gingrich: Gene Therapy for Prostate Cancer Cairns P, Sidransky D, Herman JG, and Isaacs WB (1997) Deletional, mutational, and methylation analyses of CDKN2 (p16/MTS1) in primary and metastatic prostate cancer. Genes Chromosomes Cancer 19, 90-96. Kabalin JN, Hodge K, McNeal JE, Freiha FS, and Stamey TA (1989) Identification of residual cancer in the prostate following radiation therapy: Role of transrectal ultersound guided biopsy and prostate specific antigen. J Urol 142, 326341. Kaighn ME, Narayan KS, Ohnuki Y, Lechner JF, and Jones LW (1979) Establishment and characterization of a human prostatic carcinoma cell line (PC-3). Investigative Urology 17, 16-23. Kawakita M, Rao GS, Ritchey JK, Ornstein DK, Hudson MA, Tartaglia J, Paoletti E, Humphrey PA, Harmon TJ, and Ratliff TL (1997) Effect of canarypox virus (ALVAC)-mediated cytokine expression on murine prostate tumor growth. J Natl Cancer Inst 89, 428-436. Kim M, Wright M, Deshane J, Accavitti MA, Tilden A, Saleh M, Vaughan WP, Carabasi MH, Rogers MD, Hockett RD Jr, Grizzle WE, and Curiel DT (1997) A novel gene therapy strategy for elimination of prostate carcinoma cells from human bone marrow. Hum Gene Ther 8, 157-170. Kim SJ, Gardner TA, Ko S-C, Kao C, Balian G, and Chung LWK (1999) Enhancement of bystander prostate cancer cell killing by the utilization of genetically engineered bone stromal cells that confer either cytosine deaminase or thymidine kinase gene (Dallas, Texas). pp. 58. Kleinerman DI, Zhang WW, Lin SH, Nguyen TV, von Eschenbach AC, and Hsieh JT (1995) Application of a tumor suppressor (C-CAM1)-expressing recombinant adenovirus in androgenindependent human prostate cancer therapy: A preclinical study. Cancer Res 55, 2831-2836. Ko SC, Gotoh, A, Thalmann, GN, Zhau, H., Johnston DA, Whang W-W, Kao C, and Chung LWK (1996) Molecular therapy with recombinant p53 adenovirus in an androgen-independent, metastatic human prostate cancer model. Human Gene Ther 7, 1683-1691. Ko SC, Cheon J, Kao C, Gotoh A, Shirakawa T, Sikes RA, Karsenty G, and Chung LW (1996) Osteocalcin promoterbased toxic gene therapy for the treatment of osteosarcoma in experimental models. Cancer Res 56, 4614-4619. Landis SH, Murray T, Bolden S, and Wingo PA (1999) Cancer statistics, 1999. CA Cancer J Clin. 49, 8-31. Lange P, Ercole C, Lightner D, Fraley E, and Vessella R (1989) The value of serum prostate specific antigen determinations before and after radical prostatectomy. J Urol 141, 873. Lee CH, Liu M, Sie KL, and Lee MS (1996) Prostate-specific antigen promoter driven gene therapy targeting DNA polymerase-alpha and topoisomerase II alpha in prostate cancer. AntiCancer Res 16, 1805-1811. Liggett WH Jr, and Sidransky D (1998) Role of the p16 tumor suppressor gene in cancer. J Clin Oncol 16, 1197-1206. Logothetis C, Hossan E, Pettaway C, Evans R, Steiner M, Wood C, Troncoso P, McDonnell T, Fenstenmacer M, and Pisters L (1999) AD-p53 intraporstatic gene therapy preceding radical prostatectomy (RP): An in vivo model for targeted therapy development (Dallas, TX). vol. 161, pp. 297. Lu Y, Carraher J, Zhang Y, Armstrong J, Lerner J, Rogers W, and Steiner M (1999) Delivery of adenoviral vectors to the prostate for gene therapy. Cancer Gene Therapy 6, 64-72. Lu Y, Zhang Y, Farooq F, Lerner J, Wang Y, and Steiner M (1999) Adenoviral vector containing wild type p16 suppresses

246

prostate cancer growth and prolongs survival (Philadelphia, PA). vol. 40, pp. 482. Lu Y, Zhang Y, Farooq F, Wang Y, Lerner J, and Steiner M (1998) Expression of tumor suppressor gene p16 effectively inhibits the growth of prostate cancer cells (San Diego, CA). vol. 5, pp. S29. Lubaroff D, Elzey B, Ratliff T, Scholm J, and Tartaglia J (1999) PSA-encoding adenovirus induces potent antitumor immunity in a mouse prostate tumor model (Dallas, TX). vol. 161, pp. 53. Malkowicz S, and Johnson J (1998) Gene therapy for prostate cancer. Hematol/Oncol Clin N Amer 12, 649-664. Marcelli M, Cunningham GR, Walkup M, He Z, Sturgis L, Kagan C, Mannucci R, Nicoletti I, Teng B, and Denner L (1999) Signaling pathway activated during apoptosis of the prostate cancer cell line LNCaP: Overexpression of caspase-7 as a new gene therapy strategy for prostate cancer. Cancer Res 59, 382390. Martiniello-Wilks R, Garcia-Aragon J, Daja MM, Russell P, Both GW, Molloy PL, Lockett LJ, and Russell PJ (1998) In vivo gene therapy for prostate cancer: preclinical evaluation of two different enzyme-directed prodrug therapy systems delivered by identical adenovirus vectors. Hum Gene Ther 9, 16171626. Matsubara A, Kan M, Feng S, and McKeehan WL (1998) Inhibition of growth of malignant rat prostate tumor cells by restoration of fibroblast growth factor receptor 2. Cancer Res 58, 1509-1514. Moody DB, Robinson JC, Ewing CM, Lazenby AJ, and Isaacs WB (1994) Interleukin-2 transfected prostate cancer cells generate a local antitumor effect in vivo. Prostate 24, 244-251. Muller WJ, Lee FS, Dickson C, Peters G, Pattengale P, and Leder P (1990) The int-2 gene product acts an epithelial growth factor in transgenic mice. EMBO J 9, 907-913. Naitoh J, and Belldegrun A (1998) Gene therapy--the future is here: A guide to the practicing urologist. Urology 51, 367-380. Naitoh J, Kaboo R, Tso C, Hinkel A, Stiles A, Figlin R, and Belldegrun A (1998) Immunologic effects of interleukin-2 (IL-2) gene therapy in prostate cancer (CAP) (San Diego, CA). vol. 159, pp. 12. Nielsen LL, Lipari P, Dell J, Gurnani M, and Hajian G (1998) Adenovirus-mediated p53 gene therapy and paclitaxel have synergistic efficacy in models of human head and neck, ovarian, prostate, and breast cancer. Clin Cancer Res 4, 835846. Oefelein MG (1998) Time to normalization of serum testosterone after 3-month luteinizing hormone-releasing hormone agonist administered in the neoadjuvant setting: implications for dosing schedule and neoadjuvant study consideration. J Urol 160, 1685-1688. Ou Y, Gardner TA, Ko S-C, Zhau HE, Kao C, and Chung LWK (1999) Expression of osteocalcin in canine metastatic prostate cancer: A ideal prostate cancer animal model for osteocalcin promoter-based toxic gene therapy (Dallas, Texas). pp. 131. Pang S, Dannull J, Kaboo R, Xie Y, Tso CL, Michel K, deKernion JB, and Belldegrun AS (1997) Identification of a positive regulatory element responsible for tissue-specific expression of prostate-specific antigen. Cancer Res 57, 495-499. Pang S, Taneja S, Dardashti K, Cohan P, Kaboo R, Sokoloff M, Tso CL, Dekernion JB, and Belldegrun AS (1995) Prostate tissue specificity of the prostate-specific antigen promoter isolated from a patient with prostate cancer. Hum Gene Ther

Gene Therapy and Molecular Biology Vol 4, Page 247 6, 1417-1426. Partin AW, and Oesterling JE (1994) The clinical usefulness of prostate specific antigen: Update 1994. J Urol 152, 1358-1368. Partin AW, Pound CR, Clemens JQ, Epstein JI, and Walsh PC (1993) Serum PSA after anatomic radical prostatectomy. The Johns Hopkins experience after 10 years. Urol Clin North Am 20, 713-725. Patel B, Naitoh J, Stiles A, Tso C-L, Figlin R, and Belldegrun A (1999) Effect of interleukin-2 (IL-2) gene therapy on porstate specific antigen in patients undergoing neoadjuvant or adjuvant gene therapy for prostate cancer (Dallas, TX). vol. 161, pp. 337. Price MR, Sekowski M, Ladanyi A, Uray K, Ma Y, Durrant L, and Tendler SJ (1993) Immune recognition of human colonictumour-associated MUC-2 mucins using an anti-peptide antibody. Int J Cancer 55, 753-759. Rodriguez R, Lim HY, Bartkowski LM, and Simons JW (1998) Identification of diphtheria toxin via screening as a potent cell cycle and p53-independent cytotoxin for human prostate cancer therapeutics. Prostate 34, 259-269. Rodriguez R, Schuur ER, Lim HY, Henderson GA, Simons JW, and Henderson DR (1997) Prostate attenuated replication competent adenovirus (ARCA) CN706: A selective cytotoxic for prostate-specific antigen-positive prostate cancer cells. Cancer Res 57, 2559-2563. Rodriguez R, Yang Q, Kruszewski M, and Carducci M (1999) Androgen-insensitive prostate-specific gene therapy constructs are inducible with the differentiation agent, phenylbutyrate (Dallas, TX). vol. 161, pp. 126. Sanda M, Restifo N, Walsh J, Kawakami Y, Nelson W, Pardoll D, and Simons J (1995) Molecular characterization of defective antigen processing in human prostate cancer. J Natl Cancer Inst 87, 280-285. Sanda MG, Ayyagari SR, Jaffee EM, Epstein JI, Clift SL, Cohen LK, Dranoff G, Pardoll DM, Mulligan RC, and Simons JW (1994) Demonstration of a rational strategy for human prostate cancer gene therapy. J Urol 151, 622-628. Sanda MG, Smith DC, Charles LG, Hwang C, Pienta KJ, Schlom J, Milenic D, Panicali D, and Montie JE (1999) Recombinant vaccinia-PSA (PROSTVAC) can induce a prostate-specific immune response in androgen-modulated human prostate cancer. Urol 53, 260-266. Sanford MA, Hassen W, Atkinson G, and Hall SJ (1999) Interleukin-12 gene therapy induced cell death through induction of FAS/FAS ligand-mediated death in metastatic mouse prostate cancer (Dallas, Texas) pp. 57. Scardino PT, Adler HL, Herman JR, Woo SLC, Aguilar-Cordova E, Timme TL, and Thompson TC (1998) In Herpes simplex virus-thymidine kinase (HSV-tk) mediated gene therapy is safe and active against locally recurrent prostate cancer after irradiation therapy (Los Angeles:Lippincott Williams and Wilkins). vol. 17 (suppl), pp. 308a. Schellhammer PF, and Lynch DF (1997) Are monotherapy options reasonable for T3 prostate cancer? Semin Urol Oncol 15, 207214. Schmidt W, Schweighoffer T, Herbst E, Maass G, Berger M, Schilcher F, Schaffner G, and Birnstiel ML (1995) Cancer vaccines: The interleukin 2 dosage effect. Proc Natl Acad Sci USA 92, 4711-4714. Schroder FH (1995) Screening, early detection, and treatment of prostate cancer: A European view. Urology 46, 62-70. Segawa T, Takebayashi H, Kakehi Y, Yoshida O, Narumiya S, and Kakizuka A (1998) Prostate-specific amplification of

247

expanded polyglutamine expression: A novel approach for cancer gene therapy. Cancer Res 58, 2282-2287. Sikora K, and Pandha H (1997) Gene therapy for prostate cancer. Br J Urol 79 Suppl 2, 64-68. Simons JW, and Marshall FF (1998) The future of gene therapy in the treatment of urologic malignancies. Urol Clin. North Am 25, 23-38. Simons JW, Jaffee EM, Weber CE, Levitsky HI, Nelson WG, Carducci MA, Lazenby AJ, Cohen LK, Finn CC, Clift SM, Hauda KM, Beck LA, Leiferman KM, Owens AH Jr, Piantadosi S, Dranoff G, Mulligan RC, Pardoll DM, and Marshall FF (1997) Bioactivity of autologous irradiated renal cell carcinoma vaccines generated by ex vivo granulocytemacrophage colony-stimulating factor gene transfer. Cancer Res 57, 1537-1546. Simons J, Zhong H, Ramakrishna N, Drucker B, Rioseco-Camacho N, Baccala A, Rodriguez R, Yu D, Grous J, Henderson D, and DeWeese T (1998a) Cytoreductive gene therapy for human prostate cancer (Indian Wells, CA). Simons JW, Carducci MA, Weber CE, DeMarzo A, Baccala A, Cohen L, Clift SM, Mikhak B, Piantadosi S, Partin AW, Carter HB, Levitsky HI, Marshall FF, Mulligan RC, and Nelson WG (1998b) In Bioactivity of autologous irradiated prostate cancer vaccines generated by ex vivo GM-CSF gene transfer (Los Angeles: Lippincott Williams & Wilkins). vol. 17 (suppl), pp. 313a. Simons J, Baccala A, and Lim H (1999a) In Ex vivo and in vivo gene therapy strategies for prostate cancer: Emerging clinical pharmacology, E Lattime and S Gerson, eds. (San Diego, CA: Academic Press), pp. 333-345. Simons J, Mikhak B, Chang J, DeMarzo A, Lim M, Carducci M, Goemann M, Marshall F, Weber C, Piantodosi S, and Nelson W (1999b) Clinical activity and broken immunologic tolerance from ex vivo GM-CSF gene transduced prostate cancer vaccines (Dallas, Texas) vol. 161, pp. 51. Simons JW, Baccala AA, and Lim HY (1999c) In Gene therapy of cancer : Translational approaches from preclinical studies to clinical implementation, EC Lattime and SL Gerson eds. (San Diego: Academic Press), xvii, pp. 406. Soh S, Kattan MW, Berkman S, Wheeler TM, and Scardino PT (1997) Has there been a recent shift in the pathological features and prognosis of patients treated with radical prostatectomy? J Urol 157, 2212-2218. Soloway M, Sharifi R, Wajsman Z, McLeod D, Wood Jr D, PurasBaez A, and Group LDNS ( 1997) Radical prostatectomy alone vs. radical prostatectomy preceded by androgen blockage in cT2b prostate cancer - 24 month results. J Urol, part 2 157, 160. Stamey TA, and McNeal JE (1992) In Adenocarcinoma of the prostate, PC Walsh, AB Retik, TA Stamey, and ED Vaughan Jr. eds. (Philadelphia: W.B.Saunders Co.), pp. 1159-1199. Steiner MS, Satterwhite DJ, and Moses HL (1995) Molecular insights into altered cell cycle regulation and genitourinary malignancy. Urol Oncol 1, 3-17. Steiner MS, Anthony CT, Lu Y, and Holt JT (1998a) Antisense cmyc retroviral vector suppresses established human prostate cancer. Hum Gene Ther 9, 747-755. Steiner MS, Lerner J, Greenberger M, Carraher J, Smith JJ, Koch M, Obermiller P, and Holt J (1998b) Clinical phase I gene therapy trial using BRCA1 retrovirus is safe. J Urol 159, 133. Steiner M, Lerner J, Greenberger M, Carraher J, Smith Jr J, Koch M, Obermiller P, and Holt J (1998c) Clinical Phase I gene therapy trial using BRCA1 retrovirus is safe (San Diego,

Steiner and Gingrich: Gene Therapy for Prostate Cancer CA) pp. 132. Steiner M, Zhang Y, Allay J, Reed C, Wang Y, Farooq F, Lerner J, and Lu Y (1999a) Preclinical studies of adenoviral vector containing wild type p16 (GTx-001) on its suppression of prostate cancer growth and prolongs survival in animal model (Washington, DC). Steiner M, Zhang Y, Carraher J, and Lu Y (1999b) In vivo expression of prostate-specific adenoviral vectors. Cancer Gene Therapy (In press). Tutrone RF, Ball RA, Ornitz DM, Leder P, and Richie JP (1993) Benign prostatic hyperplasia in a transgenic mouse: A new hormonally sensitive investigatory model. J Urol 149, 633639. Vieweg J, Boczkowski D, Roberson K.M, Edwards DW, Philip M, Philip R, Rudoll T, Smith C, Robertson C, and Gilboa E (1995) Efficient gene transfer with adeno-associated virus-based plasmids complexed to cationic liposomes for gene therapy of human prostate cancer. Cancer Res 55, 2366-2372. Vieweg J, and Gilboa E (1995) Gene therapy approaches in urologic oncology. Surg Oncol Clin N Am 4, 203-218. Vieweg J, Rosenthal FM, Bannerji R, Heston WD, Fair WR, Gansbacher B, and Gilboa E (1994) Immunotherapy of prostate cancer in the Dunning rat model: use of cytokine gene modified tumor vaccines. Cancer Res 54, 1760-1765. Walsh PC (1994) Prostate cancer kills: Strategy to reduce deaths. Urology 44, 463-466. Wheeler JA, Zagars GK, and Ayala AG (1993) Dedifferentiation of locally recurrent prostate cancer after radiation therapy. Evidence for tumor progression. Cancer 71, 3783-3787. Wieder JA, and Soloway MS (1998) Incidence, etiology, location, prevention and treatment of positive surgical margins after

Mitchell S Steiner

radical prostatectomy for prostate cancer. J Urol 160, 299-315. Williams B, Acree D, Venable D, and Eastham J (1998) Expression of the translation initiation factor elF4E modulates prostate tumor vascularization (Seattle, WA). pp. 117a. Wilson D, Lapadat-Tapolsky M, Gjerset R, Lebedeva S, Colletier P, and Meyn R (1999) Treatment of prostate cancer with Ad5CMV-p53 as a single agent or in combination with DNA damaging agents (Philadelphia, PA). vol. 40, pp. 6. Wood D, Sakr WA, Grignon DJ, Swiriduk M, Novalo M, Cher M, Powell IJ, and Pontes JE (1997) "Down staging" with neoadjuvant hormone treatment: The more organ confined prostate cancer (PCa) does not translate into lower PSA failure. J Urol, part 2 157, 392. Yin L, Fu S, Peng X, Wang W, Cote R, Crystal R, Poo W, Peschel R, Pizzorno G, and Deisseroth A (1998) In Cytosine deaminase (CDA) adenoviral vector (ADV) and administration of the non toxic prodrug 5-fluorocytosine (5FC) used for therapy sensitization (TS) of prostate cancer cells. Yoshimura I, Heston W, Gansbacher B, and Fair W (1996) Cytokine mediated immuno-gene therapy in rat prostate cancer model (Orlando, FL). vol. 155, pp. 510A. Zagars GK, von Eschenbach AC, Ayala AG, Schultheiss TE, and Sherman NE (1991) The influence of local control on metastatic dissemination of prostate cancer treated by external beam megavoltage radiation therapy. Cancer 68, 2370-2377. Zhang J-F, Kasper S, Koch M, Smith JJ, and Matusik R (1998) Gene therapy vectors for prostate cancer (San Diego, CA). vol. 159, pp. 10.

Jeffrey R Gingrich

248

Gene Therapy and Molecular Biology Vol 4, page 249 Gene Ther Mol Biol Vol 4, 249-260. December 1999.

The HSV-TK/GCV gene therapy for brain tumors Review Article

Naoto Adachi1,2,3, Dilek L KĂśnĂź2,3, Karl Frei2,3, Peter Roth2, Yasuhiro Yonekawa2,3 1 Department of Neurosurgery, Yamaguchi University School of Medicine 2 Department of Neurosurgery, University Hospital ZĂźrich 3 GLI-328 Study Group __________________________________________________________________________________________________ Correspondence: Naoto Adachi, MD&PhD. Department of Neurosurgery, Yamaguchi University School of Medicine, MinamiKogushi 1-1-1, Ube, Yamaguchi 755-8505, Japan. Tel +81-836-22-2295; Fax +81-836-22-2294; E-mail nadacchi@po.cc.yamaguchiu.ac.jp Abbreviations: HSV-TK, herpes simplex virus thymidine kinase; GCV, ganciclovir; Tf, transferrin; VPC, vector producer cell; GFAP, glial fibrillary acidic protein; CNS, central nervous system; RCR replication competent retrovirus; MR, magnetic resonance Key Words: HSV-TK/GCV, suicide gene therapy, brain tumor, kinase, retroviral vectors, vector producer cells, glioblastoma, bystander effect Received: 11 May 1999, accepted: 26 May 1999

Summary Herpes simplex virus-thymidine kinase (HSV-TK) gene transduction followed by ganciclovir (GCV) administration is widely used for cancer treatment as a "suicide gene" therapy. The present review describes the HSV-TK/GCV gene therapy for brain tumors (gliomas) through the eyes of a clinician, including a review of preclinical and clinical studies. In particular, the latest data on the first clinical trial are discussed and, moreover, the surgical procedures are depicted. The surgical technique is the most important for neurosurgeons; therefore, its improvement would be beneficial for further clinical developments as well as biological innovation of retroviral vectors or vector producer cells. that gene therapy must be safe and feasible for its potential applications in humans (Crystal, 1995).

I. Introduction The great advance in molecular biology has been contributing to the development of gene therapy. The gene therapy offers several novel approaches to many diseases, which are difficult to be cured by conventional treatments. In 1990, a 3-year-old girl received retrovirally transduced lymphocytes that produced the enzyme adenosine deaminase, which was absent in her own genome and had rendered her susceptible to life-threatening infections. This was the first case treated with performing the "gene therapy" in a human being. To date there have been more than 200 clinical trials of "gene therapy" in humans (Human Gene Marker/Therapy Clinical Protocols, NIH 1998).

In the cancer gene therapy, a number of target genes have been used or are expected to be applied, for example, oncogenes, tumor suppressor genes, immune related genes. Among these, the prodrug genes ("suicide genes") offer a new system for cancer gene therapy. Here the HSV-TK/GCV (herpes simplex virus thymidine kinase/ganciclovir) gene therapy for brain tumor is reviewed as one of prodrug "suicide gene" therapies. In particular, we emphasize the clinical aspects of this approach including a clinical trial and surgical technique which is expected to become the most important for the clinical application, especially for neurosurgeons.

There are many potential advantages of the gene therapy over conventional drug therapies. Gene therapy could eliminate the need for repeated drug administration, risks of immunogenicity, pharmacological tolerance or dose toxicity, and it could be less costly. The great advantage may develop the ability of a target therapy in any disease. On the other hand, an important caution is

II. Brain Tumors A. Glioma and glioblastoma multiforme Astrocytic tumors are the most common type of intracranial tumor among adults. Astrocytic tumors are classified into four grades according to WHO: (i) pilocytic astrocytoma (grade 1); (ii) astrocytoma (grade 2); (iii) 249

Adachi et al: HSV-TK/GCV gene therapy for brain tumors

A. "Suicide gene" therapy (see chapter IV)

anaplastic astrocytoma (grade 3); and (iv) glioblastoma multiforme (grade 4). Glioblastoma is the most common of all gliomas and occurs throughout the central nervous system in all decades of life, but most frequently in the fifth decade or above. The incidence rate of glioblastoma is reported to be 2.36 patients per 100,000 persons in USA (Davis et al, 1996). The prognosis of patients with newly diagnosed as well as recurrent glioblastoma is poor. Survival of most patients diagnosed with glioblastoma is less than two years and the tumor recurs most certainly following the initial treatments with the combination of surgery, irradiation and chemotherapy (Ammirati et al, 1987). The median survival time of recurrent glioblastoma is approximately seven months (Ammirati et al, 1987).

Several systems have been developed for gene therapy of brain tumor (Zlokovic and Apuzzo, 1997). The most feasible and well established approach is the suicide gene therapy, which uses a drug susceptible gene for selective destruction of the tumor cells.

B. Oncogene related therapy (antisense or replacement of genes) Progression of different brain tumors to more malignant phenotypes involves numerous molecular genetic alterations. The genes affected by these alterations are considered to be those responsible for cell cycling, apoptosis, signal transduction and angiogenesis. Especially used genes could be classified into oncogenes and tumor suppressor genes. Another group contains genes related to DNA replication and repair. Complementary DNA (cDNA) of these genes is being introduced into tumor cells in sense or antisense orientation with the purpose of leading to genetic recovery. Tumor suppressor genes, such as p53, Rb, p21, are expected to have a great effectiveness against gliomas. However, several issues of gene delivery, selectivity, efficacy and even selection of target genes are still unresolved.

B. Therapy of glioblastomas Up to now the selection of an appropriate therapy for glioblastoma involves surgery, radiation therapy and chemotherapy. Despite a multimodality approach with these therapies, the prognosis remains poor. The first significant improvement in therapy was achieved thanks to the introduction of radiation (Walker et al, 1980). Radiation improves the median survival time from 14 weeks in patients treated with surgery alone to 40 weeks in those treated with surgery and radiation. Subsequently, chemotherapy has been used during and after radiation therapy. The addition of chemotherapy to surgery and radiation was evaluated in a recent metaanalysis (Fine et al, 1993). Adjuvant chemotherapy did improve survival time compared with surgery and radiation without chemotherapy. However, the survival benefit occurred in patients with earlier-diagnosed and lower grade tumors. In addition, the combinations of chemotherapeutic agents have been studied. Cisplatin showed synergy with nitrosoureas in preclinical in vitro models (Durand, 1990). In clinical studies, cisplatin showed activity in high grade gliomas with response rates ranging from 13 to 28% (Spence et al, 1992). Moreover the administration of BCNU and cisplatin with accelerated hyperfractionated radiation therapy is expected to offer higher therapeutic efficacy against high-graded gliomas (Rajkumar et al, 1999).

C. Other approaches A number of additional approaches of gene therapy, include anti-angiogenesis, immunotherapy, and toxic gene therapy.

IV. Prodrug "suicide gene" therapy "Suicide" genes have been introduced into cancer cells to allow their elimination. This approach is also called the enzyme/prodrug system because the suicide gene encodes an enzyme that modifies a nontoxic prodrug into a toxic molecule in the cell. Only the cells bearing the suicide gene will be killed upon the subsequent prodrug treatment. The strategy of prodrug gene therapy could be summarized as follows: (i) the introduction and selective expression of a gene encoding drug-metabolizing enzyme in target cells; (ii) the prodrug is administered systemically; (iii) the non-toxic prodrug is converted to its active and toxic form in transduced target cells; (iv) the toxic drug damages and finally kills only target cells. The requirements for an ideal prodrug-enzyme system are a combination of a harmless prodrug and a toxic active drug, both of which should be compounds that have been well studied in humans. In addition, the drug-metabolizing enzyme should be of non-mammalian origin, or only expressed in very small quantities in normal human cells. The suicide gene therapy is supposed to have a number of

On the other hand, the regional therapy using the affinity to transferrin (Tf) receptors has been developed (Laske et al, 1997). The targeted protein toxin was transferrin-CRM107 (Tf-CRM107) containing diphtheria toxin (CRM107). Tf-CRM107 potently and specifically kills cells expressing Tf receptors including glioblastoma. At least a 50% reduction in tumor volume occurred in 9 of 15 patients with malignant brain tumors.

III. Gene therapy for brain tumors 250

Gene Therapy and Molecular Biology Vol 4, page 251

A. Vector producer cell (VPC)

benefits for cancer therapy including: (i) selective sensitivity to a toxic drug; (ii) short-term gene expression; (iii) induction of a "bystander effect"; and (iv) stimulation of an immune response.

The structural genes (gag, pol and env) of retrovirus used to create the vector producer cell (VPC) are derived from Moloney murine leukemia retrovirus (MoMLV). The MoMLV retrovirus is frequently used for gene therapy because its biology is relatively well understood, the vector systems can produce high titers and this virus can infect human cells efficiently. The HSV-TK gene is introduced on the vector plasmid and becomes stably integrated into the packaging cell resulting in a VPC (Figure 1). The VPC has all the components to transduce dividing cells with replication-incompetent virions containing the HSV-TK gene. It should be noted that once the recombinant vector with the HSV-TK gene transduces tumor cells, it cannot replicate because it no longer has access to replication proteins (gag, pol and env) (Ramesh et al, 1998).

A number of approaches for prodrug-enzyme system have been established to date. The most widely used enzyme/prodrug system is the HSV-TK/GCV, which is currently being tested in 40 clinical trials, mostly for the treatment of cancers (Human Gene Marker/Therapy Clinical Protocols, NIH 1998). There are many other prodrug-enzyme systems with different mechanisms of actions, including, for example, cytosine deaminase/5fluorocytosine and varicella zoster virus thymidine kinase/6-methoxypurine arabinoside (Rigg and Sikora, 1997).

V. Principle of the HSV-TK/GCV system

Figure 1. The principle of HSV-TK/GCV gene therapy. 1) The VPCs are injected into the tumor-resected cavity and inoculated in the brain parenchyma, where residual tumor cells are expected to be found. 2) The VPCs produce retroviral vectors containing HSV-TK gene. 3) The retroviral vectors infect only tumor cells specifically in the proliferative phase. 4) The HSV-TK gene is integrated into the genome DNA of tumor cells. 5) The anti-virus drug, GCV, is administered intravenously. 6) GCV delivered to HSV-TK positive cells is phosphorylated into GCV triphosphate. 7) The toxic GCV triphosphate kills only HSV-TK positive tumor cells.

251

Adachi et al: HSV-TK/GCV gene therapy for brain tumors effect: (i) The transfer of toxic GCV metabolites from HSV-TK-transduced tumor cells to nearby unmodified cells (Bi et al, 1993). (ii) Release of soluble factors from GCV-exposed HSV-TK-transduced tumor cells and subsequent response by the immune system (Freeman et al, 1993). (iii) Hemorrhagic or ischemic tumor necrosis due to the transfer of the HSV-TK gene to endothelial cells lining the tumor blood vessels (Ram et al, 1993).

B. Inoculation The VPCs are injected directly into the tumor itself or the brain walls around the excised tumor. The proliferating tumor cells are transduced by vectors containing the HSVTK gene. Non-proliferating cells, such as neuron, are resistant to transduction. Once the HSV-TK gene is integrated in the genomic DNA of tumor cells, the enzyme thymidine kinase is expressed constitutively. GCV is phosphorylated by the viral thymidine kinase into GCV monophosphate and further phosphorylated to GCV triphosphate by the mammalian thymidine kinase. GCV triphosphate inhibits DNA replication by hindering the activity of DNA polymerase and results in tumor cell death. The viral thymidine kinase is almost 1000-fold more efficient at monophosphorylating GCV compared with the mammalian thymidine kinase (Elion et al, 1977). Bone marrow is the most sensitive to GCV and granulocytopenia and thrombocytopenia are common dose-limiting toxicities of this agent. However, GCV is nontoxic to most non-transduced cells at therapeutic concentrations of the pro-drug (Crumpacker, 1996).

VI. Preclinical studies for brain tumors A. Preclinical trials Preclinical studies (Table 1) have demonstrated that HSV-TK gene transduction induces sensitivity to GCV and results in killing of tumor cells. In the brain, specific gene transduction into tumor cells is possible because tumor cells usually divide actively. In contrast, brain cells do not usually divide and are resistant to transformation by retroviral vectors (Salmons and Gunzburg, 1993). The other cells in the brain are at minimal risk for transduction; these include glia, endothelial cells, microglia and blood derived cells (leukocytes). Moreover, the lack of a strong immunological response in the brain allows for sufficient survival of xenogeneic VPCs and subsequent transduction of the tumor cells.

C. "Bystander effect" In this therapy, indirect antitumor effects occur and non-transduced tumor cells are killed. This phenomenon is termed the "bystander effect" (Freeman et al, 1993). In fact, the rates of tumor cell destruction with HSVTK/GCV gene therapy exceed those expected with the transduction rates (Samejima et al, 1995). In vitro studies show that HSV-TK-negative tumor cells can be killed when as few as 10% of the population of cultured cells are HSV-TK transduced cells (Freeman et al, 1993). There are several potential mechanisms explaining the bystander

Moolten (1986) first reported that HSV-TK geneinduced GCV sensitivity could be demonstrated in vivo using mouse fibroblast 10E2. The TK-positive and negative cells were tested to determine the response exposed to GCV in vivo. The TK-positive tumors treated with GCV demonstrated complete tumor regression. In contrast, the TK-negative tumors after GCV treatment exhibited tumor progression and all mice died.

Table 1 Preclinical trials of HSV-TK/GCV gene therapy in glioma cells; antitumor response (glioma cells containing the TK gene)

Author

Model

Glioma Cell

Response

Ezzeddine (91)

in vivo

mouse/C6(TK+)

growth inhibition

Barba (93)

in vivo

rat/9L(TK+)

100% survival at 90 days

Kato (94

in vivo

T98, U251(TK+)

cytotoxicity to GCV (1000-fold)

Kim (94)

in vivo

U251(TK+)

induction of sensitivity to radiation (1.9-fold)

LeMay (98)

in vivo

rat/C6(TK+)

induction of GCV permeability by RMP-7 (1.7-2.6-fold)

Moriuchi (98)

in vivo

U87MG(TK/TNF+)

TNF enhances cytotoxicity

Vandier (98)

in vivo

U251, C6(TK+)

GFAP promoter enhances sensitivity to GCV (2-fold)

252

Gene Therapy and Molecular Biology Vol 4, page 253 procedure. In addition, the median survival was prolonged 1.6 times for treated rats compared to untreated ones. Ram et al (1993a, 1994) showed the dose dependency of the number of VPCs. Rats transplanted with 9L gliosarcoma cells were inoculated with HSV-TK positive VPCs followed by administration of GCV. Five of 13 (38%), 13 of 18 (72%) and 10 of 12 (83%) rats treated with 1.8x106, 3x106 and 5x106 VPCs experienced complete tumor regression, respectively (Ram et al, 1993a). In addition, the number of tumor vasculature was also reduced by in vivo TK gene transfer and GCV treatment (Ram et al, 1994).

In regard to treatment of glioma, TK-positive glioma cells were sensitive to GCV and tumor growth was reduced in both mice and rats (Ezzeddine et al, 1991; Barba et al, 1993). In vitro studies showed that cytotoxicity to GCV was enhanced by 100- to 1000-fold by introducing TK gene (Kato et al, 1994). In addition, TK gene introduction retained sensitivity to radiation (Kim et al, 1994). Recently, induction of sensitivity to GCV has been developed using advanced molecular techniques. Expression of TK gene driven by GFAP (glial fibrillary acidic protein) promoter enhanced sensitivity to GCV by 2-fold (Vandier et al, 1998). The tandem expression of both TK and TNF! (tumor necrosis factor alpha) genes also induced cytotoxicity to GCV by 80% (Moriuchi et al, 1998). On the other hand, a bradykinin analog (RMP-7) induces permeability of GCV into brain parenchyma through the blood-brain barrier (LeMay et al, 1998). Taken together, TK gene transduction into glioma cells induces sensitivity to GCV, which leads a feasibility of the HSV-TK/GCV system for the treatment of glioma.

2. Adenovirus-mediated transfer Some study groups used adenovirus instead of retrovirus. The advantage of using adenovirus is the high efficacy of infection rate into the target cells, approaching nearly 100%. However, the major problem of adenovirus is the induction of a strong immune response against the virus itself as well as the transduced cells expressing viral proteins. In both rats and mice, tumor regression and elongation of survival time were observed (Chen et al, 1994; Perez-Cruet et al, 1994; Maron et al, 1996). Vincent et al (1996) compared the antitumor response between retrovirus and adenovirus. Adenoviral gene transfer showed longer survival rates (39 days vs 26 days). According to Human Gene Marker/Therapy Clinical Protocols of NIH 1998, two protocols of adenovirusmediated HSV-TK/GCV therapy for brain tumor were approved for clinical trials to date (Human Gene Marker/Therapy Clinical Protocols 98).

B. Biological gene transfer (see Table 2) 1. Retrovirus-mediated transfer Moolten and Wells (1990) demonstrated that retroviral vectors produced by VPCs could successfully transmit the HSV-TK gene into tumor cells and induce tumor regression in response to GCV. Short et al (1990) compared the efficiency of gene transfer using two methods; the procedure of in vivo gene transfer via VPC was much more efficient than that of the direct vector injection. Culver et al (1992) observed complete tumor regression in 11 of 14 rats treated with the same

Table 2 Preclinical trials of HSV-TK/GCV gene therapy in glioma cells; antitumor response (in vivo TK gene transfer) retrovirus-mediated gene transfer Author Model VPCs Response Short (90) rat/C6 psi2-BAG/•in vivo VPC transfer>vector injection Culver (92) mouse/9L PA317/G1NsCTK •reduction of tumor growth (5w; 80%) Ram (93) rat/9L PA317/G1TkSvNa.90 •reduction of tumor growth (28d; 83%) Barba (94) rat/9L (fibroblast/HSV-TK) •long survival (90d; 22%) Ram (94) rat/9L PAT24/G1TkSvNa.53 •reduction of tumor vasculature (14d; 80%) Vincent (96) rat/9L PA317/IGRVTK •survival prolonged (15d->26d) adenovirus-mediated gene transfer Author Model Chen (94) mouse/C6 Perez-Cruet (94) rat/9L Maron (96) rat/9L

Adeno-vector ADV/RSV-TK ADVAd.RSVTK

Vincent (96)

IG.Ad.MLP.TK

rat/9L

Response •volume reduction (1/500) •survival prolonged (22->80~120d) •reduction of tumor volume (97%) •elongation of suvival (22->101 d) • survival prolonged (16d->18d) • adeno > retro (26d>39d)

253

Adachi et al: HSV-TK/GCV gene therapy for brain tumors be relevant to human. Three doses of GCV that were employed in this study (10, 20 and 30 mg/kg per day) resulted in comparable tumor regression as assessed by tumor weight. In consequence, the relevant human GCV dose is expected 10 mg/kg per day intravenously.

C. “Bystander effect” in glioma (see Table 3) Ram et al (1993a) described that following successful transduction of 10% to 70% of glioma cells, complete tumor ablation occurred in most rats. These data suggested that some indirect effects might induce tumor cell death, termed "bystander effect". The bystander effect was observed in other types of cell lines (colon cancer, fibrosarcoma and so on) treated with HSV-TK/GCV system (Freeman et al, 1993; Bi et al, 1993). Bystander effect also contributes to size reduction of tumor and elongation of survival (Namba et al, 1998). The bystander mechanism is still unknown. The cell-to-cell contact is essential for bystander effect and any soluble factors don't influence to this mechanism (Samejima and Meruelo, 1995). In addition, bystander effect is mediated with gap junction, especially implied with connexin 43 (Mensil et al, 1996). On the other hand, GCV triphosphate accumulates in neighboring cells, which is supposed to lead bystander killing (Rubsm et al, 1999).

GCV causes granulocytopenia, thrombocytopenia, azoospermia, and a rise in serum creatinine (Crumpacker, 1996). Among the patients with cytomegalovirus retinitis treated with intravenous GCV (5mg/kg twice a day for three months), 40% developed granulocytopenia (neutrophil count <1000/ml) and 15% developed thrombocytopenia (platelet count<50,000/ml). These symptoms disappeared after GCV administration was discontinued. It was suggested that in case of GCV administration for two weeks in clinical trials, these toxicity symptoms might become less severe.

E. Cytotoxicity of VPC in vivo Ram et al (1993b) estimated the safety of the injection of VPCs containing HSV-TK gene or lacZ gene. They concluded that this system is not associated with significant toxicity to the brain and remote organs in mice, rats or monkeys. It should be noted that even in case all of the injected VPCs were able to cross the blood-brain barrier following intracerebral administration, there would be fewer vector particles in relation to the large number of cell receptors in the body and thus the risk to tissues beyond the CNS would be minimal. In regard to the random integration of the provirus into the host genome, it is unlikely that normal cells would undergo insertional mutagenesis. Furthermore, cells containing the HSV-TK gene would be destroyed following GCV administration. In fact, no such mutagenesis was observed.

Lyons et al. (1995) demonstrated bystander effect observed in C57BL/6 mice model injected with a mixture of tumor cells and VPCs followed by GCV treatment. In mice received a 10:1 ratio of tumor cells to VPCs, 29% of tumor cells was ablated, even though only 18% of tumor cells were transduced with HSV-TK gene. In addition, retrovirus-mediated gene transfer followed by GCV treatment induced apoptosis in neighboring cells (Colombo et al, 1995; Hamel et al, 1996).

D. Optimal dose of GCV Ram et al (1993b) also analyzed the optimal dose of GCV because previous studies used high doses as 300 mg/kg per day intraperitoneally. Such doses would be toxic and not

Table 3. Preclinical trials of the HSV-TK gene therapy in glioma cells; "bystander effect"

glioma cell containing TK gene Author

Cell Line

Bystander Mechanism

Samejima (95)

U118MG(TK+) C6(TK+)

• cell-to-cell contact • not mediated with soluble factors

Namba (98)

rat/9L(TK+)

• (survival elongation) • (size reduction)-

Rubsm (99)

U251(TK+)

• GCV triphosphate accumulation

in vivo TK gene transfer Author

Cell Line

VPCs

Bystander Mechanism

Colombo (95)

U87

PA317/G1NsCVTK

• apoptosis (nuclear segmentation)

Hamel (96)

PA317/HSVTK

• apoptosis (inhibited by BCL2)

254

Gene Therapy and Molecular Biology Vol 4, page 255

VII. Clinical trial for malignant brain tumor

1. Studies in mice using intravenous administration of VPCs

Up to date 226 protocols of clinical trials using gene therapy have been approved by NIH. Eight out of 226 are protocols using the HSV-TK/GCV gene therapy for brain tumors (Table 4). The data of the first study (phase I; GLI-0100) were published in 1997 being the only one for malignant brain tumors (Ram et al, 1997).

Intravenous administration of VPCs in mice was used to determine whether VPCs could be trapped in lungs and cause transduction of the respiratory epithelium and diffuse destruction of lung parenchyma after GCV treatment (Ram et al, 1993b). Histological examination of lung, spleen, thymus, liver, intestine and bone marrow did not reveal toxicity nor necrosis and inflammation. Some of the treated mice were followed for more than seven months and no toxicity was observed.

A. Surgical procedure administration of VPC

for

stereotactic

Stereotactic administration permits direct treatment of the initial tumor after biopsy. However, this procedure is limited to small-size tumors (Ram et al, 1997). Ram described that the antitumor response was observed in small tumors (1.4Âą0.5 ml) but not in larger-size tumors (7.4Âą5.8 ml). To determine an efficacy for large-size tumors, the second clinical trial of injecting VPCs into the resected tumor (newly diagnosed) cavity is conducted (phase III).

2. Studies in rats using direct brain injection of VPCs The HSV-TK VPCs were injected into the deep white matter of rat cerebrum, treated with GCV and sacrificed (Ram et al, 1993b). A mild and transient response including weakness, weight loss, somnolence and dehydration was evident during the initial 24 to 48 hours after VPC injection; however, this effect was reversible. A similar response was seen at the first day of GCV therapy. This response is expected to arise from an immune reaction to the xenogeneic cells. Histological examination of brain tissue showed moderate edema around the injection site. Moreover, administration of dexamethasone completely abolished this response in all treated rats.

B. Preparation of VPC suspension The VPCs, suspended in Plasma Lyte-A or Ringer's Lactate Solution at a concentration of 1x108 cells/ml, are inoculated slowly at sites distributed as evenly as possible around the tumor or to the wall of the resection cavity. To avoid settling of cells, the cell suspension should be gently and frequently massaged prior to drawing up into the syringe for injection.

3. Studies in monkeys using direct brain injection of VPCs Five monkeys were used to investigate the impact of intracerebral injection of VPCs followed by GCV therapy (Ram et al, 1993b). The monkeys received a stereotactic intracerebral injection of 1x107 HSV-TK VPCs in the deep white matter of the right frontal lobe. GCV was administered to monkeys, and physical and neurological examinations were performed daily. In addition, magnetic resonance imaging, blood analysis and cerebrospinal fluid analyses were conducted. There was no evidence of CNS toxicity following neurological examinations, nor changes in motor or behavioral activity. A few endothelial cells at the injection site were transduced. At the injection site, mild reactive gliosis without edema or pathological changes in surrounding brain tissue was observed. Localized demyelination was limited to the injection site and did not increase in size in response to GCV administration. There was no evidence of VPC proliferation in the brain. The VPCs were observed in the brain at two weeks but not at three weeks, similar to the rat model. Analysis of CSF revealed normal levels of protein and glucose, and negative bacteriological cultures. At 270 days post-injection of VPCs, no detrimental effects or alterations of baseline neurological function were noted.

C. Study design of the first clinical trial (GBI0100) (Ram et al, 1995, 1997) The first clinical trial using HSV-TK/GCV system was designed to determine: (i) the safety of intratumoral delivery of VPCs into human brain tumors; and (ii) the short and long term efficacy. This study was a small study of 15 patients with malignant brain tumors; 12 had malignant glioma (9 with glioblastoma) and 3 had secondary metastatic brain tumors (2 with melanoma and 1 with breast cancer). Patients received intratumoral stereotactic injection of TK-positive VPCs (murine PA317/G1TkSvNa.53). Seven days after the inoculation, GCV was administered at 5 mg/kg twice a day for 14 days.

D. Antitumor response on clinical trial Thirteen of 15 patients (87%) were included in this trial with 16 evaluable tumors (Ram et al, 1995, 1997). Five lesions (31%) in four patients (29%) had either a partial response (three lesions) or complete response (two lesions). The enhancing volume of these five small tumors 255

Adachi et al: HSV-TK/GCV gene therapy for brain tumors

XI. Surgical technique of injection

decreased from 1.4±0.5 ml before therapy to 0.4±0.5 ml after therapy. The five tumors were smaller at treatment than the nonresponsive tumors (1.4 ml vs 7.4 ml).

In contrast to the stereotactic injection, a multi-injection into the wide surface of tumor is needed for larger tumors. In general, the size of glioblastoma at diagnosis is usually about 30 ml. This multi-injection will be more convenient for the practical technique of VPC inoculation. Here we demonstrate a) surgical procedure for multi-injectional administration of VPC, b) characteristic MRI features of VPC implantation and c) injection technique and pitfalls, from our own experience of a phase II clinical trial (GLI-B201).

E. Clinical safety of the trial Two patients had intratumoral hemorrhage and neurological deficits from the procedures used to take biopsies and implant VPCs (Ram et al, 1995, 1997). One of two had completely recovery of deficits. Another patient required removal of hemorrhage. Two more patients had an increase in the frequency of preexisting seizures immediately after VPC injection. One patient was suffering from seizures in the immediate period after VPC implantation. Anticonvulsant medications were able to control seizures in three patients.

A. Surgical procedure for multi-injectional administration of VPC (Adachi, 1998) (Figure 2) 1. Preparation: This procedure is the same as that of the stereotactic injection (see chapter VII-B).

F. Biosafety of clinical trials Replication competent retrovirus (RCR) was not detected in any of the 79 blood samples from the 13 patients in this trail (Ram et al, 1995; 1997). In addition, the vector DNA was not detected in any of blood samples using PCR assay. Antibodies reacting with retrovirus core protein p30 were not detected in any of the same samples. Multiple serum samples from each of 15 patients were analyzed for the presence of antibodies to the injected murine VPC using flow cytometric analysis. Ten of 15 patients produced a readily detectable increase in VPC-binding antibody, which peaked at 66 days after VPC inoculation.

2. Marking: It is required to create a grid with clips to mark 1-cm apart line on the tumor cavity. This grid would be helpful for surgeons to inject at 1-cm intervals. In our experiences, it should be better to use Dubtamp® sheet, which is easy to see 1-cm2 square at a glance, to place on the cavity surface and to save the operation time. 3. Sites: In general, the number of injection is needed 50 to 100-site in 30 cm3-volume tumor cavity. It is recommended 100sites to increase an efficacy of infection to tumor cells. However, the number of sites depends on the tumor size, tumor location, operation time even more technique of an operator.

VIII. Conclusion The HSV-TK/GCV gene therapy promises one of the advanced therapies for brain tumors as well as for a variety of other cancers. As with all types of gene therapy, this approach might develop as a safe and effective clinical modality.

4. Volume: The volume per injection site is determined by dividing the total volume of cells by the number of injection sites. Actually, 0.1 ml of VPC suspension is usually injected into a site, which means 1x107 cells per one site. The depth of the injection should be approximately 1 to 2 cm to the resection margin without injury of eloquent region or entering the ventricle. It is recommended that the elastic needle should be marked every 1 cm.

One of the important points of a gene therapy approach is to select a target gene, which could become the most effective molecular tool against a tumor. Another point is to choose a gene delivery protocol, which could achieve specific and selective integration of a target gene into tumor cells. In our opinion, the best advantage would be to use a single target gene for gene therapy because of convenience of manipulation. In this point of view, the suicide gene therapy using HSV-TK gene is expected to be almost ideal. However, this approach has several points for improvement before it becomes a practical treatment modality for brain tumors. In the near future, many problems might be resolved establishing gene therapy as a strong treatment not only for brain tumors but for other malignancies alike.

5. Reflux: At the surgeon's discretion, a hemostatic sponge may be placed over the opening of the cavity to impede extravasation of VPCs into the subarachnoid space. 6. Infiltration: The histological examination showed the infiltration of VPCs along the needle tract. At four days after injection, fibroblast cells were seen 3 to 4 mm from the injection tract (Ram 97). Injection of VPC suspension at 1-cm intervals allows proximity to tumor cells for transduction.

256

Gene Therapy and Molecular Biology Vol 4, page 257

Figure 2. Injection technique. The VPCs suspension is directly injected into the tumor-resected cavity. The tip of the needle is usually placed at the depth 1.5~2.0 cm from the cavity edge.

Figure 3. Typical MRI finding after VPC inoculation. Gd-enhanced MR image shows needle tracts, which radiate from the tumorresected cavity. This finding is termed "Uni" sign ("Uni" means sea urchin in Japanese).

257

Adachi et al: HSV-TK/GCV gene therapy for brain tumors

Figure 4. Pitfalls of VPC injection.. Left: The unwilling injection into ventricle or basal ganglia. The great care should be taken for the depth of injection near to ventricle or basal ganglia. If the ventricle is opened, the VPCs suspension flows away in CSF space which results in low inoculation. Middle: The blind corner. The cortex side is difficult to be injected because of the blind space for surgeons. This can be resolved by bending the needle. Right: "Reflux". During the injection procedure, a significant quantity of the VPC suspension is lost through "reflux". The reflux phenomenon could be prevented by slowing and pumping injection while pulling out the syringe.

B. Characteristic MRI features of VPC implantation

2. Blind corner of cortex Glioma grows in the white matter and occupies subcortex regions. Neurosurgeons remove tumor in the white matter, which lead to make the sharp angle corner under the cortex. It is difficult that VPCs suspension is injected into this region using conventional method with the straight needle. We always use the angled needle for these blind corners.

A transient increase in the volume of the enhanced tumor is observed on MRI in most patients immediately after intratumoral VPC placement. Contrast enhancement of the needle tracts in the brain which the needles passed to inject the VPCs occurred and subsided 4-8 weeks after the injection. This finding is termed "tract sign" or "Uni" sigh (Uni=sea urchin in Japanese), which shows the radiation shape of several tracts from the resected cavity (Figure 3).

3. "Reflux" During the injection procedure, a significant quantity of the VPC suspension is lost through "reflux". In general, as much as 50% of the injected volume is refluxed out into the tumor cavity, however, it depends on technique of operators. The reflux phenomenon could be prevented by slowing and pumping injection while pulling out the syringe. This method actually takes at least 30 sec per one injection. In the near future, an injection technique should be improved and developed to less labor, less time and more homogeneous inoculation.

C. Pitfalls of injection (Figure 4) In case of the multi-injection, one must pay an attention to following procedures; 1) injection into ventricle or basal ganglia, 2) injection into the blind corner and 3) reflux of the VPC suspension.

1. Ventricle and basal ganglia Glioblastoma invades diffusely into the white matter, especially near the ventricle or basal ganglia. If retroviral particles could infiltrate into these regions, invaded tumor cells could be killed. The deeper injection is better for transfection, however, one must not penetrate ventricles to avoid diffusion of VPCs or not inject suspension into basal ganglia not to influence neurological deficits.

Acknowledgments We thank Dr. Nicholas Shand (Oncology Clinical Research, Novartis Pharma), GLI-328 Novartis study group and Novartis Pharma Ltd. (Basel, Switzerland) for their support, and Rosmarie Frick and Roland Stillhard (University Hospital Z端rich) for technical assistance. N Adachi is grateful to Emiko Adachi for her encouragement throughout this work. This work was supported by Grants258

Gene Therapy and Molecular Biology Vol 4, page 259 Fine HA, Dear KB, Loeffler JS, Black PM, Canellos GP. (1993) Meta-analysis of radiation therapy with and without adjuvant chemotherapy for malignant gliomas in adults. Cancer 71, 2585-2597

in-Aid for Abroad Scientific Research from the Ministry of Education, Science and Culture of Japan (#8-J-216) and Yamanouchi Foundation for Research on Metabolic Disorders in Japan.

Hamel W, Magnelli L, Chiarugi VP, Israel MA. (1996) Herpes simplex virus thymidine kinase/ganciclovir-mediated apoptotic death of bystander cells. Cancer Res 56, 26972702

References Adachi N. (1998) HSV-TK/ganciclovir gene therapy in relapsed glioblastoma multiforme: results of an international multicentre study. Gene Ther Mol Biol 2, 380-381

Kato K, Yoshida J, Mizuno M, Sugita K, Emi N. (1994) Retroviral transfer of herpes simplex thymidine kinase gene into glioma cells causes targeting of gancyclovir cytotoxic effect. Neurol Med Chir (Tokyo) 34, 339-344

Ammirati M, Galicich JH, Abrit E, Liao Y. (1987) Reoperation in the treatment of recurrent intracranial malignant gliomas. Neurosurgery 5, 607-614

Kim JH, Kim SH, Brown SL, Freytag SO. (1994) Selective enhancement by an antiviral agent of the radiation-induced cell killing of human glioma cells transduced with HSV-tk gene. Cancer Res 54, 6053-6056

Barba D, Hardin J, Ray J, Gage FH. (1993) Thymidine kinasemediated killing of rat brain tumors. J Neurosurg 79, 729735

Laske DW, Youle RJ, Oldfield EH. (1997) Tumor regression with regional distribution of the targeted toxin TF-CRM107 in patients with malignant brain tumor. Nature Med 3, 13621368

Barba D, Hardin J, Sadelain M, Gage FH. (1994) Development of anti-tumor immunity following thymidine kinasemediated killing of experimental brain tumors. Proc Natl Acad Sci USA 91, 4348-4352

LeMay DR, Kittaka M, Gordon EM, Gray B, Stins MF, McComb JG, Jovanovic S, Tabrizi P, Weiss MH, Bartus R, Anderson WF, Zlokovic BV. (1998) Intravenous RMP-7 increases delivery of ganciclovir into rat brain tumors and enhances the effects of herpes simplex virus thymidine kinase gene therapy. Hum Gene Ther 9, 989-995

Bi WL, Parysek LM, Warnick R, Stambrook PJ. (1993) In vitro evidence that metabolic cooperation is responsible for the bystander effect observed with HSV tk retroviral gene therapy. Hum Gene Ther 4, 725-731 Chen SH, Shine HD, Goodman JC, Grossman RG, Woo SL. (1994) Gene therapy for brain tumors: regression of experimental gliomas by adenovirus-mediated gene transfer in vivo. Proc Natl Acad Sci USA 91, 3054-3057

Lynos RM, Forry-Schaudies S, Otto E, Wey C, Patil-Koota V, Kaloss M, McGarrity GJ, Chiang YL. (1995) An improved retroviral vector encoding the herpes simplex virus thymidine kinase gene increases antitumor efficacy in vivo. Cancer Gene Ther 2, 273-280

Colombo BM, Benedetti S, Ottolenghi S, Mora M, Pollo B, Poli G, Finocchiaro G. (1995) The "bystander effect": association of U-87 cell death with ganciclovir-mediated apoptosis of nearby cells and lack of effect in athymic mice. Hum Gene Ther 6, 763-772

Maron A, Gustin T, Le Roux A, Mottet I, Dedieu JF, Brion JP, Demeure R, Perricaudet M, Octave JN. (1996) Gene therapy of rat C6 glioma using adenovirus-mediated transfer of the herpes simplex virus thymidine kinase gene: long-term follow-up by magnetic resonance imaging. Gene Ther 3, 315-322

Crumpacker CS. (1996) Ganciclovir. N Engl J Med 335, 721729 Crystal RG. (1995) Transfer of genes to humans: early lessons and obstacles to success. Science 270, 404-410

Mesnil M, Piccoli C, Tiraby G, Willecke K, Yamasaki H. (1996) Bystander killing of cancer cells by herpes simplex virus thymidine kinase gene is mediated by connexins. Proc Natl Acad Sci USA 93, 1831-1835

Culver KW, Ram Z, Wallbridge S, Ishii H, Oldfield EH, Blaese RM. (1992) In vivo gene transfer with retroviral vectorproducer cells for treatment of experimental brain tumors. Science 256, 1550-1552

Moolten FL. (1986) Tumor chemosensitivity conferred by inserted herpes thymidine kinase genes: paradigm for a prospective cancer control strategy. Cancer Res 46, 52765281

Davis FG, Malinski N, Haenszel W, Chang J, Flannery J, Gershman S, Dibble R, Bigner DD. (1996) Primary brain tumor incidence rates in four United State regions, 19851989: a pilot study. Neuroepidemiology 15, 103-112

Moolten FL, Wells JM. (1990) Curability of tumors bearing herpes thymidine kinase genes transfered by retroviral vectors. J Natl Cancer Inst 82, 297-300

Durand RE. (1990) Cisplatin and CCNU synergism in spheroid cell subpopulations. Br J Cancer 62, 947-953

Moriuchi S, Oligino T, Krisky D, Marconi P, Fink D, Cohen J, Glorioso JC. (1998) Enhanced tumor cell killing in the presence of ganciclovir by herpes simplex virus type 1 vector-directed coexpression of human tumor necrosis factoralpha and herpes simplex virus thymidine kinase. Cancer Res 58, 5731-5737

Elion GB, Furman PA, Fyfe JA, de Miranda P, Beauchamp L, Schaeffer HJ. (1977) Selectivity of action of an antiherpetic agent, 9-guanine. Proc Natl Acad Sci USA 74, 5716-5720 Ezzeddine ZD, Martuza RL, Platika D, Short MP, Malick A, Choi B, Breakefield XO. (1991) Selective killing of glioma cells in culture and in vivo by retrovirus transfer of the herpes simplex virus thymidine kinase gene. New Biol 3, 608-614

Namba H, Tagawa M, Iwadate Y, Kimura M, Sueyoshi K, Sakiyama S. (1998) Bystander effect-mediated therapy of experimental brain tumor by genetically engineered tumor cells. Hum Gene Ther 9, 5-11

Freeman SM, Abboud CN, Whartenby KA, Packman CH, Koeplin DS, Moolten FL, Abraham GN. (1993) The "bystander effect": tumor regression when a fraction of the tumor mass is genetically modified. Cancer Res 53, 52745283

Perez-Cruet MJ, Trask TW, Chen SH, Goodman JC, Woo SL, Grossman RG, Shine HD. (1994) Adenovirus-mediated gene

259

Adachi et al: HSV-TK/GCV gene therapy for brain tumors therapy of experimental gliomas. J Neurosci Res 39, 506511

Walker MD, Green SB, Byar DP, Alexander EJ, Batzdorf U, Brooks WH, Hunt WE, MacCarty CS, Mahaley MSJ, Mealey JJ, Owens G, Ransohoff J II, Robertson JT, Shapiro WR, Smith KRJ, Wilson CB, Strike TA. (1980) Randomized comparison of radiotherapy and nitrosoureas for the treatment of malignant glioma after surgery. N Eng J Med 303, 1323-1329

Rajkumar SV, Buckner JC, Schomberg PJ, Pitot HC, Ingle JN, Cascino TL. (1999) Phase I evaluation of preirradiation chemotherapy with carmustine and cisplatin and accelerated radiation therapy in patients with high-grade gliomas. Neurosurgery 44, 67-73

Zlokovic BV, Apuzzo MLJ. (1997) Cellular and molecular neurosurgery: Pathways from concept to reality- part I: target disorders and concept approaches to gene therapy of the central nervous system. Neurosurgery 40, 789-804

Ram Z, Culver KW, Walbridge S, Blaese RM, Oldfield EH. (1993) In situ retroviral-mediated gene transfer for the treatment of brain tumors in rats. Cancer Res 53, 83-88 Ram Z, Culver KW, Walbridge S, Frank JA, Blaese RM, Oldfield EH. (1993) Toxicity studies of retroviral-mediated gene transfer for the treatment of brain tumors. J Neurosurg 79, 400-407 Ram Z, Walbridge S, Shawker T, Culver KW, Blaese RM, Oldfield EH. (1994) The effect of thymidine kinase transduction and ganciclovir therapy on tumor vasculature and growth of 9L gliomas in rats. J Neurosurg 81, 256-260 Ram Z, Culver KW, Oshiro EM, Viola JJ, DeVroom HL, Otto E, Long Z, Chiang Y, McGarrity GJ, Muul LM, Katz D, Blaese RM. (1995) Summary of results and conclusions of the gene therapy of malignant brain tumors: clinical study. J Neurosurg 82, 343A Ram Z, Culver KW, Oshiro EM, Viola JJ, DeVroom HL, Otto E, Long Z, Chiang Y, McGarrity GJ, Muul LM, Katz D, Blaese RM, Oldfield EH. ( 1997) Therapy of malignant brain tumors by intratumoral implantation of retroviral vector-producing cells. Nature Med 3, 1354-1361 Ramesh R, Marrogi AJ, Freeman SM. (1998) Tumor killing using the HSV-tk suicide gene. Gene Ther Mol Biol 1, 253263 Rigg A, Sikora K. (1997) Genetic prodrug activation therapy. Mol Med Today 359-366 Rubsam LZ, Boucher PD, Murphy PJ, KuKuruga M, Shewach DS. (1999) Cytotoxicity and accumulation of ganciclovir triphosphate in bystander cells cocultured with herpes simplex virus type 1 thymidine kinase-expressing human glioblastoma cells. Cancer Res 59, 669-675 Salmons B, Gunzburg WH. (1993) Targeting of retroviral vectors for gene therapy. Human Gene Ther 4, 129-141 Samejima Y, Meruelo D. (1995) "Bystander killing" induces apoptosis and is inhibited by forskolin. Gene Ther 2, 50-58 Short MP, Choi BC, Lee JK, Malick A, Breakefield XO, Martuza RL. (1990) Gene delivery to glioma cells in rat brain by grafting of a retrovirus packaging cell line. J Neurosci Res 27, 427-439 Spence AM, Berger MS, Livingston RB, Ali-Osman F, Griffin B. (1992) Phase II evaluation of high-dose intravenous cisplatin for treatment of adult malignant gliomas recurrent after chloroethylnitrosourea failure. J Neurooncol 12, 187191 Vandier D, Rixe O, Brenner M, Gouyette A, Besnard F. (1998) Selective killing of glioma cell lines using an astrocytespecific expression of the herpes simplex virus-thymidine kinase gene. Cancer Res 58, 4577-4580 Vincent AJ, Vogels R, Someren GV, Esandi MC, Noteboom JL, Avezaat CJ, Vecht C, Bekkum DW, Valerio D, Bout A, Hoogerbrugge PM. (1996) Herpes simplex virus thymidine kinase gene therapy for rat malignant brain tumors. Hum Gene Ther 7, 197-205

260

Gene Therapy and Molecular Biology Vol. 4, Page 261 Gene Ther Mol Biol Vol 4, 261-274. December 1999

Establishment of tumor cell lines by transient expression of immortalizing genes Review Article

Liangping Li Max-Delbrück Centrum for Molecular Medicine, Robert-Rössle Strasse 10, 13122 Berlin-Buch, Germany __________________________________________________________________________________________________ Correspondence: Liangping Li, Ph.D., Max-Delbrück Centrum for Molecular Medicine, Robert-Rössle Strasse 10, 13122 Berlin-Buch, Germany. Tel: +49-30-9406 2687; Fax: +49-30-9406 2453; llp@orion.rz.mdc-berlin.de Key Words: tumor cell line, immortalizing genes, cell culture, transient expression Received: 27 March 1999; accepted: 20 June 1999

Summary Many basic and clinical studies on human cancers require establishing of tumor cell lines from fresh tumor tissues in a highly reproducible fashion. However, this goal has not been achieved since the first aneuploid epithelial cell line, HeLa, was established from cervix adenocarcinoma about 5 decades ago. A widely accepted concept is that malignant tumor cells are almost immortal so the difficulty in establishing new tumor cell lines is believed to arise from problems of cell culture techniques. Evidence from our experiments demonstrated that this difficulty has a genetic origin: many primary cancer cells have not completely lost their ‘tumor suppressor’ pathways. In this review the immortalizing genes, transient expression systems of foreign genes in mammalian cells, and the potential applications of transient expression of immortalizing genes are summarized.

and difficulty of the establishment of tumor-derived cell lines and new ways to overcome this difficulty as well as the future prospects of these new technologies.

I. Introduction Tumor cell lines have profound significance for the development of cancer cell biology and have made a great contribution to a variety of cancer researches (Stoker, 1996). Most stable human cell lines are established from cancer tissues and are widely used in the study of all fields of cancer. Some clinical trials such as gene-modified cancer vaccines demand growing a number of proliferative autologous tumor cells (Blankenstein, 1996). However, despite improved cell culture conditions it is still difficult to establish permanent cell lines from human primary tumors in a reproducible fashion. Growing human cancer cells in primary culture is hard work that requires patience, intuition, care and experience. The reason for the difficulty in growing human primary cancer cells is not completely understood.

II. Historical Background A. Old and new problem: Difficulties for establishing long-term tumor cell lines Since the first human tumor cell line HeLa was established from a cervical adenocarcinoma (Gery, 1951), a variety of permanent (‘immortalized’) tumor cell lines have been established from different human tumor tissues (primary, invasive, metastatic or recurring) of most types of tissues (Fogh, 1975). However, the establishment of long-term tumor cell lines is hard, time-consuming, and unpredictable work with a very low success rate.

The advances in DNA recombinant technology and gene transfer have made it possible to develop new methods for establishing permanent tumor cell lines from clinical tumor samples by the introduction of immortalizing genes into primary tumor cells (Pantel, 1995; Li, 1997). This review focuses on the background

The success rates of establishing permanent tumor cell lines vary notably from one tumor cell type to another, ranging from 50% or more of malignant melanoma to 1% or less of breast cancers (Stamps, 1992) and is related to tumor type and clinical stage.

261

Li et al: Establishment of Tumor Cell Lines Among human cancers, malignant melanoma cells are easiest to establish from primary lesions, metastases and effusions (Moor, 1996). Melanoma cells, in a malignant effusion (chest or abdomen), are the easiest to establish, over 50% success; then cells from metastases, 20-50%; and lastly primary cancers, 10-25%. Human breast carcinomas are one of tumors that are the most difficult to culture. Only 10% breast cancer cells proliferate in vitro and less than 1% of them are able to be established as long-term cell lines (O’Hare, 1991).

technical problem because primary tumor cells may be more dependent on specific growth factors than normal cells.

III. Improvement for establishing new tumor cell lines A. Modified cell culture methods The traditional method for establishment of tumor cell lines is tissue culture. Basic procedure is that tumor biopsies are dissociated with mechanical and enzymatic treatment (Speirs, 1996) and isolated tumor cells are in 520% FBS medium. The improvement of conditions of tissue treatment and culture medium can enhance the success rate of establishing tumor cell lines.

A lot of evidence showed that most tumor cell lines were established from the later stage cancer tissues. The early stage cancer cells are much more difficult to grow in vitro. Of the 136 analyzed tumor cell lines at tumor cell bank of American Type Culture Collection (ATCC), we found that most cell lines (76/136, 56%) were established from metastatic or invasive lesions. In primary tumors one third of the cell lines (23/60) derived from blastoma and sarcoma, which are often more malignant. Edington, et al. ( 68) compared cultures from different stages of head and neck squamous carcinoma and found that the immortal cultures were not obtained from normal (0/4) and pre-malignant (0/4) tissue cultures but rather from T1/T2 carcinoma (2/7), T4 carcinoma (6/10), and metastases (2/3). All recurrent tumors were immortal.

Modification of preparations of tumor samples for primary culture: the preparation of tumor samples with mechanical disaggregation or enzymatic digestion is the first and important step to establish successful primary culture. The tumor tissues lack an integrate basement and are sensitive to enzymatic digestion. It has been confirmed that complete digestion of tumor tissues into single cells can not produce successful primary culture. Use of partial enzymatic dissociation can improve the viability of tumor cells (Dairkee, 1997). In colon carcinoma the mechanical disaggregation may result in higher rates of primary cultures (Dillman, 1993).

Wilson (1996) summarized the literature concerning at least 70 ovarian tumor cell lines. The majority of these cell lines derived from ascitic fluids (n=59) rather than solid tumors (n=14) and were also most frequently established from patients with a poor prognosis. There was a lack of cell lines from well-differentiated or benign tumors. Review of the literature indicates that there are no particular growth factors or media that result in cell line development.

1. Modification of growth media We have analyzed the primary culture growth media for human tumor tissues in 136 tumor cell lines of tumor cell bank at ATCC. The most common used media are RPMI 1640, DME, Eagle’s EME, and L-15 supplemented with 10-20% fetal bovine serum. Similar results were obtained from reviewing the literature of the last decade. The most successful primary culture and long-term cultures of human tumor cells are malignant melanomas, which grow well in RPMI 1640 media supplemented with 10-20% fetal bovine serum (Semple, 1982; Marincola, 1996). The most commonly used supplement for in vitro culture to most of established cell lines is bovine serum, which provides hormones, nutrients and promotes cell growth. However, in serum-containing medium, contaminating fibroblasts impose a big obstacle and often grow over tumor cells. Fibroblasts are easy to grow in serum-containing media but normal and neoplastic epithelial cells are not. The culture of epithelial cells in serum-supplemented media is prone to terminal differentiation and always poses the problem of fibroblast overgrowth in cultures; indeed, serum does not contain enough basic growth factors and growth-inhibitory or differentiation-inducing factors for particular epithelial cells (Miyazaki, 1989). Masui, et al. (1986) found that

B. Malignant tumors are often more difficult to culture than normal cells Earlier observations showed that some cancer-derived epithelial cells such as breast carcinoma, nasopharyngeal carcinoma (Li, 1994) and prostate cancer, etc., were even more difficult to grow in vitro than the normal epithelial cells from same tissue type. Human breast carcinomas are the representative examples. The normal epithelial cell of breast could be passaged 3-4 times while those derived from cancer only survived one passage (Smith, 1984). Using improved media normal epithelial cells can be grown up to 10-20 passages but carcinoma-derived cells only proliferated 3-5 passages (our unpublished data). It is not clear so far why most of so called ‘immortal’ primary tumor cells are difficult to grow in vitro and even more difficult than normal epithelial cells. The failure to culture some tumor cell types was considered due to

262

Gene Therapy and Molecular Biology Vol. 4, Page 263 type beta transforming growth factor is the primary differentiation-inducing serum factor for normal human bronchial epithelial cells.

use the immortalizing techniques of mammalian cells to establish new tumor cell lines.

IV. In vitro immortalization of human cells

2. Serum-free and low serum medium for human epithelial cells

A. Immortalization methods

The development of defined serum-free media greatly improved the in vitro culture of a variety of different types of human epithelial cells. Many specific media are designed for particular epithelial cell types, such as keratinocyte growth media MCBD 153 (Boyce and Ham, 1985; Boisseu, 1992), mammary epithelial cell growth media MCBD 170 (Hammond, 1984; Weaver, 1995) or DFCI-1 (Band, 1989), and bronchial epithelial cell growth media LHC9 (Lechner, 1989). Serum-free media have advantages in culture over serum-supplemented media by providing defined reproducible systems without inhibitory factors and permitting selective growth of tumor cells from fresh clinical specimens. For instance, the rate of successful primary cultures from tumor-derived tissues of colon (van der Bosch, et al. 1981) and lung (Carney, et al. 1981) was higher in serum-free media than in serumcontaining media (Collodi, et al. 1991).

A wide variety of human tissues can be cultured and maintained in vitro just for a limited period time. All of these cells finally enter a â&#x20AC;&#x2DC;crisisâ&#x20AC;&#x2122; phase, that is, cell division ceases and the culture becomes senescent except that they obtain mutations and become immortalized. With the exception of fibroblast most primary human cells such as epithelial cells have a short lifespan in vitro. A formal definition of an immortal cell line is that growing indefinitely in vitro. A normal fetal human fibroblast typically reaches senescence at 60 population doublings (PD) so an operational definition is that a fetal human fibroblast exceeds 100 generations or population doublings in culture, or shortly after a colony derived from single-cell cloning can be propagated to 106 to 107 cells (MacDonald, 1990). In order to obtain permanent cell lines a number of immortalization methods have been developed.

However, using serum-free media to culture human tumor cells has not given the expected success as normal epithelial cells, from which most human malignant tumors originate. The serum-free media do not support the longterm growth of human tumor cells from primary cultures. For instance the primary and early-passage culture of human mammary carcinoma using defined serum-free or low serum media (MCDB 170, DFCI-1 and CDM3) results in outgrowth of normal epithelial cells or fibroblasts (Band, 1995). The establishment of permanent tumor cell lines has not been significantly improved with serum-free or low serum media.

(i). Treatment with carcinogenic chemicals. (ii). Exposure to a DNA tumor virus such as SV40 virus, EBV, and papilloma virus. (iii). Cell fusion between the cell with a limited lifespan and a permanent cell line. (iv). Viral and cellular immortalizing genes / oncogene transfer. With the development of recombinant DNA techniques and the advent of gene transfer immortalization of mammalian cells by the introduction of immortalizing genes became an important method for developping new cell lines from normal cells for animal cell biotechnology. A variety of cell lines from different species and types have been established via infection or transfection of viral and cellular oncogenes (MacDonald, 1990).

B. Xenograft The discovery of immunodeficient nu (nude) (Pantelouris, 1968) and scid (Bosma, 1983) mutant mice has provided new methods to establish tumor cell lines by xenogeneic transplantations. Many tumor cell lines have been established using xenograft in these immunodeficient mice. The take rates are dependent on the type of tumor tissues (Arnold, 1996). Establishment of human tumor cell lines by xenograft of tumor biopsies into immunodeficient mice is time-consuming and expensive and is not sufficient for the construction of autologous tumor vaccines.

B. Immortalizing genes There are many genes involving in vitro cellular immortalization (Duncan, 1993). Two main types of immortalizing genes are summarized based on the targeting genes. 1) Anti-tumor suppressor immortalizing genes: this kind of immortalizing genes involves the inactivation of tumor suppressors such as pRB and p53 and includes SV40 large T antigen; HPV16 E6 and E7; Adenovirus E1A and E1B; Polyoma large T; and Mutant p53, etc. 2) Oncogene-relating immortalizing genes: c-

Because of the common difficulty of establishing long-term or permanent cultures of human tumor cells from fresh surgical tumor samples new techniques are needed to solve these problems. We have been trying to 263

Li et al: Establishment of Tumor Cell Lines myc, bmi-1, c-Ha-ras etc. 3) Other genes: telomerase gene.

acids 533-626 are important for regulating the interaction. e). Region involved in conformation: a. Zinc finger (amino acid 302-318) The mutations of this region alter the overall conformation of the protein and result in the decrease of foci formation in primary mouse embryo fibroblasts and of viral DNA replication. b. Hydrophobic region (amino acids 571-589) this region maintains the general stability of protein and is indirectly important for transformation.

1. Anti-tumor suppressor immortalizing genes: SV40 large T antigen SV40 large T antigen gene is one of the most successful immortalizing genes. This gene encodes a nuclear phosphoprotein containing 708 amino acids. T antigen has a variety of biochemical activities, including viral DNA replication and transformation functions. Large T can be divided into many different functional regions of: specific DNA binding, ATPase activity, nuclear location signal, etc. (for review, Fanning, 1992; Bryan, 1994; Manfredi, 1994). To date, several transformation regions of SV 40 large T antigen have been identified. a) Amino-terminal transforming domain (amino acids 1-82) binds to DNA polymerase and the p300 protein. b) Rb-binding region (amino acids 101-118), binding to pRb and p107. c) Nuclear location signal resides in amino acid 126-132. d) p53-binding domain (amino acids 351-450 and 533-626): A lot of mutation research suggest that p53 binding domain is located in a region between residues 272 and 626. Amino acids 351450 are important for direct contact with p53 and amino

Normal human cells immortalized by SV40 large T antigen: A variety of human primary cell types have been immortalized with SV40 virus, SV40 genomic DNA, and plasmid DNA or retrovirus expressing SV40 large T antigen (Table 1). For a long time SV 40 large T antigen has been introduced to a variety of cells which are difficult to culture in vitro in order to obtain cell line with longterm growth. SV40 large T transformed human fibroblasts routinely have an extension of the lifespan of the normal parental cell but still eventually cease to proliferate. Approximately 25% of cloned transformants yielded immortalized cell lines with frequencies ranging from one per 103 to 107 cells. Keratinocytes and epithelial cells may generate immortalized derivatives at a higher frequency.

Table 1. Examples of normal human cells immortalized by SV40 Cell Type

Method

Ref.

Fibroblast

Ori+ plasmid

Shay and Wright, 1989

Ori- plasmid

Banerjee et al., 1992

Keratinocytes

Virus

Taylo-Papadimitriou et al., 1982

Bronchial epithelium

Virus

Reddel et al., 1988, 1995

ori- plasmid

Cozens et al., 1992

SV40 Virus,

Chang et al., 1982

Retrovirus

Bartek et al.,1990, 1991

Ori-plasmid

Berthon et al., 1992

Prostatic epithelium

Ori-plasmid

Hayward, et al., 1995

Esophgeal epithelium

Ori-plasmid

Stoner et al., 1991

Uroepithelium

Virus

Christian et al., 1987

Liver epithelium

SV 40 T-plasmid

Miyazaki, 1993; Schippers, et al., 1997

Retinal epithelium

Ori+plasmid

Dutt et al., 1994

Bone marrow stromal cells

SV40-adenovirus

Aizawa, et al., 1994

Human osteoblastic cells

tsA58

Hariss, et al., 1995

Monocyte/macrophage

Ori-DNA

Nagata et al., 1983; Kreuzburg, 1994

Breast epithelium

Microglial cells

Janabi, et al., 1995

Granulosa-lutein cells

SV40 T-plasmid

Lie, et al., 1996

Endothelial cells

SV40 T-plasmid

Fitzgerald, et al.,1994

Thymic nurse cells

SV40 virus

Pezzano, 1991

Dendritic cells

Large T gene

Volkmann, 1996

Endometrial cells

Merviel, et al., 1995

264

Gene Therapy and Molecular Biology Vol. 4, Page 265 p53 mutants R175H and R249s immortalized human mammary epithelial cells in a reproducibly way, whereas R248W and R273H mutants induced an extended life span but not immortalization (Cao, 1997).

Tumor cells immortalized by SV40 large T: Pantel, et al. (1995) immortalized the micrometastatic tumor cells of prostate cancer, renal cell cancer, lung cancer, breast cancer, and colorectal cancer with microinjection of SV 40 large T plasmid DNA. After microinjection of the propagated cells with T-antigen DNA, permanent cell lines were obtained and showed no notable changes in the pattern of expressed epithelial antigens and were able to disseminate into bone marrow in SCID mice.

B. Oncogene-relating immortalizing genes 1. C-myc Myc protein plays a critical role in the normal control of proliferation and differentiation. Ectopic expression of Myc is enough to drive many cells into the cell cycle in the absence of external mitogens(Bouchard, 1998). Cmyc or v-myc gene can immortalize primary rodent fibroblasts (Kelekar, 1987) and human mammary epithelial cells (Valverius, 1990). Myc and raf oncogenes cooperate with absence of p53 to immortalize hematopoietic cells (Metz, 1995).

2. Human papilloma virus E6 and E7 genes HPV E7 proteins bind and inactivate the retinoblastoma protein RB (Pei, 1998), while E6 increases the degration of p53 (Van Dyke, 1994). The introduction of both HPV-16 or -18 E6 and E7 can efficiently immortalize human keratinocytes from the cervix or foreskin (Hawley-Nelson, 1989), and mammary epithelial cells (Shay, 1991).Hhigh expression of HPV-16 E7 alone can immortalize human keratinocytes at low frequency compared with a similar virus containing HPV-16 E6 and E7 (Halbert, 1991).

2. Bmi-1 The bmi-1 is a transcriptional repressor belonging to the mouse Polycomb group and cooperates with c-myc in the generation of mouse lymphoma. Overexpression of bmi-1 allows fibroblast immortalization immediately or after a slow growth period, downregulates expression of p16 and p19, and, in combination with H-ras, leads to neoplastic transformation (Jocobs, 1999).

3. Adenovirus E1A and E1B A human kidney epithelial cell lines was established from the adenovirus DNA transfection (Graham, 1977). E1A proteins bind pRB (Whyte, 1988) and other cellular proteins p107, p130 and p300 (van Dyke, 1994), while E1B inactivates p53 protein (Sarnow, 1982).

3. ras and raf The ras and raf oncogenes can not immortalize mammalian cells alone. On the contrary, they induce premature senescence (Serrano, 1997). But these oncogenes cooperate with other oncogenes such as myc and HPV E6 to immortalize some types of cells such as fibroblasts, microglial cells (Blasi, 1990), epithelial cells (Compere, 1989).

4. Polyoma virus large T A lot of mammalian cells have been immortalized with polyoma virus large and middle T antigen such as human fetal sinusoidal liver cells (Hering, 1991), human fibroblast (Jat, 1986; Strauss, 1990), mice astroglial cells (Galiana, 1990), etc. The large T antigen of polyoma virus can bind to tumor suppressor pRb protein (Dilworth, 1990).

4. EBV genes Epstein-Barr virus can efficiently immortalize human B lymphocytes. 10%-100% of infected B cells become immortal (Sugden, 1989). The mechanism of EBV-induced immortalization is not completely to be understood. Several viral genes and the altered expression of cellular genes are required for the efficient immortalization of resting B cells. EBV-encoded EBNA-1, EBNA-2, LMP1 and LMP2 proteins are involved in the B cell immortalization.

5. Mutant p53 P53 is one of most important tumor suppressors and a transcription factor, which functions as a cell cycle G1 checkpoint molecule. P53 protein transmits damaging cellular stress signals (such as DNA damage) to cell cycle and induces cell growth arrest or apoptosis. Thus, cells that lose p53 protein are prone to be immortal. A germ line mutation of p53 gene exists in the breast cancer-prone Li-Fraumeni syndrome (Shay, 1995). Normal epithelial cells from this syndrome are spontaneously immortalized in vitro. Gao, et al. (1996) reported that a single-amino acid deletion mutant (del239) of p53 is able to immortalize primary human mammary epithelial cells. Further studies showed that

C. Telomerase It has been reported that telomere shortening is a key mechanism for cellular senescence (Shay, 1991, Bodnar, 1998, Smith, 1996). The causal relation

265

Li et al: Establishment of Tumor Cell Lines between senescence and gradual loss of chromosome ends was proposed about 25 years ago. The chromosome ends consist of telomeres, which are noncoding G-rich repeats and maintain the chromosomal stability. The incomplete replication of 5’ ends of linear DNA molecules leads gradual chromosome shortening. Telomerase can restores telomere length through adding telomeric repeats (Kim, 1994). Bodnar, et al. (1998) found that the introduction of telomerase into normal fibroblasts resulted in the extension of life-span. Telomerase-expressing clones had elongated telomeres, divided vigorously, and showed reduced staining for !galactosidase, a biomaker for senescence.

which let cells escape M2 (Holt, 1996; Bodnar, et al. 1998).

B. Yin-yang model of cellular proliferation, senescence, and immortalization Cellular immortalization is defined as a process that cells can ‘indefinitely’ proliferate in vitro. The regulation of cell proliferation is dominated by two groups of genes (Dyson, 1999): proliferation-stimulating genes (proto-oncogenes) such as ras, myc, and bcl-2, etc and proliferation-inhibiting genes (tumor suppressor genes) such as p53, Rb, and p16, etc. The cell proliferation is delicately regulated through the complicate interaction among the products from these two groups of genes. The positive and negative signals form a large feedback loop and maintain a dynamic balance. This dynamic balance of positive and negative signals determines whether cells proliferate or are arrested. Deregulation of these positive and negative regulator results in abnormal apoptosis, senescence, immortalization, or uncontrolled cell proliferation (cancer). We proposed that cellular immortalization not only involves in the inactivation of tumor suppressors (Shay, 1995) but also in the changes of oncogene expression (Fig. 1) (Li, 1998). Zinky, 1998 reported that myc signaling regulates immortalization through ARF tumor suppressor.

However, a number of evidences showed that telomerase is not enough to immortalize cells, although telomere maintenance is essential for immortalization (Carman, 1998). Some telomerase-positive cells also senescence and some senescent cells expressed telomerase activity (Carman, 1998). Recently, Kiyono, et al. (1998) reported that telomerase activity is not sufficient for immortalization of human keratinocyte or mammary epithelial cells. However, inactivation of the Rb/p16 pathway in combination with telomerase activity, is able to immortalize epithelial cells efficiently.

V. Mechanism of immortalization A. Immortalization induced by SV40 large T antigen The SV40 large T antigen-induced immortalization is a two step process (Wright, et al., 1989; Shay, et al., 1991), which was first found in human fibroblasts (Girardi, et al., 1965) and later confirmed in breast epithelial cells (van der haegen, 1993). Wright, et al. (1989) immortalized IMR normal human diploid fibroblasts with a steroid inducible mouse mammary tumor virus-deriven SV40 large T antigen and found that after dexamethasone removal the cells stop growing during the precrisis extension of life-span and after immortalization. According to these data a two-stage model for cellular senescence have been put forward. Mortality stage 1(M1) is responsible for the precrisis cell growth arrest and mortality stage 2(M2) for the failure of cell division during crisis. For cellular immortalization the first step is an extension of culture lifespan (about 20 PD after the senescence of the normal cells), which is dependent on the binding of SV40 large T antigen to the tumor suppressor p53 and Rb. And then the cells enter the second step and undertake ‘crisis’, in which cell death increases. A few cells escape the crisis and become colonies of immortalized cells. This step involves in an entirely independent mechanism. Recent studies showed that telomerase is an important enzyme

VI. Transient expression system of foreign genes The continuous presence of large T or other immortalizing genes may change the antigenicity of the cells or alter gene expression or cause mutation not specific for the tumor. It is necessary to develop controllable expression system for immortalizing gene, such as SV40 large T antigen. The transient expression system of foreign genes can be obtained with several methods such as inducible promoter system, temperature-sensitive mutants, and site-specific recombination system, etc. The former two systems pose the risk of ‘leakiness’ or not total absence of the protein and can not remove the targeting gene from genome. The site-specific recombination system can precisely cut out the gene from the genome (Dale, 1991; Bergeman, 1995).

A. Inducible promoter system Early inducible expression vectors based on the inducible promoter and its control agents, e.g. heat shock protein and heat control, metallothionine promoters and heavy metal ion control or steroid regulatory promoters

266

Gene Therapy and Molecular Biology Vol. 4, Page 267

Fig. 1 Dynamic dual-signal (Yin-Yang) feedback / interaction model of cellular proliferation: The cell proliferation is controlled by positive and negative signals. Most of positive signals for cell proliferation are specific growth factors or growth-stimulating agents from the outside of cells, and the negative include growth-inhibiting factors, differentiation-induced factors, or abnormal stresses, etc.

characteristics of hPR891 can be used as a switch and regulatory system of protein function (Wang, 1994). The fusing gene between the C-terminal hormone binding domain and targeting gene expresses a protein, whose activity is hormone-dependent.

and steroid control (Yarranton, 1992). These traditional inducible systems can regulate some level of controlled gene expression (10-20 fold induction) but often have higher background expression (‘leaky expression’). Use of bacterial transcriptional control system in mammalian cells can obtain higher specificity and induction ratios because the bacterial operator is unlikely to bind the promoter sequence of a mammalian genes. The well-characterized inducible systems that based on bacterial control elements are E.coli lac repressor-operator and tetracycline responsive expression system. Extremely high induction ratio up to 10 5 -fold was achieved in the teton and tet-off system (Hofman, 1996).

C. Temperature-sensitive mutants Earlier studies found a temperature-sensitive (ts) mutant of SV40, which can transform primary cells at the permissive temperature(33°C) but is inactive at the nonpermissive temperature (39°C) (Tegtmeyer,1975; Petit,1983). This mutant SV40 virus encode a thermolabile large T antigen tsA 58 which loss about 500 nucleotides at its C-terminal (Jat, 1989).

B. Human progesterone receptor mutant (hPR891) and RU486 system

D. Site-specific recombination system

The human progesterone receptor mutant hPR891 loses 42 amino acid in its C-terminal and ability to bind progesterone or other endogenous hormones, but still can bind the synthetic antagonist RU486 (Vegeto, 1992). This

The site-specific recombination is different in mechanism and efficiency from homologous recombination, which occurs at any homologous sequence with a low rate and is dependent on the endogenous

267

Li et al: Establishment of Tumor Cell Lines recombination mechanism of the cell (Sauer, 1993, 1994). In contrast, site-specific DNA recombination that relies on exogenous recombinase is conservative and occurs at specific sites on the DNA molecule with high rate. The well-characterized site-specific systems are Cre recombinase of bacteriophage P1 and FLP recombinase from Saccharomyces cerevisiae. The former is more common used.

E. Cre/LoxP system

site-specific

bp dyad-symmetric DNA sequence, which is composed of two 13 bp inverted repeats, separated by an 8 bp spacer molecule. Cre recombinase binds to the two 13 bp inverted repeats of LoxP sites and catalyzes precise recombination between the asymmetric 8-bp core region of two LoxP sites. The result of recombination is dependent on the direction of core region of two LoxP sites in one DNA sequence. Recombination between two directly repeated sites on the same DNA molecule results in deletion of the DNA fragment lying between two loxP sites. For a molecule with two inverted LoxP sites Cre recombinase catalyzes the inversion of the intervening DNA fragment.

recombination

Cre/Lox system is one of well-characterized sitespecific recombination systems and consists of two basic components: Cre recombinase and LoxP site (Figure 2). These two components are sufficient to carry out sitespecific recombination in vitro and in vivo (Gu, 1993).

Inducible Cre/LoxP system: the inducible expression system and Cre/LoxP site-specific deletion system have been combined to obtain controlled expression of Cre recombinase. K端hn et al. (1995) developed a method for inducible gene targeting in mice using interferonresponsive promoter to control the expression of Cre recombinase.

Cre recombinase is a 38.5 KD protein of bacteriophage P1 and catalyzes conservative site-specific recombination between two LoxP sites. LoxP site is a 34

LoxP site sequence: 13 bp inverted repeat 8 bp core region 13 bp inverted repeat

Fig. 2 Cre/Lox site-specific recombination system (A) DNA deletion: On a molecule with two same directional LoxP sites the Cre recombinase catalyzed the site-specific deletion of DNA sequence between LoxP sites. (B) DNA inversion: Cre/LoxP site-specific recombination results in the inversion of the intervening DNA fragment between two inverted LoxP sites. Black triangle, loxP sites; A, gene A; B, gene B.

268

Gene Therapy and Molecular Biology Vol. 4, Page 269 from same patient we found that all epithelial cells from normal and tumor tissues lose proliferation in vitro after 510 passages but could be immortalized by SV40 large T antigen in a similar way after a period of â&#x20AC;&#x2DC;selectionâ&#x20AC;&#x2122;. These data showed that primary breast carcinoma cells maintain functional tumor suppressor pathway (Rb or p53) and that SV40 large T inactivate these tumor suppressor pathway to induce the immortalization of primary breast cancer cells. These evidences suggest that the difficulty to establish new tumor cell lines from some clinical cancer samples is not just a technique problem but has genetic reason and that the transfer of immortalizing genes into primary tumor cells should be a new feasible method to establish long-term cancer cell lines.

Another inducible Cre/LoxP system uses a fusion protein of Cre recombinase and human steroid receptor such as progesterone (Kellendonk, 1996) or estrogen (Metzger, 1995) receptor. Normal progesterone or estrogen receptor has been fusioned with Cre recombinase, which are controlled by progesterone and estrogen. However, endogenous (in vivo) or media-containing (in vitro) steroids result in some background expression of Cre recombinase. Use of human progesterone receptor mutant hPR891 can overcome this problem. Kellerdonk et al. (1996) fused Cre recombinase and hPR891 gene and showed that the recombination activity of fusion Cre recombinase was highly dependent on RU486. The recombinant rate was up to 50% to 80%. The uninduced recombination (background expression of Cre recombinase) varied from 2% to 5.5%.

VIII. Potential application of transient expression of immortalizing genes A. Use of SV40 LT amplified autologous tumor cells as vaccine

Self-deleting retrovirus vector: Russ et al. (1996) constructed a complicated self-deleting retrovirus vectors with Cre/LoxP site-specific recombination. This system based on the natural life cycle of retrovirus, involving duplication of terminal control region U5 and U3 to generate long terminal repeats (LTRs) and Cre sitespecific recombination. But insertion of foreign gene into the U3 region of retrovirus results in low virus titer. Cre adenovirus: Adenovirus vector was used to express Cre recombinase (Anton, 1995). High titer of adenovirus let it to infect in vivo and in vitro cells efficiently.

The motivation for attempting immortalization of many types of normal and tumor cells was to obtain a permanent line in order to study the characteristics of the original normal tissue or to use them as a vaccine for the original tumor. Immortalization of primary normal cells: For normal tissues it is desirable that the immortalized cell line keep the differentiated properties of the primary cell type. The maintenance of differentiated properties has been showed to varying degrees. For instance three murine peritoneal macrophase-like cell lines have been established by transfecting primary cells with SV40 origin-deleted DNA (Kreuzburg, et al., 1994). These cell lines show to express many macrophage-specific properties, e.g. Fc receptor and staining for non-specific esterase. The cell lines phagocytosed IgG-coated partices, they were positive for murine macrophage-specific marker F4/80 and they showed antigen-presentation function.

VII. Establishment of tumor cell lime by transient expression of SV40 large T antigen mediated with Cre/LoxP deletion recombination In order to overcome the difficulty of establishing tumor cell lines from fresh tumor tissues, we chose SV40 large T antigen as stimulator agent of cell proliferation to obtain long-term proliferating cell line and Cre/LoxPmediated site-specific deletion recombination as controlled expression system of SV40 gene. Using this transient expression system of SV40 large T antigen, we found that retroviral large T gene transfer allowed rapid expansion of some primary tumor cells without significant cell crisis and that subsequent elimination of T antigen resulted in cell growth arrest in a breast cancer line grown for more than a year. Remarkably, these cells changed morphology and stopped proliferation comparable to the cells obtained from biopsy, demonstrating the requirement of large T for growth (Li, 1997).

On the other hand, most SV40 LT-positive cell lines lose many of their differentiated properties and take on a more transformed phenotype. During the immortalization of keratinocytes, the capacity of the cells to differentiation progressively decreased (Taylo-Papadimitriou, et al. 1982). Immortalization of primary tumor cells: The immortalization of primary tumor cells is different from that of primary normal cells. (i) The aim to immortalize tumor cells is to use them as a source of tumor vaccine. The most important things for this purpose are that SV 40 large T immortalized tumor cells keep tumor antigen(s) and retain an ability to present tumor antigens to immune cells (Boon, 1994). (ii) Tumor cells have no function in

Later we confirmed these observations in other 10 cases of mammary carcinoma (Manuscript in preparation). Comparing in vitro growth and SV40 large T-induced immortalization of normal epithelial cells and cancer cells 269

Li et al: Establishment of Tumor Cell Lines vivo and possess a lot of antigenic or non-antigenic mutations (Hepper, 1998).

stromal cell lines established by a recombinant SV40adenovirus vector. Exp Hemato 22, 482-487.

For cell-based tumor cell vaccine it is necessary that transduced tumor cells must maintain tumor antigens and are able to present them to immune response cells. Previous studies showed that SV40 large T antigen immortalized normal cells express cell surface major histocompatibility complex class I molecules. The SV40 large T transformed intestinal epithelial cells were able to response the treatment of interferon-_ and express a high level of class II molecules (Vidal, 1993). These cells had an ability to process and present native protein antigens to specific CD4 + T cell hybridomas, via functional class II molecule. Pantelâ&#x20AC;&#x2122;s (1995) studies showed that the expression of molecules relevant to an efficient immune response, such as MHC class I molecules and intercellular adhesion molecule-1 (ICAM-1) is not down-regulated in the genome of micrometastatic cells expressing SV40 large T antigen. So far there are no direct experimental data that show whether or not SV 40 large T antigen influence the expression of tumor antigens and their immunity.

Anton M and Graham FL (1995). Site-specific recombination mediated by an adenovirus vector expressing the Cre recombinase protein: a molecular switch for control of gene expression. J Virol 69(8), 4600-6. Arnold W, KĂśpf-Maier P, Micheel B (1996). Introduction: Immunodeficient animals-Still important tools for cancer research. In Immunodeficient Animals : Models for Cancer Research. Contrib Oncol. Basel, Karger, vol 51, ppIX-XIII. Band V (1995). Preneoplastic transformation of human mammary epithelial cells. Seminars in Cancer Biology 6, 185-192. Band V and Sager R (1989). Distinctive traits of normal and tumor-derived human mammary epithelial cells expressed in a medium that supports long-term growth of both cell types. Proc Natl Acad Sci USA 86, 1249-1253. Banerjee A, Srivatsan E, Hashimoto T, Takahashi R, Xu HJ, Hu SX and Benedict WF (1992). Immortalization of fibroblasts from two patients with hereditary retinoblastoma. Anticancer Res 12(5), 1347-1354.

B. Reversible transformation of human cells

Bartek J, Bartkova J, Kyprianou N, Lalani EN, Staskova Z, Shearer M, Chang S and Taylor-Papadimitriou J (1991). Efficient immortalization of luminal epithelial cells from human mammary gland by introduction of simian virus 40 large tumor antigen with a recombinant retrovirus. Proc Natl Acad Sci USA 88, 3520-3524.

The transient amplification and immortalization of primary normal and tumor cells by controlled expression of immortalizing gene or oncogenes such as SV40 large T has a lot of potential applications for basic and clinical studies (Westerman, 1996).

Bartek J, Bartkova J, Lalani EN, Brezina V and TaylorPapadimitriou J (1990). Selective immortalization of a phenotypically distinct epithelial cell type by microinjection of SV40 DNA into cultured human milk cells. Int J Cancer 45(6), 1105-1112. Bergeman J, KĂźhlcke K, Fehse B, Ratz I, Ostertag W and Lother H (1995). Excision of specific DNA-sequences from integrated retroviral vectors via site-sepecific recombination. Nucleic Acids Res 23, 4451-4456.

The Cre/LoxP-mediated controllable SV40 transformation and immortalization system may facilitate investigations of cell differentiation, oncogenesis, cell cycle, and senescence by allowing controlled cell proliferation and accurate on/off studies of various oncogenes in primary cells. This system also can be used to compare the phenotype of primary normal and tumor cells. From our observations the primary cells of normal and tumor cells transformed by SV40 large T showed some different phenotype and morphology. Because the stringent positive and negative selection yielded highly pure cell populations having permanently excised the transferred oncogene, this method may also prove especially adapted for the safety concerns of cell and gene therapies. Taking advantage of the cell expansion phase to perform gene targeting by homologous recombination may make genetic disease amenable to correction rather than gene addition.

Berthon P, Goubin G, Dutrillaux B, Degeorges A, Faille A, Gespach C and Calvo F (1992). Single-steep transformation of human breast epithelial cells by SV40 large T oncogene. Int J Cancer 2(1), 92-7. Blankenstein T, Cayeux S and Qin Z (1996). Genetic approaches to cancer immunotherapy. Rev Physiol Biochem Pharmacol 129, 1-49. Blasi E, Barluzzi R, Bocchini V, Mazzolla R and Bistoni F (1990). Immortalization of murine microglial cells by a vraf/v-myc carrying retrovirus. J Neuroimmunol 27, 229237. Bodnar AG, Ouellette M, Frolkis M, Holt SE, Chiu CP, Morin GB, Harley CB, Shay JW, Lichtsteiner S and Wright WE (1998). Extension of life-span by introduction of telomerase into normal human cells. Science 279, 349-352. Boisseu AM, Donatien PH, Surleve-Bazeille JE, Amedee J, Harmand MF and Taieb A (1992). Production of epidermal sheets in a serum free culture system: a further appraisal of the role of extracellular calcium. J Dermato Science 3, 111120.

References Aizawa S, Yaguchi M, Nakano M, Toyama K, Inokuchi S, Imai T, Yasuda M, Nabeshima R and Handa H (1994). Hematopoitic supportive function of human bone marrow

270

Gene Therapy and Molecular Biology Vol. 4, Page 271 Boon T, Cerottini JC, Van den Eynde B, van der Bruggen P and Van Pel A (1994). Tumor antigens recognized by T lymphocytes. Annu Rev Immunol 12, 337-365.

Dilworth SM (1990). Cell alterations induced by the large Tantigen of SV40 and polyoma virus. Semin Cancer Biol 1, 407-417.

Bosma GC, Custer RP and Bosma MJ (1983). A severe combined immune immunodeficiency mutation in the mouse. Nature 301, 527-530.

Duncan EL Whitaker NJ, Moy EL and Reddel RR (1993). Assignment of SV40-immortalized cells to more than one complementation group for immortalization. Exp Cell Res 205(2), 337-344.

Bouchard C, Staller P and Eilers M (1998). Control of cell proliferation by Myc. Trends in Cell Biology 8, 202-206. Boyce ST and Ham R (1985). Normal human epidermal keratiocytes. In In Vitro Models for Cancer Research, M. M. Weber and L. Sekely, eds. (Boca Raton, Florida: CRC Press), pp. 245-274.

Dutt K, Scott M, Wang M, Semple E, Sharma GP and Srinivasan A (1994). Establishment of a human retinal cell line by transfection of SV40 T antigen gene with potential to undergo neuronal differentiation. DNA Cell Biol 13(9), 909921.

Bryan, T. M., and Reddel, R. R. (1994) SV40-induced immortalization of human cells. Crit. Rev. Oncog. 5(4), 331-357

Dyson N and Balmain A (1999). Oncogenes and cell proliferation. Current Opinion in Genetics & Development 9, 11-14.

Cao, Y., Gao, Q., Wazer, D. E. and Band, V. (1997) Abrogation of wild-type p53-mediated transactivation is insufficient for mutant p53-induced immortalization of normal human mammary epithelial cells. Cancer Res. 57, 5584-5589

Edington KG, Loughran OP, Berry IJ and Parkinson EK (1995). Cellular immortality: a late event in the progression of human squamous cell carcinoma of the head and neck associated with p53 alteration and a high frequency of allele loss. Mol Carcinog 13, 254-265.

Carman TA, Afshari CA and Barret JC (1998). Cellular senescence in telomerase-expressing Syrian Hamster embryo cells. Exp Cell Res 244, 33-42.

Fanning E and Knippers R (1992). Structure and function of simian virus 40 large tumor antigen. Annu Rev Biochem 61, 55-85.

Carney DN, Bunn PA (Jr), Gazdar AF, Pagan JA and Minna JD (1981). Selective growth in serum-free hormonesupplemented medium of tumor cells obtained by biopsy from patients with small cell carcinoma of the lung. Proc Natl Acad Sci USA 78(5), 3185-3189.

Fitzgerald U, McLean LS and MacDonald C (1994). Production of immortal human endothelial cell lines by strontium phosphate transfection and electroporation of SV40 sequences. In Vitro Cell Dev Biol Anim 30A, 236-242. Fogh J and Trempe G (1975). New human tumor cell lines. In Human Tumor Cells In Vitro, J. Fogh, eds.( New York, London: Plenum Press), pp. 115-154.

Chang SE, Keen J, Lane EB and Taylor-Papadimitriou J (1982). Establishment and characterization of SV40-transformed human breast epithelial cell lines. Cancer Res 42(5), 20402053.

Foster SA and Galloway DA (1996). Human papillomavirus type 16 E7 alleviates a proliferation block in erly passage human mammary epithelial cells. Oncogene 12, 1773-1779.

Christian BJ, Loretz LJ, Oberley TD and Reznikoff CA (1987). Characterization of human uroepithelial cells immortalized in vitro by simian virus 40. Cancer Res 50, 4779-4786.

Galiana E, Borde I, Marin P, Rassoulzadegan M, Cuzin F, Gros F, Rouget P and Evrard C (1990). Establishment of permanent astroglial cell lines, able to differentiate in vitro, from transgenic mice carrying the polyoma virus large T gene: an alternative approach to brain cell immortalization. J Neurosci Res 26, 269-277.

Collodi P, Rawson C and Barnes D (1991). Serum-free media. In Human Cancer in Primary Culture: a Handbook, J. R.W. Masters, eds. (Dordrecht, Boston, London: Kluwer Academic Publishers), pp. 29-53. Compere SJ, Baldacci PA, Sharpe AH and Jaenisch R (1989). Retroviral transduction of the human c-Ha-ras1 oncogene into midgestation mouse embryos promotes rapid epithelial hyperplasia. Molecular and Cellular Biology 9, 6-14.

GaoQ, Hauser SH, Liu XL, Wazer DE, Madoc-Jones H and Band V (1996). Mutant p53-induced immortalization of primary mammary epithelial cells. Cancer Res 56, 3129-3133. Ge, GO, Coffman WD and Kubicek MT (1951). Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res 12, 264-265.

Cozens AL, Yezzi MJ, Chin L, Simon EM, Finkbeiner WE, Wagner JA and Gruenert DC (1992). Characterization of immortal cystic fibrosis tracheobronchial gland epithelial cells. Proc Natl Acad Sci USA 89(11), 5171-5.

Girardi AJ, Jensen FC and Koprowski H (1965). SV40-induced transformation of human diploid cells: crisis and recovery. J Cell Comp Physiol 65, 69-84.

Dairkee SH, Paulo EP, Traquina P, Moore DH, Ljung BM and Smith HS (1997). Partial enzymatic degradation of stroma allows enrichment and expansion of primary breast tumor cells. Cancer Res 57, 1590-1596.

Gossen M and Bujard H (1992). Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl Acad Sci USA 89, 5547-5551.

Dale EC and Ow DW (1991). Gene transfer with subsequent removal of the selection gene from the host genome. Proc Natl Acad Sci USA 88(23), 10558-10562.

Graham FL, Smiley J, Russell WC and Nairn R (1977). Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36, 59-74.

Dillman RO, Nayak, SK and Beutel L (1993). Establishing in vitro cultures of autologous tumor cells for use in active specific immunotherapy. J Immunother 14(1), 65-69.

Gu H, Zou YR and Rajewsky K (1993). Independent control of immunoglobulin switch recombination at individual switch

271

Li et al: Establishment of Tumor Cell Lines regions evidenced through Cre-loxP-mediated gene targeting. Cell 73(6), 1155-1164.

activity by the synthetic steroid RU 486. Nucleic Acids Res 24(8), 1404-1411.

Halbert CL, Demers GW and Galloway DA (1991). The E7 gene of human papillomavirus type 16 is sufficient for immortalization of human epithelial cells. J Virology 65, 473-478.

Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PL, Coviello GM, Wright WE, Weinrich SL and Shay JW (1994). Specific association of human telomerase activity with immortal cells and cancer. Science 266, 2011-2015.

Hammond SL, Ham RG and Stampfer MR (1984). Serum-free growth of human mammary epithelial cells: rapid clonal growth in defined medium and extended serial passage with pituitary extract. Proc Natl Acad Sci USA 81, 5435-5439.

Kiyono T, Foster SA, Koop JI, McDougall JK, Galloway DA and Klingelhutz AJ (1998). Nature 396, 84-88. Kreuzburg-Duffy UC and MacDonald C (1994). Establishment and characterization of murine macrophage-like cell lines following transformation with simian virus 40 DNA deleted at the origin of replication. J Immunol Methods 174(1-2), 33-51.

Harris SA, Enger RJ, Riggs BL and Spelsberg TC (1995). Development and characterization of a conditionally immortalized human fetal osteoblastic cell line. J Bone Miner Res 10, 178-186.

K端hn R, Schwenk F, Aguet M and Rajewesky K (1995). Inducible gene targeting in mice. Science 269, 1427-1429.

Hawley-Nelson P, Vousden KH, Hubert NL, Lowy DR and Schiller JT (1989). HPV 16 E6 and E7 proteins cooperate to immortalize human foreskin keratinocytes. EMBO J 8, 3905-3910.

Lechner JF and LaVck MA (1985). A serum-free method for culturing normal human bronchial epithelial cells at clonal density. J Tissue Culture Methods 9, 43-48.

Hayward SW, Dahiya, R, Cunha GR, Bartek J, Deshpande N, Narayan P (1995). Establishment and characterization of an immortalized but non-transformed human prostate epithelial cell line: BPH-1. In Vitro Cell Dev Biol Anim 31(1), 14-24.

Li LP and Chen JJ (1990a). Clonal growth and serial cultivation of human fetal nasopharyngeal epithelial cells in a serumfree medium. Acad J Sun Yat-Sen Univ Med Sci 4, 307311.

Heppner G and Miller F (1998). The cellular basis of tumor progression. Int Rev Cytology 177, 1-56.

Li LP and Chen JJ (1990b). Serial cultivation of human adult epithelial cells of nasopharyngeal biopsy samples. Chinese J Cancer 6, 340-345.

Hering S, Griffin BE and Strauss M (1991). Immortalization of human fetal sinusoidal liver cells by polyoma virus large T antigen. Exp Cell Res 195, 1-7.

Li LP, Schlag PM and Blankenstein T (1997). Transient expression of SV40 large T antigen by Cre/LoxP-mediated site-specific deletion in primary human tumor cells. Human Gene therapy 8, 1695-1700.

Hofman A, Noolan GP and Blau HM (1996). Rapid retroviral delivery of tetracycline-inducible genes in a single autoregulatory cassette. Proc Natl Acad Sci USA 93, 51855190.

Liangping Li, Schlag PM and Blankenstein T (1998). Establishment of new tumor cell line by genetic methods. Gene therapy & molecular biology. 2, 428-429.

Holt SE, Shay JW and Wright WE (1996). Refining the telomere-telomerase hypothesis of aging and cancer. Nature Biotech 14, 836-839.

Lie BL, Leung PC and Auersperg N (1996). Long-term growth and steroidogenic potential of human granulosa-lutein cells immortalized with SV40 large T antigen. Mol Cell Endocrinol 120, 169-176.

Jacobs JJL, Kieboom K, Marino S, DePinho RA and van Lohuizen M (1999). The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature 397, 164-168.

MacDonald C (1990). Development of new cell lines for animal cell biotechnology. Crit Rev Biotechnol 10(2), 155-178.

Janabi N, Peudenier S, Heron B, Ng KH and Tardiu M (1995). Establishment of human microglial cell lines sfter transfection of primary cultures of embryonic microglial cells with SV40 large T antigen. Neurosci Lett 195(2), 105108.

Manfredi JJ and Prives C (1994). The transforming activity of simian virus 40 large tumor antigen. Biochim Biophys Acta 1198(1), 65-83. Marincola FM, Hijazi YM, Fetsch P, Salgaller ML, Rivoltini L, Cormier J, Simonis TB, Duray PH, Herlyn M, Kawakami Y, and Rosenberg SA (1996). Analysis of expression of the melanoma-associated antigens MART-1 and gp100 in metastatic melanoma cell lines and in in situ lesions. J Immunother Emphasis Tumor Immunol 19(3), 192-205.

Jat PS and Sharp PA (1986). Large T antigens of simian virus 40 and polyomavirus efficiently establish primary fibroblasts. J Virol 59(3), 746-750. Jat PS and Sharp PA (1989). Cell lines established by a temperature-sensitive simian virus 40 large-T-antigen gene are growth restricted at the nonpermissive temperature. Mol Cell Biol 9(4), 1672-1681.

Masui T, Wakefield LM, Lechner JF, LaVeck MA, Sporn MB, and Harris CC (1986). Type beta transforming growth factor is the primary differentiation-inducing serum factor for normal human bronchial epithelial cells. Proc Natl Acad Sci USA. 83(8), 2438-2442.

Kelekar A and Cole MD (1987). Immortalization by c-myc, Hras, and E1a oncogenes induces differential cellular gene expression and growth factor responses. Mol Cell Biol 7, 3899-3907.

Merviel P, Degeorges A, Salat-Baroux J, and Calvo F (1995). Normal human endometrial cells in culture: characterization and immortalization of epithelial and stromal cells by SV 40 large T antigen. Biol Cell 84, 187-193.

Kellendonk C, Tronche F, Monaghan AP, Angrand PO, Stewart F and Schutz G (1996). Regulation of Cre recombinase

272

Gene Therapy and Molecular Biology Vol. 4, Page 273 Metz T, Harris AW and Adams JM (1995). Absence of p53 allows direct immortalization of hematopoietic cells by the myc and raf oncogenes. Cell 82, 29-36.

Schippers IJ, Moshage H, Roelofsen H, Muller M, Heymans HS, Ruiters M, Kuipers F (1997). Immortalized human hepatocytes as a tool for the study of hepatocytic (de-) differentation. Cell Bio Toxicol 13, 375-386.

Metzger D, Clifford J, Chiba H and Chambon P (1995). Conditional site-specific recombination in mammalian cells using a ligand-dependent chimeric Cre recombinase. Proc Natl Acad Sci USA 92(15), 6991-6995.

Sempl, TU, Moore GE, Morgan RT, Woods LK and Quinn LA (1982). Multiple cell lines from patients with malignant melanoma: morphology, karyology, and biochemical analysis. J Natl Cancer Inst 68 365-380.

Miyazaki H, and Horio T (1989). Growth inhibitors: Molecular diversity and roles in cell proliferation. In vitro 25, 866-872.

Serrano M, Lin AW, McCurrach ME, Beach D and Lowe SW (1997). Oncogenic ras provokes premature cell senscence associated with accumulation of p53 and p16INK4a. Cell. 88, 593-602.

Miyazaki M, Mihara K, Bai L, Kano Y, Tsuboi S, Endo A, Seshimo K, Yoshioka T and Namba M (1993). Immortalization of epithelial-like cells from human liver tissue with SV40 T-antigen gene. Exp Cell Res 206, 27-35.

Shay JW, Van Der Haegen BA, Ying Y, Wright WE (1989). The frequency of immortalization of human fibroblasts and mammary epithelial cells transfected with SV40 large Tantigen. Exp Cell Res. 209, 45-52.

Moor GE (1996). Derivation of tumor-derived cell lines (Melanoma cells): an overview. In Cell & Tissue Culture: Laboratory Procedures, A. Doyle and D.G. Newell, eds. (Chichester, New York, Brisbane, Toronto, Singapore: John Wiley and Sons), pp.11C: 1.1-11C: 1.12.

Shay JW, Wright WE, Brasiskyte D and Der Haegen V (1993). E6 of human papillomavirus type 16 can overcome the M1 stage of immortalization in human mammary epithelial cells but not in human fibroblasts. Oncogene 8, 1407-1413.

Nagata Y, Diamond B and Bloom BR (1983). The generation of human monocyte/macrophage cell lines. Nature 306(5943), 597-599.

Shay JW, Wright WE and Werbin H (1991). Defining the molecular mechanisms of human cell immortalization. Biochim Biophys Acta 1072, 1-7.

Oâ&#x20AC;&#x2122;Hare MJ (1991). Breast cancer in Human Cancer. In Primary Culture, A Handbook, ed. J. R. W. Masters, published by Kluwer Academic Publishers, p 271-286.

Shay JW, Tomlinson G, Piatyszek MA, and Gollahon LS (1995). Spontaneous in vitro immortalization of breast epithelial cells from a patient with Li-Fraumeni syndrome. Mol Cell Biol 15, 425-432.

Pantel K, Dickmanns A, Zippelius A, Klein C, Shi J, HoechtlenVollmar W, Schlimok G, Weckermann D, Oberneder R and Fanning E (1995). Establishment of micrometastatic carcinoma cell lines: a novel source of tumor cell vaccines. J. Natl. Cancer Inst. 87(15), 1162-1168.

Smith HS, Wolman SR and Hackett AJ (1984). The biology of breast cancer at the cellular level. Biochim Biophys Acta 78, 103-123.

Pantelouris, E. M. (1968) Absence of thymus in a mouse mutant Nature 217: 370-271

Smith JR, and Peraira-Smith OM (1996). Replicative senescence: implications for in vivo aging and tumor suppression. Science 273, 63-66.

Pei XF, Sherman L, Su YH, and Schlegel R (1998). HPV-16 E7 protein bypasses keratinocyte growth inhibition by serum and calcium. Carcinogenesis 19, 1481-1486.

Speirs V, Green AR, and White MC (1996). Collagenase III: A superior enzyme for complete disaggregation and improved viability of normal and malignant human breast tissue. In Vitro Cell Dev Biol Animal 32, 72-74.

Petit CA, Gardes M and Feunteun J (1983). Immortalization of rodent embryo fibroblast by SV40 is maintained by the A gene. Virology 127, 74-82.

Stampfer MR (1982). Cholera toxin stimulation of human mammary epithelial cells in culture. In Vitro 18(6), 531-537.

Pezzano M, Li Y, Yang YM and Guyden J (1991). The immortalization of thymic nurse cells by SV40 virus. Cell. Immunol. 133(2), 434-445.

Stamps AC, Gusterson BA, Oâ&#x20AC;&#x2122;Hare MJ (1992). Are tumours immortal? Eur J Cancer 28A(8-9), 1495-500.

Reddel RR, De-Silva R, Duncan EL, Rogan EM, Whitaker NJ, Zahra DG, Ke Y, McMenamin MG, Gerwin BI, and Harris CC (1995). SV40-induced immortalization and rastransformation of human bronchial epithelial cells. Int J Cancer 61(2), 199-205.

Stoner GD, Kaighn ME, Reddel RR, Resau JH, Bowman D, Naito Z, Matsukura N, You M, Galati AJ, Harris CC (1991). Establishment and characterization of SV40 T-antigen immortalized human esophageal epithelial cells. Cancer Res 51(1), 365-71.

Russ AP, Friedel C, Grez M and von Melchner H (1996). Selfdeleting retrovirus vectors for gene therapy. J Virol 70, 4927-4932.

Stoker M (1996). Fundamentals of cancer cell biology. Adv Cancer Res 70, 1-19.

Sarnow P, Ho YS, Williams J and Levine AJ (1982). The adenovirus E1b-58K tumor antigen and the SV40 large tumor antigen are physically associated with the same 54 K cellular protein. Cell 28, 287-394.

Strauss M, Herring S, Lubbe L and Griffin BE (1990). Immortalization and transformation of human fibroblasts by regulated expression of polyoma virus T antigens. Oncogene 5, 1223-1229.

Sauer B (1994). Site-specific recombination: development and applications. Current opinion in Biotech 5, 521-527.

Sugden B (1989). An intrcate route to immortality. Cell 57, 5-7. Taylor-Papadimitriou J, Purkis P, Lane EB, Mckay IA and Chang SE (1982). Effects of SV40 transformation on the cytoskeleton and behavioural properties of human keratinocytes. Cell Diff 11, 169-180

Sauer B (1993). Manipulation of transgenes by site-specific recombination: use of Cre recombinase. Methods Enzymol 225, 890-900.

273

Li et al: Establishment of Tumor Cell Lines Tegtmeyer P (1975). Function of simian virus 40 gene A in transformation infection. J Virol.15, 613-618.

Yarranton GT (1992). Inducible vectors for expression in mammalian cells. Current Opinion in Biotech. 3, 506-511.

Valverius EM, Ciardiello F, Heldin NE, Blondel B, Merlo G, Smith, G, Stampfer MR, Lippman ME, Dickson RB, Salomon DS (1990). Stromal influences on transformation of human mammary epithelial cells overexpressing c-myc and SV40T. J Cell Physiol 145(2), 207-16.

Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Sherr CJ and Roussel MF (1998). Myc signaling via the ARF tumor suppressor regulates p53-dependent apoptosis and immortalization. Gene & Development 12, 2424-2433.

van der Bosch J, Masui H and Sato GH (1981). Growth characteristics of primary tissues from heterotransplanted human colorectal carcinomas in serum-free medium. Cancer Res 41, 611-618. Van der Haegen BA and Shay JW (1993). Immortalization of human mammary epithelial cells by SV40 large T-antigen involves a two step mechanism. In Vitro Cell Dev Biol 29A(3 Pt 1), 180-182. Van Dyke TA (1994). Analysis of viral-host protein interactions and tumorigenesis in transgenic mice. Seminars in Cancer Biology 5, 47-60. Vegeto E, Allan GF, Schrader WT, Tsai M, McDonnell DP and O’Malley BW (1992). The mechanism of RU486 antagonism is dependent on the conformation of the carboxy-terminal tail of the human progesterone receptor. Cell 69, 703-713. Vidal K, Grosjean I, and Evillard JP, (1993). Immortalization of mouse intestinal epithelial cells by the SV40-large T gene: Phenotypic and immune characterization of the MODE-K cell line. J Immunol Methods 166, 63-73. Volkmann A, Neefjes J and Stockinger B (1996). A conditionally immortalized dendritic cell line which differentiates in contact with T cells or T cell-derived cytokines. Eur J Immunol 26(11), 2565-2572. Wang Y, O’Malley BW, Tsai SY and O’Malley BW (1994). A regulatory system for use in gene transfer. Proc Natl Acad Sci USA 91, 8180-8184. Weaver VM, Howlett AR, Langton-Webster B, Peterson OW, and Bissell MJ (1995). The development a functionally relevant cell culture model of progressive human breast cancer. Seminar in Cancer Bio 6, 75-184. Westerman KA and Leboulch P (1996). Reversible immortalization of mammalian cells mediated by retroviral transfer and site-specific recombination. Proc Natl Acad Sci USA 93(17), 8971-8976. Whyte P, Buchkovich K, Horowitz JM, Friend SH, Raybuck M, Weinberg RA, Harlow E (1988). Association between an oncogene and an anti-oncogene: the adenovirus E1A protein bind to the retinoblastoma gene product. Nature 334, 124129. Wilson A ( 1996). Preparation of ovarian carcinoma cell lines. In Cell & Tissue Culture: Laboratory Procedures, A. Doyle & D.G. Newell, eds. (Chichester, New York, Brisbane, Toronto, Singapore: John Wiley and Sons), pp.15C: 1.1-15C: 1.20. Wright WE, Pereira-Smith OM and Shay JW (1989). Reversible cellular senescence: implications for immortalization of normal human diploid fibroblasts. Mol Cell Biol 9(7), 30883092

274

Gene Therapy and Molecular Biology Vol 4, Page 275 Gene Ther Mol Biol Vol 4, 275-284. December 1999

Nuclear matrix and nucleotide excision repair: damage-recognition proteins are not constitutive components of the nuclear matrix Research Article

Piotr Widlak and Joanna Rzeszowska-Wolny Department of Experimental and Clinical Radiobiology, Center of Oncology, 44-100 Gliwice, Poland ____________________________________________________________________________________________________ Correspondence: Piotr Widlak, Ph.D. Phone: (48 32) 278 9804; Fax: (48 32) 231 3512; E-mail: widlak@onkol.instonko.gliwice.pl. Key Words: DNA damage, damage recognition, damaged DNA-binding proteins, nuclear matrix Received: 7 January 1999; accepted: 9 February 1999

Summary The nuclear matrix is a structure involved in organization of DNA structure and regulation of replication and transcription. It is generally believed that some enzymatic activities of nucleotide excision repair are localized in this nuclear subfraction and that the nuclear matrix anchorage affects the preferential repair of (potentially) active genes. Thus answering the question what is the role of nuclear matrix is very important to fully understand the DNA repair mechanisms. We have studied the in vitro interactions between nuclear matrices from rat liver cells and damaged DNA. A specific 36-bp DNA sequence was either UV-irradiated or damaged by benzo(a)pyrene diol epoxide and Nacetoxy-acetylaminofluorene. The data presented in this communication show that damaged DNA did not preferentially bind to nuclear matrices; damage-recognition proteins were loosely attached to the nucleoskeleton and were easily extracted from chromatin.

structures of the nucleus. The levels of DNA lesion induced by simple alkylating agents (Ryan et al, 1986), polycyclic aromatic hydrocarbons (e.g. benzo(a)pyrene) (Ueyama et al, 1981; Mironov et al, 1983; Obi et al, 1986; Perin-Roussel et al, 1988; Widlak & Rzeszowska-Wolny, 1993), aromatic amines (e.g. 2-acetylaminofluorene) (Gupta et al, 1985; Widlak & Rzeszowska-Wolny, 1994), UV-radiation (Widlak et al, 1996) and metal ions (Xu et al, 1994) were found to be higher in the matrix-attached DNA as compared to non-matrix chromatin fractions. It is suggested that preferential damage of the matrix-attached DNA reflects a specific active state of this chromatin fraction (Boulikas, 1992; MacLeod, 1995).

I. Introduction The genetic material of eukaryotic cells is organized into structural and functional domains. In the interphase chromatin such domains are frequently termed “DNA loops”. This loop organization seems to be maintained by anchorage of specific DNA sequences (MAR/SAR) to a protein network of the nucleoskeleton. The skeletal structures can be purified after removal of majority of DNA and chromatin proteins (nuclei are treated with nucleases and extracted with high salt buffers). Such a residual structure is called the “nuclear matrix”. The nucleoskeleton/nuclear matrix is thought to be involved not only in nuclear morphology but also in regulation of DNA replication and gene expression (reviewed in Bodnar 1988; Garrard 1990; Jackson et al, 1992; Boulikas 1995; Iborra et al, 1998; Davie et al, 1998).

It has been postulated that carcinogen-induced damage of the nucleoskeleton might be an important causative factor of functional and genetic instability of cancer cells (Pienta et al, 1989; Pienta & Ward, 1994). Thus, efficient repair of DNA fraction attached to this nuclear structure may be a very important task for cellular defense mechanisms. In fact, some DNA lesions can be removed faster from the matrix-bound DNA fractions than from non-matrix DNA.

Genomes of all organisms are under permanent pressure of genotoxic agents that can introduce damage into DNA structure. Many laboratories showed that genotoxic carcinogens preferentially damage DNA attached to skeletal

275

Widlak and Rzeszowska-Wolny: Nucleotide Excision Repair Such faster repair of the matrix-attached DNA fraction was shown for lesion induced by benzo(a)pyrene (Widlak & Rzeszowska-Wolny, 1993), chromium (Xu et al, 1994) and ionizing radiation (oxidative base damages like thymine glycol or 8-hydroxyguanine) (Zastawny et al, 1997). In contrary, damages induced by dimethylnitrosamine (Ryan et al, 1986) and 2-acetylaminofluorene (Gupta et al, 1985; Poirier et al, 1990) were not preferentially removed from the matrix-attached DNA.

association with the nucleoskeleton (McCready & Cook 1984; Harless & Hewitt 1987). Mullenders et al (1984, 1988, 1990) found that DNA repair patches were enriched in the nuclear matrix of human cells irradiated with low doses of UV. This effect was enhanced in cells deficient in global genome repair (xeroderma pigmentosum group C cells), but in cells deficient in transcription-coupled repair (Cockayne’s syndrome cells) the effect was lost. In contrary, we have found that unlike replicative DNA synthesis, the nuclear matrix from liver cells of rats injected with 2-aminofluorene was not enriched in newly synthesized DNA (Widlak & Rzeszowska-Wolny, 1994). This observation probably reflects the fact that 2aminoflurene-induced damage is not preferentially removed from the matrix-attached DNA and (unlike UVinduced damage) may not be preferentially removed from transcriptionally active genes in rodent cells (Tang et al, 1989). On the other hand, if the matrix localization of DNA synthesis is a general feature of NER, some differences in the experimental model used may also cause this discrepancy. As the repair patches are short (~30 nucleotides) and the rate of synthesis is rapid there is enough time during experiment for damage to be repaired at a skeleton and then for the repaired DNA to detach from it. Secondly, repaired DNA which was originally attached to the nucleoskeleton (probably through polymerases) might be subsequently detached from it during the matrix purification procedure. In fact, there is evidence showing that even in UV-irradiated cells newly synthesized repair patches of DNA can be easily released from the nucleoskeleton (Jackson et al, 1994).

The nucleotide excision repair (NER) is an important and universal pathway which removes a broad spectrum of DNA damage. NER can recognize and repair not only bulky lesion induced by UV-irradiation, polycyclic hydrocarbons or aromatic amines but also oxidative base damages. The process starts from recognition of a lesion followed by incision of DNA chain on both sides of the damage. Then, removal of the lesion-containing oligonucleotide takes place along with DNA repair synthesis and ligation to fill and close the resulting gap. In mammalian cells NER is tremendously complex and involves about 30 polypeptides. The first step of repair complex formation seems to be binding of a heterodimer XPA/RPA to damaged fragment, then other protein complexes can bind to target DNA. Binding of transcription factor TFIIH leads to unwinding of DNA helix around the damage and enables the action of endonucleases XPG and XPF/ERCC1 to incise the damaged strand on both sides of the damage. Some components of NER pathways are specific for global genome repair (XPC protein) or transcription-coupled repair (CSA and CSB proteins) (rev. in: Sancar, 1996; Friedberg, 1996; Lindahl et al,1997; DNA Wood, 1997). The XPA protein recognizes a broad spectrum of DNA damage and seems to be a general recognition factor, however other damage-recognition proteins are probably also involved in NER. Among them is XPE factor (UV-DDB protein), which shows the highest affinity to UV-induced 6-4 photoproducts (Naegeli, 1995). Different levels of NER efficiency can be distinguished: (i) slow repair of inactive genes, (ii) fast repair of transcriptionaly active (or potentially active) genes and (iii) accelerated repair of the transcribed strands, which involves transcription-coupled repair (TCR). Preferential repair of (potentially) active genes results not only from TCR involvement but is also determined by other factors (e.g. chromatin structure) (reviewed in Boulikas, 1992). Efficient removal of DNA damage from the matrix-bound fraction might simply reflect the association of preferentiallyrepaired transcriptionally active genes with the nuclear matrix. On the other hand, preferential repair of (potentially) active genes could result from their close proximity to the nuclear matrix where the repair machinery is localized (rev. in: Mullenders et al, 1990, Boulikas, 1995).

The data showing that DNA repair synthesis process is localized in the nucleoskeleton suggests the model in which damage in genomic DNA is recognized and brought into association with the nuclear matrix, where NER “machinery” is localized. Such a model has been confirmed by different lines of evidence. Park et al, (1996) showed that repair endonuclease XPG was firmly (but reversibly) associated with the nuclear matrix. Koehler & Hanawalt (1996) detected transient binding of damaged DNA to the nuclear matrix in UV-irradiated cells. Some activities involved in NER (unlike resynthesis step) are very loosely attached to chromatin (Bouayadi et al, 1997). This may suggest that the process is initiated by soluble proteins which scan the genome then mediate binding of damaged DNA to the matrix-localized repair machinery. The damaged-DNA binding (DDB) proteins are naturally best candidates for such a role. In fact, UVirradiation of mammalian cells induced translocation of UV-DDB, XPA and RP-A proteins (which all are involved in damage recognition) from a chromatin fraction loosely associated with the nuclear matrix to the tightly associated fraction (Otrin et al, 1997).

The data which suggested that nucleotide excision repair might be localized in specific nuclear compartment originally concerned DNA repair synthesis. It was shown that UV-induced unscheduled DNA synthesis occurred in

Previously we had shown that UV-damaged MAR sequences bound preferentially to the nuclear matrices from rats which were not treated with any DNA damaging

276

Gene Therapy and Molecular Biology Vol 4, Page 277 factors (Widlak et al, 1996). It might suggest that some damage-recognition proteins are constitutive components of the nuclear matrix. It has been postulated that proximity to the nuclear matrix is one of the factors involved in fast repair of (potentially) active genes (Boulikas, 1996). Thus, the nuclear matrix localization of damage-recognition proteins could partially explain the mechanisms of preferential repair of this fraction of a genome. The aim of this work was to elucidate whether damage-recognition proteins are residual components of the nuclear matrix.

II. Results and Discussion We studied the in vitro interactions between the nuclear matrices from rat liver cells and damaged DNA. A 36-bp duplex oligonucleotide was UV-irradiated or adducted by benzo(a)pyrene diol epoxide (BPDE) and Nacetoxy-acetylaminofluorene (AAAF). DNA lesions induced by the first two factors, but not the last one, are known to be removed with high efficiency from the matrixbound DNA and they may differ as a substrate for strandand gene-specific preferential repair. To evaluate the effects of DNA lesions upon protein-

DNA interactions, the in vitro binding of DNA to nuclear matrices was studied according to Cockerill & Garrard (1986) method. The nuclear matrices contained about 15% of total nuclear proteins, mainly lamins and other high molecular weight proteins but not histones. The fraction of matrix-bound DNA was separated from unbound DNA after centrifugation of matrices. The data from experiment in which matrix-DNA complexes were formed in the presence of non-damaged non-specific competitor (E. coli DNA) are shown in Figure 1A. We have found that none of the tested lesions affected the affinity of probed DNA to nuclear matrices. The binding efficiency of DNA carrying a lesion induced by AAAF, BPDE or UV was similar to that of undamaged DNA control. To confirm data obtained with a non-specific competitor, we studied formation of complexes in the presence of excess undamaged/damaged homologous competitor. DNA damaged by AAAF or BPDE was used as a radioactive probe while the same oligonucleotide or the undamaged oligonucleotide were used as nonradioactive competitors. The data which are shown in Figure 1B and 1C confirmed that both damaged and undamaged DNA have the same affinities for the nuclear matrices.

Figure 1 A. The in vitro binding of damaged DNA to the nuclear matrices. The complexes between nuclear matrix proteins and DNA probes (either undamaged or damaged) were formed in the presence of excess undamaged non-specific competitor. The matrix-binding efficiency is presented as a relative amount of the probe bound to matrices. Values are means from 3 assays.

277

Widlak and Rzeszowska-Wolny: Nucleotide Excision Repair

Figure 1 B and C. The in vitro binding of damaged DNA to the nuclear matrices. The complexes between nuclear matrix proteins and AAAF-damaged or BPDE-damaged probes were formed in the presence of excess of undamaged or damaged homologous competitor. Values are means from 3 assays (Âą S.D.). .

We also compared the distribution of a damage between fractions of DNA that did or did not bind to nuclear matrices. In this experiment about 50% of the probed DNA (in which one lesion per about 10 molecules was present) became bound to the matrices. The patterns of lesions in matrix-bound and unbound DNA fractions are shown in Figure 2. The adducts patterns in both fractions were similar for all three DNA-damaging factors. We calculated the relative levels of damage (the amount of adducts per the amount of total DNA) and compared these values between fractions. The ratio between adduct levels in the matrix-bound and the unbound DNA fractions was 1.4, 1.1 and 1.3 for AAAFdamaged, BPDE-damaged and UV-irradiated DNA, respectively (the values are means from three experiments). The data show that the nuclear matrices bound damage-carrying DNA molecules with very weak (yet detectable) preferences over undamaged DNA. We showed previously that the presence of UV-induced damage increased the efficiency of MAR DNA binding to the nuclear matrix in vitro (Widlak et al, 1996). In an experiment in which we studied binding of UV-irradiated kappa Ig MAR DNA to rat liver nuclear matrices the level of lesion in matrix-bound DNA was 5-fold higher as compared to unbound fraction. The data presented in this communication suggest that the observed phenomenon was specific for MAR sequence but not for the DNA sequence studied here (the MAR-binding proteins recognize some specific structural features of DNA which are probably affected by the UV-induced photoproducts).

Figure 2. The chromatograms of damaged DNA complexed in vitro with the nuclear matrix proteins. Shown are fractions bound to the matrices as well as unbound material. About 100 ng of each DNA was used for the assay. Denoted are major forms of lesions (according to synthetic standards). CPD cyclobutane pyrimidine dimers.

278

Gene Therapy and Molecular Biology Vol 4, Page 279 The data presented above suggest that most of damage-recognition proteins were extracted out from nuclei during the preparation of matrices. To study the presence of damage-recognition proteins in nuclear extracts we used a method based on electrophoretic mobility shift assay (EMSA). In this method, damaged DNA is complexed in vitro with soluble protein extracts in the presence of non-damaged competitor DNA. The presence of proteins that have high affinity to damaged DNA is detected by gel electrophoresis in the form of retarded bands containing protein-DNA complexes (Protic & Levine, 1993). In our experiments we have used DNA damaged by AAAF, BPDE or UV-radiation as radioactive molecular probes. The damage-recognition proteins were detected in protein fractions extracted from nuclei with increasing concentrations of NaCl, thus differing in stringency of their binding to the chromatin and nuclear matrix. The data showing the presence of proteins that preferentially bound to either UV-irradiated or AAAF-damaged DNA are shown on Figure 3. The proteins that specifically recognized UV-irradiated DNA (UV-DDB) were loosely attached to chromatin and could be eluted from the nucleus with low-salt buffers (0.15 M NaCl). The proportion of UV-DDB proteins to total extracted proteins markedly decreased in fractions extracted with 0.8 M (and higher) NaCl. The proteins which specifically recognized AAAF-damaged DNA

(AAAF-DDB) were moderately attached to nuclear structures. They were present in fractions removed from nuclei with 0.4 M (and higher) NaCl (the highest proportion of AAAF-DDB proteins to total extracted proteins was detected in 0.6 M NaCl extracts). When BPDE-damaged DNA was used as a radioactive probe any complexes specific for damaged DNA could not be detected (data not shown). However, if BPDE-damaged DNA was used as a specific competitor we found that the same proteins which recognized AAAF-induced lesions also bound to DNA damaged by BPDE, yet with much lower efficiency (see panel B in Figure 3). The amount of proteins used for reactions was adjusted according to the total protein amount in extracts. However, different amounts of proteins (in proportion to total nuclear content) were extracted with increased salt concentration. Extracts used in this experiment contained about 12, 34, 51, 86, 92 and 90 percent of the total nuclear proteins (in 0.15, 0.4, 0.6, 0.8, 1.0 and 2.0 M extracts, respectively). Thus, the observed decrease in DDB protein level in the fractions extracted with higher salt concentration was due to their â&#x20AC;&#x153;dilutionâ&#x20AC;? with other nuclear proteins (e.g. histones). As treatment of the nuclei with high salt buffers (up to 2 M NaCl) neither increased the amount nor released any new species of DDB-proteins one can conclude that damage-recognition proteins are weakly bound to chromatin or the nuclear skeleton.

Figure 3. The EMSA analysis of DNA damage-recognition proteins from rat liver nuclei. Panel A. The in vitro complexes were formed between radioactive DNA probes (UV-irradiated or damaged by AAAF) and nuclear proteins extracted with the indicated NaCl concentrations (from 0.15 to 2.0 M), in the presence of non-specific undamaged competitor. The complexes were analyzed by polyacrylamide gel electrophoresis. Denoted are positions of free probe and retarded complexes containing UV-DDB and AAAFDDB proteins. Panel B. The complexes were formed between radioactive AAAF-damaged probe and proteins extracted with 0.6 M NaCl, in the presence of excess non-radioactive oligonucleotide (undamaged and damaged by AAAF or BPDE) as a specific competitor.

279

Gene Therapy and Molecular Biology Vol 4, Page 280 The data presented above show that damagerecognition proteins interact loosely with the nuclear matrix and can be easily extracted from nuclei. As the nuclear matrices used in our experiments were purified from cells not pretreated with DNA damaging factors this left open the possibility that such a treatment would affect the nuclear distribution of damage-recognition proteins. Such an effect was observed in cultured cells, where UVirradiation caused transient translocation of proteins participating in damage recognition to chromatin fractions more tightly attached to the nucleoskeleton (Otrin et al, 1997). To determine whether carcinogen treatment induces translocation of damage-recognition proteins we purified nuclear proteins from animals treated with 2acetylaminofluorene or benzo(a)pyrene. Rats were i.p. injected with the carcinogens (50 mg/kg of body weight) then liver samples were collected at different time intervals (from 6 hours to 7 days after the treatment). The levels of adducts in hepatic DNA (assayed by the 32Ppostlabeling method) were highest 48 hours after 2-AAF treatment (about 16 fmol/Âľg DNA) and 24 hours after B(a)P treatment (about 13 fmol/Âľg DNA). We found that the nuclear matrices prepared from liver cells of animals treated with 2-AAF or B(a)P (48 and 24 hours after treatment, respectively) did not show more effective binding of DNA damaged by AAAF or BPDE (as compared to non-treated control animals, data not shown).

Another in vitro method that can be used as a supplementary analytical tool to detect DDB-proteins is the Southwestern blot (Protic & Levine, 1993). In this method the probed DNA forms complexes with membrane-bound proteins after their electrophoretic resolution (usually on SDS/polyacrylamide gels). This assay depends on the ability of tested proteins to recover their structure (thus being limited to proteins which do not form oligomers). However, it can be successfully used to study nuclear matrix proteins (Widlak et al, 1995). Figure 4 shows the binding of DNA probes to specific proteins that were either extracted from nuclei with 0.2, 0.4 and 1.0 M NaCl (low-salt extract, medium-salt extract and high-salt extract, respectively) or present in the nuclear matrices. Under the conditions of the experiment (40-fold excess of the competitor over the radioactive probe) the undamaged control DNA bound exclusively to histone proteins (both histone H1 and core histones). A similar binding pattern was seen with UV-irradiated DNA as a probe and no specific UV-DDB proteins were detected. This is in agreement with the fact that major UV-DDB protein is a heterodimer (Otrin et al, 1997). Some additional protein bands were seen on the filter tested with DNA damaged by AAAF. Such non-histone proteins which bound AAAF-damaged DNA were detected mostly in nuclear extract fractions. Some AAAF-DDB protein bands detected in the nuclear matrix probes were not matrix-specific as they were seen also in nuclear extracts.

Figure 4. The Southwestern blot analysis of DNA damage-recognition proteins from rat liver nuclei. The in vitro complexes were formed between radioactive DNA probes (undamaged or damaged by AAAF or UV-irradiated) and nuclear proteins extracted with 0.2 M (low-salt extract - LSE), 0.4 M (medium-salt extract - MSE) and 1.0 M NaCl (high-salt extract - HSE) or the nuclear matrix proteins (Matrix). Nuclear proteins were resolved by electrophoresis on 13% polyacrylamide/SDS gel. The complexes were formed in the presence of excess undamaged non-specific competitor. Denoted are positions of histones and molecular size markers.

280

Gene Therapy and Molecular Biology Vol 4, Page 281

Figure 5. The analysis of AAAF-DDB proteins in liver cells of carcinogen-treated animals. Panel A. The complexes were formed in vitro between AAAF-damaged DNA probe and nuclear proteins extracted with 1.0 M NaCl from liver cells of control animals and animals injected with 2-AAF or B(a)P (24 hours after the treatment). Two animals in each group are shown. The complexes were formed in the presence of non-specific competitor. Panel B. The complexes were formed between AAAF-damaged DNA probe and nuclear proteins extracted with buffers of increasing NaCl concentration (0.2-1.0 M) from liver cells of untreated control rats and animals injected with 2-AAF (48 hours after the treatment).

factors may affect the repair of this DNA fraction. The data presented in this communication show that damage-recognition proteins are not constitutive components of the nuclear matrix. Thus, other factors which account for preferential repair of the matrixattached DNA have to be searched for.

In another experiment we assayed the level of AAAF-damaged DNA binding proteins in 0.4 and 1.0 M NaCl nuclear extracts purified from liver cells of animals injected with 2-AAF and B(a)P. We found that treatment with neither 2-AAF nor B(a)P (at any time tested) changed the levels of AAAF-DDB proteins (see Figure 5A). To clarify whether the carcinogen treatment affected the association with nuclear components the relative levels of these proteins were assayed in nuclear hepatic fractions extracted using increasing salt concentrations from control and 2-AAFtreated (48 hours after injection) animals. Figure 5B shows that the treatment of animals with 2-AAF did not change the association of AAAF-DDB protein to the hepatic chromatin. As the treatment with the 2-AAF did not affect the nuclear distribution of AAAF-DDB proteins this may reflect the fact that repair pathways specific for these lesions and UV-induced lesions are somehow different.

Experimental Procedures A. DNA probes Synthetic double-strand 36bp-long oligonucleotides (5’AATTCGTAGG CCTAAGAGCA ATCGCACCTG TGCGCG-3’, with blunt ends) were used as molecular probes. Oligonucleotides were UV-irradiated (5 kJ/m2) using a 254 nm UV-crosslinker (Stratagene). Alternatively, oligonucleotides (at 10 µM concentration) were incubated for 4 hours at 37oC with 40 µM AAAF or BPDE (Midwest Research Institute), then purified by phenol/chloroform extraction and ethanol precipitated. In all three cases, used procedures introduced lesion into about 15% of DNA molecules (on the average), which was checked by the 32P-postlabeling method. Oligonucleotides were end-labeled with !32P ATP using T4 polynucleotide kinase and purified by polyacrylamide gel electrophoresis.

It is generally believed that nuclear matrix localization of certain activities involved in NER contributes to preferential repair of the DNA fraction associated with this nuclear structure. One can assume that the nuclear matrix localization of damage recognition 281

Widlak and Rzeszowska-Wolny: Nucleotide Excision Repair buffer consists of: 20 mM Tris-HCl pH 7.6; 5 mM MgCl2; 0.5 mM EDTA; 1 mM DTT; 5% glycerol and 150 mM NaCl (final concentration). Complexes (in final volume 20 µl) were formed in the presence of non-specific (sonicated E. coli DNA, 2µg) or specific (1µg) competitors. Complexes were resolved by electrophoresis on 6% polyacrylamide gel (in 0.5 x TBE running buffer) and detected by autoradiography.

B. Preparation of nuclear proteins Nuclei were purified from homogenized liver tissue from adult male WAG rats. To obtain extracts of nuclear proteins, nuclei were incubated for 30 minutes at 4oC with buffer consisting of: 10 mM Hepes-NaOH pH 7.9; 1.5 mM MgCl2; 0.1 mM EGTA; 0.5 mM DTT; 5% glycerol; protease inhibitors and NaCl at different molarity (ranging from 0.15 to 2.0 M). Nuclei extracted at salt concentration higher than 0.5 M were briefly (2-3 sec.) sonicated. Insoluble remnants of the nuclei were pelleted by centrifugation for 30 minutes at 16,000 rpm at 4oC. Nuclear matrices were prepared by the DNaseI/”high salt” method as in previous experiments (Widlak & Rzeszowska-Wolny, 1994; Widlak et al, 1995), without copper stabilization. Briefly, nuclei were purified by centrifugation in 2.2 M sucrose and washed with 1% Triton X100. The nuclei were then treated with DNaseI (10 µg/mg of nuclear protein) for 1 hour at 20oC in 0.1 M NaCl and next extracted with 0.5 M NaCl followed by 2 M NaCl to obtain residual matrix fraction.

F. Southwestern blot analysis The nuclear proteins (60 µg per slot) were fractionated on 13% polyacrylamide/SDS gel and electrotransferred onto PVDF membrane (Hybond-P; Amersham) in 25 mM Tris, 190 mM glycine and 20% methanol. Filter-bound proteins were renatured by incubation in the hybridization oven for 5 hours at 25oC with 25 mM Tris-HCl (pH 7.6), 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 5 mM MgCl2 and 2.5% BSA. After washing with the binding buffer (same composition as above except for 0.25 % BSA added) filters were incubated for 5 hours at 25oC it the binding buffer supplemented with 500 ng of 32P-end-labeled DNA probe and 20 µg of non-radioactive E. coli DNA (final volume 10 ml). Filters were then washed and autoradiographed.

C. Assay of DNA adducts DNA was assayed for the presence of adducts according to 32P-postlabeling method (Gupta et al, 1982). UV-induced adducts were analyzed according to Bykov et al (1995). Adducts induced by 2-AAF derivatives were enriched by butanol extraction, while adducts induced by B(a)P derivatives were enriched by nuclease P1 treatment (Widlak et al, 1996b). 32 P-labeled nucleotides were resolved by multi-dimensional thin layer chromatography. Adduct spots were visualized by autoradiography and quantitated by scintillation counting. To calculate the level of adducts in damaged oligonucleotides the number of total nucleotides was assayed according to Gupta et al (1982).

Acknowledgements The work was supported by the Polish Committee for Scientific Research (KBN, grant 6P04A04112). Special thanks to Mr. A. Sochanik for help in preparation of the manuscript.

References Bodnar JW (1988). A domain model for eukaryotic DNA organization: a molecular basis for cell differentiation and chromosome evolution. J Theor Biol 132, 479-507.

D. Complex formation between DNA and nuclear matrices

Bouayadi K, van der Leer-van Hoffen A, Balajee AS, Natarajan AT, van Zeeland AA and Mullenders LHF (1997). Enzymatic activities involved in the DNA resynthesis step of nucleotide excision repair are firmly attached to chromatin. Nucl Acids Res 25, 1056-1063.

About 25 µg of the matrix proteins were suspended in 0.1 ml of the binding buffer comprising 50 mM NaCl, 2 mM EDTA, 0.25 M sucrose, 10 mM Tris-HCl (pH 7.4), 25 µg BSA, 25 ng of 32P-end-labeled DNA probe and different amounts of non-radioactive non-specific (sonicated E. coli DNA) or specific (oligonucleotides) competitor. After 1 hour incubation at 25 oC matrices were recovered by centrifugation. To determine the amount of DNA bound to the matrices the radioactivity of pellets (matrix-bound fraction) and supernatants (unbound fraction) was quantitated in scintillation counter. In separate experiments, 25 µg of the matrix proteins were incubated with 250 ng of the oligonucleotides in the absence of non-specific competitor. Under these conditions about 50% of added DNA remained in the matrix-bound fraction. Both the pellet and the supernatant fractions were treated with proteinase K/SDS and extracted by phenol/chloroform. DNA was then recovered by ethanol precipitation and assayed for the presence of the adducts.

Boulikas T (1992). Evolutionary consequences of nonrandom damage and repair of chromatin domains. J Mol Evol 35, 156-180. Boulikas T (1995). Chromatin domains and prediction of MAR sequences. Int Rev Cytol 162A, 279-388. Boulikas T (1996). A unified model explaining the preferential repair of active over inactive genes and of the transcribed over the nontranscribed strand: a leading role for transcription factors and matrix anchorage. Int J Oncol 8, 77-84. Bykov VJ, Kumar R, Forsti A and Hemminki K (1995). Analysis of UV-induced photoproducts by 32P-postlabeling. Carcinogenesis 16, 113-118. Cockerill PN and Garrard WT (1986). Chromosomal loop anchorage of the kappa immunoglobulin gene occurs next to the enhancer in a region containing topoisomerase II sites. Cell 44, 273-282.

E. Electrophoretic mobility shift assay (EMSA) Radioactive DNA probes (25 ng) were incubated with nuclear proteins (5µg) for 30 minutes at 4oC. The binding

282

Gene Therapy and Molecular Biology Vol 4, Page 283 Davie JR, Samuel S, Spencer V, Bajno L, Sun J-M, Chen HY and Holth LT (1998). Nuclear matrix: application to diagnosis of cancer and role in transcription and modulation of chromatin structure. Gene Ther Mol Biol 1, 509-528.

Naegli H (1995). Mechanisms of DNA damage recognition in mammalian nucleotide excision repair. FASEB J 9, 10431050.

Friedberg EC (1996). Relationships between DNA repair and transcription. Annu Rev Biochem 65, 15-42.

Obi FO, Ryan AJ and Billett MA (1986). Preferential binding of the carcinogen benzo(a)pyrene to DNA in active chromatin and the nuclear matrix. Carcinogenesis 7, 907-913.

Garrard WT (1990). Chromosomal loop organization in eukaryotic genomes. In: Nucleic Acids and Molecular Biology, vol.4, eds.: F. Eckstein and D.M.J. Lilley. SpringerVerlag Berlin, pp. 163-175.

Otrin VR, McLenigan M, Takao M, Levine AS and Protic M (1997). Translocation of a UV-damaged DNA binding protein into a tight association with chromatin after treatment of mamalian cells with UV light. J Cell Sci 110, 1159-1168.

Gupta RC, Reddy MV and Randerath K (1982). 32P-postlabeling analysis of non-radioactive aromatic carcinogen DNA adducts. Carcinogenesis 3, 1081-1092.

Park MS, Knauf JA, Pendergrass SH, Coulon CH, Strniste GF, Marrone BL and MacInnes MA (1996). Ultraviolet-induced movement of the human DNA repair protein, Xeroderma pigmentosum type G, in the nucleus. Proc Natl Acad Sci USA 93, 8368-8373.

Gupta RC, Dighe NR, Randerath K and Smith HC (1985). Distribution of initial and persistent 2-acetylaminofluoreneinduced DNA adducts within DNA loops. Proc Natl Acad Sci USA 82, 6605-6608.

Perin-Roussel O, Barat N and Zajdela F (1988). Non-random distribution of dibenzo(a,e)fluoranthene-induced DNA adducts in DNA loops in mouse fibroblast nuclei. Carcinogenesis 9, 1383-1388.

Harless J and Hewitt RR (1987). Intranuclear localization of UVinduced DNA repair in human VA13 cells. Mutation Res 183, 177-184.

Pienta KJ, Partin AW and Coffey DS (1989). Cancer as a disease of DNA organization and dynamic cell structure. Cancer Res 49, 2525-2532.

Iborra F, Pombo A and Jackson DA (1998). Dedicated sites of gene expression in the nuclei of mammalian cells. Gene Ther Mol Biol 1, 495-508.

Pienta KJ and Ward WS (1994). An unstable nuclear matrix may contribute to genetic instability. Medical Hypotheses 42, 45-52.

Jackson DA, Dolle A, Robertson G and Cook PR (1992). The attachments of chromatin loops to the nucleoskeleton. Cell Biol Int Rep 16, 687-696.

Protic M and Levine AS (1993). Detection of DNA damagerecognition proteins using the band-shift assay and Southwestern hybridization. Electrophoresis 14, 682-692.

Jackson DA, Balajee AS, Mullenders L and Cook PR (1994). Sites in human nuclei where DNA damaged by ultraviolet light is repaired: visualization and localization relative to the nucleoskeleton. J Cell Sci 107, 1745-1752.

Ryan AJ, Billett MA and O'Connor PJ (1986). Selective repair of methylated purines in regions of chromatin DNA. Carcinogenesis 7, 1497-1503.

Koehler DR and Hanawalt PC (1996). Recruitment of damaged DNA to the nuclear matrix in hamster cells following ultraviolet irradiation. Nucl Acids Res 24, 2877-2884.

Sancar A (1996). DNA excision repair. Annu Rev Biochem 65, 43-81.

Lindahl T, KarranP and Wood RD (1997). DNA excision repair pathways. Curr Opin Gen Dev 7, 158-169.

Tang MS, Bohr VA, Zhang XS, Pierce J and Hanawalt PC (1989). Quantification of aminofluorene adduct formation and repair in defined DNA sequences in mammalian cells using the UVRABC nuclease. J Biol Chem 264, 1445514462.

MacLeod MC (1995). Interaction of bulky chemical carcinogens with DNA in chromatin. Carcinogenesis 16, 2009-2014. McCready SJ and Cook PR (1984). Lesion induced in DNA by ultraviolet light are repaired at the nuclear cage. J Cell Sci 70, 189-196.

Ueyama H, Matsuura T, Nomi S, Nakayasu H and Ueda K (1981). Binding of benzo(a)pyrene to rat liver nuclear matrix. Life Sci 29, 655-661.

Mironov NM, Grover PL and Sims P (1983). Preferential binding of polycyclic hydrocarbons to matrix-bound DNA in rat liver nuclei. Carcinogenesis 4, 189-193.

Widlak P and Rzeszowska-Wolny J (1993). DNA and proteins of the nuclear matrix are the main targets of benzo(a)pyrene's action in rat hepatocytes. Acta Biochim Polon 40, 559-562.

Mullenders LHF, van Kesteren AC, Bussmann CJM, van Zeeland AA and Natarajan AT (1984). Preferential repair of nuclear matrix associated DNA in xeroderma pigmentosum complementation group C. Mutation Res 141, 75-82.

Widlak P and Rzeszowska-Wolny J (1994). DNA repair is less efficient in the nuclear matrix than in non-matrix nuclear fractions in the liver of rats treated with 2-aminofluorene. Cancer Lett 78, 115-120.

Mullenders LHF, van Kesteren van Leeuwen AC, van Zeeland AA and Natarajan AT (1988). Nuclear matrix associated DNA is preferentially repaired in normal human fibroblasts, exposed to a low dose of ultraviolet light but not in Cockayneâ&#x20AC;&#x2122;s syndrome fibroblasts. Nucleic Acids Res 16, 10607-10622.

Widlak P, Rogolinski J and Rzeszowska-Wolny J (1995). Interactions of the matrix attachment region of DNA with the matrix proteins from the copper preincubated liver nuclei. Acta Biochim Polon 42, 205-210. Widlak P, Bykov VJ, Hemminki K and Rzeszowska-Wolny J (1996). The non-random distribution of UV-induced photoproducts in the nuclear matrix and non-matrix DNA fractions. Cancer Lett 108, 215-223.

Mullenders LHF, Venema J, van Hoffen A, Mayne LV, Natarajan AT and van Zeeland AA (1990). The role of the nuclear matrix in DNA repair. In: Mutation and the Environment, Part A, Wiley-Liss, Inc., pp. 223-232.

Widlak P, Grzybowska E, Hemminki K, Santella R and

283

Widlak and Rzeszowska-Wolny: Nucleotide Excision Repair Choraz M (1996b). 32P-postlabelling of bulky human DNA adducts enriched by different methods including immunoaffinity chromatography. Chem Biol Interact 99, 99-107. Wood RD (1997). Nucleotide excision repair in mammalian cells. J Biol Chem 272, 23465-23468. Xu J, Manning CR and Patierno SR (1994). Preferential formation and repair of chromium-induced DNA adducts and DNA-protein crosslinks in nuclear matrix DNA. Carcinogenesis 15, 1443-1450. Zastawny TH, Czerwinska B, Drzewiecka B and Olinski R (1997). Radiation-induced oxidative DNA base damage and its repair in nuclear matrix-associated DNA and in bulk DNA in hepatic chromatin of rat upon whole-body !irradiation. Free Radical Biol Med 22, 101-107.

Piotr Widlak

284

Gene Therapy and Molecular Biology Vol 4, Page 285 Gene Ther Mol Biol Vol 4, 285-290. December 1999.

Functional improvement in ligament scar tissue following antisense gene therapy: A model system for in vivo engineering of connective tissues Review Article

David A. Hart, N. Nakamura*, R. Boorman, L. Marchuk, H. Hiraoka, Y. Kaneda*, N.G. Shrive and C.B. Frank McCaig Centre for Joint Injury and Arthritis Research, University of Calgary, Calgary, Alberta Canada and (*) the University of Osaka, Osaka, Japan _________________________________________________________________________________________________ Correspondence: David A. Hart, Ph.D., Professor, McCaig Centre for Joint Injury and Arthritis Research, Faculty of Medicine, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta Canada T2N 4N1. Tel: 403-220-4571; Fax: 403-283-7742; E-mail: hartd@ucalgary.ca Key Words: gene therapy of ligament healing, antisense therapy, in vivo tissue engineering, connective tissue healing Received: 6 August, 1999; accepted: 17 August, 1999

Summary The present studies indicate that decorin antisense approaches and a HVJ-liposome delivery system can functionally impact on the mechanical properties of MCL scar tissue in a rabbit model. This is one of the first instances where such approaches have led to functional improvement in the tissue. While decorin antisense gene therapy has been shown to be effective at improving MCL scar tissue, the resulting tissue still remains deficient compared to normal tissue. Therefore, this study must be considered a first step towards tissue regeneration and additional experimentation is required to further move scar tissue down the path towards tissue regeneration.

to lead to altered biomechanical function of the tissue (reviewed in Frank et al, 1999). One of the consequences of compromised biomechanical integrity in the tissues discussed above is impaired function and increased risk of re-injury, induction of remodeling in other joint tissues, and development of diseases such as osteoarthritis. Therefore, development of new therapies to enhance repair or replace damaged joint tissues would have an impact on the quality of life for a large number of individuals of all ages. Considerable effort is currently on-going to develop both in vivo and in vitro methods to either improve healing of connective tissues (ie: pleiotrophic stem cells, in vivo delivery of growth factors/cytokines, drugs) or to replace damaged tissues (ie: in vitro tissue engineering, in vitro expansion of endogenous cells). In recent years, this effort has involved the use of gene therapy approaches. While much of this effort has focussed on replacing damaged tissue, not as much effort has been concentrated on improving in vivo healing to prevent development of joint disease. In part, this is likely due to the difficulties inherent in delivering

I. Introduction Overt injuries to joint tissues such as ligaments,tendons, menisci, and cartilage are major clinical problems which occur with significant frequency. In some instances, such as the extraarticular ligaments and tendons, the injured tissues heal with the formation of scar tissue (Frank, 1996; Frank et al, 1983,1994,1999). A number of the other injured tissues either heal poorly or not at all. The latter scenario may be due to the loss of a template to promote the healing process (ie: the ruptured anterior cruciate ligament (ACL) of the knee) or due to the fact that some of these tissues are either hypo- or aneural and are hypo- or avascular (cartilage, menisci). Even those tissues that do initiate scar formation following injury remain functionally compromised for extended periods of time due to the inferior mechanical and biochemical properties of the scar tissue (reviewed in Frank, 1996; Frank et al, 1983,1994,1999). Furthermore, rupture of ligaments such as the ACL often require reconstruction with autografts derived from other tissues (ie: patellar tendon) and these autografts gradually become infiltrated with scar tissue, a process which appears 285

Hart et al: Antisense gene therapy for connective tissue engineering effective concentrations of antisense ODN or transgenes to such hypocellular and dense connective tissues. Our initiative in the area of employing gene therapy approaches to improve in vivo connective tissue healing has involved the well characterized rabbit medial collateral ligament (MCL) injury model (Boykiw et al, 1998; Frank, 1996; Frank et al, 1983, 1992, 1994, 1999; Sciore et al, 1998). We chose an antisense gene therapy approach to transiently modulate the expression of specific molecules during the early stages of healing where there is significant hypercellularity and the density of the scar tissue is less than that of normal tissue. We did not want to permanently alter the target tissue, but instead wanted to temporarily influence the scar tissue until it had regained more normal properties. This model has been subjected to multidisciplinary characterization (histology, biochemistry, molecular biology and biomechanical assessment) and therefore lends itself to complete evaluation of the functional impact of the gene therapy interventions. In the present report, we present the results of studies using antisense ODN directed toward the small matrix proteoglycan, decorin. Decorin has been implicated in extracellular matrix formation and organization both directly via collagen fibril assembly (Vogel et al, 1984; discussed in Nakamura et al, 1998b) and indirectly through its ability to bind and functionally inactivate growth factors such as TGFbeta (Harper et al, 1984; Lysiak et al, 1995). The results indicate that is is possible to functionally improve the properties of ligament scar tissue using decorin-specific antisense ODN and a hemmaglutinating virus of Japanliposome (HVJ-liposome) delivery system (discussed in Nakamura et al, 1996, 1998, 1998a, 1998b; Tomita et al, 1997).

for decorin, decorin protein levels in the tissue, morphologic methods (light and transmission electron microscopy), and well established biomechanical techniques to measure both low and high load behavior of the scar tissue (reviewed in Frank et al, 1999). In a second set of experiments, rabbits with bilateral MCL injuries were exposed chronically (21 days, 2.5 ng TGF-beta/day) to exogenous human TGF-beta (Dr. Paul Gladstone, Bristol Myers-Squibb, Seattle, WA) or saline delivered via implanted ALZET pumps from day 0 to day 21, or to a single bolus injection of 7 Âľg TGF-beta delivered to the scar tissue at 3 weeks post-injury. Animals were sacrificed at 6 weeks post injury and assessed as above.

III. Influence of decorin antisense ODN on decorin mRNA and protein levels in MCL scar tissue in vivo Total RNA was isolated from MCL tissue by the TRIspin method (Reno et al, 1997) and assessed for decorin mRNA levels by semiquantitative RT-PCR as described previously (Sciore et al, 1998, Boykiw et al, 1998, Reno et al, 1998). Treatment of MCL scar tissue with decorin antisense ODN led to significant depression of decorin mRNA levels compared to those in the sense ODN treated samples and the injection controls (p < 0.025). There was no significant difference between decorin mRNA levels in the sense ODN treated samples and the injection controls. Protein was extracted from MCL scar tissue, separated by SDS-PAGE, transblotted and assessed semiquantitatively using Western Blot analysis and an anti-decorin antibody (Dr. Paul Scott; University of Alberta; Edmonton, AB). Decorin protein levels were also signficantly depressed in the antisense treated scar tissue compared to the sense ODN treated samples (p = 0.045). Thus, the decorin antisense ODN treatment led to a persistant depression in decorin mRNA levels even at 4 weeks post-injection of the tissue and this was correlated with a depression in decorin protein levels in the scar tissue.

II. The model The rabbit bilateral MCL gap injury model has been well characterized over the past 15 years using multidisciplinary methodology (reviewed in Frank et al, 1983, 1994, 1999). In the present study, skeletally mature female animals had 4 mm gap injuries to their MCL surgically induced at T = 0. Two weeks later, each MCL was injected with either a decorin antisense ODN (10 uM ODN in HVJ-liposomes; 5'-GGA-TGA-GAG-TTG-CCG-TCA-TG3'), a decorin sense ODN (10 uM ODN in HVJ-liposomes; 5'-CAT-GAC-GGC-AAC-TCT-CAT-CC-3'), or was "poked" with a needle the same number of times as the antisense and sense treated animals (injection controls for reinjury to the scar tissue). The optimal sequence and concentration of ODN was determined in preliminary experiments with rabbit MCL scar cells in vitro and the delivery system was that reported previously (Nakamura et al, 1998b). Four weeks posttreatment (6 weeks post-injury), the animals were sacrificed and the scar tissue assessed using semiquantitative RT-PCR

IV. Effect of decorin antisense ODN on collagen fibril diameters in MCL scar tissue Collagen fibril diameters were assessed by transmission electron microscopy and image analysis as described previously (Cunningham et al, 1999; Frank et al 1992). Normal MCL tissue from skeletally mature animals express a bimodal distribution of both large and small diameter collagen fibrils (reviewed in Frank et al, 1992, 1994, 1999). MCL scar tissue expresses a unimodal population of small diameter collagen fibrils and these persist out to approximately 2 years post-injury when a few large diameter fibrils are again detected (Frank et al, 1992).

286

Gene Therapy and Molecular Biology Vol 4, Page 287

Figure 1. Distribution of collagen fibril diameters in MCL tissues. Fibril diameters were assessed by TEM.

In the sense ODN treated scars, as well as the injection control scars, only a unimodal population of small diameter collagen fibrils were detected (Figure 1). In contrast, some large diameter collagen fibrils were detected in 6 week scar tissue exposed to the decorin antisense ODN (F i g u r e 1 ). Collagen fibril diameter analysis demonstrated that the average fibril diameter in antisense treated scars, sense treated scars, injection controls and normal MCL tissue was, 104.7 +/- 51.1 (n = 20246), 74.8 +/- 11.0 (n = 4465), 78.2 +/11.9 (n = 4054) and 189.1 +/-104 (n = 2156), respectively. Approximately 38% of the fibrils in normal MCL tissue had diameters >125 nm, while 14% of the fibrils in the antisense treated scars had diameters >125 nm. In contrast, collagen fibril diameters > 125 nm were not detected in the sense ODN treated scars or the injection control scars. To determine whether there was a correlation between mean collagen diameter and mRNA levels for decorin, a correlation analysis was performed. There was a significant inverse correlation between the mean collagen fibril diameter and the expression level of decorin mRNA (p = 0.000073). Therefore, the antisense-mediated depression in decorin mRNA levels was significantly correlated with the observed changes in collagen fibril diameter.

287

V. Influence of decorin antisense ODN on the biomechanical properties of MCL scar tissue To evaluate the functional impact of the depression in decorin mRNA and the increase in large diameter collagen fibrils observed in the antisense treated scar tissue, ligament scar creep (low load behavior) (Thornton et al, in press) and stress at failure (high load behavior) (discussed in Frank et al, 1983, 1999) were assessed using established methods in an MTS material testing machine (MTS Systems, Inc.; Minneapolis, MN). Normal MCL tissue exhibits significantly less creep than does early MCL scar tissue (Thorton et al, in press). As shown in Figure 2, treatment of the MCL scars with decorin antisense ODN resulted in the tissue exhibiting significantly less creep than either the sense treated scars or the injection controls. Such a decrease in creep behavior could indicate an early scar tissue which would be more resistant to "stretching out" (ie: more normal ligament behavior). As this viscoelastic/creep behavior is likely related to matrix proteoglycans and tissue organization, these observed changes following antisense treatment may relate to decorin-mediated alterations in the mixture and organization of these important molecules. Furthermore, the antisense treated scars also exhibited a significantly higher stress at failure than did the sense treated scars or the injection controls (Figure 2). As this high load behavior is, at least partially, likely related to collagen

Hart et al: Antisense gene therapy for connective tissue engineering organization and collagen fibril diameters, this increase in high load behavior is consistant with the observed increase in the large diameter collagen fibrils observed in the antisense treated scars (discussed above). These results indicate that the antisense treatment led to a functional improvement in the MCL scar tissue at 6 weeks post-injury. Preliminary experiments assessing scar function at 14 weeks post-injury and 12 weeks post-antisense treatment have indicated that the scar tissue continues to be functionally better than the sense treated or the injection controls.

Administration of decorin to the lungs of bleomycin treated rats can inhibit development of TGF-beta mediated pulmonary fibrosis (Giri et al, 1997). Therefore, it was possible that decreasing decorin levels in the antisense treated scars led to increased levels of functional TGF-beta, and the observed functional effects were indirect and actually mediated by TGF-beta. It is known that TGF-beta mRNA levels are elevated in MCL scar tissue at 3 weeks post-injury (Sciore et al, 1998) and therefore, there is likely an increased supply of this growth factor in early scar tissue. To address this issue, two sets of experiments were performed. In the first, MCL scar tissue was exposed to 2.5 ng TGF-beta/day from day 0 to day 21 via implanted ALZET pumps. The purpose of this experiment was to assess the effect of TGF-beta on very early scar development. However, when the MCL tissues were assessed at 6 weeks post-injury, the TGF-beta exposure had no effect on either collagen fibril diameters or the biomechanical properties of the tissue (both low and high load behavior). In the second set of experiments, animals were exposed to a bolus of 7 Âľg of TGF-beta injected into the MCL scar tissue at 3 weeks post-injury and then assessed at 6 weeks post-injury. The rationale for this experiment was that increases in biologically active tissue TGF-beta levels would likely occur by 1 week post-exposure to the decorin antisense ODN and that this was, therefore, the interval when exogenous TGF-beta may also be seen to have a positive effect on healing of the tissue. However, when assessed at 6 weeks, there was again no detectable effect of the TGF-beta on either the biomechanical properties of the scar tissue or collagen fibril diameters.

VI. Influence of exogenous TGF-beta on MCL scar tissue The above decribed effects of decorin antisense therapy could be a direct influence of decorin on collagen fibril assembly, and subsequent alteration in tissue biomechanics. Alternatively, since decorin has a number of biological activities, the observed changes in collagen assembly could be an indirect effect of decreasing decorin levels in the scar tissue. Thus, a decrease in decorin levels could lead to a "cascade" effect in which the activity of other biologically important molecules were modified and the antisense therapy effect due to these molecules rather than decorin itself. One possibility in this regard is the growth factor transforming growth factor beta. As mentioned earlier, decorin binds TGF-beta and can functionally decrease the availability of this growth factor in tissues (Giri et al, 1997; Harper et al, 1984; Lysiak et al, 1995; Peters et al, 1997; Stander et al, 1999).

Figure 2. Creep strain (total and irrecoverable) (Panel A) and stress at failure (Panel B) for antisense, sense and injection control tibia-MCL scar-femur complexes. All scars failed in the scar tissue.

288

Gene Therapy and Molecular Biology Vol 4, Page 289 From these experiments one can conclude that the influence of decorin antisense on MCL scar tissue is apparently not an indirect effect via TGF-beta. Furthermore, the experiments also indicate that endogenous levels of TGFbeta in the MCL scar tissue of healthy skeletally mature adult rabbits are optimal for its role in extraarticular ligament healing.

VII. Current studies While the studies discussed above demonstrate that decorin antisense gene therapy can functionaly impact on MCL scar tissue at multiple levels, the outcome still remains a tissue with less than normal properties and characteristics. Therefore, additional experimentation is required to continue development of a regenerated tissue. This will require assessment of additional molecular targets for the gene therapy and possibly, development of an antisense "cocktail" consisting of antisense ODN directed towards a number of specific targets [ie: matrix molecules (proteoglycans, collagens), proteinases and inhibitors, regulatory enzymes) or pleiotropic targets (ie: growth factors, transcription factors and other intracellular regulators) that can influence a number of cellular molecules. We are currently evaluating a number of possibilities (minor collagens, specific proteinases, inflammation regulated transcription factors, etc). Obviously, the timing for delivery of such a combination of modifiers of multiple cellular targets will have to be optimized. A second area where new developments are required is that of the delivery system. The present delivery system is functional and leads to prolonged alterations in the transfected cells, but it has limitations with regard to the uniformity of liposome delivery to all relevant cells in the tissue and clinical utility (an important consideration if the methodology is to be translated to clinical practice). While early scar tissue is more cellular (Frank et al, 1983) and more vascularized (discussed in Frank et al, 1994) than normal tissue, delivery via the regional vascularity is limited (Nakamura et al, 1998b). Therefore, new methods will be required to effect delivery in a clinically acceptable manner. A critical element is the diffusion of the liposomes within the scar matrix. We are currently exploring additional methods which overcome some of these limitations. As such limitations are not restricted to MCL scars, improvements in this area will likely influence application to other connective tissues.

VIII. Future directions In the clinical realm, MCL healing is not as critical as healing/repair of other ligaments such as the ACL, or other connective tissues such as tendons and menisci. A question of critical importance is whether the effectiveness of antisense gene therapy in the MCL model can be extrapolated 289

to these other connective tissues. Some studies have indicated that this liposome approach can be used in models of patellar tendon healing (Nakamura et al, 1996, 1998a) and can be used to affect gene transfer to cartilage (Tomita et al, 1997). Therefore, it is likely that the approach and methodology will have application in these tissues, as well. However, it remains to be determined whether generalizations can be made for healing/regeneration of most connective tissues or whether each tissue will have individual, or specific requirements in this regard because of unique mechanical environments, unique cellular aspects, or tissuespecific biochemical requirements.

Acknowledgements The authors thank Judy Crawford for excellent secretarial assistance in the preparation of the manuscript. The studies reported were supported by grants from The Arthritis Society (Canada), the Medical Research Council of Canada, and the London Life Insurance Company. DAH is the Calgary Foundation-Grace Glaum Professor in Arthritis Research, NN and RB were supported by Fellowships from the Alberta Heritage Foundation for Medical Research (AHFMR), NGS is a Killam Professor at the University of Calgary, and CBF is an AHFMR Scientist and the McCaig Professor in Joint Injury Research.

References Boykiw R, Sciore P, Reno C, Marchuk L, Frank CB, and Hart DA ( 1 9 9 8 ) Altered levels of extracellular matrix molecule mRNA in healing rabbit ligaments. M a t r i x B i o l 17, 371378. Cunningham KD, Musani F, Hart DA, Shrive NG, and Frank CB (1 9 9 9 ) Collagenase degradation decreases collagen fibril diameters-an in vitro study of the rabbit medial collateral ligament. C o n n e c t i v e T i s s u e R e s 40, 67-74. Frank CB (1 9 9 6 ) Ligament healing: current knowledge and clinical applications. J Am Acad Orthop Surg 4, 74-83. Frank CB, Bray RB, Hart DA, Shrive NG, Loitz BJ, Matyas JR, Wilson JE ( 1 9 9 4 ) Soft tissue healing. In: Knee Surgery. Fu FH, Harner CD, Vince KG (eds), Williams and Wilkins, Baltimore, pp. 189-229. Frank CB, Hart DA, and Shrive NG (1 9 9 9 ) Molecular biology and biomechanics of normal and healing ligaments-a review. O s t e o a r t h r i t i s C a r t i l a g e 7, 130-140. Frank CB, McDonald D, Bray RB, Rangayyan R, Chimich D, and Shrive N (1 9 9 2 ) Collagen fibril diameters in the healing adult rabbit medial collateral ligament. C o n n e c t i v e T i s s u e R e s 27, 251-263. Frank CB, McDonald D, and Shrive NG (1 9 9 7 ) Collagen fibril diameters in the rabbit medial collateral ligament scar: a longer term assessment. C o n n e c t i v e Tissue R e s 36, 261-269. Frank CB, Woo SLY, Amiel D, Harwood F, Gomez M, and Akeson W (1 9 8 3 ) Medial collateral ligament healing: a multidisciplinary study in rabbits. A m J S p o r t s M e d 11, 379-389.

Hart et al: Antisense gene therapy for connective tissue engineering Giri SN, Hyde DM, Braun RK, Gaarde W, Harper JR, and Pierschbacher MD ( 1 9 9 7 ) Antifibrotic effect of decorin in a bleomycin hamster model of lung fibrosis. B i o c h e m Pharmacol 54, 1205-1216. Harper JR, Spiro RC, Gaarde WA, Tamura RN, Pierschbacher MD, Noble NA, Stecker KK, and Border WA ( 1 9 9 4 ) Role of transforming growth factor beta and decorin in controlling fibrosis. M e t h o d s E n z y m o l 245, 241-254. Lysiak JJ, Hunt J, Pringle GA, Lala PK ( 1 9 9 5 ) Localization of transforming growth factor beta and its natural inhibitor decorin in the human placenta and decidua throughout gestation. Placenta 16, 221-231. Murphy PG, Frank CB, and Hart DA (1 9 9 3 ) Cell biology of ligament and ligament healing. In: The Anterior Cr uc i a t e L i g a me n t. Jackson D, Arnoczky S, Woo S, Frank C, and Simon T (eds), New York, Raven Press, pp. 165-177. Nakamura N, Frank CB, and Hart DA (1 9 9 8 ) Gene therapy in joint repair. Curr Opin Orthop 9, 25-30. Nakamura N, Horibe S, Matsumoto N, Tomita T, Natsuume T, Kaneda Y, Shino K, and Ochi T (1 9 9 6 ) Transient introduction of a foreign gene into healing rat patellar ligament. J C l i n I n v e s t 97, 226-231. Nakamura N, Shino K, Natsuume T, Horibe S, Matsumoto N, Kaneda Y, and Ochi T (1 9 9 8 ) Early biological effect of in vivo gene transfer of platelet-derived growth factor (PDGF)B into healing patellar ligament. Gene Therapy 5, 11651170. Nakamura N, Timmerman SA, Hart DA, Kaneda Y, Shrive NG, Shino K, Ochi T, and Frank CB ( 1 9 9 8 ) A comparison of in vivo gene delivery methods for antisense therapy in ligament healing. Gene Therapy 5, 1455-1461. Peters H, Noble NA, and Border WA ( 1 9 9 7 ) Transforming growth factor-beta in human glomerular injury. Curr Opin Nephrol Hypertens 6, 389-393. Reno C, Boykiw R, Martinez ML, and Hart DA (1 9 9 8 ) Temporal alterations in mRNA levels for proteinases and inhibitors and their potential regulators in the healing medial collateral ligament. B i o c h e m B i o p h y s R e s C o m m u n 252, 757763. Reno C, Marchuk L, Sciore P, Frank CB, and Hart DA ( 1 9 9 7 ) Rapid isolation of total RNA from small samples of hypocellular, dense connective tissues. B i o T e c h n i q u e s 22, 1082-1086. Sciore P, Boykiw R, and Hart DA ( 1 9 9 8 ) Semiquantitative reverse transcription-polymerase chain reaction analysis of mRNA for growth factors and growth factor receptors from normal and healing rabbit medial collateral ligament tissue. J Orthop Res 16, 229-237. Stander M, Naumann U, Wick W, Weller M (1 9 9 9 ) Transforming growth factor-beta and p21: multiple molecular targets of decorin-mediated suppression of neoplastic growth. C e l l T i s s u e R e s 296, 221-227. Thornton GM, Leask GP, Shrive NG, and Frank CB (1 9 9 9 ) Ligament scars creep more than control ligaments in early healing: an in vitro study of the rabbit medial collateral ligament. In press. Tomita T, Hashimoto H, Tomita N, Morshita R, Lee SB, Hayashida K, Nakamura N, Yonenobu K, Kaneda Y, and Ochi T (1 9 9 7 ) In vivo direct gene transfer into articular cartilage by intraarticular injection mediated by HVJ (Sendai virus) and liposomes. Arthritis Rheum 40, 901-906.

Vogel KG, Paulsson M, and Heinegard D (1 9 8 4 ) Specific inhibition of type I and type II collagen fibrillogenesis by the small proteoglycan from tendon. B i o c h e m J 223, 587597.

David A. Hart (At the 1999 Conference in Crete)

290

Gene Therapy and Molecular Biology Volume 4, page 291 Gene Ther Mol Biol Vol 4, 291-296. December 1999.

3-aminobenzamide: a novel drug to induce in vivo DNA hypermethylation Review Article

Giuseppe Zardo1, Anna Reale2, Mariagrazia Perilli1, Adriana de Capoa3 and Paola Caiafa1

Department of Biomedical Sciences and Technologies1, University of L’Aquila, Italy. Department of Cellular Biotechnologies and Haematology2 and of Genetics and Molecular Biology3, University of Rome “La Sapienza”, Italy. ______________________________________________________________________________________________________ Correspondence: Prof. Paola Caiafa, Dipartimento di Biotecnologie Cellulari ed Ematologia, Sezione di Biochimica Clinica, Facoltà di Medicina e Chirurgia, Università di Roma "La Sapienza", Viale Regina Elena, 324 (Policlinico), 00161, Roma, Italia. Tel: +39-06-49910900, Fax: +39-06 4440062, E-mail caiafa@bce.med.uniroma1.it Key Words: histone hypoacetylation, gene silencing, histone deacetylation, 5-azacytidine, trichostatin A, chromatin Received: 9 August 1999; accepted: 19 August 1999

Summary Both DNA methylation and core histone hypoacetylation are associated with gene silencing but only recent experiments allowed the interlocking of these two processes. Through such experiments it was shown that the two processes are united in inducing gene silencing through a “shuttle-system” involving the methyl CpG binding protein (MeCP2). In this scenario, it is not clear whether DNA methylation or histone deacetylation is the leader in inducing down regulation of gene expression. Trichostatin A (TSA), a potent inhibitor of histone deacetylase, is usually used to clarify this point. As far as DNA methylation is concerned, only the 5-azacytidine (5-AzaC), able to induce hypomethylation, has been described until now. The aim of this paper is to suggest the use of 3-aminobenzamide (3ABA) as a method capable of inducing in vivo DNA hypermethylation, so that new experiments could be performed in both directions to clarify the chronology by which the influence on gene expression takes place and to pinpoint the structure of methylated condensed chromatin.

housekeeping genes - termed CpG islands (Bird et al, 1985; Bird, 1986; Bird, 1987) - are present in their unmethylated state, this condition being essential for the expression of related genes (Keshet et al, 1985).

I. Introduction DNA methyltransferase (EC 2.1.1.37) is a nuclear enzyme that, by transferring methyl groups from S-adenosyl methionine (S-AdoMet) to cytosine (C) converts these residues into 5-methylcytosine (5mC) (Bestor and Ingram, 1983), the best substrate being the cytosine located in the CpG dinucleotide (Gruenbaum et al, 1981). This epigenetic modification is proposed to have an active role in the modulation of gene expression (Keshet et al, 1985; Boyes and Bird, 1992; Li et al, 1993; Hsiet, 1994). This role was confirmed by experiments in which an anomalous methylation, caused by targeted mutation of DNA methyltransferase gene in mice, results in embryonic lethality (Li et al, 1992). The DNA methylation pattern, which is defined during embryonic development (Brandeis et al, 1993), is very important since its characteristic is that some DNA regions, located in the 5’ promoter region of

A second mechanism by which DNA methylation may be involved in down regulation of gene expression has recently been shown (Jones et al, 1998; Nan, et al, 1998) and debated (Bestor, 1998; Razin, 1998). This mechanism foresees that histone deacetylase, via its association with methyl-CpG binding protein (MeCP2) (Boyes and Bird, 1991; Meehan et al, 1992) reaches methylated DNA allowing the methylation-dependent chromatin condensation favoring gene silencing. Until now it is not clear how cytosine methylation might affect chromatin structure and much still has to be done to clarify the mechanism by which this influence takes place and to identify whether DNA methylation or histone deacetylation is the post-synthetic modification “leader” in inducing gene silencing (Selker, 291

Zardo et al: 3-aminobenzamide induces in vivo DNA hypermethylation 1998; Eden et al, 1998; Cameron et al, 1999). Trichostatin A (TSA), a potent inhibitor of histone deacetylase (Yoshida et al, 1995), is usually used to clarify this point. As far as DNA methylation is concerned, only the treatment of cells with 5-azacytidine (5-AzaC), able to induce hypomethylation, has been described until now (Adams and Burdon, 1985). The aim of this paper is to propose the treatment of cells with 3-aminobenzamide as a method to induce in vivo DNA hypermethylation so that new experiments can be performed in two directions in order to both clarify the order by which the influence on gene expression takes place and to pinpoint the structure of methylated condensed chromatin.

II. Treatment of cells with 3aminobenzamide induces in vivo DNA hypermethylation

Figure 1. Methyl-accepting ability experiment. The endogenous methyl-accepting ability of native nuclei, obtained from 6.5 x 106 L929 fibroblasts preincubated for 24 h without (control) and with 8 mM 3-ABA, was performed in the presence of 16 µM [3H]-SAdoMet. The level of methyl groups has been evaluated on the total DNA purified from cells Control DNA, whose incorporation was 2.8±0.1 pmol of [3H]-S-AdoMet, was considered as 100%. Zardo et al. (1997) Biochemistry 36, 7937-7943.

The 3-aminobenzamide is a specific inhibitor (Huletsky et al, 1989; Rankin et al, 1989) of the poly(ADP-ribose) polymerase (EC 2.4.2.30), an enzyme able to build and/or transfer ADP-ribose polymers onto chromatin proteins (Jacobson and Jacobson, 1989; de Murcia, et al, 1995). The statement that following inhibition of poly(ADPribose) polymerase DNA methyltransferase becomes able to methylate the unmethylable cytosines on DNA is based on our experiments showing that a block of poly(ADPribosyl)ation introduces an anomalous hypermethylated pattern in genomic DNA (Zardo et al, 1997; de Capoa et al, 1999). Although our research was performed in order to demonstrate that poly(ADP-ribosyl)ation is an important process involved in controlling the expression of housekeeping genes (Zardo and Caiafa, 1998), the aim of this review is to point out that the 3-ABA induced block of this enzymatic process introduces an anomalous hypermethylated pattern on genomic DNA. All experiments were carried out on L929 and NIH\3T3 mouse fibroblast cells and poly(ADP-ribose) polymerase was inhibited by treatment of cells with 2 and/or 8 mM 3-aminobenzamide for 24 hours.

More recently (de Capoa et al, 1999) we have been able to show that during the 24 hours of 3-ABA treatment, interphase nuclei had already incorporated some methyl groups. The cells were indirectly immunolabeled with anti5-methylcytosine (anti-5mC) monoclonal antibodies (de Capoa et al, 1996), microscope analysis was performed on a cell-by-cell basis and the images of the nuclei were recorded by a b/w CCD camera. A computer-assisted quantitative analysis of the methylation state of individual interphase nuclei was performed by dedicated software (de Capoa et al, 1998). Cells preincubated with 3-ABA consistently showed increased levels of anti-5mC antibody binding to heterochromatic regions, Figures 2 and 3. Thus, both the DNA methyl-accepting assay and monoclonal anti 5-methylcytosine antibodies allowed us to show that reduced levels of poly(ADP-ribosyl)ation result in DNA hypermethylation.

The first evidence came from experiments on endogenous DNA methyl-accepting ability (Zardo et al., 1997). In these experiments the methyl-accepting ability of isolated nuclei, obtained from 6.5 x 105 L929 mouse fibroblasts, previously preincubated with or without 3-ABA for 24 hours, was performed in the presence of [3H]-SAdoMet. After one hour of incubation at 37° C, we compared the ability to incorporate labeled methyl groups in their DNA in the absence of any exogenous DNA methyltransferase (i.e. by a process catalyzed by the endogenous enzyme). The level of incorporated methyl groups, evaluated on the total DNA purified from cells, was found to be 60% higher in the DNA from 3-ABA treated cells than in DNA from control cells whose methylacceptance value was taken as 100%, Figure 1.

III. Possible interpretation of the mechanism by which poly(ADPribosyl)ation controls DNA methylation These results indicate that poly(ADP-ribosyl)ation protects in some way genomic DNA from full methylation although much still has to be done to explain the molecular mechanism(s). As for the CpG islands, our recent research (Zardo and Caiafa, 1998) has shown that, at least for the Htf9 promoter region, the inhibition of poly(ADPribosyl)ation allows the new methyl groups to position 292

Gene Therapy and Molecular Biology Vol 4, page 293 themselves on DNA. Further experiments have shown that this inhibition also changes the methylation pattern of plasmid transfected in its unmethylated form (Zardo et al, 1999 in press).

Malanga et al, 1998) can be modified both in a covalent and non-covalent way, the best substrate being H1 histone (Poirier and Savard, 1980; D`Erme et al, 1996; Panzeter et al, 1992; Panzeter et al, 1993; Malanga et al, 1998).

Poly(ADP-ribose) polymerase that is dimeric in its catalytic form (Mendoza-Alvarez and Alvarez-Gonzales, 1993), has three domains which play specific roles in the poly(ADP-ribosyl)ation process. The N-terminal domain contains the zinc-finger motifs which are responsible for binding to DNA (Gradwohl et al, 1990), the C-terminal domain contains the catalytic site (de Murcia and Menissier de Murcia, 1994; Rolli et al, 1997) and the central domain is the domain that undergoes automodification (Desmarais et al, 1991). The enzyme starts its automodification following binding of the enzyme to DNA and needs breaks on DNA strands to be activated (de Murcia and Menissier de Murcia, 1994). During the automodification process the ADP-ribose polymers - up to 200 residues – are built in the 28 automodification sites (Kawaichi et al, 1981; Desmarais et al, 1991) located in this domain. Following automodification, the enzyme can start heteromodification reactions allowing interactions between ADP-ribose polymers and chromatin proteins (Boulikas, 1989; Scovassi et al, 1993). Several proteins (Wesierska-Gadek et al, 1996;

Our in vitro findings show that the H1 histone, in its poly(ADP-ribosyl)ated isoform (Zardo et al, 1997) and through its genic variant H1e (Santoro et al, 1995; Zardo et al, 1996) could be a nuclear proteic trans-acting factor involved in maintaining the unmethylated state of CpG islands. We cannot exclude that other presently unknown proteic factor(s) could also play a regulatory role in the control of DNA methylation by means of this post-synthetic modification. Acknowledgements This work was supported by the Italian Ministry of University and Scientific and Technological Research (40% Progetti di Interesse Nazionale, 60% Ricerca Scientifica Università di L’Aquila e di Roma, “La Sapienza”) and by the Consiglio Nazionale delle Ricerche (CNR). We thank Alessandra Spanò for technical assistance.

Figure 2. Increased levels of heterochromatin methylation in 3-ABA treated nuclei from mouse fibroblast cell lines as shown by indirect immunolabeling with anti-5MeC antibodies. de Capoa et al. (1999) The FASEB J. 13, 89-93; b/w CCD camera images of control (a,b) and treated (c,d) nuclei from L929 (left) and NIH/3T3 cell lines (right).

293

Zardo et al: 3-aminobenzamide induces in vivo DNA hypermethylation

Figure 3. Methylation levels of control and 3-ABA treated mouse fibroblasts in samples of 20 nuclei each. Left: L929 and, right: NIH/3T3 cells. Upper rows: Examples of pseudocolored heterochromatic regions in some control (a, b) and treated (c, d) nuclei. A computer-assisted quantitative analysis of methylation levels in b/w CCD camera images of control and treated nuclei was performed. For each nucleus the methylation state was expressed as area of the methylated regions (Âľ2) and different levels of optical densities (ODs, 0-185 in the gray scale range). Blue, yellow and red staining indicate, respectively, the heavily, medium and lightly methylated regions per nucleus. Lower rows: Percentages of differentially labeled areas and different optical densities in control and treated nuclei from each sample. Blue, yellow and red staining indicate, respectively, the heavily, medium and lightly methylated regions per sample.

Sci USA 80, 5559-5563.

References

Bird AP (1986). CpG islands and the function of DNA methylation. Nature 321, 209-213.

Adams RLP and Burdon RH (1985). Molecular biology of DNA methylation. Springer Series in Molecular Biology, (Rich, A. ed), Springer-Verlag, New York.

Bird AP (1987). CpG islands as a gene markers in the vertebrate nucleus. Trends Genet 3, 324-347.

Bestor TH (1998). Methylation meets acetylation. Nature 393, 311-312.

Bird AP, Taggard M, Frommer M, Miller OJ and Macleod D (1985). A fraction of the mouse genome that is derived from islands of nonmethylated CpG-rich DNA. Cell 40, 91-99.

Bestor TH and Ingram VM (1983). Two DNA methyltransferases from murine erythroleukemia cells: purification, sequence specificity and mode of interaction with DNA. Proc Natl Acad

Boyes J and Bird AP (1991). DNA methylation inhibits

294

Gene Therapy and Molecular Biology Vol 4, page 295 transcription indirectly via a methyl-CpG binding protein. Cell 64, 1123-1134. Boyes J and Bird AP (1992). Repression of genes by DNA methylation depends on CpG density and promoter strength: evidence for involvement of a methyl CpG binding protein. EMBO J 11, 327-333.

Kawaich, M, Ueda K and Hayaishi O (1981). Multiple autopoly(ADP-ribosyl)ation of rat liver poly(ADP-ribose syntetase. J Biol Chem 256, 9483-9489. Keshet I, Yisraeli J and Cedar H (1985). Effect of regional DNA methylation on gene expression. Proc Natl Acad Sci USA 82, 2560-2564.

Brandeis M, Ariel M and Cedar H (1993). Dynamics of DNA methylation during development. BioEssays 15, 709-713.

Li E, Beard C and Jaenisch R (1993). Role of DNA methylation in genomic imprinting. Nature 366, 362-365.

Cameron EC, Bachman KE, Myohanen S, Herman JG and Baylin SB (1999). Synergy of demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer. Nature Gen 21, 103-107.

Li E, Bestor TH and Jaenisch R (1992). Targeted mutation of DNA methyltransferase gene results in embryonic lethality. Cell 69, 915-926.

D’Erme M, Zardo G, Reale, A and Caiafa P (1996). Co-operative interactions of oligonucleosomal DNA with the H1e histone variant and its poly(ADP-ribosyl)ated isoform. Biochem J 316, 475-480. de Capoa A, Febbo FR, Giovannelli F, Niveleau A, Zardo, G, Marenzi S and Caiafa P (1999). Reduced levels of poly(ADPribosyl)ation result in chromatin compaction and hypermethylation as shown by cell-by-cell computer-assisted quantitative analysis. The FASEB J 13, 89-93. de Capoa A, Menendez F, Poggesi I, Giancotti P, Grappelli C, Marotta MR, Di Leandro M, Reynaud C and Niveleau A (1996). Cytological evidence for 5-azacitidine-induced demethylation of the heterochromatic regions of human chromosomes. Chromosome Res 4, 271-276 de Capoa A, Menendez F, Poggesi I, Giancotti P, Grappelli C, Marotta MR, Di Leandro MR, Spano A, Rocchi M, Archidiacono N and Niveleau A (1998). Computer-assisted analysis of methylation status of individual interphase nuclei in human cultured cells. Cytometry 31, 85-92. de Murcia G and Mènissier de Murcia J (1994). Poly(ADP-ribose) polymerase: a molecular nick sensor. Trends Biochem Sci 19, 172-176. Desmarais Y, Menard L, Lagneux J and Poirier GG (1991). Enzymological properties of poly(ADP-ribose) polymerase: characterization of automodification sites and NADase activity. Biochim Biophys Acta 1078, 179-186. Eden S, Hashimshony I, Keshet I and Cedar H (1998). Nature 394, 842. Gradwohl G, de Murcia JM, Molinete M, Simonis F, Koken M, Hoeijmakers JH and de Murcia G (1990). The second zincfinger domain of poly(ADP-ribose) polymerase determines specificity for single stranded breaks in DNA. Proc Natl Acad Sci 87, 2990-2994. Gruenbaum Y, Stein R, Cedar H. and Razin A (1981). Methylation of CpG sequences in eukaryotic DNA. FEBS Lett 124, 67-71. Hsiet CL (1994). Dependance of transcriptional repression on CpG methylation density. Mol Cell Biol 14, 5487-5494. Huletsky A, de Murcia G, Muller S, Hengartner M, Menard L, Lamarre D and Poirier GG (1989). The effect of poly(ADPribosyl)ation on native and H1-depleted chromatin. J Biol Chem 15, 8878-8886. Jones PL, Veenstra GJC, Wade PA, Vermaak D, Kass SU, Landsberger N, Stouboulis J and Wolffe AP (1998). Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nature Gen 19, 187-191.

295

Malanga M, Pleschke JM, Kleczkowska HE and Althaus F (1998). Poly(ADP-ribose) binds to specific domains of p53 and alters its DNA binding functions. J Biol Chem 273, 11839-11843. Malanga M, Atorino L, Tramontano F, Farina B and Quesada P (1998). Poly(ADP-ribose) binding properties of histone H1 variants. Biochim Biophys Acta 1399, 154-160 Meehan RR, Lewis JD and Bird AP (1992). Characterization of MeCP2, a vertebrate DNA binding protein with affinity for methylated DNA. Nucl Acids Res 20, 5085-5092. Mendoza-Alvarez H and Alvarez-Gonzales R (1993). Poly(ADPribose) polymerase is a catalytic dimer and the automodification reaction is intermolecular. J Biol Chem 268, 22575-22580. Nan X, Ng H-H, Johnson CA, Laherty CD, Turner BM, Eisenman RN and Bird AP (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone deacetylase complex. Nature 393, 386-389. Panzeter PL, Realini CA and Althaus FR (1992). Noncovalent interactions of poly(adenosine diphosphate ribose) with histones. Biochemistry 31, 1379-1385. Panzeter PL, Zweifel B, Malanga M, Waser SH, Richard MC and Althaus F (1993). Targeting of histone tails by poly(ADPribose)polymerase. J Biol Chem 268, 17662-17664. Poirier GG and Savard P (1980). ADP-ribosylation of pancreatic histone H1 and other histones. Can J Biochem 58, 509-515. Rankin PW, Jacobson EL, Benjamin RC, Moss J and Jacobson MK (1989). Quantitative studies of inhibitors of ADP-ribosylation in vitro and in vivo. J Biol Chem 264, 4312-4317. Razin A (1998). CpG methylation, chromatin structure and gene silencing a three-way connection. The EMBO J 17, 49054908. Rolli V, O’Farrel M, Ménissier-de Murcia J and de Murcia G (1997). Random mutagenesis of poly(ADP-ribose) polymerase catalytic domain reveals amino acids involved in polymer branching. Biochemistry 36, 12147-12154. Santoro R, D’Erme M, Mastrantonio S, Reale A, Marenzi S, Saluz HP, Strom R and Caiafa P (1995). Binding of histone variants H1e-c to CpG-rich DNA correlates with the inhibitory effect on enzymic DNA methylation. Biochem J 305, 739-744. Selker EU (1998). Trichostatin A causes selective loss of DNA methylation in Neurospora. Proc Natl Acad Sci USA 95, 9430-9435. Wesierska-Gadek J, Schmid G and Cerni C (1996) ADPRibosylation of wild-type p53 protein to specific p53 consensus sequence prevents its modification. Biochem Biophys Res Commun 224, 96-102.

295

Zardo et al: 3-aminobenzamide induces in vivo DNA hypermethylation Yoshida M, Horinouchi S and Beppu T (1995). Tricostatin A and tropoxin: novel chemical probe for the role of histone acetylation in chromatin structure and function. BioEssays 17, 423-430. Zardo G, Dâ&#x20AC;&#x2122;Erme M, Reale A, Strom R, Perilli M and Caiafa P (1997). Does poly(ADP-ribosyl)ation regulate the DNA methylation pattern? Biochemistry 36, 7937-7943. Zardo G and Caiafa P (1998). The unmethylated state of CpG islands in mouse fibroblasts depends on the poly(ADPribosyl)ation process. J Biol Chem 273, 16517-16520. Zardo G, Marenzi S and Caiafa P (1998). H1 histone as a transacting factor involved in protecting genomic DNA from full methylation. Biol Chem 379, 647-654. Zardo G, Marenzi S, Perilli S and Caiafa P (1999). Inhibition of poly(ADP-ribosyl)ation introduces an anomalous methylation pattern in transfected foreign DNA. The FASEB J 13, 15181522. Zardo G, Santoro R, D`Erme M, Reale A, Guidobaldi L, Caiafa P and Strom R (1996) Specific inhibitory effect of H1e histone somatic variant on in vitro DNA methylation process. Biochim Biophys Res Comm 220, 102-107.

(From left) Anna Reale, Paola Caiafa & Giuseppe Zardo

296

Gene Therapy and Molecular Biology Vol 4, page 297 Gene Ther Mol Biol Vol 4, 297-301. December 1999.

Mechanically stretching single chromatin fibers Research Article

Sanford H. Leuba2, Mikhail A. Karymov2, Yanzhang Liu3, Stuart M. Lindsay3 and Jordanka Zlatanova1* 1

Argonne National Laboratory, Argonne, IL 60439. National Cancer Institute, NIH, Bethesda, MD 20892. 3 Department of Physics and Astronomy, Arizona State University, Tempe, AZ 85287. ___________________________________________________________________________________________________ 2

* Correspondence: Prof. Jordanka Zlatanova, Biochip Technology Center, Argonne National Laboratory, Argonne, IL 60439, USA, Tel: (630) 252 7860; Fax: (630) 252 3947; E-mail: zlatanoj@everest.bim.anl.gov Key words: Atomic Force Microscopy, chromatin, histone Received: 30 June 1999; accepted: 13 July 1999

Summary We have used the recently developed MAC Mode Atomic Force Microscope (AFM) that operates in aqueous solution to mechanically stretch single chicken erythrocyte chromatin fibers. The fibers contained the full complement of histones, or, alternatively, were depleted of linker histones. The AFM was used to produce the socalled force curves, by monitoring the cantilever deflection (proportional to force) as the distance between the AFM tip and the sample was experimentally manipulated. To that end, the AFM tip was pushed into the chromatin sample and then withdrawn, to mechanically stretch the fiber that was physically adsorbed to the tip. Pulling of the chromatin fiber produced complex sawtooth-like patterns of peaks that were characterized by unexpectedly large forces, in the range of several hundred picoNewtons. The distribution of forces in the linker histone-containing and linker histone-depleted fibers was remarkably different, possibly indicating that linker histone binding significantly changes the mechanical properties of the chromatin fiber. may be considered as a molecular motor creating forces to help evict the histone octamer from the DNA. In order to gain insight into the forces that maintain the integrity of chromatin structure, we made an attempt at directly measuring these forces with the help of the Atomic Force Microscope (AFM). Chromatin fibers were physically adsorbed onto glass coverslips, the AFM tip was pushed into the chromatin sample, and then pulled away. The deflection of the cantilever was then monitored as a function of its distance from the surface, producing the socalled force-extension curves. The cantilever deflection can be used to directly determine forces by multiplying its magnitude by the spring constant of the cantilever (Hooke's law). Such curves have been widely used in recent years to study intermolecular interaction forces between pairs of interacting molecular partners (Moy et al, 1994; Florin et al, 1994; Lee et al, 1994a, b; Dammer et al, 1996; Allen et al, 1997). In such curves, the interaction between the molecular partners (e.g. antigen-antibody) is revealed as hysteresis between the approach and retraction curve (see Figure 1 , for a schematic of such curves). The magnitude of tip deflection in the so-called adhesion peak

I. Introduction and theoretical background In the eukaryotic nucleus, processes such as transcription, replication, and repair that use DNA as a template take place in the context of chromatin. The machineries performing these reactions have to gain access to the DNA by initially disrupting the higher-order structure of the chromatin fiber (van Holde and Zlatanova, 1996), and by further moving the nucleosomes out of the way of the enzymes. The mechanisms involved in rendering the DNA accessible are not well understood. The notion that the histone octamers that form the protein core of the nucleosomal particles are removed from the DNA by an applied tension-based mechanism has attracted considerable attention, especially after the discovery that movement of the polymerase along the DNA during transcription by itself can create the necessary tension. It has been suggested, and later proven experimentally, that transcription is accompanied by the formation of relatively high levels of positive supercoiling in front, and negative supercoiling in the wake of the passing polymerase (Liu and Wang, 1987). In such a scenario, the polymerase itself 297

Leuba et al: Atomic Force Microscopy of chromatin fibers is taken as an indication of the magnitude of interaction rupture forces. More recently, force curves have been used to determine the forces maintaining the native conformation of single-chain biological polymers, such as multi-domain proteins (Rief et al, 1997a; Oberhauser et al, 1998) or polysaccharides (Rief et al, 1997b; Marszalek et al, 1998). Stretching of multi-domain proteins produces curves with multiple, sawtooth-like discontinuities thought to reflect unfolding of individual folded domains. Such discontinuities occur at points during stretching at which the cantilever restoring force F exceeds the molecular interaction forces that are responsible for the integrity of the domain structure. Thus, the magnitude of the vertical jumps in the force curves can be used to estimate interaction forces.

(Figure 2a) and in buffer (Figure 2b). In both cases, the fibers had an irregular three-dimensional appearance, similar to the previously published tapping mode images (Leuba et al, 1994). Typically, the force curves taken on such material were rather complex, exhibiting multiple snap-off peaks separated from each other at certain intervals (Figure 3). Similar experiments were also performed on linker histone-depleted fibers. Such fibers are known to lose their three-dimensional organization when imaged in the AFM (Leuba et al, 1994; Yang et al, 1994), and resemble artificial chromatin constructs containing only a piece of DNA to which core histones were added (Allen et al, 1993). The linker histone-depleted fibers also produced multi-peak curves, superficially similar to those obtained on 'native' fibers. Figure 4 presents the distribution of forces measured from large sets of force curves. As can be seen, the forces in the linker histone-depleted fibers are rather broadly distributed, with a mean ~ 400 pN. In contrast, the native fibers had a narrower distribution, centered around 100 pN. In order to analyze the complex force curves obtained on chromatin fibers, we switched to artificial chromatin constructs that are of a known length and contain a known number of nucleosomes (on average), more or less regularly spaced along the DNA length. Such artificial constructs were obtained by in vitro reconstitution of tandemly repeated cloned fragments of the 5S rRNA gene (twelve repeats of a 208 bp sequence) from the sea urchin Lytechinus variegatus with purified core histones.

II. Imaging and stretching chromatin fibers A. Native or linker histone-depleted chicken erythrocyte chromatin fibers In our analysis of chromatin, we first imaged the fibers and then monitored the force-extension curves, in low ionic strength buffers that favor the extended fiber conformation (Zlatanova et al, 1998). Imaging was important in two aspects: first, we could be sure, on the basis of the density of the material on the surface, that in each individual pulling event we were stretching single fibers; second, the morphology observed and the quantitative measurements done on imaged fibers helped us to interpret the results. Chicken erythrocyte chromatin fibers containing all histones were imaged both in air

Figure 1. Schematic of a typical contact mode force curve.

298

Gene Therapy and Molecular Biology Vol 4, page 299

Figure 2. Chicken erythrocyte chromatin fibers imaged in air (A) and low ionic strength buffer (5 mM triethanolamine, pH 7.0, 0.1 mM EDTA) (B). (B) is adapted from Leuba and Bustamante, 1999. The height of features above the surface is denoted in different shades of gray, in a range from 0 to 15 nm, with the higher features being lighter. Scan size 500 nm x 500 nm. Chicken erythrocyte chromatin fibers were obtained as described in Leuba et al. (1994).

images taken in air were close to the numbers expected from the known loading of the fibers with nucleosomes; in contrast, the number of nucleosomes in buffer images was always much less than the known number on the fibers, indicating that the attachment of the fiber to the surface in liquid did not involve the entire length of the fiber).

This nucleosomal array, first introduced in the laboratory of R. Simpson (Simpson et al, 1985), is widely used in the chromatin structure field in view of the regularity of its structure. For these constructs, the force distribution was similar to that of the linker histonedepleted chromatin fiber (Leuba, Zlatanova, Karymov, Liu, R. Bash, D. Lohr, R. Harrington, Lindsay, in preparation). The relative simplicity of this system, well characterized in biochemical terms, will hopefully help us interpret the chromatin data presented here.

III. Analysis of possible structural transitions in chromatin fibers during mechanical stretching How do we interpret these results? What are the events occurring during stretching of the chromatin fibers that give rise to the sawtooth pattern of peaks? While a detailed interpretation is not possible at this point, the pulling curves clearly reflect nucleosomal-related events. Such an assertion is based on two main observations: (i) When naked DNA was deposited and pulled under the same experimental conditions, no peaks were observed (including the first 'adhesion' peak), indicating that DNA does not stick to the surface under these conditions. (ii) There was a good correlation between the number of (presumably connected) nucleosomes in the images taken under liquid of linker histone-lacking fibers and the number of peaks observed in the retraction curves (Leuba et al, in preparation). (The number of nucleosomes in

Figure 3. Typical force curves obtained on pulling chicken erythrocyte chromatin fibers. In each set of curves, the upper curve is obtained upon moving the AFM tip downward, the lower upon withdrawing the tip.

299

Leuba et al: Atomic Force Microscopy of chromatin fibers actually measured to break DNA (Bensimon et al, 1995; Noy et al, 1997; Rief et al, 1999). It must be noted, however, that in our pulling experiments we are far from thermodynamic equilibrium conditions since we are applying the pulling force at rather large loading rates (typically between 4x104 and 4x106 pN/second). It has been recently demonstrated that the forces monitored in pulling experiments are strongly dependent on the pulling rate (Evans and Ritchie, 1997; Rief et al, 1997a; Merkel et al, 1999). We believe that further improvements in instrumentation will allow us to apply much lower loading rates, thus approaching thermodynamic equilibrium conditions.

References Allen MJ, Dong XF, O'Neill TE, Yau P, Kowalczykowski SC, Gatewood J, Balhorn R and Bradbury EM (1993) Atomic force microscope measurements of nucleosome cores assembled along defined DNA sequences. Biochemistry 32, 8390-8396. Allen S, Chen X, Davies J, Davies MC, Dawkes AC, Edwards JC, Roberts CJ, Sefton J, Tendler SJ and Williams PM (1997) Detection of antigen-antibody binding events with the atomic force microscope. Biochemistry 36, 7457-7463. Bensimon D, Simon AJ, Croquette V and Bensimon A (1995) Stretching DNA with a receding meniscus: Experiments and Models. Phys. Rev. Lett. 74, 4754-4757. Cluzel P, Lebrun A, Heller C, Lavery R, Viovy JL, Chatenay D and Caron F (1996) DNA: an extensible molecule. Science 271, 792-794. Dammer U, Hegner M, Anselmetti D, Wagner P, Dreier M, Huber W and Guntherodt HJ (1996) Specific antibody/antigen interactions measured by force microscopy. Biophys. J. 70, 2580-2587. Figure 4. Frequency distribution of forces measured from chromatin force curves: (A) native fibers, (B) linker histonedepleted fibers. Chicken erythrocyte chromatin fibers were depleted of linker histones as described in Leuba et al. (1994).

Evans E and Ritchie K (1997) Dynamic strength of molecular adhesion bonds. Biophys. J. 72, 1541-1555.

In principle, the peak events may reflect either (or both) of two things: detachment of individual nucleosomes from the surface during the pulling-off, or unraveling (unwrapping) of the DNA from around the histone octamer. It is not possible to distinguish between these two processes on the basis of the available chicken chromatin fiber data. Further analysis will obviously be needed to interpret these complex chromatin fiber curves. Even in the absence of a strict interpretation, our data show remarkably large forces upon mechanically stretching chromatin fibers. Moreover, the unexpectedly large changes in force upon removal of linker histone may reflect significant differences in the mechanical properties of linker histonecontaining and -depleted chromatin fibers. Our findings are rather surprising in view of the large forces observed upon application of tension to the chromatin fiber. These forces are very close to the forces

Han W, Lindsay SM and Jing T (1996) A magnetically-driven oscillating probe microscope for operation in liquids. Appl. Phys. Letts. 69, 4111-4114.

Florin EL, Moy VT and Gaub HE (1994) Adhesion forces between individual ligand-receptor pairs. Science 264, 415417.

Lee GU, Kidwell DA and Colton RJ (1994a) Sensing discrete streptavidin-biotin interactions with atomic force microscopy. Langmuir 10, 354-357. Lee GU, Chrisey LA and Colton RJ ( 1994b) Direct measurement of the forces between complementary strands of DNA. Science 266, 771-773. Leuba SH and Bustamante C (1999) Scanning force microscopy for the analysis of chromatin. Meth. Mol. Biol. Chapter 10, in press. Leuba SH, Yang G, Robert C, Samori B, van Holde K, Zlatanova J and Bustamante C (1994) Three-dimensional structure of extended chromatin fibers as revealed by tapping-mode scanning force microscopy. Proc. Natl. Acad. Sci. USA 91, 11621-11625. Liu LF, Wang JC (1987) Supercoiling of the DNA template during transcription. Proc. Natl. Acad. Sci. USA 84, 7024-

300

Gene Therapy and Molecular Biology Vol 4, page 301 7027. Liu YZ, Leuba S and Lindsay SM (1999) Relationship between stiffness and force in single molecule pulling experiments. Langmuir 15, 8547-8548. Marszalek PE, Oberhauser AF, Pang YP, Fernandez JM (1998) Polysaccharide elasticity governed by chair-boat transitions of the glucopyranose ring. Nature 396, 661-664. Merkel R, Nassoy P, Leung A, Ritchie K and Evans E (1999) Energy landscapes of receptor-ligand bonds explored with dynamic force spectroscopy. Nature 397, 50-53. Moy VT, Florin EL and Gaub HE (1994) Intermolecular forces and energies between ligands and receptors. Science 266, 257-259. Noy A, Vezenov DV, Kayyem JF, Meade TJ and Lieber CM (1997) Stretching and breaking duplex DNA by chemical force microscopy. Chem. & Biol. 4, 519-527. Oberhauser AF, Marszalek PE, Erickson HP and Fernandez JM (1998) The molecular elasticity of the extracellular matrix protein tenascin. Nature 393, 181-185. Rief M, Clausen-Schaumann H and Gaub HE (1999) Sequence窶電ependent mechanics of single DNA molecules. Nature Str. Biol. 6, 346-349. Rief M, Gautel M, Oesterhelt F, Fernandez JM and Gaub HE (1997a) Reversible unfolding of individual titin immunoglobin domains by AFM. Science 276, 1109-1112. Rief M, Oesterhelt F, Heymann B and Gaub HE (1997b) Single molecule force spectroscopy on polysaccharides by atomic force microscopy. Science 275, 1295-1297. Simpson RT, Thoma F and Brubaker JM (1985) Chromatin reconstituted from tandemly repeated cloned DNA fragments and core histones: a model system for study of higher order structure. Cell 42, 799-808. Van Holde K and Zlatanova J (1996) What determines the folding of the chromatin fiber? Proc. Natl. Acad. Sci. USA 93, 10548-10555. Yang G, Leuba SH, Bustamante C, Zlatanova J and van Holde K (1994) Role of linker histones in extended chromatin fibre structure. Nature Str. Biol. 1, 761-763. Zlatanova J, Leuba SH and van Holde K (1998) Chromatin fiber structure: morphology, molecular determinants, structural transitions. Biophys. J. 74, 2554-2566.

Jordanka Zlatanova (left) with Paola Caiafa (At the 1999 Conference in Crete) 301

Gene Therapy and Molecular Biology Vol 4, page 303 Gene Ther Mol Biol Vol 4, 303-312. December 1999

Gene potentiation: Forming long-range open chromatin structures Review Article

Susan M. Wykes1 and Stephen A. Krawetz1,2 2

Department of Obstetrics and Gynecology and 1The Center for Molecular Medicine and Genetics, Wayne State University School of Medicine, 253 C.S. Mott Center, 275 E Hancock, Detroit, MI 48201 __________________________________________________________________________________________________ Correspondence: Stephen A. Krawetz; Tel: (313)-577-6770; Fax: (313)-577-8554; E-mail: steve@compbio.med.wayne.edu Key Words: chromatin domain, spermatogenesis, hematopoiesis, gene expression, chromatin structure Received 2 August 1999; accepted 11 October 1999.

Summary Gene potentiation is the process of opening a chromatin domain, which in turn renders genes accessible to the various factors requisite for their expression. The formation of an open chromatin structure is central to the establishment of cell fate and tissue-specific gene expression. Both hematopoiesis and spermatogenesis serve as excellent models for examining gene potentiation. Each developmental pathway is governed by a unique differentiative program, which specifies a subset of potentiated genes enabling expression. A discussion of these contrasting potentiative cascades is presented illustrating that cell fate is ultimately determined by the selective opening and closing of gene containing domains. Elucidating the mechanism, which governs these perturbations in chromatin structure, will provide valuable insight into how differentiative decisions are made and whether commitment to a particular phenotype can be modified.

I. Introduction Many eukaryotic genes are organized into functional chromatin domains. This facilitates their coordinate regulation during development (Reviewed by Dillon and Grosveld, 1993; Vermaak et al., 1998). The ability of individual cells to regulate the genes contained within such chromatin domains is key to the establishment of cell fate and tissue-specific gene expression (Reviewed by Bonifer et al., 1997). Perturbations in chromatin structure can act both locally to alter the accessibility of trans-acting factors to cis-regulatory elements and globally to affect the opening and closing of entire chromatin domains (Bodnar et al., 1996; Vermaak et al., 1998). Gene potentiation is the process of opening a chromatin domain (Choudhary et al.,1995), which then renders genes accessible to the various trans-acting factors required for their expression (Reviewed by Higgs, 1998). As such, transcriptionally active genes are found in regions of open chromatin. Potentiated regions of the genome replicate early in S-phase and are preferentially confined to discreet chromosomal territories within the interphase nucleus (Kurz et al., 1996; Lamond et al., 1998; Wei et al.,1998). These transcriptionally competent domains exhibit a 10 fold enhanced general nuclease sensitivity (Weintraub and

Groudine, 1976). Routinely, increased DNase I sensitivity is used as a diagnostic indicator of gene potentiation. For example, the chicken ovalbumin domain is part of a multigenic, coordinately-expressed locus that exists as a single DNase I-sensitive, potentiated domain in hen oviduct where its members are expressed but remains in a DNase I-insensitive, non-potentiated configuration in all non-expressing cells (Lawson et al ., 1982). A potentiated chromatin domain may also contain small hypersensitive sequences which are approximately 100 fold more sensitive to DNase I digestion than bulk chromatin (Stalder, 1980). These hypersensitive regions often demarcate sites of interaction between specific effector proteins and cis-regulatory elements (Elgin, 1984). The transition from a closed to an open chromatin conformation is a necessary event, but alone is not sufficient to ensure transcription (Reviewed by Krawetz et al.,1999). Consider the contrasting environments of the human !-globin domain on chromosome 11 and its co-regulated family member, the "-globin domain found near the telomere of chromosome 16. Both domains assume a DNase I-sensitive, potentiated conformation in erythroid cells where their respective globin genes are expressed. However, unlike the !globin domain which forms a DNase I-insensitive, closed configuration in non-erythroid cells, the "-globin domain

303

Wykes and Krawetz: Gene potentiation in chromatin domains remains constitutively potentiated in all cell types (Craddock et al., 1995). This can can be reconciles as the "-globin domain lying within a chromosomal region containing several widely expressed genes including the constitutively expressed 3-methlyadenine DNA glycosylase gene (Vyas et al., 1992; Vickers et al., 1993; Flint et al., 1997). Although the "-globin domain remains in a potentiated conformation in non-erythroid cells, the globin genes remain transcriptionally silent because the necessary factors for its expression are lacking. It is also interesting to note that the human !-globin domain exists in a potentiated, open conformation prior to commitment to the erythroid lineage (JimĂŠnez et al., 1992) but requires additional elements and factors for appropriate expression (Calzolari et al., 1999). Taken together, these data provide supportive evidence that while the formation of a potentiated chromatin domain is a necessary event for transcription, additional levels of control are required to ensure proper spatial and temporal gene expression. The potentiated state of a gene can also be influenced by alterations in the local chromatin environment. For example, many eukaryotic genes are differentially expressed by altering their methylation status. These genes are largely unmethylated in cells where they are transcribed, but fully methylated in all non-expressing cells (Reviewed by Cedar, 1988). Histone acetylation also acts on the local gene environment during the transition from the 30 nm fiber to the more open structure that can be likened to a 10 nm fiber, stabilizing the more relaxed open structure (Reviewed by Davie and Hendzel, 1994). It has also been postulated that DNA methylation patterns may serve to modulate histone acetylation thereby maintaining local chromatin states. Both DNA methylation and histone acetylation render increased accessibility of ubiquitous and tissue-specific trans-acting factors to cis-regulatory elements, facilitating transcriptional activation (Eden et al., 1998). In addition to enhanced general nuclease sensitivity, transcriptionally active genes are often associated with the nuclear matrix (Ciejek et al., 1983). This interaction has been postulated to represent the means by which potentiated chromatin domains are organized within the eukaryotic nucleus (Reviewed by Bode et al., 1996). Matrix attachment regions (MARs) have been shown to demarcate the boundary elements of DNase I-sensitive chromatin domains of many eukaryotic loci including the human apolipoprotein-B gene, the chicken lysozyme gene and the human PRM1#PRM2#TNP2 multigenic locus (Kalos and Fournier, 1995; Stief et al ., 1989, Kramer and Krawetz, 1996). MARs can also function as insulators against the position effects of neighboring chromatin (Zlatanova and van Holde, 1992), and have been shown to interact directly with enhancer elements to extend the accessibility of a chromatin domain (Jenuwein et al., 1997; Forrester et al., 1994). Taken together, these and other

observations have led to the suggestion that MARs can be divided into discrete functional classes and may provide a means to tag genic domains for coordinate expression (Kramer and Krawetz, 1996). While it is apparent that gene potentiation involves alterations that affect both the local environment of individual genes and the physical structure and organization of large chromatin domains containing multigene families, the actual mechanism(s) governing this process remain unclear.

II. Models for Examining Gene Potentiation A. Hematopoiesis Hematopoiesis is the differentiative pathway by which pluripotent hematopoietic stem cells give rise to the various erythroid, lymphoid and myeloid blood cell lineages. A schematic representation of hematopoiesis is shown in Figure 1. This process initiates when a pluripotent hematopoietic stem cell differentiates to form a myeloid or lymphoid stem cell. Both of these multipotent stem cells are capable of self renewal and differentiation. Upon appropriate stimuli, the lymphoid stem cell differentiates to form the committed pre-B and pre-T progenitor cells which in turn differentiate to form mature B and T lymphocytes respectively. In contrast, the myeloid stem cell forms an intermediate stem cell, CFUGEMM (colony forming unit- granulocyte, erythrocyte, monocyte, megakaryocyte), which differentiates to produce the corresponding unipotent CFU-progenitors for these lineages. Subsequently, these committed precursors, influenced by various cytokines and growth factors, terminally differentiate into their respective mature erythroid, megakaryocytic, monocytic, neutrophilic, eosinophilic and basophilic cell types (Carr and Rodak, 1999). Hematopoiesis originates in the embryonic yolk sac, temporarily shifts to the fetal liver, and from the fourteenth week of gestation throughout adult life occurs primarily in the bone marrow (Yawata, 1996). Concomitantly, many hematopoietic-specific genes are coordinately regulated and differentially expressed during development (Reviewed by Orkin, 1995). Two of the most extensively characterized examples are the human " and ! globin gene clusters which encode the respective globin chains of hemoglobin, the oxygen transport protein found in erythrocytes (Reviewed by Karlson and Nienhuis, 1985). The human !-globin gene cluster is a well-established model for examining the functional role of chromatin in gene regulation. The human !globin locus is a multigenic domain containing five globin genes arranged in the order of their sequential expression during development (Reviewed by Andrin and Spencer, 1994). The entire !-globin gene cluster exists in a single DNase Isensitive, open conformation in cells of the erythroid lineage, but remains in a DNase I-insensitive, closed conformation in all non-erythroid cells (Groudine et al., 1983). However, the extent of the human !-globin domain remains to be clearly

304

Gene Therapy and Molecular Biology Vol 4, page 305

Figure 1 Hematopoiesis: The differentiation and maturation of blood cells. During, hematopoiesis multipotent stem cells, influenced by numerous cytokines and growth factors, divide and differentiate to form committed progenitors for the various erythroid, lymphoid and myeloid blood cell lineages. These unipotent precursors in turn differentiate to produce their respective mature blood cell types. Each step of this differentiative pathway involves the selective restriction of chromatin domains ultimately defining a lineage-specific subset of genes which remains potentiated for expression.

displaying classical enhancer activity in transient transfection assays (Tuan et al., 1989) and by its ability to confer high-level, position-independent expression of a transgene in mice (Philipsen et al.,1990; Lui et al., 1992 and Pawlik et al., 1995). HS-3 possesses a dominant chromatin opening and remodeling activity that is separate and distinct from its ability to direct !-globin expression in transgenic mice (Ellis et al., 1996). HS-5 is a constitutive site and although it possesses no enhancing activity (Fraser et al., 1993), it does contain a matrix attachment region (Jarman et al., 1988). In this manner, HS-5 may serve as an insulator against the position effects of neighboring chromatin (Li et al., 1994). These observations suggest that the LCR elements act cooperatively to ensure the correct spatial and temporal regulation of the !-globin gene cluster. The full LCR has been postulated to function synergistically as a holocomplex, to regulate the sequential expression of the !-globin genes through specific HSpromoter interactions (Bresnick et al., 1997 and Ellis et al., 1996). While several looping models have been proposed for this developmental switching mechanism (Reviewed by

delineated. Upstream from the gene cluster is a series of five DNase I-hypersensitive sites (HS1-HS5). Hypersensitive sites 1-4 are commonly referred to as the !-globin locus control region, LCR (Reviewed by Grosveld et al., 1993). The importance of the !-globin LCR has been widely demonstrated through the characterization of its individual hypersensitive sites (Reviewed by Dillon and Grosveld, 1993) and by examining the effects of naturally occurring deletions which result in the various !-thalassemias (Van der Ploeg et al., 1980; Driscol et al., 1989). These de-novo mutations not only lead to reduced levels of gene expression, but have also been demonstrated to alter both the timing of replication of the !-globin locus and its sensitivity to DNase I (Forrester et al., 1990). Dissection and characterization of individual hypersensitive sites has shown that the !-globin LCR functions in a tissue-specific manner and is necessary for high level expression of the globin genes. Hypersensitive sites 1-4 act as erythroid-specific enhancers and possess multiple binding sites for various transcription factors (Reviewed by Engel, 1993; Wood, 1996). HS-2 has been shown to be functionally equivalent to the full LCR both by 305

Wykes and Krawetz: Gene potentiation in chromatin domains Stamatoyannopoulos and Nienhuis, 1994), others suggest that the temporal regulation of the !-globin genes is LCRindependent, relying solely on promoter proximal elements and gene arrangement (Martin et al., 1996). Recent studies examining the organization and temporal regulation of the !globin genes have further demonstrated that inversion of gene order with respect to the LCR significantly alters their expression (Tanimoto et al., 1999). Although it has been well established that the !-globin LCR plays an essential role in the transcriptional activation and temporal regulation of the !-globin locus, its function in long range chromatin opening is debated. The !-globin LCR was the first element reported to confer tissue-specific, position-independent and copy number-dependent expression of a transgene (Grosveld et al., 1987). It has since been held that the !-globin LCR facilitates the creation of an open chromatin environment by altering the topology of the !-globin domain thereby maintaining it in a potentiated configuration (Reviewed by Martin et al., 1996). Conversely, others have suggested that while the LCR ensures the high-level, tissue-specific expression of the !globin genes, it does not function to form or maintain the open chromatin conformation that is necessary for the expression of the locus (Reitman et al., 1993; Reik et al., 1998; Epner et al., 1998). This raises the intriguing possibility that the ability of the !-globin LCR to overcome position effects in transgenic mice may be independent of the mechanism that governs the potentiation of the !-globin domain during development (Reviewed by Higgs, 1998). Interestingly, it has been now shown that the human and mouse !-globin loci reside in a cluster of functional odorant receptor genes (Bulger et al., 1999). The authors postulate that these two overlapping gene families may share some of the same regulatory elements involved in mediating their different expression patterns. Understanding how such multigenic domains are potentiated will provide valuable insight into the functional role of chromatin in the establishment of cell fate and tissue-specific gene expression.

spermatogonia differentiate to form these cell types, they are then committed to the spermatogenic pathway. Type B spermatogonia in turn differentiate to form primary spermatocytes, which subsequently enter meiosis (Reviewed by Dym, 1994). The development of meiotic spermatocytes begins with the primary or pre-leptotene spermatocyte. DNA replication occurs at this stage resulting in a genome content of 4N. After DNA synthesis, chromatin condensation is initiated, signaling the start of meiosis. In humans, approximately 22 out of the 64 days of spermatogenesis are spent in meiosis and the majority of that in prophase I (Reviewed by Willison and Ashworth, 1987). During this process, the chromosomes undergo pairing and the formation of the synaptonemal complex. This is followed by genetic recombination between the homologous pairs. Subsequent to the exchange of genetic material, the first meitotic division occurs resulting in the formation of two secondary spermatocytes. This is quickly followed by meiosis II, reduction division, producing four haploid round spermatids (Gardner and Snustad, 1984). Some of the genes expressed during meiosis include testes-specific variants of somatic cell proteins (Reviewed by Kierzenbaum, 1994). One of the best-characterized examples is Pgk2, phosphoglycerate kinase 2, and the testes specific isozyme of Pgk1. Pgk1 is a constitutively expressed, X-linked enzyme that catalyzes the conversion of 1,3 diphosphoglycerate to 3-phosphoglycerate and ATP during glycolysis (Lee et al., 1972). However, because it is subject to X-inactivation (Lifschytz and Lyndsay, 1972), the expression of the autosomal, testes-specific Pgk2 gene is initiated to compensate for the reduced levels of this essential enzyme during spermatogenesis (McCarrey et al., 1992). The final stage of spermatogenesis, termed spermiogenesis, is characterized by the morphological differentiation of round spermatids into mature spermatozoa (Clermont and Leblond, 1955). In humans, like other mammals, the round spermatid stage marks the initial expression of the haploid-specific packaging proteins the transition proteins and the protamines (Wykes et al., 1995). These proteins facilitate the remodeling of the chromatin during the morphologic transformation from round spermatid to mature spermatozoa (Reviewed by WoutersTyrou, 1998). This process involves the initial disruption of the nucleohistone structure by the transition proteins and their final replacement by the protamines to compact and condense the DNA into a species-specific shaped nucleus (Reviewed by Dadoune, 1995). Subsequent to their expression these genes are again suppressed as part of the genome-wide silencing that yields the transcriptionally quiescent sperm nucleus (Reviewed by Balhorn, 1989; Oliva and Dixon, 1991). Accordingly, many testes-specific genes, including the transition proteins and the protamines, are synthesized relatively early then placed under extensive translational control (Reviewed by Eddy and Oâ&#x20AC;&#x2122;Brien, 1998).

B. Spermatogenesis Spermatogenesis serves as another excellent model for examining gene potentiation because of the vast array of testes-specific genes that are coordinately regulated and expressed in a specific temporal manner during the formation of the male gamete (Reviewed by McCarrey, 1998). This entire differentiative pathway from the uncommitted stem cell to the mature spermatozoa occurs within the testes. The important cellular and molecular events of spermatogenesis are summarized in Figure 2. There are three stages to mammalian spermatogenesis: mitosis, meiosis and spermiogenesis (Junqueira et al., 1986). During mitosis, primitive type A spermatogonia either actively divide, renewing themselves or differentiate to form the intermediate and type B spermatogonia. Once type A 306

Gene Therapy and Molecular Biology Vol 4, page 307

Figure 2 The three stages of human spermatogenesis. Human spermatogenesis is divided into three stages, mitosis, meiosis and spermiogenesis. Some of the important cellular and molecular events are indicated. During mitosis primitive type A spermatogonia either actively divide renewing themselves or differentiate to form type B spermatogonia and thus commit themselves to the spermatogenic pathway. Type B spermatogonia in turn differentiate to form primary or preleptotene spermatocytes which then enter meiosis. Meiosis initiates after DNA replication and is characterized by chromosomal pairing and genetic recombination. Subsequent to the exchange of genetic material, meiotic spermatocytes undergo two successive cell divisions producing four round haploid spermatids. During the final stage, spermiogenesis, these round spermatids then morphologically differentiate to form mature spermatozoa. This process is facilitated by the transition proteins and the protamines which mediate the remodeling of the chromatin during the terminal differentiation of the spermatid nucleus.

Spermatogenesis is a dynamic continuum reflecting the coordinate temporal regulation of genes involved in the formation of functional male gametes. It is an exquisite model for examining the mechanism of gene potentiation.

III. Selective Potentiation of TestesSpecific Domains It is well established that the DNA in human sperm chromatin is partitioned into both a nucleohistone and nucleoprotamine fraction (Tanphaichutr et al., 1978). This organization is likely to be sequence specific and may serve to designate a particular subset of genes which have a functional role in early development (Gatewood et al., 1987). In humans, 85% of the DNA in sperm chromatin is protamine bound, while 15% remains histone bound (Tanphaichutr et al., 1978). The histone bound fraction contains several testes-specific histone variants (Reviewed by Ward, 1994; Donecke et al., 1997). This compacts the DNA to a greater extent than its somatic counterparts although histone H1 is absent and histones H3 and H4 are highly acetylated (Gatewood et al., 1990). Since both histone acetylation and the absence of histone H1 are features of active chromatin (Reviewed by Wolffe, 1994), it has been proposed that the histone-associated genes in

the human sperm nucleus may be the first genes transcribed after fertilization (Gardiner-Garden et al., 1998). Alternatively, these histone bound regions may simply serve as nucleation sites for the replacement of the protamines by the histones upon fertilization for the resumption of the somatic nucleosomal structure. The transition proteins and the protamines function solely to remodel the chromatin during late spermatogenesis. The PRM1#PRM2#TNP2 members of this sperm gene packaging family are clustered into a single 28 kb DNase Isensitive domain flanked by two regions of marked insensitivity (Choudhary et al., 1995). A diagrammatic representation of this domain is shown in Figure 3). This biophysical domain has been biologically confirmed. Transgenic mice harboring the human protamine locus express the transgene in a haploid-specific, position-independent and copy-number dependent manner (Choudhary et al., 1995; Stewart et al., 1999). These data show that this region of the genome contains all of the elements necessary for the appropriate temporal and spatial expression of this suite of genes independent of site of integration. The ends of the domain are attached to the sperm nuclear matrix (Kramer and Krawetz, 1996). The sequences of attachment at the boundaries of this domain correspond to sperm-specific matrix attachment regions, termed spMARs.

307

Wykes and Krawetz: Gene potentiation in chromatin domains

Figure 3: The human PRM1#PRM2#TNP2 domain. The human protamine locus is a multigenic domain containing three haploid expressed genes encoding the protamines PRM1, PRM2 and transition protein TNP2. This gene cluster exists as a single DNase I-sensitive domain flanked by two regions of matrix attachment. The relative position of the locus along human chromosome 16p13.13 is shown.

Figure 4: Gene potentiation during mammalian spermatogenesis. The potentiative state of individual genic domains was assessed by DNase I sensitivity in various isolated germ cell populations ranging from Type A spermatogonia to round spermatid. The alterations in the potentiative states of Pgk1 and the testes-specific Pgk2 and Prm1#Prm2#Tnp2 domains indicates that the temporal regulation of these genes for expression during spermatogenesis is mediated by the selective opening and closing of individual chromatin domains. The constitutively expressed !-actin and erythroid-specific !-globin genes provide positive and negative controls respectively.

308

Gene Therapy and Molecular Biology Vol 4, page 309 Although the functional role of the spMAR elements in gene potentiation remains to be clearly delineated recent evidence suggests that the 3â&#x20AC;&#x2122; spMAR of the PRM1#PRM2#TNP2 domain is critical for expression of this human locus (Kramer et al., 1997). The protamine gene cluster remains transcriptionally quiescent until the round spermatid stage when it is then actively transcribed in the final burst of transcription before genome-wide silencing (Fig. 2, Wykes et al., 1995). This is representative of the class of similarly expressed genes that constitute the alternative male haploid genome (Kramer and Krawetz, 1997). It has thus been argued that expression of this alternative genome is controlled by sequestering its members in a closed chromatin conformation until required (Kramer et al., 1998). To directly address this tenet, the potentiative state of the Prm1#Prm2#Tnp2 gene cluster and several other gene loci including Pgk1, Pgk2, !-actin and !-globin, was assessed throughout spermatogenesis using purified populations of germ cells (Kramer et al., 1998). The results of this study are summarized in Figure 4. It was reasoned that since, Pgk1 is a constitutively expressed gene its domain should exist in a potentiated conformation in all cell types. However, Pgk1 is also subject to X-inactivation during spermatogenesis, which is compensated by the expression of the autosomal testes-specific Pgk2. As the Pgk1 gene is silenced and its chromatin domain is closed, the closed Pgk2 domain opens to enable its transcription through to the round spermatid stage. Similarly, the multigenic Prm1#Prm2#Tnp2 domain remains closed until the pachytene spermatocyte stage when it then assumes an open potentiated chromatin conformation prior to the expression of its members in round spermatids. This demonstrated that the temporal expression of testesspecific genes is ultimately mediated by the selective opening and closing of individual chromatin domains. Interestingly, the physical domain containing the human PRM1#PRM2#TNP2 multigenic locus remains in an open conformation in mature spermatozoa (Choudhary et al., 1995).

IV. Prospects Both hematopoiesis and spermatogenesis serve as excellent model systems for examining gene potentiation. Each developmental process begins with a progenitor stem cell, which, in response to various stimuli, acquires the capacity to differentiate into individual cell types. Commitment to a particular cell fate is a gradual process and is ultimately determined by the selective potentiation of certain genes for expression in conjunction with the repression or silencing of other genes. For example, commitment to a specific hematopoietic pathway involves the selective repression of genes to restrict lineage potential from its multipotent stem cell (Hu et al., 1997). This repressive mechanism is marked by the successive

closing of chromatin domains eventually defining a lineagespecific subset of genes that remains potentiated for expression (Jimenez et al, 1992). Conversely, spermatogenic differentiation is governed by an expressive mechanism whereby the activation of testes-specific genes is mediated by the selective opening, i.e. potentiation, of individual chromatin domains (Kramer et al., 1998). Further characterization of these contrasting potentiative mechanisms which establish cell fate during hematopoiesis and spermatogenesis will provide insight into how differentiation is regulated by chromatin structure and whether commitment to a particular cell fate can be altered (Krawetz et al., 1999). The identification of potentiator elements, which facilitate the opening and closing of chromatin domains may be useful to ensure high-level, tissue-specific expression of targeted gene therapeutics. Acknowledgments This work was supported by NIH grant HD36512 to S.A.K. PreDoctoral fellowship support to S.M.W. from the WSU CMMG is gratefully acknowledged. References Andrin C, and Spencer C (1994) The intricacies of !-globin gene expression. Biochem. Cell Biol. 72, 377-380. Balhorn R (1989) Mammalian protamines: structure and molecular interactions. In: Molecular Biology of Chromosome Function. Aldolph, K. (Editor) Springer-Verlag. Pp 366-395. Bode J, Stengert-Iber M, Kay V, Schlake T, and Dietz-Pfeilsteller A (1996) Scaffold/Matrix-attached regions: Topological switches with multiple regulatory functions. Crit Rev Euk Gene Express 6, 115-138. Bodnar J, and Bradley M (1996) A chromatin switch. J. Theor. Biol. 183, 1-7. Bonifer C, JĂ¤gle U, and Huber M (1997) The chicken lysozyme locus as a paradigm for the complex developmental regulation of eukaryotic gene loci. J. Biol. Chem. 272, 26057-26078. Bresnick E, and Tze L (1997) Synergism between hypersensitive sites confers long-range gene activation by the beta-globin locus control region. Proc Natl Acad Sci USA 94, 4566-4571. Bulger M, von Doornick J, Saitoh N, Telling A, Farrell C, Bender M, Felsenfeld G, Axel R, and Groudine M (1999) Conservation of sequence and structure falnking the mouse and human !-globin loci: The !-globin genes are embedded within an array of odorant receptor genes. Proc Natl Acad Sci USA 96, 51295134. Calzolari R, McMorrow T, Yannoutsos N, Langeveld A, and Grosveld F (1999) Deletion of a region that is a candidate for the difference between the deletion forms of hereditary persistence of fetal hemoglobin and $-! Thalassemia affects ! but not % globin gene expression. EMBO J. 18, 949-958. Carr J, and Rodak B (Editors) Clinical Hematology Atlas (1999) W.B. Saunders, Philadelphia, PA, Pp 10-12. Cedar H (1988) DNA methylation and gene activity. Cell 53, 3-4. Choudhary S, Wykes S, Kramer J, Mohammed A, Koppitch F, Nelson J, and Krawetz S.A. (1995) A haploid expressed gene

309

Wykes and Krawetz: Gene potentiation in chromatin domains cluster exists as a single chromatin domain in human sperm. J. Biol. Chem. 270, 8766-8762.

alteration in chromatin structure and replication across the entire !-globin locus. Genes Dev. 4, 1637-1649.

Ciejek E, Tsai M, and O’Malley B (1983) Actively transcribed genes are associated with the nuclear matrix. Nature 306, 607-609.

Forrester W, van Genderen C, Jenuwein T, and Grosschedl R (1994) Dependence of enhancer-mediated transcription of the immunoglobulin µ gene on nuclear matrix attachment regions. Science 265, 1221-1225.

Clermont Y, and Leblond C (1955) Spermiogenesis of man, monkey, ram and other mammals as shown by the “Periodic Acid-Schiff” Technique. Am.J. Anatomy 96, 229-253. Craddock C, Vyas P, Sharpe J, Ayyub H, Wood W, and Higgs D (1995) Contrasting effects of " and ! globin regulatory elements on chromatin structure may be related to their different chromatin environments EMBO J. 14, 17181726. Dadoune J (1995) The nuclear status of human sperm cells. Micron. 26, 323-345. Davie J, and Hendzel M (1994) Multiple functions of dynamic histone acetylation. J.Cell Biochem. 55, 98-105. Dillon N, and Grosveld F (1993) Transcriptional regulation of multigene loci: multilevel control. Trends Genet 9, 134137. Doenecke D, Drabent B, Bode C, Bramlage B, Franke K, Gavenis K, Kosciessa U, and Witt O (1997) Histone gene expression and chromatin structure during spermatogenesis. Adv. Exp. Med. Biol. 424, 37-48. Driscoll M, Dobkin C, and Alter B (1989) %$ !- Thalassemia due to a de novo mutation deleting the 5’ !-globin gene activation region hypersensitive sites. Proc Natl Acad Sci USA 86, 7470-7474. Dym M (1994) Spermatogonial stem cell of the testes. Proc Natl Acad Sci USA 91, 11287-11289. Eddy E and O’Brien D (1998) Gene expression during mammalian meiosis. Curr. Topics Dev. Biol. 37, 141-200. Eden S, Hashimshony T, Keshet J, Cedar H, and Thorne A (1998) DNA methylation models histone acetylation. Nature 394, 842. Engel J (1993) Developmental regulation of human !-globin gene transcription: a switch of loyalties. Trends Genet 9, 304-309. Elgin S (1984) Anatomy of hypersensitive sites. Nature 309, 213-214. Ellis J, Tan-Un K, Harper A, Michalovich D, Yannoutsos N, Philipsen S, and Grosveld F (1996) A dominant chromatin opening activity in 5’ hypersensitive site 3 of the human !globin locus control region. EMBO J. 15, 562-568. Epner E, Reik A, Cimbora D, Telling A, Bender M, Fiering D, Enver T, Martin D, Kennedy M, Keller G, and Groudine M (1998). The !-globin LCR is not necessary for an open chromatin structure or transcription of the mouse !-globin locus. Mol.Cell 2, 447-455.

Fraser P, Pruzina S, Antoniou M, and Grosveld F (1993) Each hypersensitive site of the human !-globin locus control region confers a different developmental pattern of expression on the globin genes. Genes Dev. 7, 106-113. Gardiner-Garden M, Ballesteros M, Gordon M and Tam P (1998) Histone and protamine DNA association: conservation of different patterns within the !-globin domains in human sperm. Mol.Cell Biol. 18, 3350-3356. Gardner E, and Snustad D (Editors): Principles of Genetics. Seventh Edition. (1984) John Wiley & Sons. Pp 46-52. Gatewood J, Cook G, Balhorn R, Bradbury E, and Schmid C (1987) Sequence-specific packaging of DNA in human sperm chromatin. Science 236, 962-964. Gatewood J, Cook G, Balhorn R, Schmid C, and Bradbury E (1990) Isolation of four core histones from human sperm chromatin representing a minor subset of somatic histones. J Biol.Chem. 265, 20662-20666. Grosveld F, Blom van Assendelft G, Greaves D, and Kollias G (1987) Position-independent high-level expression of the human !-globin gene in transgenic mice. Cell 51, 975-985. Grosveld F, Dillon N, and Higgs D (1993) The haemoglobinopathies. In: Baillièris Clinical Haematology. Higgs, D. and Weathall, D. (Editors) Baillière Tindall. Pp 31-56. Groudine M, Kohwi-Shigetmatsu T, Gelinas R, Stamatoyannopoulas G, and Papayannopoulou T (1983) Human fetal to adult globin switching: changes in chromatin structure of the !-globin gene locus. Proc Natl Acad Sci USA USA 80, 7551-7555. Higgs D (1998) Do LCRs open chromatin domains. Cell 95, 299302. Hu M, Krause K, Greaves M, Sharkis S, Dexter M, Heyworth C, and Enver T (1997) Multilineage gene expression precedes commitment in the hematopoietic system. Genes Dev. 11, 774785. Jarman A, and Higgs D (1988) Nuclear Scaffold attachment sites in the human globin gene complexes. EMBO J. 7, 3337-3344. Jenuwein T, Forrester W, Fernandez-Herrero L, Laible G, Dull M, and Grosschedl R (1997). Extension of chromatin accessibility by nuclear matrix attachment regions. Nature 385, 269-272. Jiménez G, Griffithe S, Ford A, Greaves M, and Enver T (1992) Activation of the !-globin locus control region precedes commitment to the erythroid lineage. Proc Natl Acad Sci USA 89, 10618-10622. Junqueira L, Carneiro J, and Long J (Editors): Basic Histology. Fifth Edition.(1986) Appleton-Century-Crofts. Pp 468-484.

Flint J, Thomas K, Micklem G, Raynham H, Clark K, Doggett N, King A, and Higgs D (1997) The relationship between chromosome structure and function at the human telomeric region. Nature Genetics 15, 252-257.

Kalos M, and Fournier R (1995) Position-independent transgene expression mediated by boundary elements from the apolipoprotein B chromatin domain. Mol. Cell. Biol. 15, 198207.

Forrester W, Epner E, Driscoll M, Enver T, Brice M, Papayannopoulou T, and Groudine M (1990) A deletion of the human !-globin locus activation region causes a major

Karlson S and Nienhuis A (1985) Developmental regulation of the human globin genes. Ann. Rev. Biochem. 54, 1071-1108.

310

Gene Therapy and Molecular Biology Vol 4, page 311 Kierszenbaum A (1994) Mammalianm spermatogenesis in vivo and in vitro: a partnership of spermatogenic and somatic cell lineages. Endoc. Rev. 15, 116-134. Kramer J, and Krawetz, S. A. (1996) Nuclear matrix interactions within the sperm genome. J. Biol. Chem. 271, 1161911622. Kramer, J, and Krawetz, S.A. (1997) RNA in sperm: implications for the alternative haploid genome. Molecular Human Reproduction 3:473-478. Kramer J, Zhang S, Yaron Y, Zhao Y, and Krawetz, S. A. (1997) Genetic testing for male infertility: a postulated role for mutations in sperm nuclear matrix attachment regions. Genetic Testing 1, 125-129. Kramer J, McCarrey J, Djakiew D, and Krawetz S.A. (1998) Differentiation: the selective potentiation of chromatin domains. Development 125, 4749-4755. Krawetz S.A., Kramer J, and McCarrey J (1999) Reprogramming the male gamete genome: a window to successful gene therapy. Gene 234, 1-9. Kurz A, Lampel S, Nickolenko J, Brandl J, Benner A, Zirbel R, Cremer T, and Lichter P (1996) Active and inactive genes localize preferentially in the periphery of chromosome territories. J. Cell Biol. 135, 1195-1205. Lamond A, and Earnshaw (1998) Structure and function in the nucleus. Science 280, 547-553. Lawson G, Knoll B, March C, Woo S, Tsai M, and O’Malley B (1982) Definition of 5’ and 3’ structural boundaries of the ovalbumin multigene family. J. Biol. Chem. 257, 15011507. Lee C, Niesel D, Pegoraro B, and Erikson R (1980) Immunological and structural relatedness of isozymes and genetic variants of 3-phoshoglycerate kinase from the mouse. J. Biol. Chem. 255, 2590-2595. Li Q, and Stamatoyannopoulos G (1994) Hypersensitive-site 5 of the human !-locus control region functions as a chromatin insulator. Blood 84, 1399-1401. Lifschytz E and Lindsley D (1972). The role of X-chromosome inactivation during spermatogenesis. Proc Natl Acad Sci USA 69, 182-186.

Orkin S (1995) Transcription factors and hematopoetic development. J. Biol. Chem. 270, 4955-4958. Pawlik K, and Townes T (1995) Automonous, erythroid-specific DNase I hypersensitive site formed by the human !-globin locus control region, LCR 5’HS2 in transgenic mice. Dev. Biol. 169, 728-732. Philipsen S, Talbot D, Fraser P, and Grosveld F (1990) The !-globin dominant control region: Hypersensitive site 2. EMBO J. 9, 2159-2167. Reik A, Telling A, Zitnik G, Cimbora D, Epner E, and Groudine M (1998) The locus control region is necessary for gene expression in the human !-globin locus but not the maintenance of an open chromatin structure in erythroid cells. Mol. Cell. Biol. 18, 5992-6000. Reitman M, Lee E, Westphal H, and Felsenfeld G (1993) An enhancer/locus control region is not sufficient to open chromatin. Mol. Cell. Biol. 13, 3990-3998. Stalder J, Larsen A, Engel J, Dolan,M, Groudine M, and Weintraub H (1980) Tissue-specific DNA cleavages in the globin chromatin domain introduced by DNase I. Cell 20, 451-460. Stamatoyannopoulos G, and Nienhuis A (1994) Hemoglobin switching. In: The Molecular Basis of Blood Diseases. Stamatoyannopoulos, G. (Editor). Philadelphia. WB Saunders. Pp 107-156. Stewart K, Kramer J, Evans, M, and Krawetz S.A. (1999) Temporal expression of the transgenic human protamine gene cluster. Fertility and Sterility 71, 739-745. Stief A, Winter D, Str!tling W, and Sippel A (1989) A nuclear DNA attachment element mediates elevated and position-independent gene activity. Nature 341,343-345 Tanimoto K, Liu Q, Bungert J, and Engel J (1991) Effects of altered gene order or orientation of the locus control region on human !-globin gene expression in mice. Nature 398, 344-347.

Martin D, Fiering S, and Groudine M (1996) Regulation of !globin gene expression: straightening out the locus. Curr.Opin. Gen. Dev. 6, 488-495. McCarrey J, Berg W, Paragioudakis S, Zhang P, Dilworth D, Arnold B, and Rossi J (1992) Differential transcription of Pgk genes during spermatogenesis in the mouse. Dev. Biol. 154, 160-168.

Tanphaichitr N, Sobhon P, Taluppeth N, and Chalermisarachai P (1978) Basic nuclear proteins in testicular cells and ejaculated spermatozoa in man. Exp. Cell Res. 117, 347-356. Tuan D, Solomon W, London I, and Lee D (1989) An erythroidspecific developmental stage-independent enhancer far upstream of the human “!-like globin” genes. Proc Natl Acad Sci USA 86, 2554-2558. Van der Ploeg L, Konings A, Oort M, Roos D, Bernini L, and Flavell R (1980) %-! Thalassemia studies showing that deletion of the % and $ genes influence !-globin gene expression in man. Nature 283, 637-642. Vermaak D, and Wolffe A (1998) Chromatin and chromosomal controls in development. Dev. Genetics 22, 1-6. Vickers M, Vyas P, Harris P, Simmons D, and Higgs D (1993) Structure of the human 3-methyladenine DNA glycosylase gene and localization close to the 16p telomere. Proc Natl Acad Sci USA 90, 3437-3441.

McCarrey J, (1998) Spermatogenesis as a model system for developmental analysis of regulatory mechanisms associated with tissue-specific gene expression. Seminars Cell Dev. Biol. 9, 459-466.

Vyas P, Vickers M, Picketts D and Higgs D (1995) Conservation of position and sequence of a novel widely expressed gene containing the major human alpha-globin regulatory element. Genomics 29, 679-689.

Oliva R, and Dixon G ( 1991) Vertebrate protamine genes and the histone to protamine replacement reaction. In: Progress in Nucleic Acid Research and Molecular Biology. Cohn, W. and Moldave, K. (Editors) Academic Press. Pp 26-96.

Ward S (1994) The structure of the sleeping genome: implications of sperm DNA organization for somatic cells. J. Cell Biochem. 55, 77-82.

Liu D, Chang J, Moi P, Liu W, Kan Y and Curtin P (1992) Dissection of the enhancer activity of the !-globin 5’ hypersensitive site 2 in transgenic mice. Proc Natl Acad Sci USA 89, 3899-3903.

311

Wykes and Krawetz: Gene potentiation in chromatin domains Wei X, Samarabandu J, Devdhar R, Siegel A, Acharya R, and Berezney R (1998) Segregation of transcription and replication sites into higher order domains. Science 281, 1502-1505. Weintraub H, and Groudine M (1976) Chromosomal subunits in active genes have an altered conformation. Science 193, 848-856. Willison K, and Ashworth A (1987) Mammalian spermatogenic gene expression. Trends Genet 3, 351-355. Wolffe A (1994) Inheritance of chromatin states. Dev. Genet. 15, 463-470. Wood W (1996) The complexities of !-globin gene regulation. Trends Genet 12, 204-207. Wouters-Tyrou D, Martinage A, Chevaillier P, and Sautiere P (1998) Nuclear basic proteins in spermiogenesis. Biochimie 80, 117-28. Wykes S, Nelson J, Visscher D, Djakiew D, and Krawetz S.A. (1995) Coordinate expression of the PRM1, PRM2 and TNP2 multigene locus in human testes. DNA Cell Biol. 14: 155-161. Yawata Y (Editor) (1996) Atlas of Blood Diseases: Cytology and histology Martin Dunitz Ltd. London. Pp 3-6. Zlantanova J, and van Holde K (1992) Chromatin loops and transcriptional regulation. Crit Rev Euk Gene Express 3, 211-224.

Stephen A. Krawetz 312

Gene Therapy and Molecular Biology Vol 4, page 313 Gene Ther Mol Biol Vol 4, 313-322. December 1999.

Glycine clock: Eubacteria first, Archaea Protoctista, Fungi, Planta and Animalia at last

next,

Research Article

Edward N. Trifonov Department of Structural Biology, The Weizmann Institute of Science, Rehovot 76100, Israel __________________________________________________________________________________________________ Correspondence: E. N. Trifonov, Department of Structural Biology, The Weizmann Institute of Science, Rehovot 76100, Israel. Fax: +972 8 934 2653; E-mail edward.trifonov@weizmann.ac.il Key words: evolutionary trees, triplet code, earliest proteins, amino-acid chronology, amino-acid composition, molecular evolution, codon chronology, primordial soup, multiple alignments, Received: 15 April 1999; accepted 25 April 1999

Summary Twenty-five different single-factor criteria and hypotheses about chronological order of appearance of amino acids in the early evolution are summarized in consensus ranking. All available knowledge and thoughts about origin and evolution of the genetic code are thus combined in a single list where the amino acids are ranked in descending order, starting with the earliest ones: G, A, D, V, P, S, E, L, T, I, N, F, H, K, R, Q, C, M, Y, W O n e m a y e x p e c t that i n the composition o f the ancient proteins the earliest amino acids would dominate. Indeed, when homologous prokaryotic and eukaryotic protein sequences are aligned, the most frequent residue amongst matching amino acids (presumably, what remains o f the common ancestor sequence) is glycine that makes about 14% vs. glycine content of 6-7% in modern proteins. T h e g l y c i n e c o n t e n t o f t h e m a t c h i n g r e s i d u e s m a y , t h e n , s e r v e a s a m e a s u r e o f t h e time (glycine clock) since the separation of compared species. This approach is applied to 370 pairwise alignments of protein sequences from over 100 species of 6 major kingdoms. The evolutionary tree is derived, where the kingdoms separate consecutively from the central stem in the order: Eubacteria (13.5% G at the moment of separation), Archaea (11.5%), Protoctista (10.5%), Fungi (9%), Planta/Animalia (8%), largely consistent with common knowledge on the evolution of the kingdoms. The glycine content, thus, may serve as a time label that allows the tracing back of the separation of any two species with potential accuracy of the order of 50 to 100 million years, all the way to the very origin of species.

desirable to find some internal property(ies) of the sequences that would indicate their evolutionary age. One such property is suggested by the recently derived chronological ranking of amino acids, order of their appearance on the early evolutionary scene (Trifonov and Bettecken, 1997; Trifonov, 1999). The earliest amino acids should have been overrepresented in the earliest proteins, in which case mere amino-acid composition could serve as the indicator of the age of the protein. This approach, however, can not be used in as straightforward way, since all extant proteins are of the same age, if one assumes that the proteins originate from their immediate and distant ancestors, rather than formed de novo (Zuckerkandl, 1976). One way to evaluate the aminoacid composition of the proteins of the distant past is to compare (align) related sequences from evolutionary distant species and take the composition of shared residues. As it is

I. Introduction The molecular clocks of which many sophisticated versions had been developed since original suggestion by Zuckerkandl and Pauling (1962), suffer from numerous drawbacks (see, e. g., Doolittle, 1997; Ayala et al., 1998), especially when applied to very early molecular events. In particular, the evolutionary rates are not constant, the distance estimates are influenced by horizontal transfer, and double (multiple) replacements are difficult to account for. The quantitative evaluations of similarity in the sequence comparisons become unreliable when too little of a common ancestor is left in the sequences. Moreover, the sequence dissimilarity indicates evolutionary distance between the sequences, but the time direction remains uncertain, resulting in so-called unrooted evolutionary trees. It would be highly

313

Trifonov: Glycine clock described below, the "common" composition of eukaryotic and prokaryotic sequences (evolving separately about 3 Gyrs), indeed, is strongly biased towards the earliest amino acids, in particular, glycine. This suggests to use the glycine content as measure of time (glycine clock) passed since separation of the species, to construct the rooted evolutionary tree.

II. Results and discussion A. Amino-acid composition of early proteins The earliest form of the triplet code has been recently reconstructed, consisting of 10 codons and 7 respective amino acids: ala, asp, gly, pro, ser, thr and val (Trifonov and Bettecken, 1997). The reconstruction was based on natural expandability of (GCT)n sequences, and on universal (GCU)n pattern hidden in mRNA sequences (Lagunez-Otero and Trifonov, 1992). This suggested that the very first triplets were GCU and itâ&#x20AC;&#x2122;s 9 point change derivatives. The reconstruction of the above list of the earliest amino acids was based on the experiments of S. L. Miller (1987), on chemical simplicity of the amino acids and on association with more ancient class II aminoacyl-tRNA synthetases. Inspection of the table of the triplet code revealed a striking correspondence between all these residues and the GCUderived codons (Trifonov and Bettecken, 1997). This gives reason to believe that the earliest proteins, perhaps, long time before the separation of eukaryotes from prokaryotes, had been built from the above 7 ancient residues. At later stages, with appearance of other amino acids the domination of the seven, surely, was compromised. However, one could expect that even at the stage of separation eukaryotesprokaryotes some of the ancient residues still prevailed. Further insight into the amino-acid chronology is provided by adding to the analysis four more criteria of the aminoacids' evolutionary age, in addition to the above three: frequency of occurrence of various amino acids in modern proteins, stability of the codon-anticodon interactions, chemical inertness of amino acids, and the GCU triplet-based list of the amino acids, as an independent criterion. Ranking analysis of the seven "chronologies" suggested by these criteria (Trifonov, 1999) resulted in the following list of the amino acids, in descending order of their appearance on the evolutionary scene: ala, gly, ser, pro, val, thr, leu, asp, ile, glu, asn, phe, lys, arg, gln, cys, his, met, trp and tyr. The earliest proteins, therefore, would be expected to contain less of the latest amino acids, say, gln, cys, his, met, trp and tyr. As a matter of fact, these residues, indeed, are least frequent even in extant proteins (see Figure 1), but the early proteins, perhaps, had even less of these residues. This is checked by alignment of prokaryotic and eukaryotic sequences and comparing amino-acid composition of the common parts (points) to the composition of modern eukaryotic and prokaryotic proteins. In extension of an earlier work (Trifonov, 1998) this analysis is performed on 70 arbitrarily chosen functionally different aligned sequence pairs (Table 1), scoring total 5551 matching residues. The actual

314

scores and amino-acid compositions in % are presented in the Table 2 and in the Figure 1 (under "common"). In this Figure the composition values for prokaryotic and eukaryotic proteins (two upper plots) are taken from Arques and Michel, (1996). The histograms presented in the F i g u r e 1 show, first of all, that in the common (about 3 billion years old) eukaryotic-prokaryotic material the gly residues are significantly more frequent (about twice) than in modern proteins. This major bias is observed even when only 10 sequence pairs are taken for the analysis. In the T a b l e 2 amino-acid compositions for 7 different sets, 10 sequence pairs each, are presented (sequence Nos. 1-10, 11-20, 21-30, ... 61-70 of the T a b l e 1). In all cases the domination of glycine is obvious: 12.7 to 15.5 % versus 6 - 7% in modern proteins. To be sure that the bias is not due to overrepresentation of some species, E. coli in particular (33 sequence pairs), two sets have been assembled, one dominated by E. coli sequences (set 7) and another one - with E. coli sequences underrepresented (set 6). The content of gly is found to be high in both cases. Total of 27 different prokaryotic species and 32 eukaryotic species are represented in the 70 sequence pairs analyzed (Table 1). The effect, therefore, is general, apparently reflecting, indeed, the aminoacid composition of the proteins at the moment of separation between prokaryotes and eukaryotes. If the ratios of the occurrences in "common" to the occurrences in prokaryotes and in eukaryotes are considered, then two more amino acids appear on the top: asp and pro (about 20% excess). All three including glycine belong to the earliest alphabet. That is, the earliest amino acids have been still overrepresented at the time of separation eukaryotes-prokaryotes. Glycine, aspartic acid and proline are known to be the most specific residues for the turns of folded polypeptide chains (Kwasigroch et al., 1996). Their unusual conservation, thus, indicates that the turns are no less important in maintaining conserved protein structure than alpha-helices and beta-sheets. Another conspicuous feature of the "common" distribution (F i g u r e 1 ) is an abrupt drop of composition values for the amino acids tyr, asn, his, gln, met, trp and cys. Five of them belong to the latest in the amino-acid chronology (Trifonov, 1999, and manuscript in preparation). It appears, thus, that about 3 billion years back these "young" residues have been just entering the scene being, therefore, substantially less numerous than the "older" residues. Their share in the total, according to our data, was 10.7%, versus 30% for even distribution of amino acids. No such step in the amino-acid composition is observed in case of modern proteins (Figure 1, upper plots) though the "young" residues are underrepresented here as well. It appears, thus, that since the time of separation eukaryotes-prokaryotes the proportion of the "young" residues increased, apparently, in the process of their gradual accommodation and optimization of the protein composition. The proportion of the latest residues as well as excess of the earliest glycine residues may, thus, potentially serve for timing of the evolutionary bifurcations.

Gene Therapy and Molecular Biology Vol 4, page 315

Figure 1. Amino-acid composition of matching residues in alignments of related prokaryotic and eukaryotic protein sequences ("common") as compared to modern proteins of prokaryotes and eukaryotes.

B. The amino-acid and codon chronology

Exceptional status of glycine in molecular evolution has been indicated earlier in the study on the correlation of the evolutionary rate with the amino-acid composition (Graur, 1985). An "almost uninterchangeable" glycine was found to be "one of the most conserved amino acids". This also suggests higher content of glycine in the older, conserved proteins. Being the smallest amino acid glycine serves very much as a hinge in the polypeptide chain providing it with high flexibility. The conformational versatility would be of high importance in the early stages of protein evolution. Later on, perhaps, with advance in sophistication of the protein structure rather stability of the evolved conformations became important, and the glycine content eventually came down to the modest present level.

More extended analysis involving 25 different amino-acid age criteria (manuscript in preparation) arrives to the chronology very similar to the one listed above. A vertical column on the left of the Figure 2 represents the order of the amino acids, in which they, presumably, appeared on the evolutionary scene. All available knowledge and thoughts about origin and evolution of the genetic code are combined in this single list where the amino acids are ranked in descending order, starting with the earliest ones. The ranking is inevitably of rather poor accuracy. The typical differences in the calculated ranks as compared with the earlier 7-criteria list are 1-2 ranks.

315

Trifonov: Glycine clock Table 1. Aligned prokaryotic-eukaryotic protein sequence pairs.

Species

Protein (gene)

Reference

1. Escherichia coli human

thymidilate synthase --“--

Gene 150, 221, 1994

2. Halobact. cutirubrum C. elegans 3. Bacteroides fragilis maize 4. Flavobact. meningosepticum pig 5. Escherichia coli rabbit 6. Bacillus circulans Brugia malayi (nematode) 7. Enterococcus faecium carrot 8. Agrobact. tumefaciens X. laevis 9. Escherichia coli human 10. Escherichia coli human 11. Escherichia coli mouse 12. Escherichia coli tomato 13. Clostridium acetobutylicum C. elegans 14. Alcaligenes Arabidopsis thaliana 15. Escherichia coli Arabidopsis thaliana 16. Pseudomonas Aspergillus nidulans 17. Escherichia coli rat 18. Escherichia coli mouse

hypothetical G-protein --“-pyruvate dikinase --“-prolyl endopeptidase

Gene 151, 153, 1994

phosphofructokinase ATP-dep phosphofructokinase chitinase A3 chitinase dihydrofolate reductase --“-Arginase --“-ribosomal protein S1 --“-, repeat 2 glutathione reductase --“-ribose 5-phosphate isomerase --“-RNase I RNase LE 3-hydroxyacyl CoA dehydrogenase --“-(F54C8.6) Nitrilase --“-adenine phosphorybosyltransferase --“-NAD-dep. formate dehydrogenase --“-arginyl-tRNA synthetase --“--

Gene 152, 181, 1995

RNA polymerase subunit ! RNA polymerase I/III AC40

Gene 167, 203, 1995

19. Escherichia coli C. elegans

RNA polymerase subunit ! RNA polymerase III AC16 valine-tRNA synthetase --“-thermolysin microsomal endopeptidase inosine monophosphate dehydrogenase --“-methylenomycin A resistance protein glucose transporter type I NARK nitrate transporter CRNA nitrate transporter SapT (sakacin synthesis) multidrug resistance protein S-adenosylhomocysteine hydrolase --“-3-methyladenine DNA glycosylase --“--

Gene 172, 211, 1996

20. B. stearothermophilus Plasmodium knowlesi 21. B. cereus rabbit 22. B. subtilis mouse 23. B. subtilis human 24. Escherichia coli Aspergilus nidulans 25. Lactobacillus sake Chinese hamster 26. Rhodobacter capsulatus Triticum aestivum 27. B. subtilis rat

Gene 151, 173, 1994 Gene 152, 103, 1995

316

Gene 153, 147, 1995 Gene 154, 7, 1995 Gene 154, 115, 1995 Gene 155, 231, 1995 Gene 156, 123, 1995 Gene 156, 191, 1995 Gene 158, 203, 1995 Gene 160, 309, 1995 Gene 161, 15, 1995 Gene 161, 81, 1995 Gene 162, 99, 1995 Gene 164, 347, 1995

Gene 173, 137, 1996 Gene 174, 135, 1996 Gene 174, 209, 1996 Gene 175, 223, 1996 Gene 175, 223, 1996 Gene 176, 55, 1996 Gene 177, 17, 1996 Gene 177, 229, 1996

Gene Therapy and Molecular Biology Vol 4, page 317 28. Escherichia coli rice 29. Escherichia coli red alga 30. B. subtilis Arabidopsis thaliana 31. Pseudomonas putida human 32. Escherichia coli mouse 33. Escherichia coli rabbit 34. Zymomonas mobilis human 35. Zymomonas mobilis human 36. B. megaterium human 37. Escherichia coli Brassica napus 38. B. subtilis rat 39. B. subtilis X. laevis 40. Escherichia coli dog 41. Rhizobium leguminosarum D. discoideum 42. B. subtilis human 43. Escherichia coli Schistosoma mansoni 44. Streptococcus mutans tobacco 45. Staphylococcus xylosus human 46. Escherichia coli human 47. Escherichia coli human 48. Synechococcus barley 49. Escherichia coli D. melanogaster 50. P. aeruginosa T. bruce i 51. B. subtilis Geodia cydonium 52. Legionella pneumophila bovine 53. Thermus aquaticus mouse 54. M. genitalium tobacco 55. Synechococcus elongatus Chlamydomonas reinhardtii

Mrp (ATPase) EST D25016 (ATPase) 3-ketoacyl-acyl carrier prot. synthase --“-protoporphyrinogen oxidase --“-glyoxalase I --“-spermidine synthase --“-glutaredoxin --“-glyceraldehyde-3-phosphate DH --“-phosphoglycerate kinase --“-triosephosphate isomerase --“-phosphoenolpyruvate carboxykinase --“-peptidylprolyl cis-trans isomerase --“-Arginase --“-signal peptidase I --“-orotate phosphorybosyltransferase --“-myo-inositol 2-dehydrogenase biliverdin reductase cold-shock protein CSPA Y-box binding protein non-phosphorylating GAPN --“-histone deacetylase (acuC) --“-(HDm) heat-shock protein HSP 60 --“-porphobilinogen deaminase --“-HemL protein --“-RNA helicase --“-mercuric reductase trypanothione reductase alcohol dehydrogenase AidB-like protein Cu, Zn superoxide dismutase --“-DNA polymerase (5'-3' exonucl.domain) flap endonuclease (FEN-1) uracil phpsphoribosyltransferase --“-photosystem II RC domain --“--

317

Gene 178, 97, 1996 Gene 182, 45, 1996 Gene 182, 169, 1996 Gene 186, 103, 1997 Gene 187, 35, 1997 Gene 188, 23, 1997 Gene 188, 221, 1997 Gene 188, 221, 1997 Gene 188, 221, 1997 Gene 192, 235, 1997 Gene 193, 65, 1997 Gene 193, 157, 1997 Gene 194, 249, 1997 Gene 195, 329, 1997 Gene 196, 209, 1997 Gene 198, 5, 1997 Gene 198, 237, 1997 Gene 198, 275, 1997 Gene 199, 83, 1997 Gene 199, 231, 1997 Gene 199, 231, 1997 Gene 199, 241, 1997 Gene 200, 163, 1997 J. Mol. Evol. 47, 343, 1998 J. Mol. Biol. 274, 408, 1997 J. Biol. Chem. 272, 28531, 1997 EMBO J. 17, 3219, 1998 J. Mol. Biol. 280, 1998

Trifonov: Glycine clock 56. Rhodobacter capsulatus tobacco 57. Streptomyces hydrogenans Drosophila lebanonensis

uroporphyrinogen decarboxylase --“--

EMBO 17, 2463, 1998 J. Mol. Biol. 282, 383, 1998

58. Escherichia coli C. elegans 59. T. thermophilus human 60. Escherichia coli D. melanogaster 61. Escherichia coli human

3!,20"-hydroxysteroid dehydrogenase alcohol dehydrogenase transition metal transporter --“-histidyl-tRNA synthetase --“-pspE HSP67Bb GTP-binding protein (FtsY) --“-(SR!)

62. Escherichia coli rabbit 63. Escherichia coli D. melanogaster 64. Escherichia coli rat 65. Escherichia coli Brugia malayi 66. Escherichia coli human 67. Escherichia coli rice 68. Escherichia coli human 69. Escherichia coli Flaveria trinervia 70. Escherichia coli human

trehalase --“-parvulin Dodo protein aminopeptidase N --“-asparaginyl-tRNA synthetase --“-glutathione S-transferase --“-thioredoxin glutaredoxin glutaredoxin thioredoxin phosphoenolpyruvate carboxylase --“-periplasmic cyclophilin cyclophilin A1

Gene 202, 69, 1997

J. Biol. Chem. 272, 28485, 1997 J. Mol. Biol. 280, 847, 1998 J. Mol. Biol. 282, 195, 1998 Gene 201, 37, 1997

Gene 203, 89, 1997 Biochemistry 37, 686, 1998 EMBO J. 17, 2947, 1998 J. Mol. Biol. 271, 135, 1998 J. Mol. Biol. 281, 949, 1998 J. Mol. Biol. 281, 949, 1998 J. Mol. Evol. 46, 107, 1998 EMBO J. 17, 2463, 1998

of the list. Actually, they appear simultaneously, within the accuracy of the ranking (manuscript in preparation). The apparent contradiction, however, rather suggests a correction to the GCU-model. As it was indicated in the paper on the GCU theory (Trifonov and Bettecken, 1997), the GCC triplet and its point change derivatives correspond to the same seven earliest amino acids. The first codons, thus, could be, indeed, GCC and GGC, for alanine and glycine, respectively, in accordance with the chronology displayed in the Figure 2. This pair of codons has been suggested as the earliest ones 20 years ago by Eigen and Schuster (1978). What is important for the elaboration in the next section - the glycine is one of the earliest amino acids. It apparently took over at some time in the early evolution becoming a dominant residue (see Figure 1).

Despite this uncertainty, due to consensus nature of the chronology it has several important properties not visible in individual rankings by any of the initial criteria. The conclusion of the earlier GCU-based theory on the structure of the earliest code is confirmed: all 7 earliest amino acids are, indeed, found at the top of the consensus chronology (G, A, D, V, P, S and T). Ten amino acids of the Miller's imitation of primordial soup are all ranked as topmost (G, A, D, V, P, S, E, L, T, I). This result is especially important, since it confirms that, indeed, the experimental conditions chosen by Miller are close to the primordial ones, and that the first amino acids acquired by the emerging life were synthesized abiotically. The consensus order of appearance of the 20 amino acids on the evolutionary scene also reveals a unique and simple chronological organization of 64 codons, that could not be figured out from individual criteria: new codons appear in complementary pairs, with the complement recruited from the codon repertoire of the earlier or simultaneously appearing amino acids. The resulting codon chronology also reveals that of alternative codon-anticodon pairs the most stable ones appear first, if not all together. Contrary to the GCU-based theory of the origin of the code, it is glycine rather than alanine that appears at the top

C. Glycine clock and evolutionary tree for six major kingdoms. The calculations similar to those made for the prokaryotes and eukaryotes, as presented in the T a b l e s 1 and 2, are performed for sequence pairs from 6 major kingdoms: eukaryotes (Protoctista, Fungi, Planta and Animalia) and prokaryotes (Eubacteria and Archaea). Total 370 sequence pairs are analyzed, and the average contents of 318

Gene Therapy and Molecular Biology Vol 4, page 319 the glycine amongst the shared residues are calculated for each of 15 groups of the kingdom-to-kingdom sequence comparisons. The functionally diverse sequences are taken from literature, basically, on the random basis. They represent as large variety of species, as exemplified by the Table 1. In the Table 3 the derived values are presented, together with actual scores (in brackets, glycine/total). The number of sequence pairs used for the analysis is indicated as well (italics). The errors are calculated on the assumption that the scatter in the actual scores of glycines follows normal distribution with STD equal to square root of the score. The highest contents of glycine among the shared residues of the aligned sequences is observed for Eubacteria (see Table 3). The respective % GLY values vary between 12.1 ± 1.2% and 14.8 ± 0.6% with the average 13.7 ± 0.3%. If only eukaryotes are taken for the alignments with the eubacterial protein sequences, as in the T a b l e 2, the average % GLY value from the new set of the sequences is 1460/10602 = 13.8 ± 0.4%, to compare with 14.3 ± 0.5% for the earlier set (T a b l e 2), indistinguishable within the error bars. The % GLY values for Archaea, compared to four eukaryotic kingdoms, vary between 11.3 ± 1.0% and 13.3 ± 1.5%, with the average 11.7 ± 0.6%, clearly lower than the above average value for Eubacteria. That would correspond to a later separation of the Archaea from eukaryotes, some time after Eubacteria. The % GLY value for separation ArchaeaEubacteria, on the other hand, is close to the separation level for Eubacteria, as it would be expected, 12.8 ± 0.9% vs.

13.7 ± 0.3%. Similarly, the % GLY values for later separations of Protoctista, Fungi and Planta are progressively lower, while comparisons of their sequences with older kingdoms give higher % GLY values, corresponding, respectively, to the separation times of the latter. The % GLY values are arranged in the Table 3 in such a way that the line averages of the values provide the branching level of % GLY for respective kingdoms. Of 15 kingdom-to-kingdom % GLY values only 3 (< 32% of 15) are more than 1 STD off the respective averages, which, thus, justifies the assumed normal distribution of the % GLY estimates. The evolutionary tree based on the % GLY values presented in the Table 3 is shown on the Figure 3. This tree is very much consistent with the trees derived from molecular clock calculations (Feng et al., 1997; Doolittle, 1997; Otsuka et al., 1999). If the time separation between branchings of plants and of Eubacteria is taken equal 2 Gyrs, 1% GLY corresponds to about 350 Myrs. This provides an approximate calibration of the glycine clock. At this early stage of the development of the glycine clock the linear calibration is an understandable simplification. Both the Table 3 and the Figure 3 represent the first estimates of the branchings of the major kingdoms, based on only 370 sequence pairs. The number of the sequences can be substantially increased (say, to many thousands), so that the tree would be subject of further improvements towards better accuracy. However, as the current error bars indicate, the overall topology of the basic tree will most likely stay unchanged.

Table 2. Amino-acid composition of common residues in eukaryotic-prokaryotic sequence alignments

319

Trifonov: Glycine clock

F i g u r e 2 . Chronology of 32 codon pairs. The amino-acid chronology is calculated as average ranking based on 25 different criteria. The codon chronology is one simple way of arranging the 64 triplets in accordance with the amino-acid chronology. Of alternative codons those which make most stable codon-anticodon pairs are engaged first (bold). In this case there is always a complementary triplet available, of the codon repertoires for earlier amino acids.

Table 3. Contents of shared glycine (%) in kingdom-to-kingdom protein sequence alignments ANIMALIA

PLANTA

FUNGI

PROTOCTISTA

PLANTA

8.1± 0.6 (193/2194, 25)

FUNGI

8.88±0.4 (573/6479, 70).

9.1±0.7 (179/1977, 23)

PROTOCTISTA

11.1±1.1 (98/879, 11)

9.8±0.8 (156/1595, 10)

11.4±1.0 (137/1200, 11)

ARCHEA

11.3±1.0 (128/1133, 18)

11.7±1.7 (49/418, 12)

11.3±1.0 (132/1170, 19)

13.3±1.5 (82/616, 8)

EUBACTERIA

14.8±0.6 (584/3935, 63)

13.1±0.7 (313/2381, 21)

13.4±0.6 (468/3502, 46)

12.1±1.2 (95/784, 10)

ARCHEA

Branching level 8.1± 0.6 (193/2194, 25) 8.9±0.3 (752/8456, 93)

320

10.6±0.5 (391/3674, 32) 11.7 ±0.6 (391/3337, 57) 12.8±0.9 (187/1462, 23)

13.7±0.3 (1647/12064, 163)

Gene Therapy and Molecular Biology Vol 4, page 321 It is noteworthy that the glycine clock approach (or, presumably, any other approach based on the content of the earliest amino acids) apparently provides both evolutionary distance (in % GLY time units in this case) and directionality (the larger the branching % GLY value the older the separation event). This would allow to construct a detailed rooted tree, with further subdivisions of the kingdoms and potential resolution of 50 to 100 Myrs, the higher the more sequences are taken for the alignments. The technique is especially promising in dating the earliest separations where sensitivity of the classical molecular clock is low. The tree in the Figure 3 is presented in its simplest form, with the central stem from which the respective kingdoms separate in the chronological order as indicated. Animalia rather than Planta are chosen to crown the tree, to reflect the obvious trend displayed by the tree - from the simplest to the most complex. Indeed, anuclear prokaryotes separate first, followed by the nucleated eukaryotes. The eukaryotes, on the other hand, progress from unicellular to multicellular, differentiated organisms. In a way, at each stage the simpler forms separated from the stem that continued to evolve to yet more complex forms. In that sense the common ancestor of all kingdoms though, perhaps, as simple as Eubacteria at the moment of their separation, was omnipotent having carried all elements that later evolved into the higher complexity of younger kingdoms. The higher evolutionary potential stayed associated with the main stem at every next branching. The branches of the kingdoms in the Figure 3 are not continued to the top of the tree, to the typical and common modern 67% of GLY, although this is implied, in order to better reflect the linear succession of the branching events. Apart from appealing simplicity of the glycine clock, its directionality and applicability to the earliest branchings, this technique is substantially less dependent on the effects of horizontal transfer and variations in the evolutionary rates. These are averaged over large number of sequences that are taken for the calculations.

III. Sequences and methods

Figure 3. Evolutionary tree of major kingdoms, according to glycine clock estimates. The glycine content % GLY corresponds to the proteins existing at the moment of separation of respective kingdoms. The vertical bars at the separation points indicate current uncertainty of the estimates, dependent on the amount of the sequences compared.

References

The aligned prokaryotic-eukaryotic sequence pairs are collected from literature, irrespective of the alignment technique chosen by the authors of the original papers. To ensure random choice of the sequences, all alignments published in Gene, volumes 150 to 200, have been taken for the ensemble in the T a b l e 1, total 50 sequence pairs. Additional 20 pairs are collected from various sources, on random basis as well. For all 440 sequence comparisons used in this work only those sequence pairs are taken which are part of multiple alignments of no less than 4 sequences in each. Matching residues are scored which are separated by no more than 4 non-matching residues, with no gaps (local sequence similarity # 33%). Wherever possible, the sequence pairs are taken to represent as broad variety of species as the sequence data allow.

Arques, D. G., and Michel, C. J. (1 9 9 6 ) A complementary circular code in the protein coding genes. J . T h e o r . B i o l . 182, 45-58. Ayala, F. J., Rzhetsky, A., and Ayala F. J. (1 9 9 8 ) Origin of the metazoan phyla: molecular clocks confirm paleontological estimates. P r o c . N a t l . A c a d . S c i . U S A 95, 606-611. Doolittle, W. F. (1 9 9 8 ) Fun with genealogy. P r o c . N a t l . Acad. Sci. USA 94, 12751-12753. Eigen, M., and Schuster, P. (1978) The hypercycle. A principle of natural self- organization. Part C: The realistic hypercycle. Naturwissenschaften 65, 341-369. Feng, D.-F., Cho, G., and Doolittle, R. F. (1 9 9 7 ) Determining divergence times with a protein clock: update and reevaluation. P r o c . N a t l . A c a d . S c i . U S A 94, 1302813033. Graur, D. (1 9 8 5 ) Amino acid composition and the evolutionary rates of protein-coding genes. J . M o l . E v o l . 22, 53-62.

321

Trifonov: Glycine clock Kwasigroch, J.-M., Chomilier, J., and Mornon, J.-P. (1 9 9 6 ) A global taxonomy of loops in globular proteins. J . M o l e c . B i o l . 259, 855-872. Lagunez-Otero, J., and Trifonov, E. N. (1 9 9 2 ) mRNA periodical infrastructure complementary to the proof-reading site in the ribosome. J . B i o m o l . S t r u c t . D y n a m . 10, 455-464. Miller, S. L. (1987) Which organic compounds could have occurred on the prebiotic earth. C o l d S p r . H a r b . S y m p . Q u a n t . B i o l . 52, 17-27. Otsuka, J., Terai, G., and Nakano, T. (1 9 9 9 ) Phylogeny of organisms investigated by the base-pair changes in the stem regions of small and large ribosomal subunit RNAs. J . M o l . E v o l . 48, 218-235. Trifonov, E. N. (1 9 9 8 ) How basics of protein evolution could help the gene finding. Proceedings of the First International Conference on Bioinformatics of Genome Regulation and Structure BGRS'98, Novosibirsk - Altai Mountains, August 24-31, 1998, ICG, Novosibirsk, v.2. B i o i n f o r m a t i c s o f Genome Structure, pp. 266-268 Trifonov, E. N. (1 9 9 9 ) Elucidating sequence codes: three codes for evolution. A n n a l s N Y A c a d . S c i . , in press Trifonov, E. N., and Bettecken, T. (1 9 9 7 ) Sequence fossils, triplet expansion, and reconstruction of earliest codons. Gene 205, 1-6. Zuckerkandl, E. (1 9 7 5 ) The appearance of new structures and functions in proteins during evolution. J . M o l . E v o l . 7, 1-57. Zuckerkandl, E., and Pauling, L. (1 9 6 2 ) Molecular disease, evolution and genetic heterogeneity. In: Kasha, M., and Pullman, B., (eds.) H o r i z o n s i n B i o c h e m i s t r y . Academic Press, New York, pp. 189-225.

Ed Trifonov 322

Gene Therapy and Molecular Biology Vol 4, page 323

323

Gene Therapy and Molecular Biology Vol 4, page 323 Gene Ther Mol Biol Vol 4, 323-338. December 1999.

Nuclear prostaglandin receptors Review Article

Mousumi Bhattacharya1, Daya R. Varma1, and Sylvain Chemtob1,2 1

Department of Pharmacology & Therapeutics, McGill University, Montreal, PQ, H3G 1Y6; 2Departments of Pediatrics, Ophthalmology and Pharmacology, and Research Center of H么pital Ste. Justine, University of Montreal, Montreal, PQ; H3T 1C5. __________________________________________________________________________________________________ Correspondence: Dr. Sylvain Chemtob, M.D., Ph.D., Research Center of H么pital Ste. Justine, 3175 C么te Sainte-Catherine, Montreal, Quebec H3T 1C5 Canada. Tel: (514)-345-4692; Fax: (514)-345-4801; E-mail: chemtobs@ere.umontreal.ca Key words: PGE2, EP receptors, G proteins, signal transduction, gene transcription, cyclooxygenase, prostanoid transporter Received: 20 August 1999; accepted 16 September 1999

Summary Prostaglandins and thromboxane are ubiquitous compounds and play important roles in cardiovascular homeostasis, inflammation, reproduction, respiration, mitogenesis and gene transcription and so on. These actions of prostanoids are presumed to be mediated by plasma membrane receptors belonging to the superfamily of G protein-coupled receptors. However, several lines of evidence suggest prostanoids may also act at the nuclear level. Nuclei contain cyclooxygenases and other intermediates required for prostanoid synthesis and receptor-mediated responses. This review focuses closely on various signal transduction cascades that exist in the nuclear membranes, including the presence of other nuclear G protein-coupled receptor, and discusses the discovery of functional nuclear prostaglandin E2 receptor. These data add new dimensions to the functions and signaling mediated by prostaglandin receptors.

also influences mitogenesis (Hashimoto et al., 1997; Glantschnig et al., 1996), promotes growth and metastasis of tumors (Paoletti et al., 1989; Fulton et al., 1991) and modulates the transcription of many genes (Danesch et al., 1994; Gashler and Sukhatme, 1995; Umayahara et al., 1997; Paliogianni and Boumpas, 1996; Minghetti et al., 1997).

I. Introduction Prostaglandins (PG) and thromboxane, collectively named prostanoids, are products of arachidonic acid metabolism. Prostanoids produce numerous physiologic and pathophysiologic effects, regulating cellular processes in nearly every tissue. These compounds act as local hormones, acting in the vicinity of their site of production and function in an autocrine and/or paracrine manner to maintain local homeostasis. Prostanoids are widely distributed and can be formed by nearly every tissue and cell type; the same prostanoid has the ability to provoke different responses in various tissues (Campbell and Halushka, 1996). There are five physiologically important prostanoids, PGD2, PGE2, PGF2!, PGI2 and TXA2. PGE2 in particular, has a wide spectrum of physiological and pharmacological actions in diverse tissues which include effects on the immune, endocrine, cardiovascular, renal and reproductive systems as well as the contraction and relaxation of smooth muscle (Campbell and Halushka, 1996; Negishi et al., 1993a). PGE2 is one of the most abundant prostanoids in the brain (Leffler and Busija, 1985) and plays an important role in many cerebral functions particularly in the newborn (Leffler and Busija, 1987; Chemtob et al., 1996). PGE2

II. PGE2 receptor subtypes and signaling Prostanoids exert their effects through GTP-binding protein (G protein)-coupled, rhodopsin-type receptors. The receptors for PGE 2 are termed EP, which include EP1, EP 2, EP3 and EP4 subtypes (Coleman et al., 1994). High specific PGE2 binding has been observed in the brain, kidney, uterus, liver, thymus and the adrenal medulla (Robertson, 1986). PGE2 has versatile and opposing actions due to multiple EP receptor subtypes and the coupling of EP receptor isoforms to a variety of signal transduction pathways (Table 1). Molecular cloning of these receptor cDNAs has revealed that the EP3 receptor isoforms are generated by alternative mRNA splicing (Narumiya, 1996).

323

Bhattacharya et al: Nuclear Prostaglandin Receptors Table 1 Classification of prostanoid EP receptors and their signal transduction. Ligand

Type

Subtype

Isoform

G protein

Signal Transduction

PGE2

EP1

PI", Ca2+ "

Gq (?)

Ca2+ # (in co-expression with

rEP1variant EP1) EP2

cAMP "

EP4

cAMP "

EP3A

cAMP #

EP3B

cAMP "

EP3C

cAMP "

EP3D

Gi/s/q

cAMP # , cAMP", PI"

EP3

Modified and summarized from Coleman et al., (1994) and Narumiya, (1996). Data obtained from receptors of various species; the EP3 isoforms from bovine; the rEP1 variant from rat; others from mouse. PI denotes phosphoinositol turnover.

EP1 receptors mediate Ca2+ mobilization (Watabe et al., 1993; Funk et al., 1993; Okudu-Ashitaka et al., 1996) by activating phospholipase C (PLC) and increasing inositol 1,4,5-trisphosphate (IP3) (Katoh et al., 1995; Suba and Roth, 1987). However, PGE2 has also been shown to stimulate Ca2+ mobilization by activating EP1 without altering IP 3 formation in RCCT cells (Hebert et al., 1991), and without activating phospholipase C (PLC) in myometrial cells (Asboth et al., 1996). An alternative splice variant has been identified for rat EP1 (rEP1-v), which completely lacks a cytoplasmic carboxy terminus (Okudu-Ashitaka et al., 1996). EP1 attenuates the action of PGE2 on tissues by interfering with signaling mediated by other EP receptors and constitutes an example of crosstalk between receptor subtypes. EP2 receptors and EP4 receptors stimulate adenylyl cyclase by coupling to G s. EP3 receptors inhibit adenylyl cyclase (Namba et al., 1993) but some of its functions are mediated by other second messenger pathways. Several isoforms of EP3 receptors, produced by alternative splicing, have been found in bovine, human, rat, mouse and rabbit tissues (Narumiya, 1996). These isoforms of EP3 receptors perform different functions but differ only in the carboxy terminal tail, which is known to influence the coupling selectivities and activities of G proteins (Pierce et al., 1995). However, a recent study suggests that this is not the only determinant and that regions in both the third intracellular loop and in the carboxyl termini of the prostanoid receptors contribute to the specificity of receptor-G protein interactions (Neuschafer-Rube et al., 1997). The constitutive activity of EP3 receptor isoforms and their coupling to Gi (Negishi et al., 1996) as well as their response to prolonged exposure to agonist also differ (Negishi et al., 1993b). The EP3 receptor can also couple to voltage-sensitive and insensitive Ca2+ channels (Tanaka et al., 1998) and Clchannels (Sakai et al., 1995).

The biological actions of PGE2 have been attributed to result from its interaction with cell surface EP receptors (Coleman et al., 1994). However, several lines of evidence suggest that prostaglandins may act intracellularly and exert a direct nuclear action. Recent studies have shown that cyclooxygenases which synthesize prostanoids are located mainly in the nuclear membrane (Spencer et al., 1998; Morita et al., 1995). Phospholipase A2 (PLA2), which releases arachidonic acid is activated at nuclear membranes (Schievella et al., 1995). In addition, a transporter which mediates the influx of prostanoid has been identified (Schuster, 1998). It is thus possible that PGE2 may exert some of its effects via intracellular EP receptors as has been proposed by several workers (Goetzl et al., 1995; Morita et al., 1995; Smith, 1997). This article reviews the literature suggesting the nuclear action of prostanoids and the discovery of functional G proteincoupled receptors, such as the prostanoid EP receptors, at the nuclear membrane.

III. Mitogenic effects of PGE2 and modulation of gene transcription Substantial evidence suggests that PGE2 affects gene transcription and regulates growth and cell proliferation. Endogenous PGE 2 is important in regulating the growth of epithelial cells (Konger et al., 1998). PGE2 exhibits mitogenic activities in bone cells and stimulates DNA synthesis (Glantschnig et al., 1996). PGE2 is also involved in the growth and metastasis of tumors (Paoletti et al., 1989; Fulton et al., 1991). The inhibition of prostaglandin synthesis has been shown to result in growth retardation of tumors in experimental animals (Lupulescu, 1978). The importance of PGE2 in neoplastic development is also suggested by the association of a decreased risk of colon 324

Gene Therapy and Molecular Biology Vol 4, page 325 cancer and the use of nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin (Williams et al., 1997). It is of interest to note that PGE2 inhibited the antiproliferative effects of aspirin on growth factor-stimulated DNA synthesis (Castano et al., 1997). Cellular immediate-early genes such as c-fos and erg1 serve as nuclear couplers of early cytoplasmic events to long term alterations in gene expression and these genes are closely related to proliferation and/or differentiation (Gashler and Sukhatame, 1995). PGE2 regulates c-fos and erg-1 gene expression (Simonson et al., 1994; Danesch et al., 1994; Glantschnig et al., 1996). PGEs are also immunomodulatory agents and PGE2 inhibits nuclear transcription of interleukin 2 (IL-2) by decreasing the binding of transcription factors AP-1 and NF-AT. Agents that elevate cAMP also inhibit transcription of IL-2 gene; however, the IL-2 promoter lacks a cAMP response element (CRE) and therefore the mechanism remains to be shown (Paliogianni and Boumpas, 1996). The transcription factor CCAAT/Enhancer-binding protein $ (C/EBP $) is a PGE2 activated transcription regulator of the insulin-like growth factor-1 gene; however, C/EBP $ is not phosphorylated by protein kinase A (PKA) and the mechanism of activation is unclear (Umayahara et al., 1997). Other genes modulated by PGE2 include the inducible nitric oxide synthase (iNOS) (Milano et al., 1995; Minghetti et al., 1997) and the constitutive endothelial nitric oxide synthase (eNOS) via EP3 receptors (Dumont et al., 1998; 1999). The precise mechanism by which PGE2 affects gene expression is not fully known. Recently PGE2 was shown to inhibit lipogenic gene expression through a pertussis toxin (PTX)-sensitive G protein signaling cascade (Mater et al., 1998). Other studies showed that PGE2 signals through a novel cAMP response element binding protein/CRE pathway, which appears to be independent of cAMP generation (Audoly et al., 1999). The cAMP (Paliogianni and Boumpas, 1996; Audoly et al., 1999), protein kinase A (PKA) (Glantschnig et al., 1996; Umayahara et al., 1997), or protein kinase C (PKC) (Glantschnig et al., 1996) pathways are not obligatory for mediating the effects of PGE2. Recent studies have identified several aspects to EP3 receptor function distal to the G protein. An EP3 receptor-dependent activation of MAP kinase, which is followed by its translocation into the nucleus, has been shown by Burkey and Regan (1995). The involvement of the small GTPase Rho in the EP3 receptor-mediated stress fiber formation in kidney cells (Hasegawa et al., 1997) and neurite retraction (Katoh et al., 1996) has been identified. The latter occurs through a pathway, distinct from adenylyl cyclase inhibition or protein kinase C activation (Katoh et al., 1996). More recent studies have demonstrated that the stimulation of EP3 receptors induces the translocation of the transcription factor NF%B to the nucleus (Meyer-Kirchrath et al., 1998).

The discovery of nuclear EP receptors might now provide possible explanations for some of these observations.

IV. Nuclear phospholipase A2 (PLA2) and cyclooxygenase enzymes The presence in the nuclear region of the machinery for prostanoid synthesis favors the possibility of their nuclear actions. Prostaglandin synthesis is initiated by activation of PLA2. PLA 2 is one of the growing family of enzymes, which catalyze the hydrolysis of phospholipids at the sn-2 position, liberating free fatty acids including arachidonic acid, the precursor of platelet -activating factor, and other lysophospholipids (Campbell and Halushka, 1996). This arachidonate release step is the major site of regulation of prostanoid biosynthesis and activation of PLA2 is the rate-limiting step in this process. PLA2 exist in both calcium-dependent and independent isoforms. Among the various groups of PLA2, the importance of cytosolic PLA2 (cPLA2) in mediating the generation of prostanoids has been suggested (Mukarami et al., 1997). The most convincing data have come from studies using transgenic mice deficient in cPLA2; these studies demonstrate that PLA2 is essential for both the calcium- and lipopolysaccharide-induced PGE2 production (Bonventre et al., 1997). cPLA2 is a ubiquitously distributed enzyme which requires Ca2+ in nanomolar range for its activity and is activated via increases in intracellular Ca2+ (Kramer and Sharp, 1997). The Nterminal calcium-dependent lipid binding (CaLB) domain is responsibile for Ca2+-dependent translocation of cPLA2 from the cytosol to the perinuclear and endoplasmic reticular membranes. Translocation to nuclear membranes results in loss of arachidonic acid from this site verifying functional activation of cPLA2 at the nuclear membrane (Schievella et al., 1995; Peters-Golden et al., 1996). The extracellular or secretory PLA2 (sPLA2) may be involved in arachidonic release and prostaglandin production (Reddy and Herschman, 1996). sPLA 2 enzymes require millimolar concentrations of Ca2+ to exert their enzymatic action. Thus sPLA 2 is activated continuously by the levels of Ca2+ found in the extracellular environment and can mediate transcellular prostanoid biosynthesis since they are secreted extracellularly and can bind to cell surfaces of neighbouring cells (Mukarami et al., 1999). Data derived from sPLA2 -deficient mice reveal this enzyme does not play a crucial developmental role (Kennedy et al., 1995). The initial step in the synthesis of prostanoids from arachidonic acid is mediated by cyclooxygenase (COX, also called prostaglandin H synthase or prostaglandin enderoperoxide synthase) (Campbell and Halushka, 1996). COX converts arachidonic acid to PGH2, which is then acted upon by discrete prostaglandin synthases to yield different prostanoids (Smith, 1997).

325

Bhattacharya et al: Nuclear Prostaglandin Receptors COX enzymes are membrane-bound hemoproteins and include two isozymes, a constitutive form (COX-1) and an inducible form (COX-2); these isoforms function independently (Smith, 1997; Vane et al., 1998). Aspirin and other currently available NSAIDs inhibit both enzymes. At the subcellular level, both enzymes are located on the luminal surface of the endoplasmic reticulum and in inner and outer nuclear membranes (Morita et al., 1995; Spencer et al., 1998). COX-2 has also been shown to translocate to the nucleus in response to growth factors (Coffey et al., 1997). To date, a nuclear localization sequence (NLS) for COX has not been identified; it has been proposed that these proteins reach the nuclear membrane via lateral diffusion following synthesis in the endoplasmic reticulum (Spencer et al., 1998). The two COX isoforms play distinct roles in regulating arachidonic metabolism (Smith, 1997; Smith et al., 1997), and have distinctive roles in human biology and during development. COX-1 is constitutively and almost ubiquitously expressed and is responsible for the low prostaglandin synthesis required for cell homeostasis. COX-2 is an inducible enzyme, which is de novo synthesized in response to a wide range of extracellular and intracellular stimuli (cytokines, growth factors and tumor promoters) in the course of inflammation or other cellular stresses; it contributes to the generation of prostanoids at sites of inflammation and at certain stages of cell proliferation and differentiation. The induction of COX-2 expression and prostaglandin formation has been associated with the activation of mitogen-activated protein kinase (MAPK) and c-Jun-N-terminal kinase (JNK) pathways (Guan et al., 1998). Oxidant stress may also induce COX-2 since it can be activated by intracellular peroxides (Shitashige et al., 1998) whereas much higher peroxide levels are needed to activate COX-1 (Kulmacz and Wang, 1995). The brain is one of the few organs where COX-2 is constitutively expressed; it is expressed exclusively in neurons (Kaufmann et al., 1996) and is the primary isoform in the brains of the neonate (Peri et al., 1995). Prostanoids synthesized by COX-2 are essential for the survival of fetuses since the majority of offspring born to homozygous COX-2 knockout mice do not survive (Morham et al., 1995); whereas COX-1-null mice survive well (Langenbach et al, 1995). COX-2-null mice exhibit renal abnormalities, cardiac fibrosis and corpora lutea defects implicating the importance of this enzyme in the development of diverse organ systems (Morham et al., 1995). COX-2 expression has been shown to be protective against apoptosis (Von Knethen and Brune, 1997) and oxidant-induced injury of cardiomyocytes (Adderley and FitzGerald, 1999). However, elevated COX-2 expression has been associated with carcinogenesis (Kutchera et al., 1996; Tucker et al., 1999).

Cyclooxygenases utilizes different pools of arachidonic acid for synthesizing prostanoids; low concentrations of arachidonic acid are utilized predominantly by COX-2 whereas high concentrations are utilized preferentially by COX-1 (Reddy and Herschman, 1996; Shitashige et al., 1998). Two kinetically different prostanoid generating pathways, the immediate and delayed phases have been elucidated, implying the recruitment of different sets of biosynthetic enzymes, expression and activation which are tightly regulated by distinct transmembrane signalings (Mukarami et al., 1997; Naraba et al., 1998). The immediate phase of prostanoid biosynthesis occurs within several minutes of stimulation and is elicited by agonists that mobilize intracellular Ca2+; it is characterized by a burst release of arachidonic acid, and is mediated predominantly by COX-1. In the delayed phase, COX-2 dependent prostanoid biosynthesis proceeds over several hours in parallel with the induction of COX-2 expression following growth or proinflammatory stimuli (Reddy and Herschmann, 1997). Moreover, preferential coupling between particular PLA2 and COX has been suggested (Reddy and Herschman, 1997; Murakami et al., 1999). Thus, the perinuclear COX-2 plays a role in the prolonged generation of prostanoids that may act at nuclear sites and would be expected to modify nuclear events associated with cell differentiation and replication (Goetzl et al., 1995).

V. Transport of prostanoids across membranes Prostanoids are charged anions at physiological pH and diffuse poorly across biological membranes. Unlike hormones, these substances are not stored and they function as autocoids, binding to specific receptors on the same or nearby cells, signaling a wide variety of physiological functions. The plasma half-life of prostanoids is short (< 3 min) and their signal must be terminated locally since a single prostanoid molecule can signal diverse biological events, depending on the cell type. A prostaglandin transporter (PGT), which plays a primary role in mediating prostanoid transport and metabolic clearance, has been recently identified (Schuster, 1998). The PGT plays a role in the uptake of newly released prostanoids thus acting as a carrier across the plasma membrane before intracellular oxidation (Kanai et al., 1995; Lu et al., 1996). Moreover, PGT can facilitate intracellular actions of circulating as well as intracellularly produced prostanoids. The PGT mRNA is broadly expressed in diverse tissues, but its expression is most abundant in the lung, liver, brain and kidney. The PGT preferentially transports PGE2, PGE1, PGF2!, PGD2, with high affinity and to a lesser extent, TXB2 and PGI2 (Lu et al., 1996).

326

Gene Therapy and Molecular Biology Vol 4, page 327 signal (NLS) and is translocated to the nucleus only upon stimlution with angiotensin II; this was the first example of nuclear targeting of a G protien-coupled receptor upon binding of ligand. However, angiotensin II failed to cause nuclear translocation of AT1 in vascular smooth muscle cells or in astroglial cells (Lu et al., 1998). It has been suggested that the bradykinin B 1 receptors and the endothelin ETA and ET B receptors may also exist at the nuclear membrane and may modulate nuclear Ca2+ concentration (Bkaily et al., 1997a, b; Naik et al., 1998). Other G protein-coupled receptors have been detected in the perinuclear region, usually as a consequence of agonist-induced internalization, such as reported for somatostatin (Krisch et al., 1998), substance P (Grady et al., 1995), neurotensin (Faure et al., 1995; Castel et al., 1992) and other neuropeptides (Beaudet et al., 1998). Recent studies have also identified functional opioid receptors in isolated myocardial nuclei and these modulate opioid peptide gene transcription via PKC (Ventura et al., 1998).

VI. Signal transduction pathways at the nucleus It has generally been assumed that the signal transduction cascades are initiated at the plasma membrane leading to the stimulation of activities in target organelles such as the nucleus, and that the nuclear envelope played a passive role in this cascade. However, it is becoming increasingly clear that the nuclear envelope plays a major role in signaling cascades as summarized in the following discussion.

A. Nuclear G proteins and G proteincoupled receptors Receptors coupled to heterotrimeric G proteins comprise the largest known family of cell surface receptors (Gudermann et al., 1997). The significance of G proteins as a crucial link between plasma membrane receptors and intracellular events is well known and has been extensively studied (see reviews, Clapham, 1996; Hamm, 1998; Lefkowitz, 1998). Several recent studies have provided evidence for a similar role of both heterotrimeric and low molecular weight G proteins in nuclear signaling. Rubins et al. (1990) identified G proteins in the nuclear envelope of isolated rat liver nuclei. This finding was later confirmed by Takei et al. (1994) who showed that these nuclear G proteins were PTX-sensitive and involved in a pathway of nuclear protein transport. The presence of the small G protein RhoA in isolated nuclei has also been demonstrated (Balboa and Insel, 1995). Saffitz et al. (1994) showed the localization of Gs! in the nucleus of S49 lymphoma cells by immunoelectron microscopy. Nuclear translocation of Gi! and its association with nuclear chromatin has been shown to occur in response to insulin as well as serum and epidermal growth factors (Crouch, 1991). It has been recently reported that the heterotrimeric Gi protein translocates to the nucleus in response to growth factors where it regulates mitosis (Crouch and Simson, 1997). Until recently, there were a limited number of studies on functional G protein-coupled receptors that have been localized at the nucleus. The presence of muscarinic acetylcholine receptors was demonstrated by radioligand binding studies in isolated nuclei from rabbit corneal and Chinese hamster ovary cells (Lind and Cavanagh, 1993; 1995). A similar approach was used by two groups of researchers (Booz et al., 1992; Tang et al., 1992) to identify AT 1 subtype of angiotensin II receptors in isolated rat liver nuclei. Recently, the AT1 receptor has also been detected in nuclei of rat cardiomyocytes by immunocytochemistry techniques (Fu et al., 1998). Stimulation of nuclear angiotensin II receptors induces transcription of renin and angiotensinogen mRNA (Eggena et al., 1993). Lu et al. (1998) demonstrated that the AT1 receptor in brain neurons contains a nuclear localization

B. Nuclear calcium and inositol cycles Ca2+ signals in the nucleus control a variety of nuclear functions including gene transcription, DNA synthesis and repair, nuclear envelope breakdown or its reconstruction, protein transport and apoptosis (Malviya and Rogue, 1998). Ca2+ does not passively diffuse across the nuclear envelope, suggesting that regulatory mechanisms exist to modulate intranuclear Ca2+ levels (Nicotera et al., 1989). Nuclear Ca2+ signals are generated not only by Ca2+ transport into and out of the nucleoplasmic space, but also into and out of the nuclear envelope, which serves as the pool for nuclear Ca2+. Nuclear and cytosolic Ca2+ signals are differentially regulated and are independent of each other because of the presence of nucleocytoplasmic barrier to Ca2+ movement (Malviya and Rogue, 1998; Badminton et al., 1998). Increases in nuclear Ca2+ activate gene transcription by a mechanism that is distinct from gene regulation by cytoplasmic calcium signals. For example the expression of c-fos is differentially regulated upon increasing cytosolic or nucleoplasmic Ca 2+ (Hardingham et al., 1997). Nuclear Ca2+ concentration specifically controls Ca2+activated gene expression mediated by the cyclic-AMPresponse element (CRE) and the CRE-binding protein, CREB, which function as a nuclear Ca2+-responsive transcription factor. There is crosstalk between cytosolic and nuclear Ca2+ pools. The concentration of Ca2+ in the nucleus is lower (Al-Mohanna et al., 1994) or higher (Przywara et al., 1991; Waybill et al., 1991) than that in the cytosol depending upon the cell system studied, the method used, or the physiological state. In the cytosol, Ca2+ signals are produced by the release of Ca2+ from intracellular storage sites, mainly the endoplasmic reticulum; this mediated by IP3 and inositol

327

Bhattacharya et al: Nuclear Prostaglandin Receptors al., 1996). Numerous Ca2+ binding proteins such as calmodulin, calreticulin and calpain have been have been identified in the nucleus (Malviya and Rogue, 1998). Ca2+/calmodulin-dependent protein kinase (CaM kinase), which control gene expression through the phosphorylation of key regulatory sites on nuclear transcription factors such as CREB, has been detected within the nucleus. These kinases may serve to decode Ca2+ signals to the nucleus (Heist and Schulman, 1998).

1,3,4,5-tetrakisphosphate (IP4) (Berridge, 1993). The IP3 receptor acts as an IP3-gated Ca 2+ channel; it is not known if the IP 4 receptor functions as a Ca2+ channel (Humbert et al., 1996). Ca2+ mobilization can also be mediated by another second messenger, cyclic adenosine diphosphate ribose (cADP ribose) via ryanodine receptors, which are also an intracellular Ca2+ channels and present in the endoplasmic reticulum. The nucleus contains the necessary machinery for IP3 production and possesses distinct phosphoinositide cycles (Divecha and Irvine, 1995). Nuclear phosphoinositide cycles act independently from that of plasma membrane in that its activation takes place when the cytoplasmic cycle is not affected in response to external stimuli (Martelli et al., 1992). IP3 is generated by the breakdown of phosphatidylinositol 4,5-biphosphate (PIP2). Two interconnected pools of PIP2 have been identified in the nucleus, one in the nuclear membrane and the other located within the nucleus (Malviya and Rogue, 1998). Other components of the nuclear inositide cycles, including phospholipase C (&1 and $4 isoforms) which hydrolyze PIP2 to generate IP3 and diacylglycerol, are present within the nucleus (see below). Functional IP 3 and ryanodine receptors are located on the inner nuclear membrane; the outer nuclear membrane, which constitutes a continuum with the endoplasmic reticulum, is the site of the location of IP4 receptors and Ca2+ pump ATPase (Humbert et al., 1996). The nuclear IP3 receptors are not identical with the microsomal IP3 receptors (Matter et al., 1993) and play a role in the meiotic process in the mouse oocyte (Pesty et al., 1998). Upon binding to their receptors on the inner nuclear membrane, IP3 or cADP ribose open the Ca2+-release channels allowing selective release of Ca2+ into the nucleoplasm (Humbert et al., 1996; Gerasimenko et al., 1996). The nuclear IP3 receptor is phosphorylated by nuclear PKC, which is activated by diacylglycerol (Matter et al., 1993). The nuclear Ca2+-ATPase is an ATPmediated nuclear Ca 2+-transporter, which is responsible for filling the nuclear calcium pool and is functionally distinct from the endoplasmic reticulum Ca2+-ATPase (Humbert et al., 1996). Nuclear calcium uptake is also mediated by IP4, which is generated from IP3 by phosphorylation via a Ca2+/calmodulin-dependent IP3-3-kinase. High affinity nuclear IP4 receptors have been localized on the outer nuclear membrane and these are different from other IP4 receptors documented so far (Malviya and Rogue, 1998). Besides calcium channels (Bkaily et al., 1997a, b; Gerasimenko et al., 1996), zinc channels have also been identified in the inner nuclear membrane (Longin et al., 1997). Ca2+-activated K+ channels are localized in the outer nuclear membrane in isolated pancreatic nuclei which are only sensitive to changes in the nuclear Ca2+ concentration, and are activated by a relatively high Ca2+ concentrations (approximately 200 ÂľM) (Gerasimenko et

C. Nuclear protein kinase C (PKC) PKC is a family of serine/threonine kinases that is involved in the transduction of a wide variety of cellular signals (Nishizzuka, 1992; Buchner, 1995). PKC family comprises of at least eleven isoforms that can be divided into three groups, namely the Ca2+-dependent or conventional PKC, Ca2+-independent isoforms or novel PKC, and the atypical PKC. The activity of the conventional PKCs that are composed of the !, &, and ' isoforms is dependent on Ca2+, phospholipids and diacylglycerol, whereas the novel PKCs, $, (, ), and * require diacylglycerol and phospholipids but are Ca2+independent. A third branch of the family, atypical PKCs, consisting of +, ,, Âľ and - has also been identified and are characterized by lacking one of the two cysteine-rich zincfinger regions present in the other isoforms; their activity is independent of Ca2+, phospholipid, diacylglycerol and phorbol esters. The presence of PKC isoforms has been described for many cellular compartments and there are many examples of translocation to other compartments including the cell nucleus, where upon stimulation, they complex with and phosphorylate specific protein substrates (Malviya and Block, 1993). In addition to the stimulus-dependent translocation of PKC into the nucleus (Leach et al., 1992; Martelli et al., 1992), a constitutive localization of PKC in the nuclear compartment has been described (Buchner, 1995). Both Ca2+-dependent (!, &, and ') and Ca2+independent isoforms ($, (, and )) are located in isolated nuclei from various cells and the + isoenzyme is associated with nuclei isolated from brain. PKC isoforms can also be differentially distributed within the subnuclear compartments. Intracellular targeting mechanisms of PKC isoforms in the nucleus are unknown; No known NLS sequences of PKC have been identified (Buchner, 1995; Schmalz et al., 1998). It has been suggested that PKC may be directed to their intracellular sites of action through specific interactions with a growing family of docking proteins (Mochly-Rosen, 1995) such as the A kinase-anchoring protein, AKAP79 (Faux and Scott, 1997), and the RACK protein (Buchner, 1995). Specific lipid components within the nuclear membrane play a key role in PKC signaling by stimulating PKC activity at the nucleus. Nuclear

328

Gene Therapy and Molecular Biology Vol 4, page 329 diacylglycerol can also activate nuclear PKC (Topham et al., 1998).

VII. Nuclear EP receptors Until recently, it was believed that the diverse effects of PGE 2 could only be mediated by plasma membrane EP receptors. However, there is no direct evidence that effects of PGE2 on modulation of gene transcription (Danesch et al., 1994; Simonson et al., 1994; Paliogianni and Boumpas; Minghetti et al., 1997), mitogenesis (Pasricha et al., 1992; Hashimoto et al., 1997) and growth and metastasis of tumors (Paoletti et al., 1989; Fulton et al., 1991; Yano et al., 1997) results exclusively from its interaction with cell surface receptors. The possible existence of nuclear EP receptors is suggested by the occurrence of enzymes involved in the biosynthesis of prostanoids (COX-1, COX-2, PLA2) at the nuclear envelope (Spencer et al., 1998; Schievella et al., 1995). It has been proposed that the inducible COX-2 provides prostaglandins for nuclear eicosanoid signaling system, which participate in cellular growth, replication and differentiation (Goetzl et al., 1995; Minghetti and Levi, 1998). Thus COX-2 could play a role in prolonged generation of prostanoid which would modify nuclear events (Reddy and Herschmann, 1997). Moreover, the identification of a prostanoid transporter, which mediates the influx of prostanoid (Kanai et al., 1995), provided an efficient means of delivering circulating PGE2 to the nuclear site. It has been demonstrated that other prostanoids, namely PGD2 and its metabolite PGJ2, can activate the peroxisome proliferator-activated receptors (PPARs), which are members of the nuclear receptor superfamily of ligand-dependent transcription factors; however, PPARs are not responsive to PGE2 (Kliewer et al., 1995; Hertz et al., 1996) and therefore distinct from nuclear EP receptors identified by us (Bhattacharya et al., 1998, 1999). Also, as discussed above, nuclear membranes contain various intermediate factors involved in EP receptor-mediated signal transduction systems, such as G proteins, calcium, PLC, adenylyl cyclase, PKA and PKC. Our recent studies have demonstrated that functional EP receptors are indeed localized in nuclear membranes in a variety of cells and tissues (Bhattacharya et al., 1998, 1999). Nuclear PGE2 binding sites were detected by radioligand binding studies in both newborn (porcine brain) and adult (porcine myometrium and rat liver) tissues; the PGE2 binding kinetics to nuclear and plasma membrane fractions were similar. All subtypes of EP receptors (EP1, EP2, EP3! and EP4) were detected in nuclear membranes. However in each tissue, the relative distribution of EP receptors in plasma membrane and nuclear membrane differed. Since the molecular weights and pharmacological characteristics EP receptors in plasma and nuclear membranes were similar, it would appear that the receptors in both cellular compartments are highly homologous or identical. The presence of EP receptors in the nuclear membrane was verified by immunoelectron microscopy of primary cultures of porcine newborn cerebral

D. Nuclear adenylyl cyclase and protein kinase A cAMP and cAMP-dependent protein kinase A (PKA) mediate signal transduction pathways that regulate a variety of physiological responses in cells. Compartmentalization of cAMP and PKA control specific signal functions (Yabana et al., 1995). Localization of functional adenylyl cyclase has been demonstrated in nuclear envelopes in human cardiomyocytes (Yamamoto et al., 1998). There is evidence for cAMP accumulation in the nucleus (Barsony and Marx, 1990). The nuclear localization of the catalytic subunit of PKA, which does not contain an endogenous nuclear localization signal, has also been reported (Wiley et al., 1999).

E. Nuclear phospholipase C Phospholipase C (PLC) signaling occurs in the nucleus (Divecha and Irvine, 1995; Cocco et al., 1999). PLC&1 is the major isoform in the nuclei of various cells but other isoforms, PLC&2, PLC&3 and PLC&4 are also present. The PLC& family is regulated by heterotrimeric G proteins. Nuclear PLC&1 is activated independently of its plasma membrane counterpart, and increases during cell growth and decreases during differentiation (Divecha and Irvine, 1995). The long carboxy-terminal region was shown to be necessary for the nuclear localization of PLC&1 (Kim et al., 1996).

F. Nuclear phospholipase D Phospholipase D (PLD) is a ubiquitous enzyme that catalyzes the hydrolysis of cellular phospholipids, particularly phosphatidylcholine, in response to a variety of hormones, neurotransmitters and growth factors (Exton, 1997). Phosphatidic acid, the primary lipid product of PLD possesses growth factor-like properties and can act as a second messenger in certain cell types. Low molecular weight G proteins, ADP-ribosylation factor (ARF) and RhoA play key roles in agonist-induced PLD activation (Exton, 1997). PLD is enriched in plasma membrane and cytosol but is also present with high activity in isolated nuclei (Exton, 1997; Balboa and Insel, 1995). PLD is a component of a novel nuclear signaling cascade, defined as nuclear envelope signal transduction (NEST), involving the induction of specific nuclear lipid metabolism (Baldassare et al., 1997). Mitogen induced activation of nuclear PLD is mediated by the translocation of RhoA to the nucleus; the activated PLD leads to the production in the nuclear envelope of the signaling molecule, phosphatidic acid.

329

Bhattacharya et al: Nuclear Prostaglandin Receptors microvascular endothelial cells using specific rabbit antiEP receptor antibodies (Zhao et al., 1995) (Fig. 1). Moreover, the presence of nuclear EP receptors was also detected in vivo , in adult rat brain cortex endothelial cells and neurons (Bhattacharya et al., 1998); both these tissues are important sources of brain prostanoids (Li et al., 1994; Parfenova et al., 1997; Mingetti and Levi, 1998). Over-expression of human EP1, EP3! and EP4 receptors cDNA in human embryonic kidney (HEK) 293 cells that do not express prostanoid receptors (Boie et al., 1997) revealed a perinuclear localization of EP receptors by indirect immunofluorescence (Bhattacharya et al., 1998; 1999); a similar distribution was observed by expressing EP 1 receptor fused to green fluorescent protein in HEK 293 cells. These data once again suggest similar

identity between plasma membranes and nuclear envelope EP receptors. Although the stimulation of nuclear EP receptors by prostaglandin analogs did not modify the generation of second messengers, cAMP or IP3, it affected gene transcription and nuclear calcium signals (Bhattacharya et al., 1998; 1999). The stimulation of nuclear EP1 receptors in isolated nuclei by prostaglandin analogs was found to modulate c-fos gene transcription and nuclear calcium transients. Stimulation of intact nuclei isolated from primary cultures of porcine brain endothelial cells with the EP3 receptor agonist M&B 28,767 (0.1 ÂľM) increased transcription of iNOS (Fig. 2a) to a greater extent than after stimulation of whole cells as determined by dot RNA hybridization.

Figure 1. Immunogold localization of EP3! and EP4 receptors in porcine cerebral microvascular endothelial cells (primary culture from newborn brain) by electron microscopy (see arrows). (a) Anti-rabbit gold-conjugated IgG alone; note absence of immunostaining when primary antibody is omitted. EP 3! immunoreactivity on (b) plasma membrane, and (c) nuclear membrane. EP4 immunoreactivity on (d) Golgi vesicles, (e) plasma membrane, and (f) nuclear membrane. Scale bar in each represents 0.5 Âľm.

330

Gene Therapy and Molecular Biology Vol 4, page 331

Figure 2. Effects of nuclear EP receptor stimulation on iNOS gene transcription and nuclear calcium transients. (a) Effects of EP3 agonist M&B 28,767 (0.1 µM) on iNOS transcription in porcine cerebral microvascular endothelial cells (primary culture from newborn brain) as determined by dot blot hybridization of RNA; MB and C refer to M&B 28,767 and control (unstimulated) respectively. &-actin dot blot indicates equal loading. One representative dot blot of three is shown. (b) Effect of M&B 28,767 (1 µM) on 45Ca2+ uptake by isolated liver nuclei; control refers to absence of drug. The free calcium concentration was 400 nM. The movement of 45Ca2+ transient after a given time was defined as radioactivity at a given time minus the radioactivity at time 0. (c) Typical tracings showing effects of PGE2 analog 16, 16-Dimethyl PGE2 (1 µM) and EP 3 agonist M&B 28,767 (0.01-1 µM) on intranuclear calcium concentrations ([Ca2+]n) in isolated liver nuclei loaded with fura-2 AM; arrow shows the time of application of test agents. (d) Peak increases in isolated liver intranuclear calcium ([Ca2+]n) after addition of M&B 28,767 (1 µM) in the presence or absence of PTX preincubation (20 µg/ml, 20 min at 25 °C). Experiments (b-d) were carried out on three independent isolations of intact nuclei each one performed in duplicate.

increasing nuclear Ca2+ levels (Fig. 2c). EP3 couples mainly to G i or G o (Namba et al., 1993), which are known to affect Ca 2+ mobilization (Namba et al., 1993; Kojima et al., 1986; Huges et al., 1987; Hescheler et al., 1988); such G proteins are detected in rat liver nuclei (Takei et al., 1994). Pretreatment of isolated nuclei with PTX, an inhibitor of Gi or Go markedly attenuated the stimulatory effect of M&B 28,767 on intranuclear calcium levels suggestive of the involvement of a PTX-sensitive Gprotein in mediating the effects of nuclear EP 3 receptors (Fig. 2d). These findings are consistent with other studies which have demonstrated that G proteins, especially Gi (Hescheler et al., 1988; Rosenthal et al., 1988) may directly control Ca 2+ channels independent of cAMP or IP3 (Kojima et al., 1986; Hughes et al., 1987). Our studies indicated that functional nuclear EP receptors were present in newborn porcine brain (Bhattacharya et al., 1998, 1999; Figs. 1 and 2). PGE2 is

Previous studies have reported that iNOS expression is modulated in an opposite way by endogenous and exogenous PGE2; endogenous PGE2 has a stimulatory effect on iNOS as opposed to the inhibitory effect of exogenous PGE2 (Milano et al., 1995; Minghetti et al., 1997). It is thus possible that endogenous PGE2 might act via EP nuclear receptors as speculated (Minghetti et al., 1997). Application of EP3 agonist M&B 28,767 to intact isolated liver nuclei also caused rapid nuclear uptake of 45 Ca2+ (Fig. 2b). In addition, M&B 28,767 produced a dose-dependent increase in rat liver nuclear calcium transients as determined by fura-2 AM, a fluorescent dye which localizes in the nuclear envelope space (Gerasimenko et al., 1996) (Fig. 2c). At an equivalent concentration, M&B 28,767 was nearly as effective as the non-selective EP agonist 16,16-dimethyl PGE2 in

331

Bhattacharya et al: Nuclear Prostaglandin Receptors one of the most abundant prostanoids in the brain during the perinatal period (Leffler and Busija, 1985) and plays an important role in many cerebral functions, such as in the regulation of cerebral blood flow (see Chemtob, et al., 1996) and modulating the gene transcription of nitric oxide synthase (Dumont et al., 1998). However, in the newborn brain and cerebral microvasculature, high levels of prostaglandins have been shown to result in downregulation of plasma membrane EP receptors and associated functions (see Chemtob et al., 1996; Li et al., 1994). On the other hand, PGE 2 increases the expression of nitric oxide synthase via stimulation of EP3 receptors in the neonate (Dumont et al., 1998). This effect was prevented in the presence of an inhibitor of the prostanoid transporter (Dumont et al., 1999), which supports a role of nuclear EP receptors in modulating NOS gene transcription. The details of the mechanism by which nuclear EP receptors activate gene transcription of c-fos and iNOS need to be elucidated. Our studies revealed that selective stimulation of nuclear EP1 receptors increased the transcription of c-fos; Simonson et al., (1994) also showed that PGE 2 elevated c-fos expression by cAMP-independent but PKC-dependent mechanisms. The increase in nuclear calcium transients following a stimulation of nuclear EP receptors might be the mechanism by which calcium regulates the transcription of many genes (Malviya and Rogue, 1998) including c-fos (Hardingham et al., 1997; Malviya and Rogue, 1998). The transcription of iNOS is predominantly activated by the ubiquitous transcription factor NF-%B (Xie et al., 1994) and EP3 receptors have been shown to modulate NF-%B-dependent cellular signaling (Meyer-Kirchrath et al., 1998). Recent studies indicate that there may be constitutive expression of NF%B in the nucleus (Hunot et al., 1997; Delfino and Walker, 1998) and this may suggest a possible mechanism by which nuclear EP3 receptors may modulate iNOS transcription. To date, the presence of a NLS sequence has not been reported in the literature in any of the cloned EP receptor subtypes (Narumiya, 1996). A search of the human EP receptor protein sequence database against a ProfileScan database revealed the presence of a consensus sequences with high degree of similarity to bipartite NLS signals in the C-terminals tails of human EP1 (Funk et al., 1993) and EP3 (Kotani et al., 1997) isoforms (amino acids 236-253 for EP1 and 350-367 for EP3). This places prostanoid receptors together with the AT1 receptor, in a distinct class of G protein-coupled receptors, which function at the cell surface as well as at the nucleus, and revolutionizes our current understanding of signal transduction mediated via G protein-coupled receptors. Nevertheless, an important distinction between the AT 1 and prostanoid EP receptors is that the latter were detected at the nuclear envelope without prior stimulation with PGE2. Angiotensin II receptors have been detected in chromatin fragments, and it has been suggested that

Angiotensin II alone or complexed to its receptor, might be involved in the regulation of gene expression through interactions with nuclear DNA (Re et al., 1984; Eggena et al., 1996). There are other reports of nuclear peptides capable of binding DNA (Castel et al., 1992). In contrast, there have been no reports of PGE2 binding chromatin; only the prostaglandins of the J series and their receptors, the PPARs, have been detected within the nucleus (Narumiya and Fukushima, 1986; Kliewer et al., 1995; Hertz et al., 1996). Interestingly, it has been reported that putative NLS sequences have been detected in other G protein-coupled receptors such as the human platelet activating factor receptor and the human M1, M3 and M5A muscarinic acetylcholine receptors (Lu et al., 1998). However, whether or not these receptors are actually localized at the nucleus and are functional is yet to be established. The details of the mechanism of intracellular trafficking that target EP receptors to the nucleus remain to be elucidated. Since over-expression of the EP receptor cDNA in HEK 293 (Bhattacharya et al., 1998, 1999) and Swiss 3T3 cells or transfection of EP1-GFP fusion protein in HEK 293 cells revealed a perinuclear localization (Bhattacharya et al., 1998), it is very possible that translational or post-translational mechanism are operative in determining trafficking of the receptors to either plasma membranes or nuclear membranes. Segments of EP receptors may contain docking sites for interaction with proteins that may function to target the receptors to defined subcellular locations as reported for the T-cell protein tyrosine phosphatase (TCPTP) by Tiganis et al. (1997). TCPTP contains both an endoplasmic reticulum-targeting motif and a bipartite NLS sequence in the C-terminal but is associated exclusively with the endoplasmic reticulum, which indicates that the endoplasmic reticulum targeting event is the dominant event. A cytoplasmic protein has been shown to interact with residues in the C terminus and directs the enzyme to the endoplasmic reticulum; the binding site of this protein overlaps residues in the NLS cluster and therefore might control the accessibility of the NLS to nuclear import factors. Another protein that is located in the plasma membrane but is also targeted to the nucleus is the myelin basic protein; this too possesses putative NLS sequences (Pedraza et al., 1997). Putative NLS sequences have so far not been detected on EP2 or EP4 receptors. In fact, known NLS sequences are lacking in various proteins that are targeted to the nucleus such as COX (Spencer et al., 1998), PKC (see Buchner, 1995; Schmalz et al., 1998) or the kinase Raf-1 (Lu et al., 1998). There may be some additional targeting motifs, which direct some of these proteins to the nuclear membranes. Nuclear translocation of Raf-1 may be accomplished by its binding to a carrier protein (Lu et al., 1998) and nuclear targeting of PKC may involve its interaction with docking proteins (Buchner, 1995; Faux and Scott, 1997). EP receptors may be directed to the nucleus by a piggy-back mechanism that involves 332

Gene Therapy and Molecular Biology Vol 4, page 333 additional protein(s) containing a basic nuclear localization sequence as demonstrated for I-%B!, which is constitutively transported to the nucleus when it is not bound to NF-%B (Turpin et al., 1999).

Baldassare JJ, Jarpe MB, Alferes L, and Raben DM (1997) Nuclear translocation of RhoA mediates the mitogen-induced activation of phospholipase D involved in nuclear envelope signal transduction. J Biol Chem 272, 4911-4914. Barsony J, and Marx S (1990) Immunocytology on microwavefixed cells reveals rapid and agonist-specific changes in subcellular accumulation patterns for cAMP or cGMP. Proc Natl Acad Sci USA 87, 1188-1192. Beaudet A, Nouel D, Stroh T, Vandenbulcke F, Dal-Farra C, and Vincent JP (1998) Fluorescent ligands for studying neuropeptide receptors by confocal microscopy. Braz J Med Biol Res 31, 1479-1489. Berridge MJ (1993) Inositol triphosphate and calcium signaling. Nature 361, 315-325. Bhattacharya M, Peri K, Almazan G, Ribeiro-da-Silva A, Shichi H, Durocher Y, Abramovitz M, Hou X, Varma DR, and Chemtob S (1998) Nuclear localization of prostaglandin E2 receptors. Proc Natl Acad Sci USA 95, 15792-15797. Bhattacharya M, Peri K, Ribeiro-da-Silva A, Almazan G, Shichi H, Hou X, Varma DR, and Chemtob S (1999) Localization of functional prostaglandinE2 receptors EP3 and EP4 in the nuclear envelope. J Biol Chem 274, 15719-15792. Bkaily G, Jaalouk D, Jacques D, Economos D, Hassan G, Simman M, Regoli D, and Pothier P (1997a) Bradykinin activates R-, T- and L-type Ca 2+ channels and induces a sustained increase of nuclear Ca2+ in aortic vascular smooth muscle cells. Can J Physiol Pharmacol 75, 652-660. Bkaily G, Pothier P, D'Orleans-Juste P, Simman M, Jacques D, Jaalouk D, Belzile F, Hassan G, Boutin C, Haddad G, and Neugebauer W (1997b) The use of confocal microscopy in the investigation of cell structure and function in the heart, vascular endothelium and smooth muscle cells. Mol Cell Biochem 172, 171-194. Boie Y, Stocco R, Sawyer N, Slipetz D, Ungrin M, NeuschaferRube F, Puschel GP, Metters KM, and Abramovitz M (1997) Molecular cloning and characterization of the four rat prostaglandin E 2 receptor subtypes. Eur J Pharmacol 340, 227-241. Bonventre JV, Huang Z, Taheri MR, O'Leary E, Li E, Moskowitz MA, and Sapirstein A (1997) Reduced fertility and postischaemic brain injury in mice deficient in cytosolic phospholipase A2. Nature 390, 622-625. Booz GW, Conrad KM, Hess AL, Singer HA, and Baker KM (1992) Angiotensin-II-binding sites on hepatocyte nuclei. Endocrinology 130, 3641-3649. Buchner K (1995) Protein kinase C in the transduction of signals toward and within the cell nucleus. Eur J Biochem 228, 211-221. Burkey TH, and Regan JW (1995) Activation of mitogenactivated protein kinase by the human prostaglandin EP3A receptor. Biochem Biophys Res Commun 211, 152-158. Campbell WB, and Halushka PV (1996) Lipid-derived autocoids: eicosanoids and platelet activating factor. In: Gileman AG, Rall TW, Nies AS, Taylor P (eds) The Pharmacological Basis of Therapeutics. 9th edition. Pergamon Press, New York, pp 601-614. Castano E, Dalmau M, Marti M, Berrocal F, Bartrons R, and Gil J (1997) Inhibition of DNA synthesis by aspirin in Swiss 3T3 fibroblasts. J Pharmacol Exp Ther 280, 366-372. Castel MN, Woulfe J, Wang X, Laduron PM, and Beaudet A (1992) Light and electron microscopic localization of

VIII. Conclusions The discovery of nuclear EP receptors (Bhattacharya et al., 1998, 1999) proposes new avenues for the intracellular actions of prostanoids. In addition, these studies shed new light to the complex field of signaling via G protein-coupled receptors. Further studies are needed to clarify the details of the molecular mechanisms involved in this action of prostaglandins via nuclear EP receptors. It would also be interesting to determine whether there are nuclear receptors for other prostanoids (FP, DP, TP, IP). The findings that will result from these studies will provide valuable insight in the field of prostanoid pharmacology. Given that prostanoids are widely distributed in body tissues and play a major role in various physiological and pathophysiological conditions, these studies may lead to new therapeutic strategies for a number of pathological conditions.

Acknowledgements We thank Dr. S. Ferguson at the John P. Robarts Research Institute, London, Ontario for pointing out the NLS sequence for EP receptors. Studies on nuclear EP receptors reported in this review were supported by grants from the Medical Research Council of Canada. M. Bhattacharya and S. Chemtob are recipients of the Doctoral Research and Scientist awards respectively, from the Medical Research Council of Canada.

References Adderley SR and FitzGerald J (1999) Oxidative damage of cardiomyocytes is limited by extracellular regulated kinases 1/2-mediated induction of cyclooxygenase-2. J Biol Chem 274, 5038-5046. Al-Mohanna FA, Caddy KW, and Bolsover SR (1994) The nucleus is insulated from large cytosolic calcium ion changes. Nature 367, 745-750. Asboth G, Phaneuf S, Europe-Finner GN, Toth M, and Bernal AL (1996) Prostaglandin E2 activates phospholipase C and elevates intracellular calcium in cultures myometrial cells: involvement of EP1 and EP3 receptor subtypes. Endocrinology 137, 2572-2579. Audoly LP, Ma L, Feoktistov I, De Foe S, Breyer MD, and Breyer RM (1999) Prostaglandin E-prostanoid-3 receptor activation of cyclic AMP response element-mediated gene transcription. J Pharmacol Exp Ther 289, 140-148. Badminton MN, Kendall JM, Rembold CM, and Campbell AK (1998) Current evidence suggests independent regulation of nuclear calcium. Cell Calcium 23, 79-86. Balboa MA, and Insel PA (1995) Nuclear phospholipase D in Madin-Darby canine kidney cells. J Biol Chem 50, 2984329847.

333

Bhattacharya et al: Nuclear Prostaglandin Receptors retrogradely transported neurotensin in rat nigrostriatal dopaminergic neurons. Neuroscience 50, 269-282. Chemtob S, Li D-Y, Abran D, Hardy P, Peri KG, Varma DR (1996) The role of prostaglandin receptors in regulating cerebral blood flow in the perinatal period. Acta Paediatrica 85, 517-524. Clapham DE (1996) The G-protein nanomachine. Nature 379, 297-298. Cocco L, Rubbini S, Manzoli L, Billi AM, Peruzzi D, Matteucci A, Artico M, Gilmour RS, and Rhee SG (1999) Inositides in the nucleus: presence and characterization of the isozymes of phospholipase C beta family in NIH 3T3 cells. Biochim Biophys Acta 1438, 295-299. Coffey RJ, Hawkey CJ, Damstrup L, Graves-Deal R, Daniel VC, Dempsey PJ, Chinery R, Kirkland S, DuBois R, Jetton TL, and Morrow JD (1997) Epidermal growth factor receptor activation induces nuclear targeting of cyclooxygenase-2, basolateral release of prostaglandins, and mitogenesis in polarizing colon cancer cells. Proc Natl Acad Sci USA 94, 657-662. Coleman RA, Smith WL, and Narumiya S (1994) International Union of Pharmacology. Classification of prostanoid receptors: properties, distribution, and structure of the receptors and their subtypes. Pharmacol Rev 46, 205-229. Crouch MF (1991) Growth factor-induced cell division is paralleled by translocation of G i! to the nucleus. FASEB J 5, 200-206. Crouch MF, and Simson I (1997) The G-protein Gi regulates mitosis but not DNA synthesis in growth factor-activated fibroblasts: a role for the nuclear translocation. FASEB J 11, 189-198. Danesch U, Weber P, and Sellmayer A (1994) Arachidonic acid increases c-fos and Erg-1 mRNA in 3T3 fibroblasts by formation of prostaglandin E2 and activation of protein kinase C. J Biol Chem 269, 27258-27263. Delfino F, and Walker WH (1998) Stage-specific nuclear expression of NF-kappaB in mammalian testis. Mol Endocrinol 12, 1696-1707. Divecha N, and Irvine RF (1995) Phospholipid signaling. Cell 80, 269-278. Dumont I, Peri K, Hardy P, Hou X, Martinez-Bermudez A K, Molotchnikoff S, Varma DR, and Chemtob S (1998) PGE2, via EP 3 receptors, regulates brain nitric oxide synthase in the perinatal period. Am J Physiol 275, R1812-R1821. Dumont I, Hou X, Hardy P, Peri K, Beauchamp M, Najarian T, Molotchnikoff S, Varma DR, and Chemtob S (1999) Developmental regulation of endothelial nitric oxide synthase in cerebral vessels of newborn pig by prostaglandin E2. J Pharmacol Exp Ther, in press Eggena P, Zhu JH, Clegg K, and Barrett JD (1993) Nuclear angiotensin receptors induce transcription of renin and angiotensinogen mRNA. Hypertension 22, 496-501. Exton JH (1997) New developments in phospholipase D. J Biol Chem 272, 15579-15582. Faure MP, Alonso A, Nouel D, Gaudriault G, Dennis M, Vincent JP, and Beaudet A (1995) Somatodendritic internalization and perinuclear targeting of neurotensin in the mammalian brain. J Neurosci 15, 4140-4147.

Faux MC, and Scott JD (1997) Regulation of the AKAP79protein kinase C interaction by calcium/calmodulin. J Biol Chem 272, 17038-17044. Fu ML, Schulze W, Wallukat G, Elies R, Eftekhari P, Hjalmarson A, and Hoebeke J (1998) Immunohistochmical localization of angiotensin II receptors (AT1) in the heart with anti-peptide antibodies showing a positive chronotropic effect. Receptor Channels 6, 99-111. Fulton AM, Zhang S, and Chong YC (1991) Role of the prostaglandin E2 receptor in mammary tumor metastasis. Cancer Res 51, 2047-2050. Funk CD, Furci L, FitzGerald GA, Grygorczyk R, Rochette C, Bayne MA, Abramovitz M, Adam M, and Metters KM (1993) Cloning and expression of a cDNA for the human prostaglandin E receptor EP1 subtype. J Biol Chem 268, 26767-26772. Gashler A, and Sukhatme VP (1995) Early growth response protein 1(Erg-1): prototype of a zinc-finger family of transcription factors. Prog Nucleic Acid Res 50, 191-224. Gerasimenko OV, Gerasimenko JV, Tepilin AV, and Petersen OH (1996) Calcium transport pathways in the nucleus. Eur J Physiol 432, 1-6. Glantschnig H, Varga F, and Klaushoffer K (1996) The cellular protooncogenes c-fos and erg-1 are regulated by prostacyclin in rodent osteoblasts and fibroblasts. Endocrinology 137, 4536-4541. Goetzl E, An S, and Smith W (1995) Specificity of expression and effects of eicosanoid mediators in normal physiology and human diseases. FASEB J 9, 1051-1058. Grady EF, Garland AM, Gamp PD, Lovett M, Payan DG, and Bunnett NW ( 1995) Delineation of the endocytic pathway of substance P and its seven-transmembrane domain NK1 receptor. Mol Biol Cell 6, 509-524. Guan Z, Buckman SY, Pentland AP, Templeton DJ, and Morrison AR (1998) Induction of cyclooxygenase-2 by the activated MEKK1 SEK1/MKK4 p38 mitogen-activated protein kinase pathway. J Biol Chem 273, 12901-12908. Gudermann T, Schoneberg T, and Schultz G (1997) Functional and structural complexity of signal transduction via Gprotein-coupled receptors. Annu Rev Neurosci 20, 399-427. Hamm H (1998) The many faces of G protein signaling. J Biol Chem 273, 669-672. Hardingham GE, Chawla S, Johnson CM, and Bading H (1997) Distinct functions of nuclear and cytosolic calcium in the control of gene expression. Nature 385, 260-265. Hasegawa H, Negishi M, Katoh H, and Ichikawa A (1997) Two isoforms of prostaglandin EP3 receptor exhibiting constitutive activity and agonist-dependent activity in Rho-mediated stress fiber formation. Biochem Biophys Res Commun 234, 631-636. Hashimoto N, Watanabe T, Ikeda Y, Yamada H, Taniguchi S, Mitsui H, and Kurokawa K (1997) Prostaglandn induce proliferation of rat hepatocytes through a prostaglandin E2 receptor EP3 subtype. Am J Physiol 272, G597-G604. Hebert RL, Jacobson H, and Breyer MD (1991) Triple signal transduction model for the mechanism of PGE2 actions in rabbit cortical collecting duct. Prostaglandins Leukot Essent Fatty Acids 42, 143-148.

334

Gene Therapy and Molecular Biology Vol 4, page 335 Heist EK, and Schulman H (1998) The role of Ca2+/calmodulindependent protein kinases within the nucleus. Cell Calcium 23, 103-114. Hertz R, Berman I, Keppler D, and Bar-Tana J (1996) Activation of gene transcription by prostacyclin analogues is mediated by peroxisome-activated proliferator receptor (PPAR). J Biochem 235, 242-247. Hescheler J, Rosenthal W, Hinsch KD, Wulfern M, Trautwein W, and Schultz G (1988) Angiotensin II-induced stimulation of voltage-dependent Ca2+ currents in an adrenal cell line. EMBO 7, 613-624. Hughes BP, Crofts JN, Auld AM, Read LC, and Barritt GJ (1987) Evidence that a pertussis-toxin-sensitive substrate is involved in the stimulation by epidermal growth factor and vasopressin of plasma-membrane Ca 2+ inflow in hepatocytes. Biochem J 248, 911-918. Humbert J, Matter N, Artault J, Koppler P, and Malviya AN (1996) Inositol 1,4,5-triphosphate receptor is located to the inner nuclear membrane vindicating regulation of nuclear calcium signaling by inositol 1,4,5-triphosphate. J Biol Chem 271, 478-485. Hunot S, Brugg B, Ricard D, Michel PP, Ruberg M, Faucheux BA, Agid Y, and Hirsch EC (1997) Nuclear translocation of NF-kappaB is increased in dopaminergic neurons of patients with Parkinson disease. Proc Natl Acad Sci USA 94, 75317536. Kanai N, Lu R, Satriano J A, Bao Y, Wolkoff AW, and Schuster VL (1995) Identification and characterization of a prostaglandin transporter. Science 268, 866-869. Katoh H, Watabe A, Sugimoto Y, Ichikawa A, and Negishi M (1995) Characterization of the signal transduction of prostaglandin E receptor EP 1 subtype in cDNA-transfected Chinese hamster ovary cells. Biochim Biophys Acta 1244, 41-48. Katoh H, Negishi M, and Ichikawa A (1996) Prostaglandin E receptor EP3 subtype induces neurite retraction via small GTPase Rho. J Biol Chem 271, 29780-29784. Kaufmann WE, Worley PF, Pegg J, Bremer M, and Isakson P (1996) Cox-2, a synaptically induced enzyme, is expressed by excitatory neurons at postsynaptic sites in rat cerebral cortex. Proc Natl Acad Sci USA 93, 2317-2321. Kennedy BP, Payette P, Mudgett J, Vadas P, Pruzanski W, Kwan M, Tang C, Rancourt DE, and Cromlish WA (1995) A natural disruption of the secretory group II phospholipase A2 gene in inbred mouse strains. J Biol Chem 270, 2237822385. Kim CG, Park D, and Rhee SG (1996) The role of carboxylterminal basic amino acids in Gq!-dependent activation, particulate association and nuclear localization of phospholipase C- &1. J Biol Chem 271, 21187-21192. Kliewer S, Lenhard J, Willson T, Patel I, Morris D, and Lehmann J (1995) A prostaglandin J2 metabolite binds peroxisome proliferator-activated receptor ' and promotes adipocyte differentiation. Cell 83, 813-819. Kojima I, Shibata H, and Ogata E (1986) Pertussis toxin blocks angiotensin II-induced calcium influx but not inositol triphosphate production in adrenal glomerulosa cell. FEBS Lett 204, 347-351. Konger RL, Malaviya, and Pentland AP (1998) Growth regulation of primary human keratinocytes by prostaglandin

E receptor EP2 and EP3 subtypes. Biochim Biophys Acta 1401, 221-234. Kotani M, Tanaka I, Ogawa Y, Usui T, Tamura N, Mori K, Narumiya S, Yoshimi T, Nakao K (1997) Structural organization of the human prostaglandin EP3 receptor subtype gene (PTGER3). Genomics 40, 425-434. Kramer RM, and Sharp JD (1997) Structure, function and regulation of Ca2+-sensitive cytosolic phospholipase A2 (cPLA2). FEBS Lett 410, 49-53. Krisch B, Feindt J, and Mentlein R (1998) Immunoelectromicroscopic analysis of the ligand-induced internalization of the somatostatin receptor subtype 2 in cultured human glioma cells. J Histochem Cytochem 46, 1233-1242. Kulmacz RJ, and Wang LH (1995) Comparison of hydroperoxide initiator requirements for the cyclooxygenase activities of prostaglandin H synthase-1 and -2. J Biol Chem 270, 24019-24023. Kutchera W, Jones DA, Matsunami N, Groden J, and McIntyre TM, Zimmerman GA, White RL, Prescott SM (1996) Prostaglandin H synthase-2 is expressed abnormally in human colon cancer: evidence for a transcriptional effect. Proc Natl Acad Sci USA 93, 4816-4820. Langenbach R, Morham SG, Tiano HF, Loftin CD, Ghanayem BI, Chulada PC, Mahler JF, Lee CA, Goulding EH, Kluckman KD, et al. (1995) Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration. Cell 83, 483-492. Leach KL, Buff VA, Jarpe MB, Adams LD, Fabbro D, and Raben DM (1992) Alpha-thrombin stimulates nuclear diglyceride levels and differential nuclear localization of protein kinase C isozymes in II C9 cells. J Biol Chem 267, 816-821. Leffler CW, and Busija DW (1985) Arachidonate metabolism on the cerebral surface of newborn pigs. Prostaglandins 30, 811-817. Leffler CW, and Busija DW (1987) Prostanoids and pial arteriolar diameter in hypotensive newborn piglets. Am J Physiol 252, H687-H691. Lefkowitz RJ (1998) G protein-coupled receptors. J Biol Chem 273, 18677-18680. Li D-Y, Varma DR, and Chemtob S (1994) Ontogenic increase in PGE2 and PGF2! receptor density in brain microvessels of pigs. Br J Pharmacol 112, 59-64. Lind GJ, and Cavanagh HD (1993) Nuclear muscarinic acetylcholine receptors in corneal cells from rabbit. Invest Ophthalmol Vis Sci 34, 2943-2952. Lind GJ, and Cavanagh HD (1995) Identification and subcellular distribution of muscarinic acetylcholine receptor-related proteins in rabbit corneal and Chinese hamster ovary cells. Invest Ophthalmol Vis Sci 36, 1492-1507. Longin A-S, Mezin P, Favier A, and Verdetti J (1997) Presence of zinc and calcium permeant channels in the inner membrane of the nuclear envelope. Biochem Biophys Res Commun 235, 236-241. Lu D, Raizada MK, and Shaw G (1998) Angiotensin II-induced nuclear targeting of the angiotensin type 1 (AT1) receptor in brain neurons. Endocrinology 139, 365-375.

335

Bhattacharya et al: Nuclear Prostaglandin Receptors Lu R, Kanai N, Bao Y, and Schuster VL (1996) Cloning, in vitro expression, and tissue distribution of a human prostaglandin transporter cDNA (hPGT). J Clin Inves 98, 1142-1149. Lupulescu A (1978) Enhancement of carcinogenesis by prostaglandins. Nature 272, 634-636. Malviya AN, and Block C (1993) Nuclear protein kinase C and signal transduction. Receptor 3, 257-275. Malviya AN, and Rogue PJ (1998) Tell me where is calcium bred: clarifying the roles of nuclear calcium. Cell 92, 17-23. Martelli AM, Gilmour RS, Bertagnolo V, Neri LM, Manzoli L, and Cocco L (1992) Nuclear localization and signaling activity of phosphoinositidase C beta in Swiss 3T3 cells. Nature 358, 242-245. Mater MK, Pan D, Bergen WG, and Jump DB (1998) Arachidonic acid inhibits lipogenic gene expression in 3T3L1 adipocytes through a prostanoid pathway. J Lipid Res 39, 1327-1334. Matter N, Ritz M, Freyermuth S, Rogue PJ, and Malviya AN (1993) Stimulation of nuclear protein kinase C leads to phosphorylation of nuclear inositol 1,4,5-triphosphate receptor and accelerated calcium release by inositol 1,4,5triphosphate from isolated rat liver nuclei. J Biol Chem 268, 732-736. Meyer-Kirchrath J, Kuger P, Hohlfeld T, and Schror K (1998) Analysis of a porcine EP3-receptor: cloning, expression and signal trasduction. Naunyn-Schmiedeberg's Arch Pharmacol 358, 160-167. Milano S, Arcoleo F, Dieli M, D'Agostino R, D'Agostino P, De Nucci G, and Cillari E (1995) Prostaglandin E2 regulates inducible nitric oxide synthase in the murine macrophage cell line J774. Prostaglandins 49, 105-115. Minghetti L, Nicolini A, Polazzi E, Creminon C, Maclouf J, and Levi G (1997) Inducible nitric oxide synthase expression in activated rat microglial cultures is downregulated by exogenous prostaglandin E2 and by cyclooxygenase inhibitors. Glia 19, 152-160. Minghetti L, and Levi G (1998) Microglia as effector cells in brain damage and repair: focus on prostanoids and nitric oxide. Prog in Neurobiolog 54, 99-125. Mochly-Rosen D (1995) Localization of protein kinases by anchoring proteins: a theme in signal transduction. Science 268, 247-251. Morham SG, Langenbach R, Loftin CD, Tiano HF, Vouloumanos N, Jennette JC, Mahler JF, Kluckman KD, Ledford A, Lee C, and Smithies O (1995) Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 83, 473-482. Morita I, Schindler M, Regier MK, Otto JC, Hori T, DeWitt DL, and Smith WL (1995) Different intracellular locations for prostaglandin endoperoxide H synthase-1 and -2. J Biol Chem 270, 10902-10908. Murakami M, Natakani Y, Atsumi G, Inoue K, and Kudo I (1997) Regulatory functions of phospholipases A2. Crit Rev Immunol 17, 225-283. Murakami M, Kambe T, Shimbara S, and Kudo I (1999) Functional coupling between various phospholipase A2s and cyclooxygenases in immediate and delayed prostanoid biosynthetic pathways. J Biol Chem 274, 3103-3115.

Naik RD, McNeill R, Wilson TW, and Gopalakrishnan V (1998) Nuclear Ca 2+ signaling to endothelin-1 in rat aortic smoothmuscle cells. J Cardiovasc Pharmacol 31, S199-S202. Namba T, Sugimoto Y, Negishi M, Irie A, Ushikubi F, Kakizuka A, Ito A, Ichikawa A, and Narumiya S (1993) Alternative splicing of C-terminal tail of prostaglandin E receptor subtype EP3 determines G-protein specificity. Nature 365, 166-170. Naraba H, Murakami M, Matsumoto H, Shimbara S, Ueno A, Kudo I, and Oh-ishi S (1998) Segregated coupling of phospholipase A2, cyclooxygenases and terminal prostanoid synthases in different phases of prostanoid biosynthesis in rat peritoneal macrophages. J Immunol 160, 2974-2982. Narumiya S, and Fukushima M (1986) Site and mechanism of growth inhibition by prostaglandins. I. Active transport and intracellular accumulation of cyclopentenone prostaglandins, a reaction leading to growth inhibition. J Exp Pharmacol Ther 239, 500-505. Narumiya S (1996) Prostanoid receptors and signal transduction. Prog Brain Res 113, 231-241. Negishi M, Sugimoto Y, and Ichikawa A (1993a) Prostanoid receptors and their biological actions. Prog Lipid Res 32, 417-434. Negishi M, Sugimoto Y, Irie A, Narumiya S, and Ichikawa A (1993b) Two isoforms of prostaglandin E receptor EP3 subtype: different COOH-terminal domains determine sensitivity to agonist-induced desensitization. J Biol Chem 268, 9517-9521. Negishi M, Hasegawa H, and Ichikawa A (1996) Prostaglandin E receptor EP3' isoform, with mostly constitutive Gi activity and agonist-dependent Gs activity. FEBS Lett 386, 165-168. Neuschafer-Rube F, Hanecke K, Blascheke V, Jungermann K, and Puschel GP (1997) The C-terminal domain of the Gscoupled EP4 receptor confers agonist dependent coupling to control to Gi but no coupling to Gs in a receptor hybrid with the Gi-coupled EP3 receptor. FEBS Lett 401, 185-190. Nicotera P, McConkey DJ, Jones DP, and Orrenius S (1989) ATP stimulates Ca2+ uptake and increases the free Ca2+ concentration in isolated rat liver nuclei. Proc Natl Acad Sci USA 86, 453-457. Nishizuka Y (1992) Intracellular signaling by hydrolysis of phospholipids and activation of protein kinase C. Science 258, 607-614. Okuda-Ashitaka E, Sakamoto K, Ezashi T, Miwa K, Ito S, and Hayaishi O (1996) Suppression of prostaglandin E receptor signaling by the variant form of EP1 subtype. J Biol Chem 271, 31255-31261. Paliogianni F, and Boumpas DT (1996) Prostaglandin E 2 inhibits the nuclear transcription of the human interleukin 2, but not the IL-4 gene in human T cells by targeting transcription factors AP-1 and NF-AT. Cell Immunol 171: 95-101. Paoletti P, Chiabrando C, Gaetani P, Castelli MG, Butti G, Martelli L, and Rolli M (1989) Prostaglandins in human brain tumors. J Neurosurg Sci 33, 65-69. Parfenova H, Eidson TH, and Leffler CW ( 1997) Upregulation of COX-2 in cerebral microvascular endothelial cells by smooth muscle cell signals. Am J Physiol 273, C277-C288. Pasricha PJ, Hassoun PM, Teufel E, Landman MJ, and Fanburg BL (1992) Prostaglandins E1 and E2 stimulate the proliferation of pulmonary artery smooth muscle cells. Prostaglandins 43, 5-19.

336

Gene Therapy and Molecular Biology Vol 4, page 337 Pedraza L, Fidler L, Staugaitis SM, and Colman DR (1997) The active transport of myelin basic protein into the nucleus suggests a regulatory role in myelination. Neuron 18, 579589. Peri K, Hardy P, Li D-Y, Varma DR, and Chemtob S (1995) Prostaglandin G/H synthase-2 is a major contributor of brain prostaglandins in the newborn. J Biol Chem 270, 2461524620. Pesty A, Avazeri N, and Lefevre B (1998) Nuclear calcium release by InsP3-receptor channels plays a role in meiosis reinitiation in the mouse oocyte. Cell Calcium 24, 239-251. Peters-Golden M, Song K, Marshall T, and Brock T (1996) Translocation of cytosolic phospholipase A2 to the nuclear envelope elicits topographically localized phospholipid hydrolysis. Biochem J 318, 797-803. Pierce KL, Gil DW, Woodward DF, and Regan JW (1995) Cloning of human prostanoid receptors. Trends Pharmacol Sci 16, 253-256. Przywara DA, Bhave SV, Bhave A, Wakade TD, and Wakade AR ( 1991) Stimulated rise in neuronal calcium is faster and greater in the nucleus than the cytosol. Faseb J 5, 217-222. Re RN, Vizard DL, Brown J, and Bryan SE (1984) Angiotensin II receptors in chromatin fragments generated by micrococcal nuclease. Biochem Biophys Res Commun 119, 220-227. Reddy ST, and Herschman HR (1996) Transcellular prostaglandin production following mast cell activation is mediated by proximal secretory phospholipase A2 and distal prostaglandin synthase 1. J Biol Chem 271, 186-191. Reddy ST, and Herschman HR (1997) Prostaglandin synthase-1 and prostaglandin synyhase-2 are coupled to distinct phospholipases for the generation of prostaglandin D2 in activated mast cells. J Biol Chem 272, 3231-3237. Robertson RP (1986) Characterization and regulation of prostaglandin and leukotriene receptors: an overview. Prostaglandins 31, 395-411. Rosenthal W, Hescheler J, Trautwein W, and Schultz G (1988) Control of voltage-dependent Ca2+ channels by G proteincoupled receptors. FASEB J 2, 2784-2790. Rubins JB, Benditt JO, Dickey BF, and Riedel N (1990) GTPbinding proteins in rat liver nuclear envelopes. Proc Natl Acad Sci USA 87, 7080-7084. Saffitz JE, Nash JA, Green KG, Luke RA, Ransnas LA, and Insel PA (1994) Immunoelectron microscopic identification of cytoplasmic and nuclear G s! in S49 lymphoma cells. FASEB J 8, 252-258. Sakai H, Kumano E, Ikari A, and Takeguchi N (1995) A gastric housekeeping Cl- channel activated via prostaglandin EP3 receptor-mediated Ca2+/nitric oxide/cGMP pathway. J Biol Chem 270, 18781-18785. Schievella AR, Regier MK, Smith WL, and Lin L-L (1995) Calcium-mediated translocation of cytosolic phospholipase A2 to the nuclear envelope and endoplasmic reticulum. J Biol Chem 270, 30749-30754. Schmalz D, Hucho F, and Buchner K (1998) Nuclear import of protein kinase C occurs by a mechanism distinct from the mechanism used by proteins with a classical nuclear localization signal. J Cell Sci 111, 1823-1830. Schuster VL (1998) Molecular mechanisms of prostaglandin transport. Annu Rev Physiol 60, 221-242.

Shitashige M, Morita I, and Murota S (1998) Different substrate utilization between prostaglandin endoperoxide H synthase-1 and -2 in NIH3T3 fibroblasts. Biochim Biophys Acta 1389, 57-66. Simonson MS, Herman WH, and Dunn MJ (1994) PGE 2 induces c-fos expression by a cAMP-independent mechanism in glomerular mesangial cells. Exp Cell Res 215, 137-144. Smith WL (1997) Molecular biology of prostanoid biosynthetic enzymes and receptors. In: Honn KV (ed) Eicosanoids and other bioactive lipids in cancer, inflammation and radiation injury 2. Plenum Press, New York, pp 989-1011. Smith WL, De Witt D, Arakawa T, and Spencer AG (1997) Independent prostanoid biosynthetic systems associated with prostaglandin endoperoxide synthases-1 and -2. Thrombosis and Haemostasis 78, 627-630. Spencer AG, Woods JW, Arakawa T, Singer II, and Smith WL (1998) Subcellular localization of prostaglandin endoperoxide H synthases-1 and -2 by immunoelectron microscopy. J Biol Chem 273, 9886-9893. Suba EA, and Roth BL (1987) Prostaglandins activate phosphoinositide metabolism in rat aorta. Eur J Pharmacol 136, 325-332. Takei Y, Takahashi K, Kanaho Y, and Katada T (1994) Possible involvement of a pertussis toxin-sensitive GTP-binding protein in protein transport into nuclei isolated from rat liver. J Biochem 115, 578-583. Tanaka K, Shibuya I, Kabashima N, Ueta Y, and Yamashita H (1998) Inhibition of voltage-dependent calcium channels by prostaglandin E2 in rat melaotrophs. Endocrinology 139, 4801-4810. Tang S, Rogg H, Schumacher R, and Dzau VJ (1992) Characterization of nuclear angiotensin-II-binding sites in rat liver and comparison with plasma membrane receptors. Endocrinology 131, 374-380. Tiganis T, Flint AJ, Adam SA, and Tonks NK (1997) Association of the T-cell protein tyrosine phosphatase with nuclear import factor p97. J Biol Chem 272, 21548-21557. Topham MK, Bunting M, Zimmerman GA, McIntyre TM, Blackshear PJ, and Prescott SM (1998) Protein kinase C regulates the nuclear localization of diacylglycerol kinase +. Nature 394, 697-700. Tucker ON, Dannenberg AJ, Yang EK, Zhang F, Teng L, Daly JM, Soslow RA, Masferrer JL, Woerner BM, Koki AT, and Fahey TJ (1999) Cyclooxygenase-2 expression is upregulated in human pancreatic cancer. Cancer Res 59, 987990 Turpin P, Hay RT, and Dargemont C (1999) Characterization of I%B! nuclear import pathway. J Biol Chem 274, 6804-6812. Umayahara Y, Ji C, Centrella M, Rotwein P, and McCarthy TL (1997) CCAAT/enhancer-binding protein $ activates insulinlike growth factor-1 gene transcription in osteoblasts. J Biol Chem 272, 31793-31800. Vane JR, Bakhle YS, and Botting RM (1998) Cyclooxygenases 1 and 2. Annu Rev Pharmacol Toxicol 38, 97-120. Ventura C, Maioli M, Pintus G, Posadino AM, and Tadolini B (1998) Nuclear opioid receptors activate opioid peptide gene transcription in isolated myocardial nuclei. J Biol Chem 273, 13383-13385.

337

Bhattacharya et al: Nuclear Prostaglandin Receptors Von Knethen A and Brune B. (1997) Cyclooxygenase-2 an essential regulator of NO-mediated apoptosis. Faseb J 11, 887-895. Watabe A, Sugimoto Y, Honda A, Irie A, Namba T, Negishi M, Ito S, Narumiya S, and Ichikawa A. (1993) Cloning and expression of cDNA for a mouse EP1 subtype of prostaglandin E receptor. J Biol Chem 268, 20175-20178. Waybill MM, Yelamarty RV, Zhang YL, Scaduto RC, LaNoue KF, Hsu CJ, Smith BC, Tillotson DL, Yu FT, and Cheung JY (1991) Nuclear calcium gradients in cultured rat hepatocytes. Am J Physiol 261, E49-E57. Wiley JC, Wailes LA, Idzerda RL, and McKnight GS (1999) Role of regulatory subunits and protein kinase inhibitor (PKI) in determining nuclear localization and activity if the catalytic subunit if protein kinase A. J Biol Chem 274, 6381-6387. Williams CS, Smalley W, and DuBois R (1997) Aspirin use and potential mechanisms for colorectal cancer prevention. J Clin Invest 100, 1325-1329. Xie QW, Kashiwabara Y, and Nathan C (1994) Role of transcription factor NF-%B/Rel in induction of nitric oxide synthase. J Biol Chem 269, 4705-4708. Yabana H, Sasaki Y, Narita H, and Nagao T (1995) Subcellular fractions of cyclic AMP and cyclic AMP-dependent protein kinase and the positive ionotropic effects of selective beta 1and beta 2-aderenoceptor agonists in guinea pig hearts. J Cardiovasc Pharmacol 26, 893-898. Yamamoto S, Kawamura K, and James TN (1998) Intracellular distribution of adenylate cyclase in human cardiocytes determined by electron microscopic cytochemistry. Microscopy Res Technique 40, 479-487. Yano T, Yano Y, Uchida M, Murakami A, Hagiwara K, Otani S, and Ichikawa T (1997) The moulation of vitamin E on prostaglandin E2 level and orthine decarboxylase activity at the promotion phase of lung tumorigenesis in mice. Biochem Pharmacol 53, 1757-1759. Zhao C, Fujimoto N and Schichi H (1995) Immunocytochemical localization of prostaglandin E2 receptor subtypes in porcine ocular tissues. I. Uveal tissues. J Ocul Pharmacol Ther 11, 421-435.

Sylvain Chemtob 338

Gene Therapy and Molecular Biology Vol 4, page 339 Gene Ther Mol Biol Vol 4, 339-348. December 1999

Target of rapamycin (TOR) signaling coordinates tRNA and 5S rRNA gene transcription with growth rate in yeast Review Article

Michael C. Schultz Department of Biochemistry, University of Alberta, Edmonton, Canada T6G 2H7 __________________________________________________________________________________________________ Correspondence: Tel: (780)-492-9144; Fax (780)-492-9556; E-mail: michael.schultz@ualberta.ca Key words: TOR; target of rapamycin; transcriptional regulation by TORs; yeast; ribosomal RNA; tRNA; RNA polymerase I; RNA polymerase III; microarrays; transcription factors Received 9 August 1999; accepted 11 October 1999

Summary Transcriptional regulation of genes encoding ribosomal proteins, translation factors, ribosomal RNAs, a n d t h e t R N A s p l a y s a c r i t i c a l r o l e i n c e l l u l a r p h y s i o l o g y b y modulating the availability o f key components of the protein synthetic machinery according to the need for cell growth. Recent work in yeast and mammalian systems has revealed that the target of rapamycin (TOR) signaling pathway functions i n setting the translational output o f the c e l l i n response t o nutrient/growth factor availability. The central components of the TOR pathway are the TOR kinases, which are inhibited by the macrolide antibiotic rapamycin. Using yeast as a model system we have tested if the control of translation by the TOR kinases includes an effect on transcription o f two components o f the translational apparatus, 5S rRNA and the tRNAs. Biochemical studies reveal that polymerase (pol) III transcription o f the 5 S rRNA and the tRNA genes i s regulated by TOR signaling i n yeast. Interference with TOR signaling inhibits the activity o f RNA p o l III and l i k e l y TFIIIB, core components of the pol III transcriptional machinery. The mechanism of inhibition involves an effect that is independent of the repression of translation that results when TOR signaling is impaired. We propose that the TOR kinases are components o f a signaling network that ensures appropriate expression of the protein synthetic machinery under different growth conditions. Considering the conservation of the TOR pathway and the pol III transcription machinery between yeast and human, this regulatory mechanism is likely to be conserved in eukaryotes.

I. A signaling network problem An important determinant of the cell's capacity to synthesize protein is the availability of components of the translational apparatus, particularly the ribosomal (r) RNAs, tRNAs, and the r proteins. How does the cell ensure that the availability of these components is matched to the demand for protein synthesis? Studies performed over the last 20 years using conventional genetic and biochemical techniques have demonstrated that transcription of the genes encoding the rRNAs, tRNAs and r proteins is under growth control in eukaryotes from yeast to human (Warner, 1989; Mager and Planta, 1991; Denis et al., 1993; Clarke et al., 1996; DeRisi et al., 1997; Wang et al., 1997; White, 1997; Grummt, 1999; Reeder, 1999). Measurement of relative transcription rates by DNA microarray and serial analysis of gene expression technologies has further shown that many genes encoding

translation factors are growth regulated at the level of mRNA abundance in yeast (DeRisi et al., 1997) and that translation factors and r proteins account for almost half the genes induced in colorectal epithelium cancers as compared to normal colon cells encode either (Zhang et al., 1997). Hence the transcription of genes encoding key components of the translational apparatus, including non-coding RNAs, varies according to cell growth rate. Three mechanisms, alone or in combination, could account for the coordinated transcriptional regulation of these genes (Figure 1; for simplicity this model excludes effects on protein turnover). If signaling directly regulates translation according to growth factor or nutrient availability, then effects on transcriptional machineries could indirectly result from changes in the rate of translation of a limiting and labile transcription factor (pathway A). Similarly if signaling directly regulates transcription of a polymerase (pol) II gene 339

Schultz: Target of rapamycin (TOR) signaling encoding a limiting and labile component of the pol I and pol III transcriptional machineries, then effects on rRNA and tRNA synthesis could be indirect as well (pathway B). On the other hand the three nuclear transcriptional machineries might be directly regulated in response to nutrient/growth factor signals (pathway C). The challenge is to characterize the network of signaling pathways that coordinately regulates transcription by mechanisms that might involve all effector mechanisms outlined above.

II. Transcriptional regulation of RNA components of the translational apparatus A. The yeast model Studies of model systems will likely identify conserved elements of the regulatory network that coordinates transcription of the eukaryotic rRNA, tRNA and r protein genes. The budding yeast Saccharomyces cerevisiae is expected to be a useful model because growth control of transcription of these genes is well documented in this organism (Warner, 1989; Sethy et al., 1995; Clarke et al., 1996; DeRisi et al., 1997) and sophisticated biochemical and genetic approaches can be brought to bear on the problem. Studies focusing on the tRNA and 5S rRNA genes transcribed by RNA pol III might prove particularly instructive since components of the pol III transcription machinery are highly conserved between yeast and human (T a b l e 1). The yeast pol III transcription machinery, at its most fundamental level, is comprised of the polymerase (14 subunits), TFIIIB (3 subunits) and TFIIIC (6 subunits). The fundamental steps of the pol III transcription cycle are specific binding of TFIIIC to promoter elements, recruitment of TFIIIB to the promoter by virtue of its interaction with TFIIIC, and finally, recruitment of the polymerase. TFIIIB and the polymerase are highly conserved in structure and function between yeast and human (Wang and Roeder, 1998) and the growth control of transcription acts partly at the level of TFIIIB in yeast and metazoans. Our work has therefore concerned the role that protein kinases, as potential components of signaling pathways that impinge on the pol III machinery, play in transcription of the tRNA and 5S rRNA genes of yeast. The following discussion focuses on the role of signaling by the target of rapamycin (TOR) kinases in the regulation of pol III transcription.

B. TOR signaling and the cellular response to nutrient availability The target of rapamycin (TOR) protein kinases that define TOR signaling pathways were originally identified by virtue of their sensitivity to the antibiotic rapamycin. Rapamycin, in complex with a cellular protein of the FK506 binding protein (FKBP) class, binds to and inhibits TOR phosphotransferase activity. The TORs are conserved between yeast and human and belong to the MEC/ATM family of protein kinases (Hunter and Plowman, 1997; Dennis et al., 1999). Budding yeast has two TOR kinases, Tor1p and Tor2p, whereas only one TOR has been

identified in mammalian cells. Literature pertaining to TOR signaling in yeast and higher eukaryotes has been reviewed extensively in recent years, principally with regard to TOR regulation of cytoplasmic functions (for example: Sigal and Dumont, 1992; Cardenas et al., 1994; Kay, 1996; Brown and Schreiber, 1996; Thomas and Hall, 1997; Conlon and Raff, 1999; Dennis et al., 1999; Polymenis and Schmidt, 1999). Here I focus on those features of the pathway in yeast that are most relevant to control of events in the nucleus, particularly the regulation of transcription. The only protein kinases presently known to be downstream of the TORs in yeast are protein kinase C (PKC) and components of the MAP kinase (MAPK) cascade involved in signaling to the actin cytoskeleton (Helliwell et al., 1998b). It is not yet clear if PKC/MAPK signaling is involved in TOR regulation of transcription (see Section II E.). TOR function in transcriptional regulation in yeast is critically important in the cell's global response cell to nutrient limitation. Yeast cells respond to nutrient limitation by entering a metabolically inert state called G0 or stationary phase (reviewed in Werner-Washburne et al., 1993, 1996). Entry into G0 involves biochemical reprogramming events that have two major outcomes at the cellular level: 1) lowered energy consumption, and 2) increased stress resistance. Energy consumption declines as a consequence of the repression of translation in G0, which occurs in concert with repression of transcription of the genes encoding many components of the translational apparatus, including the rRNAs, tRNAs and r proteins. The repression of rRNA, tRNA and r protein transcription is particularly important from an energetic viewpoint because the production of these RNAs accounts for most RNA synthesis in the nucleus (Nierras and Warner, 1999). Cells also improve their chance of surviving in G0 by inducing a number of stress response genes (for example heat shock proteins). Transcriptional induction of these genes is a further hallmark of the yeast response to starvation. Recent evidence suggests that the TOR pathway in yeast sets the transcriptional output of the genes encoding the G0specific proteins in response to nutrient availability. In a landmark paper in the field, Barbet et al. (1996) showed that interference with TOR signaling by treatment of yeast cells with rapamycin causes a rapid and global repression of translation as well as transcriptional induction of several G0specific pol II genes. Similar results were obtained using cells that conditionally express functional Tor2p (in a tor1! background). More generally Barbet et al. showed that Tor1p and Tor2p function redundantly to control many aspects of the cellular response to starvation, including cell cycle arrest with 1n DNA content, failure to reach START, enlargement of the vacuole, accumulation of glycogen, and acquisition of thermotolerance. Since transcriptional repression of genes encoding the RNA components of the translational apparatus is an integral component of the starvation response of yeast, it is an attractive possibility that the TOR pathway also regulates pol I and pol III transcription according to nutrient availability. We tested this possibility by biochemical methods.

340

Gene Therapy and Molecular Biology Vol 4, page 341

Figure 1: Control mechanisms whose interplay might regulate transcription of key components of the translational apparatus (box) according to nutrient supply or the availability of growth factors.

Transcriptional machinery Yeast Function Pol III Rpo31p 160 kDa subunit pol III (Rpc160) Rpc82p 82 kDa subunit pol III Rpc53p 47 kDa subunit pol III Rpc40p Rpc34p Rpc31p Rpb6p # (Rpo26) Rpc128p (Ret1) Rpb5p Rpc25p Rpc12p (Rpc10) Rpc19p# Rpb8p Rpb10p

References

Related human protein

References

YPD*/MIPS †

Sepehri and Hernandez, 1997 Wang and Roeder, 1997 Ittmann et al., 1993

40 kDa subunit pol III 34 kDa subunit pol III 31 kDa subunit pol III 18 kDa subunit pol I/II/III 130 kDa subunit pol III

" " " "

hRPC155 (pol III largest subunit) hRPC62 BN51 (BHK21) temperature sensitivity complementing pol I subunit RPA40 hRPC39 hRPC32 pol II 14.4 kDa polypeptide

Unknown

25 kDa subunit pol I/II/III 25 kDa subunit pol III 7.7 kDa subunit pol I/II/III 16 kDa subunit pol I/III 16 kDa subunit pol I/II/III 8.3 kDa subunit pol I/II/III

Unknown

" "

Unknown Unknown

" "

Unknown Unknown

Unknown

" "

YPD Wang and Roeder, 1997 Wang and Roeder, 1997 YPD

TFIIIB Spt15p

TATA binding protein (TBP) Brf1p (Pcf4p) 70 kDa subunit, related to TFIIB Tfc5p 90 kDa subunit

Willis, 1993

TBP

White et al., 1992

Willis, 1993

TFIIIB90 (TAF3C)

Wang and Roeder, 1995

Kassavetis et al., 1995

Unknown

Table 1: Conservation between yeast and human of the core components of the pol III transcriptional machinery. Only selected references from the text are cited in this table. *Yeast protein database: http://www.proteome.com/databases/YPD/index.html. † Munich Information Center for Protein Sequences: http://www.mips.biochem.mpg.de/proj/yeast/catalogues/funcat/fc04_03_01.html. # Known to exist as phosphoproteins in yeast (Thuriaux and Sentenac, 1992).

341

Schultz: Target of rapamycin (TOR) signaling

C. TOR regulation of rRNA and tRNA transcription Rapamycin was used to analyze the role of TOR signaling in the regulation of pol I and pol III transcription (Zaragoza et al., 1998). Rapamycin treatment of cells was shown to significantly repress specific transcription by pol I and pol III in crude extracts from wild type strains. Since rapamycin treatment did not repress pol III transcription in extracts from a mutant lacking Fpr1p, the protein to which rapamycin must bind in order to perturb previously defined TOR functions related to the starvation response, we concluded that signals conveyed by the TOR pathway regulate rRNA and tRNA transcription in yeast. In this way the TOR pathway is thought to ensure that production of the RNA components of the translational apparatus is coordinated with the availability of amino acids for protein synthesis. At least three mechanisms could account for the regulation of transcription by pol I and pol III according to nutrient availability (Figure 1). Importantly, because interference with TOR signaling results in a rapid global repression of translation in yeast (Barbet et al., 1996), the effect of rapamycin on pol I/III transcription could be a secondary consequence of a decreased rate of protein synthesis (pathway A in Figure 1). This notion is supported by the observation that cycloheximide treatment also represses pol III transcription in some strain backgrounds (Dieci et al., 1995). These results however do not exclude the possibility that translation-independent effects of TOR signaling impinge on the pol III transcriptional machinery (pathway C in Figure 1). Indeed, using extracts from wild type cells and a tor2 temperature sensitive mutant in temperature shift experiments we demonstrated that interference with TOR function in vitro represses pol III transcription independently of translation (Zaragoza et al., 1998). Whether or not this translation-independent mechanism accounts fully for the repression of pol III transcription seen when cells are treated with rapamycin remains to be determined. None the less we are at a point now where we can begin to identify direct target/s of the TOR signaling pathway from among the components of the pol III transcriptional machinery. Transcription complementation experiments have been used to identify, as biochemical fractions, the components of the pol III machinery that are inactivated when TOR signaling is perturbed. Add-back of TFIIIB purified from control cells significantly restores activity to rapamycin-treated extracts, suggesting that TFIIIB is a target of TOR signaling. Because TFIIIB does not fully restore the activity of rapamycin-treated extracts, another component of the transcriptional machinery is expected to be limiting in this circumstance. That other component is the polymerase, which has lower elongation (bulk) activity in rapamycin-treated as compared to control extracts, and which in partially purified form (from control extract) is able to stimulate transcription in rapamycin-treated extract supplemented with a saturating amount of TFIIIB. The

polymerase is therefore a target of a TOR pathway. The TORresponsive subunits of TFIIIB and pol III remain to be identified, as does the mechanism of their regulation by TOR signaling. In view of previous reports that TFIIIB activity is regulated according to nutrient/growth factor availability it was not surprising to discover that TFIIIB is sensitive to TOR signaling in yeast. Regulation of the pol III enzyme however has not been described in any organism. On the other hand the activity of pol III is known to be sensitive to misregulation of an enzyme involved in signaling. Thus van Zyl et al. (1992) reported that the pol III enzyme is repressed in yeast cells lacking the Tpd3p regulatory subunit of protein phosphatase 2A (PP2A). Since PP2A is a component of the TOR pathway that regulates translation according to nutrient availability in yeast (Di Como and Arndt, 1996; Jiang and Broach, 1999), the results obtained to date may reflect the existence of pathways for the direct regulation of translation and pol III transcription that share (not unexpectedly) more signaling components than just the TOR kinases.

D. TOR regulation of pol II transcription in yeast: implications for the regulation of pol III subunit transcription Because TOR signaling regulates many aspects of the cellular response to nutrient limitation it is interesting to consider our data on pol III transcription and TOR signaling in the context of what is known about genome wide-changes in pol II transcription that occur as yeast cells enter G0. Global changes in the pattern of gene expression in response to nutrient depletion have been characterized by DeRisi et al. (1997) using DNA microarray methodology. This data is available on the world wide web at the yeast protein database (http://www.proteome.com/databases/YPD/index.html). I queried this database with respect to the genes encoding the known core components of the pol III transcriptional machinery. An interesting pattern emerged from this analysis (Figure 2). From among the 21 genes for which data are available, 9 show a significant change in expression level (mRNA abundance) between the actively growing and nutrientlimited states. All genes whose expression levels change are repressed upon nutrient depletion. Intriguingly from the viewpoint of our results, all repressed genes are subunits of RNA pol III. This suggests a working model in which the repression of pol III enzyme activity observed when cells are treated with rapamycin is due to decreased transcription of pol III subunit genes and ultimately decreased abundance of the enzyme. In other words, the results are consistent with the hypothesis that TOR signaling regulates specific pol III transcription in part by regulating transcription of pol III enzyme subunits. This indirect mechanism of regulation of tRNA/5S rRNA transcription is represented as pathway B in Figure 1. From the available data we propose that two components of the pol III transcriptional machinery, TFIIIB and the polymerase, are regulated by TOR signaling. TOR signaling acts on the transcriptional machinery by a direct mechanism and perhaps an indirect mechanism involving an effect on 342

Gene Therapy and Molecular Biology Vol 4, page 343

Figure 2: Transcriptional regulation of components of the core pol III transcriptional machinery upon nutrient depletion in yeast (data from DeRisi et al., 1997). The solid gray line indicates no difference (ND) between the starting level of transcript abundance and the level of transcript abundance at the indicated time points of culture. The dashed gray line indicates the cutoff below which repression is considered to be significant in this type of analysis. For simplicity the gene names are not given for the subunits of TFIIIB and TFIIIC.

transcription of the subunits of the polymerase. We do not rule out direct regulation of pol III activity, especially since pol III is a phosphoprotein in yeast (reviewed in Thuriaux and Sentenac, 1992; Table 1). Many pol II genes required for translation, in addition to the subunits of RNA pol III, are regulated at the transcriptional level according to nutrient availability (DeRisi et al., 1997). From among these it has been clearly demonstrated that transcription of several r protein genes is regulated by TOR signaling (Powers and Walter, 1999). Powers and Walter (1999) make the reasonable argument that one or more of the transcriptional activators previously implicated in the regulation of r protein genes is likely to be under TOR control. An interesting candidate from the viewpoint of possible TOR control of pol III subunit transcription is Abf1p, which regulates expression of the Rpc40p subunit of pol III (Della Seta et al., 1990) and many r protein genes. A simple relationship between TOR signaling and transcriptional regulation by Abf1p however is not likely, since Rpo31p, another pol III subunit under Abf1p control (Della Seta et al., 1990), is not down-regulated in G0 (expression data from DeRisi et al., 1997). Since r protein genes with the same pattern of regulation in G0 do not all have the same (potential) regulatory elements in their upstream regions, it may be that the common effector of changes in transcriptional activity is a chromatin remodeling machine that can be recruited to promoters by different DNA-binding transcription factors.

E. Genetic interactions of components of the TOR signaling machinery that may be relevant to pol III transcriptional regulation As outlined above the biochemical data obtained by rapamycin treatment of yeast establish a link between pol III transcription and TOR signaling that is also suggested by evidence that the Tpd3p regulatory subunit of PP2A is involved in TOR signaling and in the regulation of pol III transcription. There are additional genetic data supporting a link between TOR signaling and pol III transcription (Figure 3). Alleles of the SSD1 gene have been recovered in screens for suppressors of polymerase mutations and mutations in various signaling molecules. Thus, conditional alleles of a number of pol III subunits are suppressed by SSD1-v (Stettler et al., 1993), which also suppresses deletion of the TOR pathway component sit4 (Sutton et al., 1991a, b). Other genetic interactions of SSD1 are intriguing in terms of the TOR signaling pathway that impinges upon the pol III transcriptional machinery. SSD1 suppresses mutants in the PKC/MAPK cascade of yeast (Costigan et al., 1992; Lee et al., 1993) and PKC has been placed downstream of the TORs in a MAPK signaling pathway that regulates cell wall biogenesis and probably other cellular functions (Helliwell et al., 1998a, b). These genetic interactions suggest a functional link between the TORs, PKC and pol III transcription, although direct evidence of this link has not been provided. Indeed, the role of PKC/MAPK signaling in the regulation of

343

Schultz: Target of rapamycin (TOR) signaling cell. Considering the further connections between PKC/MAPK and Ras/PKA signaling (reviewed in Heinisch et al., 1999), it is possible that the regulatory network that governs the rate of pol III transcription in yeast involves threeway cross-talk between TOR, PKA and PKC signaling pathways (Figure 4).

F. Related unresolved questions in yeast

F i g u r e 3 : Interactions of TOR signaling components with the pol III transcriptional machinery. Arrows denote effects of TOR signaling on pol III transcription inferred from genetic or pharmacological manipulation of the indicated TOR pathway component. MS and SL respectively indicate multicopy suppression and synthetic lethality with the nominated (boxed) mutant alleles.

pol III transcription may relate to a plasma membrane stress response rather than to the nutrient response. Among other events (Heinisch et al., 1999) the PKC/MAPK pathway of yeast regulates pol I transcription in reaction to cell membrane stretch (Nierras and Warner, 1999). Given the coregulation of pol I and pol III transcription during the yeast life cycle, the genetic evidence indicating a functional relationship between pol III transcription and the PKC/MAPK cascade possibly reflects signaling events that coordinate transcription with cell membrane synthesis required for growth and division. In this respect it is intriguing to note that in some strain backgrounds a HIS3 disruption allele of rpc53 confers a cell lysis defect at 38Ë&#x161;C (Mann et al., 1992) that is reminiscent of the lysis phenotype of pkc1 mutants (Watanabe et al., 1994). Besides TOR-dependent signaling events the cellular response to nutrient availability involves Ras/protein kinase A (PKA) signaling (see Powers and Walter [1999] for a recent discussion of the literature). Although the relationship between pol III transcription and signaling through the TORs and PKA remains to be fully characterized, it is known that SSD1 suppresses a mutation in the regulatory subunit (Bcy1p) of PKA that causes misregulation of the enzyme (Wilson et al., 1991), as well as various mutations in the transcriptional and TOR signaling machineries (Figure 3). These relationships may reflect the existence of a PKA/TOR signaling network that sets the rate of pol III transcription according to the physiological state of the

TOR signaling is likely to regulate pol III transcription by direct effects on the transcriptional machinery, and by indirect effects at the level of translation and pol II transcription of genes encoding the various components of the pol III transcriptional machinery. The indirect effects ultimately control steady-state protein abundance. Nutrientdependent changes in the abundance of proteins involved in pol III transcription have not been comprehensively analyzed in yeast. However it is clear that TBP protein levels decline dramatically in G0 (Walker et al., 1997) and that Brf1p (TFIIIB70), which is limiting for TFIIIB activity in extracts from early G0 cells (Sethy et al., 1995), is rapidly depleted when cells are treated with cycloheximide (Dieci et al., 1995). In the long term studies aimed at characterizing indirect mechanisms of TOR action on the pol III transcriptional machinery should take these observations into account, especially in view of evidence that autophagy and the turnover of a number of proteins are under TOR control in yeast (Berset et al., 1998; Noda and Ohsumi, 1998; Schmidt et al., 1998).

III. Comparing yeast with higher eukaryotes Considering the conservation of the TOR kinases and the signaling molecules that interact with the TORs (Table 2) it is not surprising that TOR signaling contributes to the regulation of pol I transcription in mammalian cells (Mahajan, 1994) as in yeast (Zaragoza et al., 1998; Powers and Walter, 1999). The notion that TOR signaling also regulates pol III transcription in mammalian cells is intriguing and is encouraged by the evident conservation, between yeast and human, of the transcriptional machinery and the fundamental steps of the transcription cycle (Section II A.). While the yeast results provide a framework in which to study TOR regulation of pol III transcription in mammalian cells such an analysis will be complicated by two facts, 1) that rapamycin inhibits progression through G1 in various mammalians cell types (Cardenas et al., 1994), and 2) that pol III transcription is repressed in G1 (White et al., 1995). The challenge will be to uncover TOR effects on the transcriptional machinery that are independent of TOR-dependent regulation of the cell cycle. In view of the available data it is apparent that a conserved function of the TORs is to regulate the overall protein synthetic capacity of the cell. The TORs effect this control in yeast partly by regulating production of the principal non-coding RNAs required for translation (the tRNAs and rRNAs) as well as transcription of genes encoding r proteins. In mammalian cells TOR signaling regulates the translational machinery partly at the level of rRNA (pol I)

344

Gene Therapy and Molecular Biology Vol 4, page 345

Figure 4: Protein kinases in yeast implicated in the control of transcription in response to nutrient availability and membrane stretch (references in text). No kinases besides those indicated have been directly implicated in the regulatory events depicted.

Signaling molecules Yeast

Function

References

Related human protein

References

Tor1p

Protein kinase

Thomas and Hall, 1997; Dennis et al., 1999

mTOR (FRAP1/RAFT)

Tor2p Pph21/22p

Protein kinase " Redundant protein Sneddon et al., 1990 phosphatase 2A (PP2A) catalytic subunits Regulatory subunit of van Zyl et al., 1992 PP2A

Thomas and Hall, 1997; Dennis et al., 1999 As above Peterson et al., 1999; YPD*

Tpd3p

Cdc55p

Pkc1p

Regulatory subunit of PP2A Phosphatase 2A-like catalytic subunit Binding partner of Sit4p and Pph21/22p Protein kinase C

Mpk1p (Slt2)

A MAP kinase

Sit4p

Tap42p

" Protein phosphatase 2 (formerly 2A), catalytic subunit, " isoform (PPP2CB) Protein phosphatase 2 (formerly 2A), regulatory subunit A (PR 65), " isoform (PPP2R1B) Protein phosphatase 2 regulatory subunit B # isoform (PPP2R2A) Serine/threonine protein phosphatase 6 (PPP6C)

YPD

Di Como and Arndt, 1996

Murata et al., 1997

Levin and Errede, 1995; Herskowitz, 1995; Helliwell et al., 1998

PKC isoforms #, " and $ (phospho-lipids, diacylglycerol, Ca2+ cofactors) BMK1 " kinase (ERK5)

YPD

Healy et al., 1991 Sutton et al., 1991

YPD YPD

YPD

T a b l e 2 : Conservation between yeast and human of TOR signaling components possibly involved in the regulation of pol III transcription. For human proteins specific references are given only when the nominated protein has been experimentally implicated in TOR signaling. Other related human proteins are as given in the yeast protein database (*http://www.proteome.com/databases/YPD/index.html).

345

Schultz: Target of rapamycin (TOR) signaling transcription, and, through regulation of the translation of specific mRNAs, also at the level of production of r proteins and other protein components of the translational machinery (reviewed in Dennis et al., 1999). Considering the shared functions of the TORs it is likely that yeast will continue to be a valuable tool for characterizing the TOR signaling network involved in the regulation of pol I and pol III transcription in eukaryotes.

Acknowledgments I thank Karen Robinson for critical comments on the manuscript. Operating support provided by the Canadian Cancer Society through the National Cancer Institute of Canada. The author is a Senior Scholar of the Alberta Heritage Foundation for Medical Research.

References Barbet, N.C., Schneider, U., Helliwell, S.B., Stansfield, I., Tuite, M.F., and Hall, M.N. (1 9 9 6 ) TOR controls translation initiation and early G1 progression in yeast. M o l B i o l C e l l 7, 25-42. Berset, C., Trachsel, H., and Altmann, M. (1 9 9 8 ) The TOR (target of rapamycin) signal transduction pathway regulates the stability of translation initiation factor eIF4G in the yeast Saccharomyces cerevisiae. P r o c N a t l Acad Sci USA 95, 4264-4269. Brown, E.J., and Schreiber, S.L. (1 9 9 6 ) A signaling pathway to translational control. C e l l 86, 517-520. Cardenas, M.E., Lorenz, M., Hemenway, C., and Heitman, J. (1 9 9 4 ) Yeast as model T cells. P e r s p D r u g D i s c o v D e s i g n 2, 103-126. Clarke, E.M., Peterson, C.L., Brainard, A.V., and Riggs, D.L. (1 9 9 6 ) Regulation of the RNA polymerase I and III transcription systems in response to growth conditions. J B i o l C h e m 271, 22189-22195. Conlon, I., and Raff, M. (1 9 9 9 ) Size control in animal development. C e l l 96, 235-244. Costigan, C., Gehrung, S., and Snyder, M. (1 9 9 2 ) A synthetic lethal screen identifies SLK1, a novel protein kinase homolog implicated in yeast cell morphogenesis and cell growth. M o l C e l l B i o l 12, 1162-1178. Della Seta, F., Treich, I., Buhler, J.M., and Sentenac, A. (1 9 9 0 ) ABF1 binding sites in yeast RNA polymerase genes. J B i o l C h e m 265, 15168-15175. Denis, M.G., Chadeneau, C., Lecabellec, M.T., LeMoullac, B., LeMevel, B., Meflah, K., and Lustenberger, P. (1 9 9 3 ) Over-expression of the S13 ribosomal protein in actively growing cells. Int J Cancer 55, 275-180. Dennis, P.B., Fumagalli, S., and Thomas, G. (1 9 9 9 ) Target of rapamycin (TOR), balancing the opposing forces of protein synthesis and degradation. C u rr Op in Ge ne t D e v 9, 49-54. DeRisi, J.L., Iyer, V.R., and Brown, P.O. (1 9 9 7 ) Exploring the metabolic and genetic control of gene expression on a genomic scale. S c i e n c e 278, 680-686. Di Como, C.J., and Arndt, K.T. (1 9 9 6 ) Nutrients, via the Tor proteins, stimulate the association of Tap42 with type 2A phosphatases. G e n e s D e v 10, 1904-1916.

Dieci, G., Duimio, L., Peracchia, G., and Ottonello, S. (1 9 9 5 ) Selective inactivation of two components of the multiprotein transcription factor TFIIIB in cycloheximide growth-arrested yeast cells. J B i o l C h e m 270, 13476-13482. Grummt, I. (1 9 9 9 ) Regulation of mammalian ribosomal gene transcription by RNA polymerase I. P r o g N u c l e i c A c i d R e s M o l B i o l 62, 109-154. Healy, A.M., Zolnierowicz, S., Stapleton, A.E., Goebl, M., DePaoli-Roach, A.A., and Pringle, J.R. (1 9 9 1 ) CDC55, a Saccharomyces cerevisiae gene involved in cellular morphogenesis, identification, characterization, and homology to the B subunit of mammalian type 2A protein phosphatase. M o l C e l l B i o l 11, 5767-5780. Heinisch, J.J., Lorberg, A., Schmitz, H.P., and Jacoby, J.J. (1 9 9 9 ) The protein kinase C-mediated MAP kinase pathway involved in the maintenance of cellular integrity in saccharomyces cerevisiae. M o l M i c r o b i o l 32, 671-680. Helliwell, S.B., Howald, I., Barbet, N., and Hall, M.N. (1 9 9 8 a ) TOR2 is part of two related signaling pathways coordinating cell growth in Saccharomyces cerevisiae. G e n e t i c s 48, 99112. Helliwell, S.B., Schmidt, A., Ohya, Y., and Hall, M.N. (1 9 9 8 b ) The Rho1 effector Pkc1, but not Bni1, mediates signalling from Tor2 to the actin cytoskeleton. C u r r B i o l 8, 12111214 . Herskowitz, I. (1 9 9 5 ) MAP kinase pathways in yeast C e l l 80, 187. Hunter, T., and Plowman, G.D. (1 9 9 7 ) The protein kinases of budding yeast, six score and more. T r e n d s B i o c h e m S c i 22, 18-22. Ittmann, M., Ali, J., Greco, A., and Basilico, C. (1 9 9 3 ) The gene complementing a temperature-sensitive cell cycle mutant of BHK cells is the human homologue of the yeast RPC53 gene, which encodes a subunit of RNA polymerase C (III). C e l l Growth Differ 4, 503-511. Jiang, Y., and Broach, J.R. (1 9 9 9 ) Tor proteins and protein phosphatase 2A reciprocally regulate Tap42 in controlling cell growth in yeast. EMBO J 18, 2782-2792. Kay, J.E. (1 9 9 6 ) Structure-function relationships in the FK506binding protein (FKBP) family of peptidylprolyl cis-trans isomerases. B i o c h e m J 314, 361-385. Kassavetis, G.A., Nguyen, S.T., Kobayashi, R., Kumar, A., Geiduschek, E.P., and Pisano, M. (1 9 9 5 ) Cloning, expression, and function of TFC5, the gene encoding the B" component of the Saccharomyces cerevisiae RNA polymerase III transcription factor TFIIIB. P r o c N a t l A c a d S c i U S A 92, 9786-9790. Lee, K.S., Irie, K., Gotoh, Y., Watanabe, Y., Araki, H., Nishida, E., Matsumoto, K., and Levin, D.E. (1 9 9 3 ) A yeast mitogenactivated protein kinase homolog (Mpk1p) mediates signalling by protein kinase C. M o l C e l l B i o l 13, 30673075. Levin, D.E., and Errede, B. (1 9 9 5 ) The proliferation of MAP kinase signaling pathways in yeast. C u r r O p i n C e l l B i o l 7 , 197-202. Mager, W.H., and Planta, R.J. (1 9 9 1 ) Coordinate expression of ribosomal protein genes in yeast as a function of cellular growth rate. M o l C e l l B i o c h e m 104, 181-187.

346

Gene Therapy and Molecular Biology Vol 4, page 347 Mahajan, P.B. (1 9 9 4 ) Modulation of transcription of rRNA genes by rapamycin. I n t J I m m u n o p h a r m a c o l 16, 711-721. Mann, C., Micouin, J.Y., Chiannilkulchai, N., Treich, I., Buhler, J.M., and Sentenac, A. (1 9 9 2 ) RPC53 encodes a subunit of Saccharomyces cerevisiae RNA polymerase C (III) whose inactivation leads to a predominantly G1 arrest. M o l C e l l B i o l 12, 4314-4326. Murata, K., Wu, J., and Brautigan, D.L. (1 9 9 7 ) B cell receptor-associated protein alpha4 displays rapamycinsensitive binding directly to the catalytic subunit of protein phosphatase 2A. Proc Natl Acad Sci USA 94, 10624-10629. Nierras, C.R., and Warner, J.R. (1 9 9 9 ) Protein kinase C enables the regulatory circuit that connects membrane synthesis to ribosome synthesis in Saccharomyces cerevisiae. J B i o l C h e m 274, 13235-13241. Noda, T., and Ohsumi, Y. (1 9 9 8 ) Tor, a phosphatidylinositol kinase homologue, controls autophagy in yeast. J B i o l C h e m 273, 3963-3966. Peterson, R.T., Desai, B.N., Hardwick, J.S., and Schreiber, S.L. (1 9 9 9 ) Protein phosphatase 2A interacts with the 70-kDa S6 kinase and is activated by inhibition of FKBP12-rapamycinassociated protein. Proc Natl Acad Sci USA 96, 4438-4442. Polymenis, M., and Schmidt, E.V. (1 9 9 9 ) Coordination of cell growth with cell division. Curr Opin Genet Dev 9, 76-80. Powers, T., and Walter, P. (1 9 9 9 ) Regulation of ribosome biogenesis by the rapamycin-sensitive TOR-signaling pathway in Saccharomyces cerevisiae. M o l B i o l C e l l 10, 987-1000. Reeder, R.H. (1 9 9 9 ) Regulation of RNA polymerase I transcription in yeast and vertebrates. Prog N u c l e i c A c i d R e s M o l B i o l 62, 293-327. Schmidt, A., Beck, T., Koller, A., Kunz, J., and Hall, M.N. (1 9 9 8 ) The TOR nutrient signalling pathway phosphorylates NPR1 and inhibits turnover of the tryptophan permease. EMBO J 17, 6924-6931. Sepehri, S., and Hernandez, N. (1 9 9 7 ) The largest subunit of human RNA polymerase III is closely related to the largest subunit of yeast and trypanosome RNA polymerase III. Genome Res 7, 1006-1019. Sethy, I., Moir, R.D., Librizzi, M., and Willis, I.M. (1 9 9 5 ) In vitro evidence for growth regulation of tRNA gene transcription in yeast. A role for transcription factor (TF) IIIB70 and TFIIIC. J B i o l C h e m 270, 28463-28470. Sigal, N.H., and Dumont, F.J. (1 9 9 2 ) Cyclosporin A, FK506, and rapamycin, pharmacologic probes of lymphocyte signal transduction. A n n u R e v I m m u n o l 10, 519-560. Sinn, E., Wang, Z., Kovelman, R., and Roeder, R.G. (1 9 9 5 ) Cloning and characterization of a TFIIIC2 subunit (TFIIIC beta) whose presence correlates with activation of RNA polymerase III-mediated transcription by adenovirus E1A expression and serum factors. Ge n e s D e v 9, 675-685. Sneddon, A.A., Cohen, P.T., and Stark, M.J. (1 9 9 0 ) Saccharomyces cerevisiae protein phosphatase 2A performs an essential cellular function and is encoded by two genes. EMBO J 9, 4339-4346.

Stettler, S., Chiannilkulchai, N., Hermann-Le Denmat, S., Lalo, D., Lacroute, F., Sentenac, A., and Thuriaux, P. (1 9 9 3 ) A general suppressor of RNA polymerase I, II and III mutations in Saccharomyces cerevisiae. M o l G e n G e n e t 239, 169176. Sutton, A., Immanuel, D., and Arndt, K.T. (1 9 9 1 a ) The SIT4 protein phosphatase functions in late G1 for progression into S phase. M o l C e l l B i o l 11, 2133-2148. Sutton, A., Lin, F., and Arndt, K.T. (1 9 9 1 b ) The SIT4 protein phosphatase is required in late G1 for progression into S phase. C o l d S p r i n g H a r b S y m p Q u a n t B i o l 56, 7581. Thomas, G., and Hall, M.N. (1 9 9 7 ) TOR signalling and control of cell growth. Curr Opin Cell Biol 9, 782-787. Thuriaux, P., and Sentenac, A. (1 9 9 2 ) Yeast nuclear RNA polymerases. C o l d S p r i n g H a r b o r L a b o r a t o r y P r e s s , Cold Spring Harbor, New York. van Zyl, W., Huang, W., Sneddon, A.A., Stark, M., Camier, S., Werner, M., Marck, C., Sentenac, A., and Broach, J.R. (1 9 9 2 ) Inactivation of the protein phosphatase 2A regulatory subunit A results in morphological and transcriptional defects in Saccharomyces cerevisiae. M o l C e l l B i o l 12, 4946-4959. Walker, S.S., Shen,W.-C., Reese, J.C., Apone, L.M., and Green, M.R. (1 9 9 7 ) Yeast TAFII145 required for transcription of G1/S cyclin genes and regulated by the cellular growth state. C e l l 90, 607-614. Wang, Z., and Roeder, R.G. (1 9 9 5 ) Structure and function of a human transcription factor TFIIIB subunit that is evolutionarily conserved and contains both TFIIB-and highmobility-group protein 2-related domains. Proc Natl Acad Sci USA 92, 7026-7030. Wang, Z., and Roeder, R.G. (1 9 9 7 ) Three human RNA polymerase III-specific subunits form a subcomplex with a selective function in specific transcription initiation. Ge ne s D e v 11, 1315-1326. Wang, Z., and Roeder, R.G. (1 9 9 8 ) DNA topoisomerase I and PC4 can interact with human TFIIIC to promote both accurate termination and transcription reinitiation by RNA polymerase III. M o l C e l l 1, 749-757. Wang, Z., Luo, T., and Roeder, R.G. (1 9 9 7 ) Identification of an autonomously initiating RNA polymerase III holoenzyme containing a novel factor that is selectively inactivated during protein synthesis inhibition. Genes D e v 11, 23712382. Warner, J.R. (1 9 8 9 ) Synthesis of ribosomes in Saccharomyces cerevisiae. M i c r o b i o l R e v 53, 256-271. Watanabe, M., Chen, C.Y., and Levin, D.E. (1 9 9 4 ) Saccharomyces cerevisiae PKC1 encodes a protein kinase C (PKC) homolog with a substrate specificity similar to that of mammalian PKC. J B i o l C h e m 269, 16829-16836. Werner-Washburne, M., Braun, E., Johnston, G.C., and Singer, R.A. (1 9 9 3 ) Stationary phase in the yeast Saccharomyces cerevisiae. M i c r o b i o l R e v 57, 383-401. Werner-Washburne, M., Braun, E.L., Crawford, M.E., and Peck, V.M. (1 9 9 6 ) Stationary phase in Saccharomyces cerevisiae. M o l M i c r o b i o l 19, 1159-1166. White, R.J. (1 9 9 7 ) Regulation of RNA polymerases I and III by the retinoblastoma protein, a mechanism for growth control? Trends Biochem Sci 22, 77-80.

347

Schultz: Target of rapamycin (TOR) signaling White, R.J., Gottlieb, T.M., Downes, C.S., and Jackson, S.P. (1 9 9 5 ) Cell cycle regulation of RNA polymerase III transcription. M o l C e l l B i o l 15, 6653-6662. White, R.J., Jackson, S.P., and Rigby, P.W. (1 9 9 2 ) A role for the TATA-box-binding protein component of the transcription factor IID complex as a general RNA polymerase III transcription factor. P r o c N a t l A c a d S c i U S A 89, 19491953. Willis, I.M. (1 9 9 3 ) RNA polymerase III. Genes, factors and transcriptional specificity. Eur J Biochem 212, 1-11. Wilson, R.B., Brenner, A.A., White, T.B., Engler, M.J., Gaughran, J.P., and Tatchell, K. (1 9 9 1 ) The Saccharomyces cerevisiae SRK1 gene, a suppressor of bcy1 and ins1, may be involved in protein phosphatase function. M o l C e l l B i o l 11, 3369-3373. Zaragoza, D., Ghavidel, A., Heitman, J., and Schultz, M.C. (1 9 9 8 ) Rapamycin induces the G0 program of transcriptional repression in yeast by interfering with the TOR signaling pathway. M o l C e l l B i o l 18, 44634470. Zhang, L., Zhou, W., Velculescu, V.E., Kern, S.E., Hruban, R.H., Hamilton, S.R., Vogelstein, B., and Kinzler, K.W. (1 9 9 7 ) Gene expression profiles in normal and cancer cells. S c i e n c e 276, 12681272.

348

Gene Therapy and Molecular Biology Vol 4, page 349 Gene Ther Mol Biol Vol 4, 349-362. December 1999

DNA structural and sequence determinants for nucleosome positioning * Review Article

Daniel J. Fitzgerald and John N. Anderson Department of Biological Sciences, Purdue University, West Lafayette, Indiana 47907-1392 *This work was supported in part by the Walther Cancer Foundation. __________________________________________________________________________________________________ Correspondence: John N. Anderson, Ph.D., Department of Biological Sciences, Purdue University, 1392 Lilly Hall, West Lafayette, Indiana 47907-1392; Tel: (765)-494-4988; Fax:(765)-494-0876; E-mail: janderso@bilbo.bio.purdue.edu Key Words: DNA, nucleosome, histone octamers, GeneBank database, chromatin Received: 10 August 1999; accepted: 16 August 1999

Summary Positioned nucleosomes are thought to be regulators of genome function but the role of DNA sequence and structure in the control of positioning is poorly understood. We examined the intrinsic curvature of DNA sequences that are known to position histone octamers at single translational sites as a first step to investigate this problem and discovered a conserved pattern of intrinsic DNA curvature that was proposed to direct the formation of nucleosomes to unique positions. The pattern consists of two 50-60 base pair regions of curved DNA separated by preferred lengths of non-curved DNA. The conserved pattern was also seen in all 57 satellite sequences present in the GeneBank database and the distances between successive pairs of curved elements in repeated arrays of satellite monomers were similar to the average spacing of nucleosomes in chromatin. To test the significance of the pattern, ten synthetic DNAs were constructed which contain two regions of curved DNA that are separated by non-curved regions of variable length. Translational mapping of in vitro reconstituted nucleosomes demonstrated that two of the fragments positioned nucleosomes at a single site while most of the remaining fragments positioned octamers at multiple sites spaced at 10 base intervals. In support of the curvaturebased model, the positioning sequences contained non-curved central regions of the same lengths that were seen in natural positioning sequences and displayed an affinity for histone octamers comparable to the strongest known natural positioning sequences. A detailed study was then carried out to identify the features that were responsible for high affinity and unique translational positioning activity. Nucleosomes assembled onto positioning fragments of different lengths shared a common upstream border suggesting that the positioning signals were located on the upstream half of these nucleosomal DNAs. In this region, the compressed minor grooves of the A-tracts did not assume the typical rotational orientation of facing the histone octamer. This unusual orientation was showed to be required for unique positioning since positioning activity was lost upon the insertion of 4 bp between the upstream tracts and the pseudo-dyad region. A permanganate hypersensitive site was also found in this region 1.5 turns from the pseudo-dyad at a site known to display DNA distortion in the nucleosome. The sequence of the hypersite contained a TA step flanked by an oligo-pyrimidine tract and the rotational orientation of the reactive TA step in the nucleosomal DNAs was such that the minor groove faces the histone octamer. Substitutions were made in the region of the hypersite and the resulting constructs tested for affinity for histone octamers and translational positioning. The results revealed that a single base change in the TA step and a few changes in the adjacent tract were sufficient to dramatically reduce affinity and positioning activity in a manner that appeared to be correlated with the presence of a permanganate hypersite. In addition, the rotational orientation of the sequence was shown to be important for function since altering the orientation of the site in a positioning fragment reduced positioning activity and octamer affinity while altering the orientation of the sequence in a nonpositioning fragment had the opposite effects. The 5S rDNA positioning sequence from L. variegatus also contained a permanganate hypersite at 1.5 turns from the pseudo-dyad and other natural positioning sequences were enriched in the sequence motifs that give rise to permanganate hypersensitivity in this location. These results suggest a model in which translational positioning is due to a concerted action between the stabilizing forces associated with the hypersite sequences occupying specific sites within the central three turns of nucleosomal DNA and destabilizing forces which appear when the upstream A- tracts with outward facing minor grooves occupy particular translational positions.

349

Fitzgerald and Anderson: Nucleosome Positioning

I. Introduction The first step in the packaging of DNA into a chromosome is the formation of a complex between 146 base pairs (bp) of DNA and an octamer of histone proteins, resulting in a particle known as a nucleosome (Van Holde, 1993). The difference in stability between potential nucleosome positions along bulk genomic DNA is usually small enough that a number of different positions are equally likely to be occupied on a given DNA sequence (Cao et al., 1998; Lowary & Widom, 1998). However, sequences have been identified that direct the deposition of histone octamers to specific locations. The locational preference is usually considered in terms of the translational positioning which refers to the position of the histone octamer along the DNA molecule and the rotational setting, which gives the local orientation of the DNA relative to the direction of curvature on the octamer surface. Positioned nucleosomes frequently occur in the vicinity of promoters, enhancers and other control sequences in the genome. Consequently, these elements have received considerable attention as potential regulators of genome function. It is now clear that positioned nucleosomes can regulate both transcription and replication by controlling the accessibility of regulatory proteins to DNA (reviewed in Wolffe, 1994; Svaren and Hรถrz, 1996; Beato and Eisfeld, 1997). For example, some regulatory factors bind their target sequence with high affinity only if this sequence is located along DNA separating two adjoining nucleosomes. Other regulatory factors can bind to nucleosomal DNA, but only if the target sequence is presented in a particular rotational orientation. The translational position of a DNA sequence within a nucleosome can also influence its accessibility for protein factor binding. Thus, a complete understanding of genomic control mechanisms in eukaryotes will require a more thorough description of the determinants of nucleosome positioning than is currently available. An understanding of mechanisms that lead to nucleosome positioning would also likely enhance our knowledge of the physical organization of DNA in the nucleosome and in solution. In addition, a strong positioning sequence could be used to control the site of nucleosome residency in vivo which might permit the development of new approaches to regulate gene activity and to further clarify relationships between nucleosome positioning and genome function in the cell. Some progress has been made in our understanding of sequence features that give rise to a preferred rotational orientation of DNA in the nucleosome. Such DNA has been viewed as an anisotropic rod, which may bend more easily in one plane than another. Sequence features that can give rise to anisotropic bendability include short A/T rich and G/C rich elements, which can bend more readily into the minor and major grooves, respectively. When these sequences are arranged in a 10.0 bp period, the A+T rich sequences are oriented such that their minor grooves face the histone surface while the minor grooves of G+C rich sequences face outward (Drew and Travers, 1985; Satchwell et al., 1986; Shrader and Crothers, 1989). Thus,

the proper rotational orientation of such sequences decreases the amount of mechanical work that is required to fold the DNA and results in an increase in nucleosome stability. Intrinsically curved DNA arises from oligo A-tracts arranged in a 10 bp period and such DNAs also generally adopts this preferred rotational orientation in the nucleosome (reviewed in Crothers et al., 1990; Harvey et al., 1995). Intrinsically curved DNA favors nucleosome formation and has been identified within many of the most stable nucleosomes in the genome (Penning et al., 1989; et al., 1990, Widlund et al., 1997; Fitzgerald and Anderson, 1999a). However, anisotropic bendability or intrinsic curvature per se is insufficient to position nucleosomes at single sites (Schrader and Crothers 1990; Fitzgerald and Anderson, 1999a). For example, nucleosomes form preferentially on kinetoplast DNA and on a circular intron segment which are both highly curved, and these sequences assume an optimal rotational setting in the nucleosome. However, nucleosome formation occurs at multiple positions along these circular DNA fragments presumably because of the monotonous bending along these segments (Costanzo et al., 1990, Fitzgerald and Anderson,1999a). A commonly held view is that multiple distinct sequence elements along the nucleosome contribute to the translational positioning of nucleosomes (Trifonov, 1980; Drew & Travers, 1985; Neubauer et al., 1986; Shrader & Crothers, 1990). Consistent with this proposition are studies which have identified sinusoidal patterns of sequence preference along bulk nucleosomal DNA which are likely to reflect particular sequences occupying favorable and avoiding unfavorable sites. For example, AA and AAA are found most often at sites where their minor grooves face the histone surface while longer Atracts are frequently observed in distal turns of the nucleosome and along linker regions (Struhl, 1985; Satchwell et al., 1986). Similarly, dinucleotides displaying low stacking energy are known to prefer particular translational positions along nucleosomal DNAs (Satchwell et al., 1986; Satchwell & Travers, 1989; Ioshikhes et al., 1996). These sequence patterns are magnified in nucleosome positioning DNAs and thus may play some role in dictating sites of nucleosome residency in the genome (Fitzgerald et. al, 1994). The non-uniform distribution of sequence elements along nucleosomal DNAs is likely to be related to the curvature demands that are imposed on the DNA during nucleosome formation and to the non-uniform shape of the DNA superhelix in the mature nucleosome. The recent high resolution crystallographic study of the nucleosome core particle (Luger et al., 1997) confirmed earlier suspicions that nucleosomal DNA does indeed take an irregular path around the histone octamer. Along this path, regions 15 and 45 bps from the pseudo-dyad exhibit marked departures from ideal base stacking. Biochemical studies have also suggested that the DNA in these areas assumes a distinct structure not seen elsewhere along the particle. For example, the region 15 bps from the pseudo-dyad is preferentially attacked by singlet oxygen (Hogan et al., 1987) and the enediyne antibiotic calicheamicin (Kuduvalli et al., 1995). This position is also a hot spot for the formation of thymidine dimers (Gale & Smerdon, 1988) and is a preferred 350

Gene Therapy and Molecular Biology Vol 4, page 351 site of action of the HIV integrase where a distorted (or distortable) DNA structure is required for the action of this enzyme (Pruss et al., 1994 A&B). A recent analysis of a nucleosome from the Adh gene of Drosophila revealed an asymmetric intra-nucleosome structure in the region 1.5 turns from the pseudo-dyad that was suggested to be a result of a unique local DNA conformation (Gao & Benyajati, 1998). In addition, an in vivo positioned nucleosome in C. Glabrata was shown to display sequence-dependent DNA distortion directly adjacent to a transcription factor binding site in the absence of the factor (Zhu & Thiele, 1996). Since a disruption in base stacking along nucleosomal DNA is likely to be both energetically costly and sequencedependent (Crick & Klug, 1975; Zhurkin, 1985; Travers, 1991), the localization of sequences that facilitate DNA distortion to distorted sites within the nucleosome might serve as a translational positioning signal (Neubauer et al., 1986; Travers & Klug, 1987; Luger & Richmond, 1998). It is evident from the above correlation studies that variations in sequence-dependent properties of DNA in the nucleosome such as intrinsic curvature and distortion could be involved in directing nucleosomes to unique sites. However, in spite of the considerable amount that is known about DNA in a nucleosome, the mechanisms responsible for translational positioning remain poorly understood. To this point, all previously published attempts by other laboratories at producing a synthetic sequence that positions nucleosomes at a single site have failed (Schrader and Crothers, 1990; Tanaka et al., 1992; Patterton and Simpson,

1995). In this article, we review our studies on identifying the features that are responsible for the translational positioning of nucleosomes.

II. Results and discussion A. Conserved patterns of curvature in satellite and nucleosome positioning DNA We described the intrinsic curvature of DNA sequences that are known to position histone octamers at single translational sites. The 11 nucleosome positioning sequences that were analyzed in this study were from satellites, promoters and gene coding regions. The computer program used in the analysis is based on the A-tract bending phenomenon and yields predicted DNA curvature values that closely reflect the results of experimental data on DNA curvature ( Eckdahl & Anderson, 1987; Van Wye et al., 1991; Wang et al., 1994; Albert et al., 1995). The analysis uncovered a conserved pattern of intrinsic curvature that was proposed to direct the formation of nucleosomes to unique positions (Fitzgerald et. al,. 1994). As summarized in Figure 1, the pattern consists of two 40-50 base pair regions of curved DNA separated by about 40-50 bp of non-curved DNA and the pseudo dyad was located in the region of low curvature. The conserved pattern was also seen in all 57 satellite sequences present in the GeneBank database and the distances between successive pairs of curved elements in repeated arrays of satellite monomer were similar to the average spacing of nucleosomes in chromatin.

Figure 1. Conserved patterns of DNA curvature in nucleosome positioning and satellite DNAs. The computer program used for this analysis is based on the wedge model for A-tract bending and calculates an index of intrinsic DNA curvature from nucleotide sequence that is called the ENDs ratio (Eckdahl and Anderson, 1997). The ENDs ratio is defined as the ratio of the contour length of a DNA segment to the shortest distance between its ends. The program was used to analyze the curvature along 11 nucleosome positioning sequences and all 57 satellite sequences present in the Gene Bank database and the averages for each sequence set are shown. Each of the sequences displayed the conserved pattern of curvature which consists of two ~ 40-50 bp regions of curvature DNA separated by a ~40-50 bp segment of noncurved DNA. The data is summarized from Fitzgerald et al., 1994.

351

Fitzgerald and Anderson: Nucleosome Positioning An analysis of the nucleotide sequences responsible for the curvature patterns revealed the expected enrichment of oligo A-tracts in a 10.0 bp period on both sides of the A-tract deficient central regions. The studies also demonstrated that the sequence phase of the oligo A-tracts on the left side of the central region was offset from that on the right side by about 3 bp as seen in Figure 2. That is, an insertion of an average of about 3 bp into the central regions was required to bring the two sides of the nucleosomal and satellite DNA into a single phase of 10 bp.

B. Unique translational positioning of nucleosomes on synthetic DNA

Figure 2. Rotational sequence patterns in nucleosome positioning and satellite DNA. Sequences of the forms (A/T)2 (N)4 (G/C) (N)3 (A/T)2 and (A/T)2 (N)3 (G/C) (N)4 (A/T)2 were identified in the nucleosome positioning (A and B) and satellite (C and D) DNA. The distribution of these motifs along the 5' 73 bp halves of the segments were used to align the 11 sequences which comprised each set. The occurrences of these elements was then summed over a window of 3 bp and plotted as a function of sequence position. In panels B and D, G's were inserted at position 73 until the motifs on the 3' side came into phase with those on the 5' side. The mean lengths + SEM of the insertions for the 11 sequences are given above the arrows. The data is from Fitzgerald et al., 1994.

To test the significance of the conserved curvature pattern, ten synthetic DNAs were constructed which contained two regions of curved DNA of the form (A5. T5)(G/C) 5)4 that were separated by non-curved regions of variable length. The fragments were named according to the phase relationship between upstream A-tracts 1-4 and downstream A-tracts 5-8 as shown in Figure 3 . For example, fragment 67 contains 67 bp from the center of tract 4 to the center of tract 5. The two major groups of sequences were the +1 series (fragments 41, 51, and 61) and the -3 series (fragments 47, 57, 67, and 77). The -3 series matches the sequence phase we previously reported as the average for natural nucleosome positioning sequences (Figure 2). Nucleosomes were assembled onto labeled DNA fragments by exchange of histones from chicken erythrocyte mononucleosomes using the salt-dilution method. In order to assess the relative reconstitution efficiencies, samples were reconstituted in the presence of varying amounts of competitor DNA and then analyzed on polyacrylamide-glycerol gels. A summary of the results of these experiments is provided in the top panel of Figure 4. Nucleosomes that formed on all of the synthetic fragments were significantly more stable than those that formed on bulk DNA and on the 5S rDNA nucleosome positioning sequence from L. variegatus . In addition, a general trend seen in the figure is that the -3 fragments form nucleosomes of greater stability than the +1 fragments. Translational mapping of in vitro reconstituted nucleosomes demonstrated that the two longest fragments in the -3 series (#67 and 77) also positioned nucleosomes at a single site as summarized in the bottom panel of Figure 4. This major position, as well as other minor positions, was also seen with the shorter members of the -3 series (fragments 57,47). In contrast, all fragments in the +1 series positioned octamers at multiple sites that were spaced at 10 base intervals. In support of the curvature-based model, the positioning sequences contained non-curved central regions of similar length and phase that were seen in natural positioning sequences and these fragments displayed an affinity for histone octamers comparable to the strongest known natural positioning sequences. In addition, both the upstream and downstream curved segments were shown to be required for unique positioning as revealed by translational mapping studies of nucleosomes assembled onto derivative 67 fragments that lacked these segments (Fitzgerald &Anderson, 1999).

352

Gene Therapy and Molecular Biology Vol 4, page 353

Figure 3. The synthetic sequences. Synthetic DNA fragments were constructed which contained two regions of curved DNA separated by non-curved regions of variable length. To generate the fragments, two 45 bp synthetic oligonucleotides of the form (A 5 .T5 )(G/C)5 )4 were inserted into the polylinker of pUC 18 to yield a recombinant parent plasmid. Insertions and deletions were then made in the region separating the two synthetic bending elements. The fragments were named according to the phase relationship between upstream A-tracts 1-4 and downstream A-tracts 5-8. For example, fragment 67 contains 67 bp from the center of tract 4 to the center of tract 5. The two major groups of sequences were the +1 series (fragments 41, 51, and 61) and the -3 series (fragments 47, 57, 67, and 77). The plasmids were used as templates in the PCR to produce end-labeled or uniformly labeled fragments that ranged in length from 178 bp to 235 bp. These fragments were then used for nucleosome reconstitution.

Figure 4. Characteristics of nucleosomes assembled onto the synthetic segments. Nucleosomes were reconstituted onto the fragments by the standard exchange-salt dilution method (1.0M NaCl to 0.1 M NaCl) using H1/H5 deficient nucleosomes from chicken erythrocytes as a source of core histones (Drew and Travers, 1985). Energies of reconstitution relative to fragment 67 were determined by competitive reconstitution analysis (Shrader and Crothers, 1990). Translational positioning was assessed by exonuclease III (Exo III) and micrococcal nuclease (MNase) digestion studies. Hydroxyl radical cleavage analysis and DNase 1 digestion studies were used to determine the rotational orientation of sequences along the nucleosomal DNAs. The hydroxyl radical cleavage data in conjunction with the results of the Exo III and MNase experiments was also used in assigning the position of the pseudo-dyad. Note that the upstream curved element containing A-tracts 1-4 assumed a variable rotational orientation between constructs that depended on the phase: the narrow minor grooves of the tracts pointed inward (IN) in the +1 series but outward (OUT) in the positioning -3 series. Data summarized from Fitzgerald and Anderson, 1998.

353

Fitzgerald and Anderson: Nucleosome Positioning All of the synthetic fragments assumed standard pitches when packaged into nucleosomes of about 10.010.3 bp/turn in distal areas and 10.5-10.7bp/turn in the central three turns (Hayes, et al., 1990, 1991). Consequently, the minor grooves of A-tracts in the +1 series should display the same rotational orientation on both sides of the nucleosomal dyad while the opposite should be true for the -3 series fragments. Hydroxy radical cleavage analysis and DNase I digestion studies confirmed these expectations as summarized in the bottom of Figure 4. These results showed that the synthetic curved element on the downstream side of the nucleosomal dyad in all fragments existed in the same rotational orientation with the narrow minor grooves of A-tracts 5-8 facing inward toward the histone surface. We therefore suggested that these tracts set the rotational orientation of all fragments. In contrast, the curved element on the other side of the nucleosome displayed variable rotational orientations which appears to be related to translational positioning. The narrow minor grooves of the A-tracts 1-4 faced inward in nucleosomes assembled onto +1 series fragments but outward in the -3 series nucleosomes.

C. The positioning signals 1. Location A detailed study was then carried out to identify the features that were responsible for high nucleosome stability and unique translational positioning activity of the fragments in the -3 series. The unique translational position in fragments 67 and 77 and the major position in 57 share a common upstream border as shown schematically in Figure 5. We therefore proposed that the translational positioning signals were located on the upstream half of these nucleosomal DNAs. One unusual feature that characterized

this upstream region was that the compressed minor grooves of the A-tracts did not assume the typical rotational orientation of facing the histone octamer as described above. An additional unusual feature in this region was detected using the oxidative reagent potassium permanganate. Previous studies have shown that permanganate oxidizes mononucleotides with a relative rate of T>>C>A=G (Hayatsu & Ukita, 1967). Double stranded DNA displays up to 20-fold lower reactivity than single stranded DNA (HĂ¤nsler & Rokita, 1993) with the highest reactivities also being observed at thymines. Since the 5-6 ! bond of pyrimidines is expected to be involved in base-stacking with the adjacent base in B-DNA, hyperreactivity of this reagent is considered to be indicative of a melted or distorted DNA structure (reviewed in Nielsen, 1990; John & Workman, 1998). Although permanganate has been used to detect DNA distortion around transcription start sites, enhancers, and other DNA binding sites (Borowiec et al., 1987; Hsieh et. al., 1993; Summers et al., 1997; Wilkins & Lis, 1999), it had not been used previously to study DNA distortion in the nucleosome. As shown in Figure 6, a single permanganate hypersensitive site was found in the upstream DNA region of nucleosomes assembled onto fragments 67 and 77. This site was not hypersensitive in the naked DNA samples which shows that the nucleosome packaging of the positioning fragments was required for formation of the hypersensitive site. There was no KMnO4 hypersensitivity on the top strands of these fragments or on either strand of fragment 61 (see below). Of crucial importance is that the prominent reactive site is located 1.5 turns away from the dyad in the single position of 67 and 77 at a site known to display DNA distortion in the nucleosome. The 5S rDNA positioning sequence from L. variegatus also contained a permanganate hypersite at 1.5 turns from the pseudo-dyad as seen in Figure 6 and other natural positioning sequences were enriched in the sequence motifs that give rise to permanganate hypersensitivity in this location as will be discussed below .

Figure 5. Upstream location of the positioning signals. The unique translational position in fragments 77, 67 and the major position in 57 are depicted by the rectangles. The A-tracts are indicated by dark squares (tracts 5-8) if their minor grooves face the protein surface and by hatched squares (tracts 1-4) if they do not. Areas of common sequence are shown by horizontal lines between the rectangles. The mapped pseudo-dyad is indicated by an arrow and presumptive sites of DNA distortion at Âą15 bp and Âą45 bp from the pseudo-dyad are indicated above the fragments. Note that the nucleosomes assembled onto each of the fragments share a common upstream border suggesting that the positioning signals are located on the upstream half of these nucleosome DNAs. The permanganate hypersite is indicated by the stars.

354

Gene Therapy and Molecular Biology Vol 4, page 355

Figure 6. The permanganate hypersite and its rotational orientation. Left: DNA distortion in the nucleosome detected by permanganate hypersensitivity-For fragments 67 and 77, lanes 1 and 2 are mononucleosomes treated with hydroxyl radical and permanganate, respectively, and lanes 3 and 4 are naked DNAs digested with permanganate and hydroxyl radical, respectively. For the 5S rDNA sequence from L. variegatus, lanes 1 and 4 are mononucleosomes treated with hydroxyl radical and permanganate, respectively, while lane 3 is naked DNA digested with permanganate. Lane 2 is a G/A sequencing lane. The bp positions of the permanganate hypersites at -15 (fragment 67) and +16 (5S rDNA) relative to the pseudo-dyad at 0 are indicated by the arrows. The hypersite in fragment 77 is also located at -15. Note that all three hypersites are located 1.5 turns away from the dyad which is a site of DNA distortion in the nucleosome. Right: Rotational orientation of the hypersite- The hyperreactive T residue at -15 bp relative to dyad along the bottom strand of 67 is indicated by the star at the top. Portions of the sequences of positioning fragment 67 and nonpositioning fragment 61 are shown at the bottom. The hyperreactive T residue is contained within a TA step which is flanked on the 5' side by Y3 (Y=pyrimidine) and on the 3' side by R5 (R = purine). The possible positions of this TA step in fragment 61 are indicated by the upward facing arrows. Hydroxyl radical cutting sites are represented by lines above and below the bases that were maximally cleaved. Note that minor groove of the hyperreactive TA in 67 faces inward toward the histone surface while the minor groove of the TA in 61 faces away and can be no closer than half a helical turn from the major distortion site at -15 bp from the dyad (from Fitzgerald and Anderson, 1999b).

355

Fitzgerald and Anderson: Nucleosome Positioning

2. Sequence characteristics and rotational orientation of the permanganate hypersite The sequence of the region containing the permanganate hypersite in the nucleosomal DNA from fragment 67 and the corresponding sequence from fragment 61 are shown in Figure 6. The hyperreactive T residue is contained within a TA step which is flanked on the 5' side by Y3 (Y=pyrimidine) and on the 3' side by R 5 (R = purine). The TA step has been shown to be a major site of proteininduced kinking (Dickerson, 1998). In contrast, purine homopolymers resist protein-induced bending due to their mechanical rigidity (Hagerman & Hagerman, 1996), and are also known to form unusual structures under a variety of conditions (Wells, 1988; Ohyama et al., 1998). An initial question we considered is why fragment 61 did not display hyperreactivity at the same site as was observed with 67 and 77 since each of these fragments share a common sequence spanning >15 bp on either side of the hyperreactive thymine. A comparison of the rotational orientation of fragments 67 and 61 as deduced by hydroxyl radical mapping studies suggested that rotational orientation might be involved in generating the permanganate hypersite (Figure 6). The hydroxy radical cleavage data demonstrated that minor grooves of the hyperreactive TA steps displayed by fragments 67, 77, and 5S rDNA faced inward toward the histone surface as is expected for sequences positioned at the -15 position in the nucleosome. In contrast, the minor groove of the permanganate-insensitive TA step in fragment 61 faced away from the histone surface. One possibility is that potentially distortable sequences in the nucleosome positioning fragments 67, 77, and 5S rDNA occupy a rotational orientation that facilitates distortion in response to strong bending forces while the same sequences in fragment 61 do not occupy a rotational orientation that allows such a conformation. In the following section we describe 11 additional constructs that contain mutations in the hypersite region of the synthetic constructs which were designed to determine sequence and rotational orientation requirements for hypersite production. These constructs were subsequently analyzed to determine the effects of the mutations in the hypersite region on nucleosome stability and translational positioning.

3. New constructs with mutations in the hypersite region Oligonucleotide-directed mutagenesis was used to modify fragments 61 and 67. For these studies, the positioning fragment 67 was renamed 1 and the nonpositioning fragment 61 was renamed 11. Permanganate reactions performed on mononucleosomes reconstituted onto the parent and derivative fragments revealed that several factors were required for permanganate hypersite induction as shown in Figure 7. Permanganate hypersensitivity along nucleosomal DNAs was only observed at particular TA steps. Among the 9 derivatives of fragment 1, substitution of the reactive TA step eliminated permanganate hypersensitivity in every case (fragments 2, 3, 4, 5, and 6) in spite of the fact that several

new T residues and TA steps were introduced into these fragments. The second factor involved in generating permanganate reactivity is the identity of sequences immediately flanking the reactive TA step. Fragment 8 contains the intact hypersensitive TA step of the parent but carries two substitutions in the sequences flanking the hypersite. The lack of hyperreactivity in this fragment suggests that flanking sequences are involved in permanganate sensitivity. Similarly, TA steps flanked by homopolymeric Y or R tracts have been shown to be hypersensitive to permanganate in other systems where protein-induced kinking is thought to occur (Borowiec et al., 1987; et al., 1996). In addition, the two prototype kinked DNAs analyzed by Dickerson (1998) occur at YR steps flanked by Y stretches of 6 and 7 bp. The third factor involved in determining permanganate reactivity is the local rotational orientation of the hypersite in agreement with our initial interpretation described above. This was seen with the hypersite displayed by nucleosomes reconstituted onto fragment 13. This fragment was produced by a combination 6 bp insertion and 6 bp deletion flanking the 5' sequence in fragment 11 (61) which altered the local rotational orientation of this sequence rendering it hypersensitive to permanganate. We reasoned that if the deformation detected by KMnO4 was important for transitional positioning and high nucleosome stability, mutations that alter permanganate reactivity should also alter these properties. A comparison of reconstitution efficiencies of fragment 1 and its derivative fragments (fragments 2-10) revealed that substitutions that eliminated permanganate reactivity in the parent positioning fragment resulted in 2-10 fold decreases in affinity for histone octamers and a partial to total loss in positioning activity while substitutions in the parent non-positioning sequence that rendered the fragment sensitive to permanganate resulted in nearly a 20-fold increase in affinity as seen in Figure 8. However, induction of a second hypersite at position +15 (fragment 9) had little effect on affinity or positioning activity. The results also demonstrated that alteration of the rotational orientation of A-tracts 1-4 (fragment 10) did not result in a significant change in octamer affinity when compared to the parent fragment but did cause a large reduction in positioning activity. With this fragment, the nucleosomes assembled at four positions with permanganate hypersite sequences occupying positions at + and/or -15 bps and to a lesser extent Âą5 bps from the pseudo-dyad. These results show that both the permanganate hypersite and the unusual rotational orientation of the A-tract region 1-4 are required for the unique translational positioning activity displayed by the synthetic fragments.

D. A model for the translational positioning of nucleosomes A general model for the positioning of nucleosomes on the synthetic fragments is shown in Figure 9. According to the model, A-tracts 5-8 with their narrow minor grooves facing the histone surface set the rotational orientation of the sequence (Fitzgerald and Anderson, 1998, 1999b). This rotational orientation was seen in this upstream region of all fragments

356

Gene Therapy and Molecular Biology Vol 4, page 357 studied in Figures 1 and 2 as well as in a fragment that contained A-tracts 5-8 but not A-tracts 1-4. All of the synthetic fragments assumed standard pitches when packaged into nucleosomes of 10.0-10.3 bp/turn in distal areas and 10.5-10.7bp/turn in the central three turns (Hayes, et al., 1990&1991). Consequently, the narrow minor

grooves of A-tracts 1-4 in the +1 series face the histone surface while the narrow minor groove of the TA step at position -15 faces away. The opposite is true in the -3 series fragments which places the translational signals (A-tracts 1-4 and the hypersite) into a rotational orientation that favors high stability and unique positioning of the nucleosome.

Figure 7. Mutations and permanganate hypersensitivity. Oligonucleotide-directed mutagenesis was used to modify fragments 67 and 61 in order to determine the importance of the unusual features to nucleosome stability and translational positioning. Left: Sequences are designated by the number to the left. The positioning fragment 67 was renamed 1 and fragments 2-10 are derivatives of this fragment. The nonpositioning fragment 61 was renamed fragment 11 and fragments 12 and 13 are derivatives of this sequence. Shown is the bottom strand (3' to 5') with the position indicated above the sequences. Bold-faced type indicates a substitution while triangles refer to insertions (points upward) or deletions (points downward). Right: Permanganate probing of reconstituted fragments 1-13. Bottom-labeled reconstituted fragments 1-13 (lanes 1-13) were modified with permanganate and the cleaved products run on a sequencing gel. The fragments were coelectrophoresed with permanganate-treated naked fragment 1 (N), the products of a fragment 1 G/A sequencing reaction (leftmost M), and a fragment 11 G/A sequencing reaction (rightmost M). A-tracts 1-8 and the permanganate hypersites of fragment 9 are indicated by the arrows (from Fitzgerald and Anderson, 1999b).

357

Fitzgerald and Anderson: Nucleosome Positioning

Figure 8. Properties of Nucleosomes Assembled onto Mutagenized Fragments. The mutation type is indicated along the top of the figure and the sequence number from Figure 7 is given along the bottom. Nucleosome stability, translational positioning, permanganate reactivity (+ or -), and rotational orientation of the TA in the hypersite sequence (inward or outward facing minor grooves) were determined as described in Figures 4 and 6. Results are summarized from Fitzgerald & Anderson, 1999b.

The mode of action of both positioning determinants is most likely related to the irregular path of the DNA around the histone octamer. Along this path, regions 15 and 45 bps from the pseudo-dyad exhibit marked departures from ideal base stacking Luger et al., 1997). Biochemical studies have also suggested that the DNA in these areas assumes a distinct structure not seen elsewhere along the particle as described in the "Introduction". Although correlation studies have implicated this irregular DNA path in nucleosome positioning, the studies described in this report provide the first direct evidence for this view.

The marked stability conferred by a functional permanganate hypersite can be attributed to either variations in histone-DNA contacts in the hypersite region, or to differences in histone-DNA contacts formed elsewhere along the sequence due to a change in the mechanical properties of the DNAs at the hypersite. We favor the second alternative because the 1 Kcal/mol drop in affinity resulting from substitutions in this 12 bp region is incompatible with studies that show substitutions generally result in changes of affinity on the order of only 100 cal/helical turn (Shrader & Crothers, 1989). Studies of the effect of sequence changes on other types of protein-DNA complexes suggest a mechanism for the propagation of histone-DNA contact changes beyond the immediate vicinity of the hypersite. 358

Gene Therapy and Molecular Biology Vol 4, page 359

Figure 9. A model for nucleosome positioning. The role of the sequence phase in dictating the rotational orientation of the positioning determinants (the hypersite and A-tracts 1-4) is shown in (A). The proposed modes of action of the determinants are shown in (B) and (C). Further details are provided in the text.

Figure 10. Y>2 TA/TAR>2 in natural positioning DNA. The frequency of the Y >2TA or TAR>2 sequence motifs in eight natural positioning sequence counted at the T residue were summed over a window of three bp and plotted as a function of distance from the dyad. Sequences were from : Drosophila Adh Promoter (Jackson & Benyajati, 1993); African green monkey a-Satellite (Neubauer et al., 1986); L. variegatus 5S rDNA (Flaus et.al.,1996; Fitzgerald and Anderson, 1999b); simian virus 40 major-late transcription site (Powers & Bina); Xenopus vitellogenin B1 promoter (Schild et al., 1993); mouse mammary tumor virus promoter (PiĂąa et al., 1990); Xenopus thyroid hormone receptor Ă&#x;A (Wong et al., 1997); S. cerevisiae pet56-his3-gioded1 ren (Losa et al., 1990).

359

Fitzgerald and Anderson: Nucleosome Positioning The CAP protein induces large kinks at two Y-R steps in its binding site (see Table 1 Dickerson, 1998), and substitutions in this vicinity cause up to 2.5 Kcal drops in the relative affinity of the protein for its binding site and an accompanying decrease in the magnitude of the deflection of the helical axis induced by CAP binding (Gartenberg & Crothers, 1988). Other studies of protein induced DNA bending also suggest that substitutions at the bend site can have a significant effect on the extent of post-binding DNA bending (Bareket-Samish et al., 1997; Hotrod & Perona, 1998). The nucleosome hypersite region may therefore act as a hinge to accommodate protein-induced bending in a manner that affects the stability of flanking contact sites as indicated in the figure. Short oligo A/T tracts are most often oriented in the nucleosome such that their narrow minor grooves face the histone surface. A-tracts 1-4 in the positioning fragments display a different rotational orientation and this unusual orientation is required for positioning. In the absence of these outward facing A-tracts as in fragments 10 and 13, nucleosomes are found at four positions with permanganate hypersite sequences occupying positions at + and/or -15 bps and to a lesser extent ±5 bps. We are left with the question regarding the mechanism by which A-tracts 1-4 fix the permanganate hypersite sequence to -15 bps from the pseudo-dyad in fragment 1 while the same element can be found at ±15 and to a lesser extent ±5 bp from the pseudodyad in fragments 10 and 13. It is conceivable that A-tracts with outward facing minor grooves are restricted to the relatively flat surface of the histone octamer present in the distal three turns of nucleosomal DNA because they cannot tolerate the high curvature demands that are found inside 40-50 bps from the pseudo-dyad. Unique translational positioning on the synthetic fragments therefore can be viewed as a concerted action between both the stabilizing forces associated with the hypersite sequences occupying specific sites within the central three turns of nucleosomal DNA and destabilizing forces which appear when A-tracts 1-4 with outward facing minor grooves occupy particular translational positions.

E. Relationship to natural sequences The synthetic sequences that position nucleosomes at single sites were modeled after natural nucleosome positioning sequences in terms of two regions of curvature and the length of the central region. Our analysis of the nucleosomes that assemble onto the synthetic sequences has revealed that the similarities extend beyond the original model. For example, the 5S rDNA nucleosome positioning sequence from X. borealis and a tobacco satellite positioning sequence share with constructs 67 and 77 an unusual rotational orientation of A-tracts on one side of the nucleosomal DNA (Hayes et al., 1990; Kralovics et al., 1995). The 5S rDNA positioning sequence from L. variegatus also contained a permanganate hypersite at 1.5 turns from the pseudo-dyad as seen in Figure 6. To determine if other positioning sequences were enriched in the sequence motifs that give rise to permanganate

hypersensitivity in this location, we analyzed other known natural positioning sequences for the frequency of TA dinucleotides flanked by pyrimidine or purine homopolymers. In vivo mapped nucleosomal DNAs were not included, nor were positioning sequences that were mapped by low resolution techniques (see Trifonov, 1993). We found 8 sequences that met the above criteria. The distribution of these sequence motifs throughout the nucleosomal DNAs is illustrated in Figure 10. Seven of the eight sequences had at least one of the Y 3TA or TAR 3 motifs within 5 bp of + or - 15 bp from the mapped pseudo-dyad. The remaining sequence had a Y2TA motif at position -15. It is evident that the TA sequence motifs cluster in the ±15 region of the positioning sequences. The likelihood that the observed number of Y3TA and TAR3 sequence motifs occurred by chance in the 10bp region flanking ±15 from the pseudo-dyad is vanishingly small (P<0.005 correcting for base composition). Studies are in progress to evaluate the functional significance of these sequence features for the positioning of nucleosomes on natural DNA.

References Albert, F. G., E. C. Bronson, D. J. Fitzgerald, and J. N. Anderson. (1995) "Circular structures in cellular and retroviral genomes." J. Biol Chem. 270, 23570-23581. Bareket-Samish, A., I. Cohen, T. Haran. (1997) "Repressor assembly at trp binding sites is dependent on the identity of the intervening dinucleotide between the binding half sites." J. Mol. Biol. 267, 103-117. Beato, M., and K. Eisfeld. (1997) “Transcription factor access to chromatin.” Nucleic Acids Res 25, 3559-3563. Borowiec, J. A., Li Zhang, S. Sasse-Swight, and J. D. Gralla. (1987) “DNA Supercoiling Promotes Formation of a Bent Repression Look in lac DNA.” J. Mol. Biol. 196, 101-111. Cao, H., H. Widlund, T. Simonsson, and M. Kubista. (1998) "TGGA repeats impair nucleosome formation." J. Mol. Biol. 281, 253260. Constanzo, G., E. DiMauro, G. Salina. (1990) "Attraction, phasing and neighbour effects of histone octamers on curved DNA." J. Mol. Biol. 216, 363-374. Crick, F. H. C. and A. Klug. (1975) "Kinky Helix." Nature 255, 530-533. Dickerson, R. E. (1998) “DNA bending: the prevalence of kinkiness and the virtues of normality.” Nucleic Acids Res. 26, 19061926. Drew, H., and A. A. Travers. (1985) “DNA bending and its relation with nucleosome positioning.” J.Mol.Biol. 186, 773-790. Eckdahl, T. T., and J. N. Anderson. (1987) “Computer modelling of DNA structures involved in chromosome maintenance.” Nucleic Acids Res. 15, 8531-8545. Fitzgerald, D. J., G. L. Dryden, E. C. Bronson, J. S. Williams, and J. N. Anderson. (1994) "Conserved patterns of bending in satellite and nucleosome positioning DNA." J. Biol. Chem. 269, 2130321314. Fitzgerald, D. J., and J. N. Anderson. (1998) "Unique translational positioning of nucleosomes on synthetic DNA." Nucleic Acids Res.. 26, 2526-2535,

360

Gene Therapy and Molecular Biology Vol 4, page 361 Fitzgerald, D. J. and J. N. Anderson. (1999a) "Selective nucleosome disruption by drugs that bind in the minor groove of DNA." J. Biol Chem. 274, 27128-27138. Fitzgerald, D. J. and J. N. Anderson. (1999b) "DNA distortion as a factor in nucleosome positioning." J. Mol. Biol. 293, 477-491. Flaus, A., K. Luger, S. Tan, and T.J. Richmond. (1996) "Mapping nucleosome position at single base-pair resolution by using site-directed hydroxyl radicals." Proc. Natl. Acad. Sci. USA 93, 1370-1375. Flaus, A. and T. J. Richmond. (1998) "Positioning and stability of nucleosomes on MMTV 3'LTR sequences." J. Mol. Biol. 275, 427-441. Gale, J., and M. Smerdon. (1988) "Photofootprint of nucleosome core DNA in intact chromatin having different structural states." J. Mol. Biol. 204, 949-958. Gao, J., and C. Benyajati. (1998) "Specific local histone-DNA sequence contacts facilitate high-affinity, non-cooperative nucleosome binding of both Adf-1 and GAGA factor." Nucleic Acids Res. 26, 5394-5401. Gartenberg, M., and D. M. Crothers, D.M. (1988) "DNA sequence determinants of CAP-induced bending and protein binding affinity." Nature 333, 824-829. Hagerman, K.R. and P. J. Hagerman. (1996) "Helix rigidity of DNA: the meroduplex as an experimental paradigm." J. Mol. Biol., 260, 207-223. Hänsler, U., and S. Rokita. (1993) "Electrostatics rather than conformation control the oxidation of DNA by the anionic reagent permanganate." J. Am. Chem. Soc. 115, 8554-8557. Harvey, S. C., M. Dlakic, J. Griffith, R. Harrington, K. Park, D. Sprous, and W. Zacharias. (1995) “What is the basis of sequence-directed curvature in DNAs containing A tracts?” J. Biomol. Struct. Dyn. 13, 301-307. Hayatsu, H., and T. Ukita. (1967) "The selective degradation of pyrimidines in nucleic acids by permanganate oxidation." Biochem. Biophys. Res. Commun. 29, 556-561. Hayes, J., D. Clark, and A. Wolffe. (1991) “Histone contributions to the structure of DNA in the nucleosome.” Biochemistry 30, 8434-8440. Hayes, J., T. Tullius, and A. Wolffe. (1990) “The structure of DNA in a nucleosome.” Proc.Natl.Acad.Sci.U.S.A. 87, 7405-7409. Hogan, M. E., T. F. Rooney, and R. H. Austin. (1987) "Evidence for kinks in DNA folding in the nucleosome." Nature 328, 554-557. Horton, N.C., and J. J. Perona, J.J. (1998) "Role of proteininduced bending in the specificity of DNA recognition: crystal structure of EcoRV endonuclease complexed with d(AAAGAT) + d(ATCTT)." J. Mol. Biol. 277, 779-787. Hsieh, D.J., S. M. Camiolo, and J. L. Yates. (1993) "Constitutive binding of EBNA 1 protein to the Epstein-Barr virus replication origin, oriP, with distortion of DNA structure during latent infection." EMBO J 12, 4933-4944. Ioshiknes, I., A. Bolshoy, K. Derenshteyn, M. Borodovsky, and E. Trifonov (1996) "Nucleosome DNA sequence pattern revealed by multiple alignment of experimentally mapped sequences." J. Mol. Biol. 262, 129-139. Jackson, J., and C. Benyajati. (1993) "DNA-histone interactions are sufficient to position a single nucleosome juxtaposing Drosophila Adh adult enhancer and distal promoter." Nucleic Acids Res. 21, 957-967.

John, S., and J. L. Workman. (1998) “Bookmarking genes for activation in condensed mitotic chromosomes.” BioEssays 20, 275-279. Kralovics, R., J. Fajkus, A. Kovarik, and M., Bezdek. (1995) "DNA curvature of the tobacco GRS repetitive sequence family and its relation to nucleosome positioning." J. Biomol. Struct. Dyn. 12, 1103-1119. Kuduvalli, P., C. Townsend, and T. D. Tullius. (1995) "Cleavage by Calicheamicin g 1 I of DNA in a nucleosome formed on the 5S RNA gene of Xenopus borealis.." Biochemistry 34, 3899-3906. Lebrun, A., and R. Lavery. (1997) “Unusual DNA conformations.” Curr.Opini. Struct. Biol. 7, 348-354. Losa, R., S. Omari, and F. Thoma. "Poly(dA).poly(dT) rich sequences are not sufficient to exclude nucleosome formation in a constitutive yeast promoter." Nucleic Acids Res. 18 (1990): 3495-3502. Lowary, P.T., and J. Widom. (1998) "New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning." J. Mol. Biol. 276, 19-42. Luger, K., and T. J. Richmond. (1998) "DNA binding within the nucleosome core." Curr.Opini. Struct. Biol. 8, 33-40. Luger, K., A. W. Mäder, R. K. Richmond, D. F. Sargent, and T. J. Richmond. (1997) “Crystal structure of the nucleosome core particle at 2.8 Å resolution.” Nature 389, 251- 260. Neubauer, B., W. Linxweiler, and W. Horz. (1986) "DNA engineering shows that nucleosome phasing on the African green monkey a-satellite is the result of multiple additive histone-DNA interactions." J. Mol. Biol. 190, 639-645. Nielsen, P.E. (1990) "Chemical and photochemical probing of DNA complexes." J. Mol. Recogn. 3, 1-25. Ohyama, T., H. Tsujibayashi, H. Tagashira, K. Inano, T. Ueda, Y. Hirota, and K. Hashimoto. (1998) "Suppression of electrophoretic anomaly of bent DNA segments by the structural property that causes rapid migration." Nucleic Acids Res. 26, 4811-4817. Patterton, H.-G., and R. T. Simpson. (1995) “Modified curved DNA that could allow local DNA underwinding at the nucleosomal pseudodyad fails to position a nucleosome in vivo.” Nucleic Acids Res. 23, 4170-4179. Piña, B., U. Brüggemeier, and M. Beato. (1990) "Nucleosome positioning modulates accessibility of regulatory proteins to the mouse mammary tumor virus promoter." Cell 60, 719-731. Pennings, S., S. Muyldermans, G. Meersseman, and L. Wyns. (1989) “Formation, stability and core histone positioning of nucleosomes reassembled on bent and other nucleosome-derived DNA.” J.Mol.Biol. 207, 183-192. Powers, J.H., and M. Bina. (1991) "In vitro assembly of a positioned nucleosome near the hypersensitive region in Simian Virus 40 chromatin." J. Mol. Biol. 221, 795-803. Pruss, D., F. D. Bushman, and A. P. Wolffe. (1994a) "Human Immunodeficiency Virus integrase directs integration to sites of severe DNA distortion within the nucleosome core." Proc. Natl. Acad. Sci. USA 91, 5913-5917. Pruss, D., R. Reeves, F. D. Bushman, and A. P. Wolffe. (1994b) "The influence of DNA and nucleosome structure on integration events directed by HIV integrase." J. Biol. Chem. 269, 25031-25041. Satchwell, S. C., H. R. Drew, and A. A. Travers. (1986) “Sequence periodicities in chicken nucleosomal core DNA.” J.Mol.Biol. 191, 659-675. Satchwell, S.C., and A. A. Travers. (1989) "Asymmetry and polarity of nucleosomes in chicken erythrocyte chromatin." EMBO J 8, 229-238.

361

Fitzgerald and Anderson: Nucleosome Positioning Schild, C., F.-X. Claret, W. Wahli, and A. P. Wolffe. (1993) "A nucleosome-dependent static loop potentiates estrogenregulated transcription from the Xenopus vitellogenin B1 promoter in vitro." EMBO J 12, 423-433. Schrader, T., and D. Crothers. (1989) “Artificial nucleosome positioning sequences.” Proc. Natl. Acad. Sci. U.S.A. 86, 7418-7422. Schrader, T., and D. Crothers. (1990) “Effects of DNA sequence and histone-histone interactions on nucleosome placement.” J.Mol.Biol. 216, 69-84. Struhl, K. (1985) "Naturally occurring poly(dA-dT) sequences are upstream promoter elements for constitutive transcription in yeast." Proc. Natl. Acad. Sci. USA 82, 8419-8423. Summers, H., A. Fleming, and L. Frappier. (1997) "Requirements for Epstein-Barr Nuclear Antigen 1 (EBNA1)-induced Permanganate Sensitivity of the Epstein-Barr Virus Latent Origin of DNA Replication." J.Biol.Chem. 272, 2643426440. Svaren, J., and W. Hörz. (1996) "Regulation of gene expression by nucleosomes." Curr. Opin. Genet. Dev. 6, 164-170. Tanaka, S., M. Zatchej, and F. Thoma. (1992) “Artificial nucleosome positioning sequences tested in yeast minichromosomes: a strong rotational setting is not sufficient to position nucleosomes in vivo.” EMBO J. 11, 1187-1193. Trifonov, E.N. "Sequence-dependent deformational anisotropy of chromatin DNA." Nucl. Acids Res. 8, (1980): 4041-4053. Trifonov, E.N. (1993) "Nucleosomal DNA sequence database." Nucleic Acids Res. 21, 4857-4859. Van Holde, K. (1993) “The omnipotent nucleosome.” Nature 362, 111-112. VanWye, J. D., E. C. Bronson, and J. N. Anderson. (1991) “Species-specific patterns of DNA bending and DNA sequence.” Nucleic Acids Res. 19, 5253-5261. Wang, Q., F. G. Albert, D. J. Fitzgerald, J. M. Calvo, and J. N. Anderson. (1994) "Sequence determinants of DNA bending in the ilvIH promoter and regulatory region." Nucleic Acids Res. 22, 5753-5760. Wells, R. D. (1988) "Unusual DNA Structures." J. Biol.Chem. 263, 1095-1098. Widlund, H. R. (1997) "Identification and characterization of genomic nucleosome-positioning sequences." J. Mol. Biol. 267, 807-817. Wilkins, R.C., and J. T. Lis. (1999) "DNA distortion and multimerization: novel functions of the glutamine-rich domain of GAGA factor." J. Mol. Biol. 285, 515-525. Wolffe, A. P. (1994) "Nucleosome positioning and modification: chromatin structures that potentiate transcription." Trends Biochem Sci 19, 240-244. Wong, J., Q. Li, B.-Z. Levi, Y.-B. Shi, and A. P. Wolffe, A.P. (1997) "Structural and functional features of a specific nucleosome containing a recognition element for the thyroid hormone receptor." EMBO J 16, 7130-7145. Zhurkin, V.B. (1985) "Sequence-dependent bending of DNA and phasing of nucleosomes." J. Biomol. Struct. Dyn. 2, 785804. Zhu, Z., and D. J. Thiele. (1996) A specialized nucleosome modulates transcription factor access to a C. glabrata metal responsive promoter. Cell 87, 459-470.

John N. Anderson

362

Gene Therapy and Molecular Biology Vol 4, page 363 Gene Ther Mol Biol Vol 4, 363-368. December 1999

Regulation of transcription by bent DNA through chromatin structure Review Article

Ryoiti Kiyama1, Yoshiaki Onishi1, Chanane Wanapirak1, and Yuko WadaKiyama2 1

National Institute of Bioscience and Human-Technology, Tsukuba, Ibaraki 305, Japan, and 2 Department of Physiology, Nippon Medical School, Sendagi, Bunkyo-ku, Tokyo 113, Japan. __________________________________________________________________________________________________ Correspondence: Ryoiti Kiyama, Ph.D.; Tel: 81-298-61-6189; Fax: 81-298-61-6190; E-mail: kiyamar@nibh.go.jp Key Words: Transcription, chromatin, bent DNA, !-LCR, nucleosome, silencer, enhancer Received: 10 August 1999; accepted: 26 August 1999

Summary A class of bent DNA is present in the genomic DNA of higher eukaryotes as a repeating unit (Kiyama, R., Gene Ther. Mol. Biol. 1, 641-647, 1998). This bent DNA appears once every 680 bp on average, and often shows periodicity, suggesting biological significance. By having a higher affinity for histone core particles, it has a role in designating a key nucleosome, which initiates nucleosome alignment. This modulates the enhancer activity of !LCR and the silencer activity of the human !-globin locus and other loci. This suggests that this type of bent DNA plays important roles in various biological functions by affecting chromatin structure.

I. Periodic bent DNA Recent advances in genome researches have made an extensive progress in the technology for identifying genes. A high attention, although less than that for identifying genes, has also been paid for the regulatory elements in the promoter regions of these genes. However, there are still a number of elements on the genomic DNA that control functions of the cell, and many exist as cryptic codes, i. e. elements that cannot be identified only from nucleotide sequence data. These include enhancers, silencers, insulators, matrix attachment regions and replication origins (Wolffe, 1995). Although the most advanced computer programs for gene identification have 80 ~ 90% accuracy in prediction or higher, the most reliable methods by which to identify these cryptic codes are experimental assays. One of the codes that are difficult to predict is the one for higher order chromatin structure. The difficulty in prediction of these codes is based mostly on the ambiguity of sequence-function relationships. There are many variations in the sequences of enhancers, for example, which confer specificity to species, genes and

timing of expression. It was a time of excitement when we first observed a very beautiful periodicity of DNA bend sites in the human "-globin gene region (Wada-Kiyama & Kiyama, 1994; see Fig. 1). One of the reasons for this was that we were able to find an orderly entity within a not-well-ordered sequence, although it is not completely random. This was made possible only by assays based on biophysical properties of the structure within polyacrylamide gels, including the circular permutation assay (Wu & Crothers, 1984). This is still true now despite the availability of highly sophisticated computer programs. However, finding the relevance or significance of this entity, and especially proving it experimentally, which are not so simple, have not yet been completed. Here, we summarize the potential functions of these bend sites and the evidence to support them.

II. Signal for nucleosome alignment DNA bend sites appear regularly and sometimes nearly periodically, as in the case of the "-globin gene, at an average interval of 680 bp in the human !-globin locus as well as other loci (Kiyama, 1998; Wada-Kiyama et al., 363

Kiyama et al: Regulation of transcription by bent DNA

Fig. 1. The human !-globin locus and the positions of DNA bend sites in !-LCR. HS: DNase I hypersensitive site. The mapped bend sites (-1 to â&#x20AC;&#x201C;28, solid boxes) are shown at the bottom. Note the periodicity between â&#x20AC;&#x201C;1 and â&#x20AC;&#x201C;9.

1999a). This interval length of 680 bp corresponds to that of four nucleosomes. Interestingly, this 680 bp interval actually exists in the human chromosomal DNA as alphoid satellite DNA occupying ~3% of the genomic DNA (Wu & Manuelidis, 1980), and therefore, is not unfamiliar as a unit of chromatin structure. In African green monkey genomic DNA, on the other hand, the equivalent of the human repeat family, # satellite DNA, is present at ~24% (Musich et al., 1980) and this is the major structural component that determines the chromatin structure in this species. This 680 bp unit is further divided into four subunits ~170 bp in length, which are the nucleosome units in this repeat family. These DNA bend sites have a high affinity for histone core particles and their deletion results in disappearance of not only the phase at the bend site but also those in the immediate vicinity (Wada-Kiyama & Kiyama,1996; Onishi et al., 1998). Furthermore, the positions of the sites have a good correlation with the positions of the dyad axis of 146 bp nucleosomal DNA (Wada-Kiyama et al., 1999b). These lines of evidence support the idea that periodic bent DNA plays a key role in chromatin folding or nucleosome formation. In addition to the rotational and translational positioning of nucleosomes, there is a third determinant for nucleosome position on the chromosomes: alignment of nucleosomes. Since chromosomal DNA in general is favorable in binding with core histones to form nucleosomes and each base pair has different affinities, there should be preference in binding driven by the differences in the nucleotide sequences of the DNA. Then, if one of the histone core particles occupies a space on DNA by having a higher affinity, there is no

way of occupying the overlapping region by other particles. This indicates that overall alignment should be determined not only by rotational and translational positions of each nucleosome but also by a preference in binding among nucleosomes in the region. This led to the suggestion of the alignment de facto by key nucleosomes (Onishi et al., 1998). Determination of the best way to obtain perfect alignment with the least number of signals would facilitate our understanding of the role of key nucleosomes. It is not difficult to imagine that each nucleosome does not need its own positioning signal, because when every other nucleosome has a signal all positions are eventually determined. However, this will not be the case if every third or fourth are determined. To resolve this, it is necessary to know two parameters: the cooperativity of core histones and the force of random mutation. The former will result in preferential localization of nucleosomes next to the key nucleosomes, including preferential dinucleosome formation, and create more freedom for sequence variation, as the force becomes larger. The latter force, in contrast, decreases the chance of retaining the signal and increases the chances of sequence variation. A survey of more than 100 mapped bend sites indicated that four nucleosomes is the average unit. This suggests that both of the positions next to the key nucleosomes will be determined by the signal and one remaining position located just in the middle of the two neighboring key nucleosomes will be uniquely determined as a result. This alignment of nucleosomes might have some effect on biological reactions and two such examples where the presence of these bend sites is closely associated with cisacting elements in transcription are shown below. 364

Gene Therapy and Molecular Biology Vol 4, page 365

Fig. 2. DNA bend sites at HS2 of !-LCR. Two DNA bend sites of the region, "B-16 and "B-17, were determined by circular permutation assay. DNA bend centers of these sites were determined by assays first with plasmids containing deletions and then with oligonucleotides. The highest degrees of bending were shown on the right for plasmid assays and the locations are shaded. The regions used for the bending assay with the concatenated oligonucleotides $16A to M16F and M17A to M17D, and their nucleotide sequences are shown below. A T5 tract in "B-16 and two A5 tracts in "B-17, which are most likely the bend centers, are shown by horizontal lines. For more details, see Onishi et al., 1998.

nucleosomes in this region were regularly phased except that located in the middle which corresponded to the precise location of HS2 and included the binding site for NF-E2 (Region II, Fig. 3). Several phases were adopted in this region in the reconstituted chromatin and in erythroid K562 cells where the globin genes are expressed, whereas only one phase was adopted in nonerythroid HeLa cells. Meanwhile, almost unique phases were adopted at the flanking bend sites in vitro as well as in vivo (Regions I and III). These observations suggest that HS2 is placed at a region of weak nucleosome phasing activity along with factor binding sites and could be influenced by the nucleosomal phases determined by those located at the bend sites (Fig. 4).

III. Modulator of enhancer activity The first example is the enhancer of the human !-globin locus, which is located 11 kb upstream of the most upstream gene, the "-globin gene, among the five (", G%, A %, &, and !) active genes in the locus (Stamatoyannopoulos et al., 1994). The enhancer is located within the region HS2, which shows tissuespecific DNase I-hypersensitivity, and contains the binding site for an erythroid-specific transcription factor, NF-E2 (reviewed in Hardison et al., 1997). The enhancer was mapped between two DNA bend sites the distance between which was longer than the average and can accommodate five nucleosomes (Fig. 2). The 365

Kiyama et al: Regulation of transcription by bent DNA

Fig. 3. In vitro and in vivo nucleosomal phases at HS2. Nucleosomal phases were determined by restriction analysis of the ~146bp core DNA fragments for all nucleosomes or by inverse PCR for Regions I to III with reconstituted chromatin (in vitro) and with nuclei isolated from K562 and HeLa cells (in vivo). The relative positions of nucleosome boundaries from nucleotides â&#x20AC;&#x201C;11,417 (Region I), -10,848 (Region II) and â&#x20AC;&#x201C;10,496 (Region III) are shown in the figure. For more details, see Onishi et al., 1998.

Fig. 4. Effects of the key nucleosomes at "B-16 and "B-17.

Interestingly, removing "B-16 or changing the distance between "B-16 and NF-E2 site affected transcription efficiency, while removing "B-17 did not (Onishi et al., unpublished results). This can be explained if the effect of the bend sites reaches the next and the second closest nucleosomes but not the third closest, and supports the idea of four nucleosomes as a unit.

IV. Modulator of silencer activity The other example is the silencers, which are present in the promoter regions of the !-like globin genes. The best characterized silencer in this locus is located 200 to 300 bp upstream of the "-globin gene (Wada-Kiyama et al., 1992) and overlaps the first DNA bend site from the cap

366

Gene Therapy and Molecular Biology Vol 4, page 367 site (Fig. 5A). This co-localization of the silencer or repressor activity was observed in the !-globin gene (Fig. 5B), and also in other genes including the erythropoietin receptor (Fig. 5C), c-myc and estrogen receptor genes (see Lang et al., 1991, DeConinck et al., 1995, Ohki et al., 1998, Wada-Kiyama et al., 1999b). Furthermore, co-localization was observed among all of the human !-like globin genes and the globin genes of the other species where the locations of the first bend site upstream from the cap site are conserved (Wanapirak et al., unpublished results). Although these sites contain transcription factor binding sites and binding sites for less well characterized proteins, the mechanism of silencing gene expression is not clear. Therefore, co-localization of DNA bend sites with silencers could yield clues for understanding the mechanism of regulation.

V. Other functions Although experimental evidence is lacking, these bend sites could have functional relationships with other functions through regulation of nucleosome positioning. One such example is recombination. As discussed previously (Kiyama, 1998; Ohki et al., 1998), the continuity of periodic bent DNA was conserved after rearrangement between the c-myc and immunoglobulin genes. Recombination is likely to be mediated by specific proteins that interact with DNA through chromatin, and thus chromatin structure is an important factor that determines the recombination efficiency in that the structure at least affects accessibility of the proteins to the binding sites. Transcription factors and the integrases require specific nucleosome positions when they recognize the motifs located within the nucleosome structure (Pina et al., 1990; Pruss et al., 1994). Determination of the precise nucleosome structure based on X-ray diffraction analysis and establishing databases of nucleosome positions will provide more insight into the biological relevance of periodic bent DNA.

Fig. 5. Co-localization of DNA bend sites with the silencers of the "-globin ( A), !-globin (B) and erythropoietin receptor (C) genes. Silencer activities were found between -274 and -177 bp ("-globin; Wada-Kiyama et al., 1992), between -338 and -233 bp ( !-globin; Berg et al., 1989) or between -194 and -116 bp (erythropoietin receptor; Chin et al., 1995) from the cap sites and overlapped the bend sites located between -221 and -142 bp ("B-1; Wada-Kiyama and Kiyama, 1994), between â&#x20AC;&#x201C;373 and -210 bp (!B-1; Wada-Kiyama & Kiyama, 1995) or between -517 and -289 bp (EPB-1; Rodley et al., 1998). These were included in the regions with several peaks in the curvature plots by TRIF analysis (Shpigelman et al., 1993; Wada-Kiyama et al., 1999a).

References Berg, P. E., Williams, D. M., Qian, R. L., Cohen, R. B., Cao, S. X., Mittelman, M. & Schechter, A. N. (1989) A common protein binds to two silencers 5' to the human !-globin gene. Nucleic Acids Res. 17, 8833-8852. Chin, K., Oda, N., Shen, K. & Noguchi, C. T. (1995) Regulation of transcription of the human erythropoietin receptor gene by proteins binding to GATA-1 and Sp1 motifs. Nucleic Acids Res. 23, 3041-3049. DeConinck, E. C., McPherson, L. A. & Weigel, R. J. (1995) Transcriptional regulation of estrogen receptor in breast carcinomas. Mol. Cell. Biol. 15, 2191-2196.

Hardison, R., Slightom, J. L., Gumucio, D. L., Goodman, M., Stojanovic, N. & Miller, W. (1997) Locus control regions of mammalian !-globin gene clusters: combining phylogenetic analyses and experimental results to gain functional insights. Gene 205, 73-94. Kiyama, R. (1998) Periodicity of DNA bend sites in eukaryotic genomes. Gene Ther. Mol. Biol. 1, 641-647.

367

Kiyama et al: Regulation of transcription by bent DNA Lang, J. C., Wilkie, N. M., Clark, A. M., Chudleigh, A., Talbot, S., Whitelaw, B. & Frame, M. C. (1991) Regulatory domains within the P0 promoter of human cmyc. Oncogene 6, 2067-2075. Musich, P. R., Brown, F. L. & Maio, J. J. (1980) Highly repetitive component alpha and related alphiod DNAs in man and monkeys. Chromosoma 80 331-348. Ohki, R., Hirota, M. Oishi, M. & Kiyama, R. (1998) Conservation and continuity of periodic bent DNA in genomic rearrangements between the c-myc and immunoglobulin heavy chain Âľ loci. Nucl. Acids Res. 26, 3026-3033. Onishi, Y., Wada-Kiyama, Y. & Kiyama, R. (1998) Expression-dependent perturbation of nucleosomal phases at HS2 of the human !-LCR: Possible correlation with periodic bent DNA. J. Mol. Biol. 284, 989-1004. Pina, B., BrĂźggemeier, U. & Beato, M. (1990) Nucleosome positioning modulates accessibility of regulatory proteins to the mouse mammary tumor virus promoter. Cell 60, 719-731. Pruss, D., Bushman, F. D. & Wolffe, A. P. (1994) Human immunodeficiency virus integrase directs integration to sites of severe DNA distortion within the nucleosome core. Proc. Natl. Acad. Sci. USA 91, 5913-5917. Rodley, P., Nishikawa, N. S., Hirota, M., Wada-Kiyama, Y., Noguchi, C. T., Oishi M. & Kiyama, R. (1998) Localization and characterization of curved DNA in the human erythropoietin receptor gene by experimental and theoretical approaches. DNA Res. 5, 349-354. Shpigelman, E. S., Trifonov, E. N. & Bolshoy, A. (1993) CURVATURE: software for the analysis of curved DNA. Computer Appl. in Biosci. 9, 435-440. Stamatoyannopoulos, G., Nienhuis, A. W., Majerus, P. W. & Varmus, H. (1994) The molecular basis of blood diseases. W. B. Saunders, Philadelphia. Wada-Kiyama, Y., Peters, B. & Noguchi, C. T. (1992) The "globin gene silencer. Characterization by in vitro transcription. J. Biol. Chem. 267, 11532-11538. Wada-Kiyama, Y. & Kiyama, R. (1994) Periodicity of DNA bend sites in human "-globin gene region: possibility of sequence-directed nucleosome phasing. J. Biol. Chem. 269, 22238-22244. Wada-Kiyama, Y. & Kiyama, R. (1995) Conservation and periodicity of DNA bend sites in the human !-globin gene locus. J. Biol. Chem. 270, 12439-12445. Wada-Kiyama, Y. & Kiyama, R. (1996) An intrachromosomal repeating unit based on DNA bending. Mol. Cell. Biol. 16, 5664-5673. Wada-Kiyama, Y., Suzuki, S. & Kiyama, R. (1999a) DNA bend sites in the human !-globin locus: Evidence for a basic and universal structural component of genomic DNA. Mol. Biol. Evol. 16, 922-930. Wada-Kiyama, Y., Kuwabara, K., Sakuma, Y., Onishi, Y., Trifonov, E. N. & Kiyama, R. (1999b) Localization of curved DNA and its association with nucleosome phasing in the promoter region of the human estrogen receptor # gene. FEBS Letts. 444, 117-124.

Wolffe, A. P. (1995) Chromatin: structure & function. Academic Press. Wu, H.-M. & Crothers, D. M. (1984) The locus of sequencedirected and protein-induced DNA bending. Nature 308, 509-513. Wu, J. C. & Manuelidis, L. (1980) Sequence definition and organization of a human repeated DNA. J. Mol. Biol. 142, 363-386.

Ryoiti Kiyama

368

Gene Therapy and Molecular Biology Vol 4, page 369 Gene Ther Mol Biol Vol 4, 369-385. December 1999

Crosstalk between intra- and extracellular factors in the development of prolactinomas in the anterior pituitary Review Article

HĂŠlĂ¨ne Courvoisier The Rockefeller University, Developmental Neurogenetics, 1230 York Avenue, New York 10021, USA __________________________________________________________________________________________________ Correspondence: Tel: (+1) 212 327 76 20; Fax: (+1) 212 327 79 23; E-mail: courvoh@rockvax.rockefeller.edu Key words: prolactinomas, anterior pituitary, prolactin, lactotroph, oncogenes, neuroendocrine tumor, signal transduction, G-protein coupled receptor, molecular pharmacology Received: 5 October 1999; accepted 20 October 1999

Summary Prolactinomas are non invasive neoplasms resulting from an abnormal proliferation of the lactotroph cells in the anterior pituitary. Although prolactinomas have clearly been shown to be monoclonal tumors, the seek for the initial causal mutation has not been successful yet. In parallel to the recent advances made in the mutation analyses, this review aims at presenting the main extracellular factors, such as hypothalamic factors, estrogens, growth factors and cytokines, that could participate to the overproliferation of the lactotroph cells. In addition to their role in the development and maintenance of prolactinomas, it is proposed that the initial causal mutation could take place in the transduction pathways of such extracellular factors. Taking in account the complex regulation of the anterior pituitary could help in designing more specific and efficient treatments, especially for patients resistant to the most commonly used bromocriptine therapy. (Riddle et al, 1933). Since, prolactin has been shown to exhibit a very wide range of physiological functions, including lactation, reproduction, osmoregulation, adaptation to stress, immunoregulation, parental behaviors, growth and development (reviewed in Bole-Feysot, 1998). Prolactin has been identified in all vertebrates and sequence comparisons suggest that it could share a common ancestral gene with two closely related hormones: growth hormone (GH) also secreted in the anterior pituitary, and placenta lactogen (PL) secreted by mammalian placenta (Miller and Eberhardt, 1983; Goffin et al, 1996). Prolactin acts via its binding to a specific membrane receptor (PRLR), which belongs to the large class 1 cytokine receptor family (reviewed in Kelly et al, 1991). Consistent with the variety of effects of prolactin, PRLR is expressed in a wide number of tissues, including skin, lung, heart, liver, gastrointestinal tract, reproductive organs, central nervous system and pituitary itself (Bole-Feysot et al, 1998). Structural variants of both PRL (Sinha, 1995) and PRLR (Kelly et al, 1991), resulting from alternative splicing and/or posttranslational modifications, have been described, and could also participate to the pleiotropic physiological effects of prolactin. The regulation of prolactin synthesis and

I. Introduction The anterior pituitary is composed of five different cell types which are defined by the hormones they produce and secrete into the bloodstream. These cell types (and their characteristic hormone) are: corticotrophs (adrenocorticotropic hormone or ACTH), gonadotrophs (luteinizing and follicle-stimulating hormone or LH and FSH), thyrotrophs (thyroid-stimulating hormone or TSH), somatotrophs (growth hormone or GH) and lactotrophs (prolactin). These cells receive, decode and transfer to peripheral endocrine organs the signals coming from the brain. The anterior pituitary thus plays a central and unique role in controlling the neuroendocrine interactions in the body. Any disregulation of the proliferation/differentiation of the cells or in the hormone production can have dramatic effects. This review focuses on prolactinomas, resulting from an overproliferation of the lactotroph cells, which produce and secrete in the bloodstream the hormone prolactin. Prolactin (PRL) was first isolated by its ability to stimulate mammary development and lactation in rabbits 369

Courvoisier: Development of prolactinomas secretion involves numerous factors, including dopamine, VIP (vasointestinal peptide), TRH (thyrotropin-releasing hormone), estrogens and growth factors (Brown, 1994), which also appear to regulate the differenciation and proliferation of the lactotroph cells (see below, and Figure 1). Among all the dysregulations leading to hyperprolactinemia, prolactinomas, the most frequent anterior pituitary tumors (Thapar et al, 1993), derive from an abnormal proliferation of the lactotroph cells. They are, in most of the case, noninvasive neoplasms. The characteristic symptoms of prolactinomas, although not always present in the same patient, can be elevated prolactin levels, amenorrhea, galactorrhea, infertility, loss of libido, headaches or visual defects due to the compression of the optic chiasma by the tumor (Katznelson and Klibanski, 1997). Prolactinomas are more frequent in women and are often microadenomas in this case (< 10 mm), whereas in men they are more likely macroadenomas (Faglia, 1993). It is however not clear whether this reveal a true gender difference and an influence of sex hormones, or if it is due to the fact that women consult more quickly a clinician, especially because of the menstrual disturbance (Katznelson and Klibanski, 1997). Most of the prolactinomas can be shrinked by dopamine agonists such as bromocriptine,

although 8-15% of patients are resistant to this treatment, or by trans-phenoidal surgery aiming at removing the tumor, but the choice between pharmacology and surgery as a first treatment, as well as the addition of radiotherapy, is still discussed and depends on many factors, including the size of the adenoma, resistance/sensitivity to bromocriptine and its side-effects, or particular situation like pregnancy (see Besser, 1993). The etiology of prolactinomas remains still poorly understood. It has first been thought that an overproduction of stimulating factors or a defect in inhibiting factors from the hypothalamus could be the origin of pituitary adenomas. However, the discovery of the monoclonality of prolactinomas suggests a intrinsic pituitary cellular defect such as a mutation in transduction mechanisms rather than an extracellular dysfunction (Herman et al, 1990). A number of such mutations in signal transduction pathways have already been identified in sommatotroph adenomas, such as a truncated GHRH receptor (Hashimoto et al, 1995) or the gsp somatic mutation in the G protein alpha-s subunit leading to a constitutively activation of adenylyl cyclase (Landis et al, 1989). Such pituitary molecular alterations have not yet been fully characterized in prolactinomas, but, like for sommatotroph adenomas, if an hormonal dysregulation would cause prolactinomas, polyclonal tumor would also ensue.

Figure 1. The organization of the hypothalamo-pituitary axis. The pituitary gland regulates target organs such as ovaries and mammary glands. Threee groups of factors can control the activity of the pituitary: (i) signals coming from the hypothalamus in the central nervous system; (ii) auto/paracrine loops of pituitary cytokines and growth factors; and (iii) feedback by hormones secreted by the target organs.

370

Gene Therapy and Molecular Biology Vol 4, page 371 Nowadays, beside the antagonism between the "hypothalamic origin" and the "pituitary mutation origin", a more multifactorial view of the development of prolactinomas, and pituitary adenomas in general, can be drawn, considering first the appearance of a somatic mutation leading to monoclonal expansion of a single transformed cell, and in a second step extracellular factors could favor the development of the mutated clones into large or invasive tumor. The best demonstration of the importance of extracellular factors is obviously the possibility to shrink most of prolactinomas with molecules that mimic the inhibitory action of hypothalamic dopamine. In this respect, it has to be stressed that prolactinomas are a very potent model for other endocrine or brain tumors. Parallely to the identification of mutated genes in prolactinomas, a better understanding of how the differenciation/proliferation and hormonal production of the lactotroph cells can be regulated, on both physiological and molecular levels, is thus needed to identify candidate molecules for the elaboration of more specific therapeutical drugs. With this perspective, this review presents the efforts that have being made to identify mutations in prolactinomas, and discusses the potential implication of different extracellular factors which regulate prolactin production and lactotrophs differentiation/ proliferation, since a dysfunction of their signaling pathways could be as well involved in the development/maintenance of the pituitary tumor.

II. Mutations in prolactinomas Based on study demonstrating the monoclonality of prolactinomas (Herman et al, 1990), it can be assumed that a somatic mutation is the first step of the development of prolactinomas, as it has already been shown in other pituitary tumors. For example, mutations in the G-protein subunit alpha-s gene (gsp mutations) are found in up to 40 % of screened human GH-secreting adenomas, and lead to a constituvely stimulation of the cAMP signaling pathway (Vallar et al, 1987; Landis et al, 1989). However those mutations seem specific to GH cell tumorigenesis, and are absent in prolactinomas (Tordjman et al, 1993; Williamson et al, 1994). Similarly, mutations in the subunit alpha of Gi2 protein, although found in other endocrine tumors (Lyons et al, 1990), have not been found in pituitary tumors (Lyons et al, 1990; Tordjman et al, 1993). Three homologous ras protooncogenes H-ras, K-ras and N-ras are structurally related to G proteins (Lochrie et al, 1985). Missense mutations which convert ras protooncogenes into active oncogenes are commonly identified in a variety of different human cancer (Bos, 1989) and both benign and malignant endocrine tumors (Namba et al, 1990). Four studies (Karga et al, 1992; Herman et al, 1993; Boggild et al, 1994; Cai et al, 1994)

examining more than 200 secreting and non-secreting pituitary tumors identified only one H-ras mutation in an aggressive prolactinoma (Karga et al, 1992). Thus ras oncogene point mutations and activation are uncommon event in pituitary tumor initiation, but may be important in aggressive tumors and in the very rare pituitary metastasis formation and growth. Similarly, mutations in two common genes associated with cancer, p53 and Rb, have not been detected yet in prolactinomas (Hollstein et al, 1991; Herman et al, 1993, Pei et al, 1994, 1995; Cryns et al, 1993; Zhu et al, 1994; Woloschak et al, 1994, 1996). Some studies have also started to investigate the signaling pathways of hypothalamic regulatory factors, without much success, since no mutation have been identified neither in the D2 dopaminergic receptor gene (Friedman et al, 1994), nor in the TRH receptor gene (Dong et al, 1996; Faccenda et al, 1996). Thus, in prolactinomas, mutations usually associated with cancer are rare and sporadic, suggesting that they have probably little or no role in pituitary tumorigenesis. They seem more likely involved in the rare cases of invasive and aggressive tumors. Thus, more subtile or tissue-specific mutations in protooncogenes, tumor suppressor genes, or in molecules involved in the transduction pathways of extracellular regulating factors have still to be identified. In the following parts, the physiological regulators of the lactotroph cells are presented, and their potential role in the induction/maintenance of prolactinomas is discussed.

III. Early development of the anterior pituitary Using the expression of Ki-67 antigen as a marker of the mitotic phase, it has been shown that pituitary tumors have a doubling time ranging from 100 to 700 days (Landolt et al, 1988; Knosp et al, 1989), prolactinomas having one of the most rapid growth and being also more frequently encountered in young patients (Haddad et al, 1991). Because of their slow growth, it can be assumed that the initial mutation take place early in life. One can thus ask whether factors regulating the embryonic development of the anterior pituitary could be at the origin of prolactinomas. The embryonic development of the anterior pituitary is a complex multi-step process. The Rathke's pouch first appears as an invagination of an ectodermal layer of cells and will then give rise to the anterior pituitary, whereas the neural lobe of the pituitary originates from the ventral hypothalamus (reviewed in Treier and Rosenfeld, 1996, 1998; WatkinsChow and Camper, 1998; Sheng and Westphal, 1999). During the embryonic development of the anterior pituitary, five endocrine cell types arise from a common pluripotent precursor in a specific spatial and temporal pattern (corticotrophs, gonadotrophs, thyrotrophs, somatotrophs and lactotrophs) (Voss and Rosenfeld, 1992). The expression of the gene encoding alpha GSU, the common alpha subunit of glycoprotein hormones (luteinizing, follicle-stimuling and 371

Courvoisier: Development of prolactinomas thyroid-stimuling hormones) represents the first discernible steps in the development of the anterior pituitary gland at e11 (Simmons et al, 1990). Subsequently, the ontogeny of the five cell types can be followed by the appearance of each cell type-specific secreted hormone (Figure 2). The first committed cells to appear at e12 are the corticotrophs, revealed by the expression of the proopiomelanocortin gene (Therrien and Drouin, 1993). The thyrotrophs expressing TSH (e14) and the gonadotrophs (e15) expressing LH and FSH emerge in turn, followed by the somatotrophs (e16) and finally the lactotrophs proliferating mainly after birth (Voss and Rosenfeld, 1992). The appearance and maintenance of these specfic cell types from a common precursor result from positive or negative control exerted by cell-type specific transcription factors. The POU domain factor Pit-1 (also called GHF-1), specific of the pituitary gland, was initially identified and cloned as a transactivator of the GH and PRL genes (Ingraham et al., 1988; Bodner et al., 1988), and later as a regulator of the TSH beta gene (Steinfelder et al., 1991). Analyses of Pit-1 expression pattern reveal that initiation of its expression correlated both spatially and temporally to the activation of the transcription of its target genes (DollĂŠ et al, 1990, Simmons et al, 1990). Later studies indicate that, in addition to activation of the GH, PRL and

TSH genes, Pit-1 might be competent to initiate other programs of gene activation required for cell differentiation/proliferation of the thyrosomatolactotroph lineage (reviewed in Andersen and Rosenfeld, 1994). Indeed, it has been shown that Pit-1 antisense oligonucleotides not only block GH and PRL transcription but also inhibit the proliferation of pituitary somatotropic (GC) and lactotropic (235-1) cell lines (Castrillo et al, 1991). Moreover, Pit-1 defective mice and humans are not only characterized by a lack in hormone genes expression but also by the failure for the three cell types to proliferate (Li et al, 1990; PfĂ¤ffle et al, 1992; Radovick et al, 1992). Regarding the crucial role of Pit1 in pituitary cell differentiation, one may ask if this transcription factor might be associated with abnormal cell proliferation or development of anterior pituitary tumors. Pit-1 mRNA and protein have been found in human normal and tumorous pituitaries, in lactotrophs, sommatotrophs and thyrotrophs cells (Pellegrini et al, 1994; Asa et al, 1993; Delhase et al, 1993). However, no mutations of Pit-1 gene have been detected in human GH, PRL or TSH secreting adenomas, and Pit-1 transcripts, correlated with the presence of the Pit-1 protein, are identical in size and sequence to those observed in normal pituitary (reviewed in Pellegrini-Bouiller et al, 1997). The results concerning the expression level of Pit1 in pituitary adenomas are less clear.

Figure 2. Ontogeny of anterior pituitary cell types. Theh five cell types of the antrior pituitary derive from a common precursor in a specific spatial and temporal pattern rregulated by transcription factors.

372

Gene Therapy and Molecular Biology Vol 4, page 373 The transcripts of Pit-1 are more abondant in lactotrophs and somatotrophs adenomas than in normal pituitary, but one must be careful in interpreting those results obtained by northern blot on whole pituitaries since the level of expression of Pit-1 has to be balanced with the predominant cell type in the specific adenoma (PellegriniBouiller et al, 1997). Since there is no difference in Pit-1 expression between somatotroph and lactotroph adenomas, other cell type-specific factors, with or without interaction with Pit-1, should be specifically involved in the abnormal proliferation of each cell type. But those mechanisms remain to be elucidated (see Treier and Rosenfeld, 1996). For example, it has been shown that Pit-1 can heterodimerize with another factor, Oct-1, to stimulate the transcription of the prolactin gene (Voss et al, 1991). On the other hand, the hormone production and cell fate in lactotrophs are regulated by many other factors and hormones in adult, as we will see below. For exemple, there is evidence that the prolactin distal enhancer requires the estrogen receptor in addition to Pit-1 for full activity (d'Emben et al, 1992; Waterman et al, 1988). The recent discovery in the pituitary of other homeobox genes, such as OTX, Prop-1 or members of the LIM family (see Treier and Rosenfeld, 1996; Sheng and Westphal, 1999), should offer in the next few years a better understanding of processes that direct organogenesis, specific cell commitment as well as regulation of cell proliferation in the anterior pituitary. Strengthened by the fact that the mutation at the origin of the prolactinoma is thought to appear early in life (Landolt et al, 1988; Knosp et al, 1989), it will have to be shown whether those multi-step events of the anterior pituitary development could also be involved in abnormal proliferation of cells in adenomas.

IV. Extracellular regulating factors In adult, regarding either hormone production or cell differentiation/proliferation, lactotrophs are controled by many factors from different origins. Based on the organization of the hypothalamo-pituitary axis, three levels of control by extracellular factors have to be distinguished, and will be followed in this review (Figure 1): (i) regulation by central signals, including classic hypothalamic stimulatory (mainly VIP and TRH) and inhibitory (mainly dopamine). These factors, via the portal veinous system, acts as classical endocrine factors, stimulating their distal specific receptors on the membrane of pituitary cells. (ii) regulation by the intrapituitary network of cytokines and growth factors. It is more and more clear that locally produced cytokines and growth factors could mediate cell division and hormone production, either directly or by regulating other specific trophic hormone expression.

(iii) regulation by the circulating hormones produced by target organs, that could act as an endocrine feeback mechanism.

A. Hypothalamic factors Like other pituitary hormones, prolactin synthesis/secretion is under the control of dual stimulatory and inhibitory factors from the hypothalamus. However, in contrast to other pituitary hormones, prolactin is predominantly under tonic inbition of dopamine, that is produced in the neurons of the tegmental ventral area of the hypothalamus, and transported to the anterior pituitary via the hypophyseal stalk circulation. Stimulatory factors, like VIP, TRH, serotonin or opioid peptides provide thus a flexible regulation of prolaction production under diverse physiological conditions (Ben-Jonathan et al, 1996). It has also to be mentionned that an intrapituitary production of hypothalamic releasing hormones (TRH, GHRH, CRH, VIP), which are able to elicit hormone secretion, induce c-fos expression and facilitate cell replication, have been documented (Hsu et al, 1989; Castro et al, 1991; Pagesy et al, 1992; Wakabayashi et al, 1992).

1. Stimulatory factors Thyrotropin- releasing hormone (TRH) is one of the most potent stimulator of prolactin synthesis and secretion in vivo (Bowers, 1971). Via its specific membrane G-protein coupled receptor TRHR, TRH increases intracellular concentration of calcium by stimulating both the entry of extracellular calcium and calcium release from intracellular store (Gershengorn and Thaw, 1985; Ashworth and Hinkle, 1996). However, no evidence of an important role of TRH in the regulation of lactotroph differentiation/proliferation has been described yet. In pituitary adenomas, TRH signalling appears intact, as evidenced by normal receptor expression level, binding and induced-PRL release from lactotrophs (Le Dafniet et al, 1987; Yamada et al, 1997). Although some splice variants of the TRH receptor have been described recently in some pituitary tumors (Yamada et al, 1997), studies failed to find any mutation in the TRH receptor gene or any direct link between the TRH signaling and prolactinomas (Dong et al, 1996; Faccenda et al, 1996). Both VIP (vasointestinal peptide) and PACAP (adenylate cyclase activating polypeptide), which share receptors (Harmar and Lutz, 1994), has also been shown to have in vivo stimulatory effects on prolactin synthesis and secretion in rat (Onali et al, 1983; Rawlings and Hezareh, 1996). However, like for TRH, their implication in the control of lactotroph proliferation and in the development of prolactinomas remain to be shown.

2. Inhibiting factors Although other prolactin inhibitory factors have been described (Schally et al, 1991), dopamine remain the major 373

Courvoisier: Development of prolactinomas inhibitor of prolactin in vivo. Dopamine is a widely distributed catecholaminergic neurotransmitter in the brain. Its pleiotropic effects are mediated by receptors, belonging to the G-protein-coupled seven transmembrane receptor family and subdivised in five different subtypes from D1 to D5 (reviewed in Jackson and WestlindDanielsson, 1994). Dopaminergic neurons of the tegmental ventral area of the hypothalamus project on the hypophyseal stalk, as well as directly to the intermediate lobe of the pituitary. It is well characterized for many years that, in normal lactotroph cells, PRL synthesis and secretion are under the tonic inhibitory control of dopamine (Ben-Jonathan, 1985). Therefore, damage to the hypothalamus or the hypophyseal stalk, by compression of a tumor other than prolactinoma for example, can impede dopaminergic inhibition and lead to hyperprolactinemia (Besser, 1993). Hyperprolactinemia can also be caused by medications interacting with dopamine secretion or effect, such as antihypertensive drugs (e.g. reserpine or alphamethyldopa) or psychotropic agents (e.g. haloperidol) (see Katznelson and Klibanski, 1997). In the anterior pituitary, the effects of dopamine are mediated by the receptor subtype D2 (D2R). D2R are present in the anterior (lactotroph cells) and in the intermediate (melanotroph cells) lobe of the pituitary. D2 receptors have been first characterized by their ability, via Gi/G0 proteins, to inhibit ADc and the cAMP/protein kinase A signaling pathway, but it appears now that they can be coupled to other transduction pathways. D2R can hyperpolarize cells via the stimulation of K+ currents, stimulate the synthesis of arachidonic acid by phospholipase A2, decrease intracellular Ca2+ concentration, by modulating membrane Ca2+ channels, as well as the release of Ca2+ from intracellular store through the inhibition of IP3 levels (see Jackson and WestlindDanielsson, 1994). The D2R gene contain eight exons, seven of which are coding exons. An alternative splicing mechanism, including or not the exon 6 in the mRNA, produces two different isoforms of the receptors: a short one D2S (415 AA) and a long one D2L (444 AA) (Dal Toso et al, 1989; Eidne et al, 1989; Giros et al, 1989; Monsma et al, 1989). These two isoforms differ in their third intracellular loop, that specify the coupling to Gprotein, suggesting that they could activate different second messengers (Picetti et al, 1997). No mutation in the D2R gene could be detected in human prolactinomas (Friedman et al, 1994), but there is evidence that dopamine can modulate not only prolactin production but also the lactotroph differentiation/proliferation. Although, this is known, empirically, for about thirty years since prolactinomas can be shrinked by dopamine agonists, little is known about the mechanisms involved in the dopaminergic control of lactotroph proliferation. Interestingly, animal models with a dopaminergic defect show abnormal phenotypes of lactotroph cells. Knock-out mice for D2R, resulting in the

blockade of the dopaminergic inhibition, show hyperprolactinemia coupled to hypoestrogenism and unaffected expression of hypothalamic stimulatory factors (VIP and TRH), as well as an hyperplasia of the anterior pituitary lobe, leading to lactotroph adenomas after about 6 months of life (Saiardi et al, 1997; Kelly et al, 1997). These results clearly show that removing the dopaminergic inhibition is sufficient to induce an overproliferation of lactotroph cells. On the other hand, knock-out mice for the dopamine transporter show low blood levels of prolactin and a reduced number of lactotrophs, since, in this case, released dopamine is no more removed, leading to an overstimulation of dopamine receptors (BossĂŠ et al, 1997). Thus, even if D2R does not seem to be directly involved in prolactinomas (excepted for resistant prolactinomas, see below), it is clear that the transduction pathway activated by dopamine plays a key role in the regulation of prolactin production and lactotrophs differentiation/proliferation.

3. Resistance to bromocriptine In most prolactinomas, the dopaminergic regulation of PRL production, via the D2R, is maintained. This explains that most prolactinomas are successfully treated with bromocriptine, that mimics the action of dopamine. However, 8-15% of prolactinomas are resistant to this treatment, even with high doses of bromocriptine (15 mg daily). The resistance is not due to difference in drug metabolism, but several elements of the dopaminergic transduction pathway have been shown to be affected in the bromocriptine-resistant tumors (reviewed in Barlier et al, 1997). First a decrease of the level of both D2R mRNA and protein has been shown. The density of the binding sites for D2R agonists in the resistant tumor was half of that in the responsive tumors, without any change in the affinity of the receptor. Interestingly the alternative splicing of the D2R mRNA is also modified in resistant tumors (and not in sensitive tumors) compared to normal pituitaries, leading to a preferential decrease of the D2R short isoform (Caccavelli et al, 1994). Although both isoforms show comparative binding and are coupled to ADc via G proteins, they seem to be coupled to different second messenger systems (Picetti et al, 1997). On the level of G proteins, a marked decrease in Gi2alpha transcripts was found in resistant prolactinomas, compared to the sensitive ones, leading to a 60% decrease in the Gi2alpha/G0alpha ratio (Caccavelli et al, 1996). This change in the balance between the two isoform transduction pathways could shift the D2R coupling from ADc to Ca2+ channels, and contribute to the dopamine resistance. Interestingly, a recent study has shown that the level of Pit-1 mRNA is highly correlated to the D2R mRNA level in 15 prolactinomas with variable response to bromocriptine (Pellegrini-Bouiller et al, 1996). Thus, in resitant prolactinomas, expression of both D2R and Pit-1 are decreased, which suggests a crossregulation of the two genes and which could explain the cell-specific role of Pit-1 in this case. Pit-1 promoter has already been shown to be negatively regulated by dopamine (Elsholtz et al, 1991; Lew and 374

Gene Therapy and Molecular Biology Vol 4, page 375 Elsholtz, 1995), but, on the other way, a direct regulation of Pit-1 on the transcription of the D2R gene seems unlikely since no Pit-1 consensus binding site has been found in the promoter region of the D2R (Minowa et al, 1992; Valdenaire et al, 1994). One can thus postulate either an indirect regulation of D2R gene by Pit-1 that remain to be elucidated, or a common regulatory factor acting on both D2R and Pit-1 gene. In the latter case, it has to be noted that a retinoic acid receptor response element has been described in the mouse Pit-1 gene upstream enhancer as well as in the rat promoter of the D2R gene (Rhodes et al, 1993; Valdenaire et al, 1994). Retinoic acid and its coregulator NZF 1 has also been shown to increase the expression of Pit-1 and co-regulate target genes with Pit-1 (Sanchez-Pacheco et al, 1995; Jiang et al, 1996). Recently, it has been shown that retinoic acid can regulate D2R expression and binding sites during embryologic development (Valdenaire et al, 1998) and that retinoic acid receptor deficient mice exhibit reduced expression of D2R (Samad et al, 1997). Knowing the key role of retinoic acid in development (Chambon, 1996), this could be a very seducing hypothesis to explain, at least part of the lactotroph-specific action of Pit-1 during development, and maybe in resistant prolactinomas. In conclusion, contrary to bromocriptine-sensitive prolactinomas, in bromocriptine-resistant prolactinomas, a decrease in the number of D2R could lead to a less efficient inhibition of dopamine on lactotroph differentiation/proliferation. The molecular mechanisms, that could involved Pit-1, dopamine and the ratio of the two isoforms of the D2 receptor, leading to an abnormal proliferation of these cells should be clarified in such an interesting model.

even if its specific expression on lactotroph cells remain to be shown, suggests a paracrine or autocrine role of the cytokine. Although such a role of IL6 has not been extensively studied, it has been shown that IL6 could stimulate the release of PRL, GH, FSH and LH from cultured rat pituicytes in vitro (Spangelo et al, 1989; Yamaguchi et al, 1990), and that IL6 levels are increased in pituitary tumors (Jones et al, 1994; Rezai et al, 1994).

2. Nerve Growth Factor (NGF) The neurotrophic NGF family includes NGF, brainderived neurotrophic factor (BDNF), and neurotrophin-3 and 4. NGF action is mediated by the Trk family of tyrosine kinase receptors (Trk A, B, C, and the low-affinity p75 receptor), which are closely related to the transforming trk oncogenes (Saltiel et al, 1994). NGF expression has been described in lactotroph cells, where it could be co-secreted with prolactin (Missale et al, 1996a). It has been shown that NGF could induce prolactin secretion and increase the number of lactotroph cells in postnatal rat pituitary cultures (Missale et al, 1995a) and in mice overexpressing a pituitary-directed NGF transgene (Borrelli et al, 1992). Interestingly, in bromocriptine-resistant human pituitary PRL-secreting tumor cells, NGF can restore dopamine responsivness, presumably by inducing D2 receptor availability (Missale et al, 1995b). Such an interaction between NGF and D2R has also been shown by the use of NGF antisense nucleotides in prolactinomas cells that result in loss of D2R expression and an increase in cell proliferation rate (Missale et al, 1996b).

3. Epidermal Growth Factor (EGF) EGF is a single-chained polypeptide (53 AA) acting via a receptor with intrinsic tyrosine kinase activity and that need dimerization to be active (Groenen et al, 1994). Both mRNA and protein as well as the receptor have been described in the rat pituitary (Halpern and Hinkle, 1983; Chabot et al, 1986). EGF can enhance the number of lactotroph cells and induce PRL production (Felix et al, 1995), presumably via its direct action on the prolactin gene promoter (Supowit et al, 1984). Its expression in pituitary cell cultures is modulated by estrogens (Mouihate and Lestage, 1995). It has also to be noted that, like NGF, EGF can induce D2R expression confering dopamine responsivness to prolactinoma cells resistant to bromocriptine (Missale et al, 1991).

B. Cytokines and growth factors Although their action on specific cell type and their physiological relevance remain to be precised in most of the cases, there is now evidence that many intrapituitary cytokines and growth factors can modulate both hormone production and cell fate in the anterior pituitary. Only the cytokines and growth factors whose actions on lactotrophs has been documented will be described in this review. For further informations about their other actions on the anterior pituitary, the recent review from Ray and Melmed (1997) is recommended.

1. Interleukins

4. Transforming Growth Factors (TGF-alpha and TGF-beta)

Although classically involved in hematopoietic and inflammatory cell functions, interleukin 6 (IL6) is also expressed in the hypothalamus and the pituitary (Rezai et al, 1994). Heterotrimerization between the specific subunit of IL6 and two gp130 signal transducer subunits (Hibi et al, 1990) is needed to activate the IL6 receptor that can be either soluble, or membrane-anchored. The presence of the IL6 receptor in the anterior pituitary (Ohmichi et al, 1992),

TGF-alpha has been implicated in a number of malignencies (Groenen et al, 1994) and its expression is sufficient to induce fibroblasts transformation in culture (Rosenthal et al, 1986). It acts via the same receptor than EGF (Groenen et al, 1994). TGF-alpha is also expressed in the pituitary, predominantly in lactotroph cells (Kobrin et al, 375

Courvoisier: Development of prolactinomas 1987). Interestingly, female transgenic mice, carrying a lactotroph-targeted TGF-alpha transgene, develop selective pituitary lactotrophs hyperplasia and PRL-containing adenoma formation (Mc Andrew et al, 1995). This sex specificity in adenoma development could be explained by the fact that TGF-alpha expression is regulated by estogens (Bates et al, 1988) and activators of protein kinase C (Mueller et al, 1989). Indeed, it has been shown that estrogen treatment both induce lactotroph hyperplasia and expression of TGF-alpha, and that this estrogen effect can be reverse by bromocriptine (Bordundvaag et al, 1992). Expression of TGF-beta1, the predominant form of TGF-beta, has been described in lactotroph cells where it inhibits prolactin gene expression (Sarkar et al, 1992; Abraham et al, 1998). TGF-beta, contrary to TGF-alpha, inhibits cell growth (Massague and Polyak, 1995) and its expression is decreased during estrogen treatment that induce lactotroph hyperplasia (Pastorcic et al, 1995). Interestingly, two putative TGF-beta inhibitory response elements in the 5'-flanking regions of the rat PRL gene have been reported (Delidow et al, 1991). Thus, TGF-beta, has to be considered as an inhibitor of tumor ell proliferation, and has effects similar to those of dopamine on lactotroph cells. In this sense, it has to be noted that TGF-beta can inhibit TGF-alpha expression (Mueller and Kudlow, 1991). A balance between the opposite effects of the two TGF could thus be regulated by estrogens.

hst mRNA and protein levels have reported in some prolactinomas, mainly invasive ones (Gonsky et al, 1991; Herman et al, 1993; Shimon et al, 1998).

6. Galanin Galanin is a small peptide (29 AA) with a limited sequence homology with other characterized peptides (Takemoto et al, 1983), although it acts though G proteincoupled receptors (Habert-Ortoli et al, 1994; Howard et al, 1997; Wang et al, 1997). It co-localizes with PRL and can be released by lactotroph cells in the pituitary (Steel et al, 1989; Wynick et al, 1993). Its stimulation of prolactin release has been documented in rodents as well as in humans (Wynick et al, 1993; 1998). Galanin seems extremely sensitive to the estrogen status of the animal. It has also been shown that its expression is increased during pregnancy (Vrontakis et al, 1992) and that it can mediate the estrogen-induced stimulation of prolaction production and lactotroph proliferation (Wynick et al, 1993; reviewed in Hyde et al, 1998). On the contrary, pituitary galanin content is dramatically decreased in ovariectomized animals (O'Halloran et al, 1990). In addition to its action on prolactin production, galanin has also been shown to be a potent mitogen to the 235-1 clonal lactotroph cell line (Wynick et al, 1993). Recently, it has been reported that mice, whose galanin gene has been disrupted, exhibit low prolactin levels associated, without changes in the pituitary content of VIP and TRH, and fail to develop hyperproliferation of lactotroph cells in response to estrogen (Wynick et al, 1998). Interestingly, an estrogen-induced increase of the galanin mRNA has been shown in several breast cancer cell lines (Ormandy, 1998). It would thus be interesting to see whether galanin could be directly involved in human prolactinoma development.

5. Fibroblasts Growth Factors (FGFs) The family of FGF consists of peptides that share the capacity to bind heparin (Mason, 1994). Basic FGF (bFGF) is found in abundance in normal pituitary tissue as well as in human pituitary adenomas (Gospodarowicz et al, 1987; Li et al, 1992). Similarly FGF receptor, that possess an intrinsic tyrosine kinase domain, is also detected in pituitary endocrine cells and in human pituitary adenomas (Li et al, 1992; Gonzalez et al, 1994). Although bFGF can enhance prolactin secretion in GH4C1 (Yajima et al, 1984) and GH3 cell lines (Black et al, 1990) as well as in lactotroph tumors (Atkin et al, 1993), the role of bFGF in pituitary cell proliferation is discussed since some studies suggest a proliferative role in pituitary adenomas (Zimering et al, 1990) whereas others show no effect or an inhibitory effect on cell proliferation (Schweigerer et al, 1987; Atkin et al, 1993). FGF-4, the protein product of the hst gene (Sakamoto et al, 1986), is a potent in vitro and in vivo mitogen for PRL-secreting cells. FGF-4 stimulates PRL synthesis and secretion and potentiate lactotrophs tumorigenesis in rat (Shimon et al, 1996). DNA sequences from human prolactinomas have been demonstrated to be transforming in the NIH-3T3 cell focus assay, and to contain part of the coding region of hst gene, suggesting indirectly that this growth factor gene may be associated with pituitary tumorigenesis in human (Gonsky et al, 1991). Abnormal

7. Prolactin Consistant with its homology with GH and with the fact that PRLR belongs to the cytokine receptor family, prolactin has been reported to influence growth and development in a wide range of species, in addition to its well-known effects in lactation and reproduction. In amphibians, PRL is best known for its antimetamorphic effects and is thus considered as a larval form of GH (Tata, 1993) Although no genetic desease has been directly associated with the gene encoding PRL or PRLR in mammals, prolactin has been associated with autoimmune deseases, as well as with certain forms of cancers and may be directly or indirectly involved in tumor growth (recently reviewed in Bole-Feysot et al, 1998). PRL is thought to increase colorectal tumor aggressivity, induce the proliferation of several lines of human breast cancer, active B lymphocytes and lymphoma cells (see Bole-Feysot et al, 1998). PRL has been shown to modulate proliferation of many cell types, including smooth muscle, pancreatic beta-cells, prostate epithelial cells, astrocytes and various cells of the immune system (see Bole-Feysot et al, 1998). PRL has also been suggested to have some functional activity in various

376

Gene Therapy and Molecular Biology Vol 4, page 377 developmental processes, such as the maturation of the lung, the differentiation of preadipocytes and maturation of germ cells (see Bole-Feysot et al, 1998). PRL influences the development of the tuberoinfundibular hypothalamic dopamine system, that will latter participate to the lactotroph regulation (Shyr et al, 1986). PRL could thus play a role in the development of the hypothalamopituitary axis, from which it is secreted and regulated. Based on the effects of PRL on the proliferation of various cell types and on the development of hypothalamic neurons,as well as the presence of its receptor on lactotrophs as well as other anterior pituitary cell types (Ouhtit et al, 1993; Morel et al, 1994), one can postulated that prolactin could have autocrine/paracrine effects on the lactotroph cells. However, such an hypothesis, emphasized by the fact that most of the factors stimulating prolactin production, stimulate also the proliferation of the lactotrophs, has not been much explored yet. Treatments with prolactin has been shown to inhibits prolactin secretion in 2B8 pituitary cell line (Herbert et al, 1979) and PRLR expression in lymphocyte cells (Di Carlo et al, 1995), but prolactin has been shown to act as an autocrine growth factor in somatolactotroph (GH3) cell line (Krown et al, 1992) and lymphocytes (Hartmann et al, 1989). Since there is more and more evidence for a para/autocrine role of prolactin in breast cancer (Fuh and Wells, 1995; Mershon et al, 1995; Vonderhaar, 1998), many studies remain to be done to investigate the potential hyperproliferative role of prolactin on lactotroph cells and in prolactinomas. The development of PRLR antagonists should help those investigations (Fuh and Wells, 1995; Kuo et al, 1998).

al, 1988) or stimulate other growth factors such as galanin (Ormandy, 1998) or TGF-alpha (Bates et al, 1988).

2. Extrapituitary PRL PRL can be found in several fluid compartments, such as blood, amniotic fluid, milk tears, sweat and cerebrospinal fluid. Indeed, in addition to being synthesized and secreted by lactotroph cells of the anterior pituitary, PRL is also produced by a wide number of other tissues, including brain, thymus, spleen, skin and mammary gland, lacrima and sweat glands, and cells, such as lymphocytes (reviewed in Ben-Jonathan, 1996). The importance of the extrapituitary PRL has been demonstrated by Nagy and Berczi (1991). They showed that, in hypophysectomized rats, biologically active PRL blood levels can return up to 50% of normal levels with time. Moreover, neutralization of circulating PRL with anti-PRL antibodies, in the same hypophysectomized rats, results in immune dysfunctions and death, showing that extrapituitary PRL is important and, at least in part, can compensate for pituitary PRL. Extrapituitary PRL can act as a growth factor, a neurotransmitter or an immune modulator, via a classic endocrine pathway or in a paracrine or autocrine manner, and could, in particular, influence also the lactotroph cells.

V. Concluding remarks Prolactinomas are, in most of the cases, non invasive neoplasms, caused by an overproliferation of the lactotroph cells. The etiology of prolactinomas remains poorly understood, and, although prolactinomas are monoclonal, no original mutation has yet been identified. In physiological conditions, the development and maintenance of lactotroph cells as well as the regulation of prolactin requires the coordination of the action of many extracellular factors from different origins and natures (hypothalamic factors, estrogens, intrapituitary growth factors and cytokines). In this respect, prolactinomas are a very potent model for the better understanding of other endocrine tumors. In this review, it is proposed that the original mutation could be part of the transduction pathways used by these extracellular factors, and/or that a dysregulation of one or several of those factors could facilitate the expansion of the tumor. Among all the regulator factors, dopamine is the major inhibitor one, since the blockade of its effects is sufficient to induce the development of lactotroph tumors. A better understanding of the target genes activated by dopamine signaling in lactotroph cells could highlight the etiology of prolactinomas, and it has to be mentionned that the study of the effect of such a neurotransmitter on cell differentiation/proliferation is certainly a very promising field in neurosciences. On the other hand, since most of the factors stimulate both lactotroph cell proliferation and prolactin production, and since prolactin is known to have strong mitogenic effects in other cell types, one could also postulate that prolactin could play a crucial role in prolactinomas.

C. Circulating hormones 1. Estrogens Estrogens are highly important physiologic stimulators of prolactin release and are responsible for the elevation of prolactin levels during gestation (Raymond et al, 1978). Estrogen positively regulate lactotroph proliferation after birth (Lieberman et al, 1983; Elias and Weiner, 1987). Exogenous administration of estrogens induces prolactin gene transcription and secretion (Lloyd, et al, 1991; Song et al, 1989; Chernavsky, 1993; Perez, 1986) and stimulates lactotroph proliferation and adenoma formation (Lloyd, 1983; Asscheman et al, 1988; Gooren et al, 1988). Recent studies have shown that only a small pool of estrogens is required to stimulate the proliferation of lactotrophs (Chun et al, 1998), and that estrogens are crucial for prolactin production and lactotroph proliferation,but not for the specification of the lactotroph cell type in mice lacking the estrogen receptor alpha (Scully et al, 1997). Estrogens can act directly on the level of the promoter of the prolactin gene via their specific nuclear receptors (Maurer and Notides, 1987; Waterman et

377

Courvoisier: Development of prolactinomas More than the wide number of physiological regulator factors (the list given in this review being not exhaustive), it has to be stressed that those factors are from different natures and origins. Altogether, they activate most, if not all, the possible intracellular signaling pathways, giving thus a high level of possible effects and regulations. For the sake of simplicity, each factor has been described independantly in this review, but many interactions among them occur. Much work remain to be done to understand how those factors can interact on extraand intracellular levels, and by which molecular processes they can co-regulate lactotroph differentiation and/or proliferation. It appears to be a crucial point for therapeutical research since a combination of different drugs could be more efficient, for example for the treatment of prolactinomas resistant to bromocriptine (Missale et al, 1995b).Two levels of interactions can be distinguished: first, a physiological level representing the interactions between the factors outside the lactotroph cells, and second, an intracellular level in the lactotroph cells representing the cross-talk between the different signaling pathways. On the physiological level, interactions are numerous, and are far from being fully understood. Some pituitary growth factors, such as galanin, TGF-alpha and TGF-beta, could mediate the action of estrogens on the lactotroph, in addition to the direct transcriptional estrogenic signal (Bates et al, 1988; Hyde et al, 1998). On the other hand, prolactin, EGF and NGF can have a direct effect on the hypothalamic dopaminergic neurons and on the expression of the D2R by lactotrophs (Shyr et al, 1986; Missale et al, 1991; Missale et al, 1996b). EGF has also been shown to regulate the expression of TRH receptors in rat pituitary cells (Monden et al, 1995). It has also to be stressed that the other cell types of the anterior pituitary secrete growth factors and that such paracrine regulation between cells could participate, in great part, to the control of prolactin production and lactotroph proliferation (reviewed in Schwartz and Cherny, 1992). Abraham et al (1996) have shown, for example, that the level of prolactin gene expression was influenced by the phenotype of the adjacent cells. Such a complex organization of the hypothalamo-pituitary axis allows multi-step and subtile regulations of hormone production and cell proliferation in the anterior pituitary, according to signals coming from inside as well as outside the body. Although not fully characterized yet, they have to be taken in account in the seek for novel therapeutic drugs. Figure 3 represents a very schematic view of the main intracellular signaling pathways activated by hypothalamic factors, estrogens, and pituitary growth factors. It is also clear now that those signaling pathways are not independant and that cross-regulations can occur.

For example, TRH has been shown to be able to activate MAP kinase activity, in addition to PLC, in GH3 cells, and, furthermore, this effect can be attenuated by dopamine, suggesting complex cross-talk (Ohmichi et al, 1994). Much work remain to be done to show and understand the different intracellular interactions occuring in lactotroph cells to regulate their differentiation/proliferation. The challenge will be to show how molecules from different natures and activating different signaling pathways can interact and coordinate to regulate cell fate. Very interesting findings are arising in this field (reviewed in Weiss et al, 1998). In fibroblasts, it has been shown that serotonin, alpha 2 adrenergic agonists and bFGF synergize to promote proliferation, via a common G-protein pathway (Seuwen et al, 1990). More recently, Lev et al (1995) have shown that the G protein-mediated activation of PLC by bradykinin could lead to the activation of the Ras/MAPK pathway via PYK2, a nonreceptor tyrosine kinase. Very little is also known about the transcription factors that are finally activated by those different pathways. In this respect, the pituitary-specific factor Pit-1 has long been thought to be the crucial determinant of the control of prolactin production and lactotroph differentiation/ proliferation, since it is the final effectors of dopamine, TRH, VIP and growth factors signalings (Andersen and Rosenfeld, 1994). However clear role of Pit-1 in prolactinomas failed to be proved. It would have to be shown wether all those signaling networks activate finally the same or different transcription factor(s) and which can be involved in the development of prolactinomas. Finally, especially because of the wide number of molecules which could potentially participate in the developpment of prolactinomas, application to prolactinomas of new technologies like differential display or microarrays DNA chips could be very promising. Indeed, a new pituitaryspecific gene, PTTG (pituitary tumor-transforming gene), which can induce tumor development, has recently been identified by using differential display on rat pituitary tumor cells and normal pituitary tissue (Pei and Melmed, 1997; Zhang et al, 1999). Such studies could help in understanding the genes whose expression is altered in prolactinomas, but also in identifiing the initial mutation since prolactinomas are monoclonal tumors.

Acknowledgments The author thanks A. Benecke for helpfull discussions and encouragement, and gratefully acknolwledges the Association pour la Recherche sur le Cancer for past and present financial support.

378

Gene Therapy and Molecular Biology Vol 4, page 379

Figure 3. Schematic representation of the intracellular pathways activated by dopamine, VIP, TRH, growth factors, and estrogens. Adc: adenyl cyclase; Ca ++: calcium; cAMP: cyclic AMP; DAG:diacyl glycerol; G: protein G; IP3: inositol triphosphate; Jak: janus kinase; MAPK: MAP kinase; PKA: protein kinase A; PKC: protein kinase C; PLC: phospholipase C; !: activation; ==!: inhibition

379

Gene Therapy and Molecular Biology Vol 4, page 380

References Abraham EJ, Faught WJ, Frawley LS (1996). Intercellular communication: relative importance of cellular adhesion and paracrine signaling to hormone gene expression. Endocrinology, 137, 4050-4053. Abraham EJ, Faught WJ, Frawley LS (1998). Transforming growth factor beta 1 is a paracrine inhibitor of prolactin gene expression. Endocrinology, 139, 5174-5181. Andersen B and Rosenfeld MG (1994). Pit-1 determines cell types during development of the anterior pituitary gland. J Biol Chem, 269 (47), 29335-29338. Asa SL, Puy LA, Lew AM, Sundmark VC, Elsholtz HP (1993). Cell type-specific expression of the human transcription activator Pit-1 in the human pituitary and pituitary adenomas. J Clin Endocrinol Metab, 77, 1275-1280. Ashworth R, Hinkle PM (1996). Thyrotropin-releasing hormoneinduced intracellular calcium responses in individual rat lactotrophs and thyrotrophs. Endocrinology, 137, 52055212. Asscheman H, Gooren LJ, Assies J, Smits JP, de Slegte R (1988). Prolactin levels and pituitary enlargement in hormone-treated male-to-female transsexuals. Clin Endocrinol, 28, 583-588. Atkin SL, Landolt AM, Jeffreys RV, Diver M, Radcliffe J, White MC (1993). Basic fibroblastic growth factor stimulates prolactin secretion from human anterior pituitary adenomas without affecting adenoma cell proliferation. J Clin Endocrinol Metab, 77, 831-837. Barlier A, Pellegrini-Bouiller I, Caccavelli L, Gunz G, MorangeRamos I, Jaquet P, Enjalbert A (1997). Abnormal transduction mechanisms in pituitary adenomas. Horm Res , 47, 227-234. Bates SE, Davidson NE, Valverius EM, Freter CE, Dickinson RB, Tan JP, Kudlow JE, Lippman ME, Salomon DS (1988). Expression of transforming growth facor-alpha and its messenger ribonucleic acid in human breast cancer: its regulation by estrogen and its pssoible functional significance. Mol Endocrinol, 2, 543-555. Ben-Jonathan N (1985). Dopamine: a prolactin-inhibiting hormone. Endocrin Rev, 6, 564-589. Ben-Jonathan N, Mershon JL, Allen DL, Steinmetz RW (1996). Extrapituitary prolactin: distribution, regulation, functions, and clinical aspects. Endo Rev, 17 (6), 639-669. Besser M (1993). Criteria for medical as opposed to surgical treatment of prolactinomas. Acta Endocrinol, 129 (suppl 1), 27-30. Black EG, Logan A, Davis JRE, Sheppard MC (1990). Basic fibroblast growth factor affects DNA synthesis and cell function and activates multiple signalling pathways in rat thyroid FRTL-5 and pituitary GH3 cells. J Endocrinol, 127, 225-259. Bodner M, Castrillo J-L, Theill LE, Deerinck T, Ellisman M, Karin M (1988). The pituitary-specific transcription factor GHF-1 is a homeobox-containing protein. Cell, 55, 505-518. Boggild MD, Jenkinson S, Pistorello M, Boscaro M, Scanarini M, Mc Ternan P, Perrett CW, Thakker RV, Clayton RN (1994). Molecular genetic studies of sporadic pituitary tumors. J Clin Endocrinol Metab, 78, 387-392.

Bole-Feysot C, Goffin V, Edery M, Binart N, Kelly PA (1998). Prolactin (PRL) and its receptor: actions, signal transduction pathways and phenotypes observed in PRL receptor knockout mice. Endocrin Rev, 19 (3), 225-268. Bordundvaag B, Kundlow JE, Mueller SG, George SR (1992). Dopamine receptor activation inhibits estrogen stimulated transforming growth factor alpha gene expression and growth in anterior pituitary, but not in uterus. Endocrinology, 130, 34533458. Borrelli E, Sawchenko PE, Evans RM (1992). Pituitary hyperplasia induced by ectopic expression of nerve growth factor. Proc Natl Acad Sci USA, 89, 2764-2768. Bos JL (1989). ras oncogenes in human cancer: a review. Cancer Res, 49, 4682-4689. BossĂŠ R, Fumagalli F, Jaber M, Giros B, Gainetdinov RR, Wetsel WC, Missale C, Caron MG (1997). Anterior pituitary hypoplasia and dwarfism in mice lacking the dopamine transporter. Neuron, 19, 127-138. Bowers CY (1971). The role of cyclic AMP in the release of anterior pituitary hormones. Ann NY Acad Sci, 185, 263-290. Brown RE (1994). An introduction to neuroendocrinology. (Cambridge: University Press). Caccavelli L, Feron F, Morange I, Rouet E, Benarous R, Dewailly D, Jaquet P, Kordon C, Enjalbert A (1994). Decreased expression of the two D2 dopamine receptor isoforms in bromocriptineresistant prolactinomas. Neuroendocrinology, 60, 314-322. Caccavelli L, Morange I, Dewailly D, Kordon C, Jaquet P, Enjalbert A (1996). Alterations of Galpha subunits mRNAs levels in bromocriptine resistant prolactinomas. J Neuroendocrinol, 8, 737-746. Cai WY, Alexander JM, Hedley-Whyte ET, Scheithauer BW, Jameson JL, Zervas NT, Klibanski A (1994). ras mutations in human prolactinomas and pituitary carcinomas. J Clin Endocrinol Metab, 78, 842-846. Castrillo J-L, Theill LE, Karin M (1991). Function of the homeodomain protein GHF-1 in pituitary cell proliferation. Science, 253, 197-199. Castro MG, Brooke J, Bullman A, Hannah, M, Glynn BP, Lowry PJ (1991). Biosynthesis of corticotrphin releasing hormone (CRH) in mouse corticotrophic tumour cells expressing the human proCRH gene: intracellular storage and regulated secretion. J Mol Endocrinol, 7, 97-104. Chabot JG, Walker P, Pelletier G (1986). Distribution of epidermal growth factor binding sites in the adult rat pituitary gland. Peptides, 7, 45-50. Chambon P (1996). A decade of molecular biology of retinoic acid receptors. FASEB, 10, 940-954. Chernavsky AC, Valerani AV, Burdman JA (1993). Haloperidol and estrogens induce c-myc and c-fos expression in the anterior pituitary gland of the rat. Neurol Res, 15, 339-343. Chun TY, Gregg D, Sarkar DK, Gorski J (1998). Differential regulation by estrogens of growth and prolactin synthesis in pituitary cells suggests that only a small pool of estrogen receptors is required for growth. Proc Natl Acad Sci USA, 95, 2325-2330. Cryns VL, Alexander JM, Klibanski A, Arnold A (1993). The retinoblastoma gene in human pituitary tumors. J Clin Endocrinol Metab, 77, 644-646.

380

Gene Therapy and Molecular Biology Vol 4, page 381 Dal Toso R, Sommer B, Ewert M, Herb A, Pritchett DB, Bach A, Shivers BD, Seeburg PH (1989). The dopamine D2 receptor: two molecular forms generated by alternative splicing. EMBO J, 8, 4025-4034. Delhase M, Vergani P, Malur A, Velkiniers B, Teugels E, Trouillas J, Hooghe-Peters EL (1993). Pit-1/GHF-1 expression in pituitary adenomas: further analogy between human adenomas and rat SMtTW tumours. J Mol Endocrinol, 11, 129-139.

free calcium in GH3 cells. Further evidence that TRH mobilizes cellular and extracellular Ca2+. Endocrinology, 116, 591-596. Giros B, Sokoloff P, Martres MP, Riou JF, Emorine LJ, Schwartz JC (1989). Alternative splicing directs the expression of two D2 dopamine receptor isoforms. Nature, 342, 923-926. Goffin V, Shiverick KT, Kelly PA, Martial JA (1996). Sequencefunction relationships within the expanding family of prolactin, growth hormone, placental lactogen, and related proteins in mammals. Endocr Rev, 17, 385-410.

Delidow BC, Billis WM, Agarwal P, White BA (1991). Inhibition of prolactin gene transcription by transforming growth factor-beta in GH3 cells. Mol Endocrinol, 5, 17161722. d'Emden MC, Okimura Y, Maurer RA (1992). Analysis of functional cooperativity between individual transcriptionstimulating elements in the proximal region of the rat prolactin gene. Mol Endocrinol, 6, 581-588.

Gonsky R, Herman V, Melmed S, Fagin J (1991). Transforming DNA sequences present in human prolactin-secreting pituitary tumors. Mol Endocrinol, 5, 1687-1695. Gonzalez AM, Logan A, Ying W, Lappi DA, Berry M, Baird A (1994). Fibroblast growth factor in the hypothalamic-pituitary axis: differential expression of fibroblast growth factor-2 and a high affinity receptor. Endocrinology, 134, 2289-2297. Gooren LJ, Assies J, Asscheman H, de Slegte R, van Kessel H (1988). Estrogen-induced prolactinoma in a man. J Clin Endocrinol Metab, 66, 444-446. Gospodarowicz D, Ferrara N, Schweigere L, Neufeld G (1987). Structural characterization and biological functions of fibroblast growth factor. Endocrin Rev, 8, 95-114. Groenen LC, Nice EC, Burgess AW (1994). Structure-function relationships for the EGF/TGF alpha family of mitogens. Growth Factors, 11, 235-257. Habert-Ortoli E, Amiranoff B, Loquet I, Laburthe M, Mayaux JF (1994). Molecular cloning of a functional human galanin receptor. Proc Natl Acad Sci USA, 91, 9780-9783.

Di Carlo R, Bole-Feysot C, Gualillo O, Meli R, Nagano M, Kelly PA (1995). Regulation of prolactin receptor mRNA expression in peripheral lymphocytes in rats in response to changes in serum concentrations of prolactin. Endocrinology, 136, 4713-4716. Dollé P, Castrillo J-L, Theill LE, Deerinck T, Ellisman M, Karin M (1990). Expression of GHF-1 protein in mouse pituitaries correlates both temporally and spatially with the onset of growth hormone gene activity. Cell, 60, 809-820. Dong Q, Brucker-Davis F, Weintraub BD, Smallridge RC, Carr FE, Battey J, Spiegel AM, Shenker A (1996). Screening of candidate oncogenes in human thyrotroph tumors: abscence of activating mutations of the Galphaq, G alpha 11, Galpha s, or thyrotropin-releasing hormone receptor genes. J Clin Endocrinol Metab, 81, 1134-1140. Eidne KA, Taylor PL, Zabavnik J, Saunders PT, Inglis JD (1989). D2 receptor, a missing exon. Nature, 342, 865. Elias KA, Weiner RI (1987). Inhibition of estrogen-induced anterior pituitary enlargement and arteriogenesis by bromocriptine in Fisher 344 rats. Endocrinology, 120, 617621.

Haddad SF, Vangilder JC, Menezes AH (1991). Pediatric pituitary tumors. Neurosurgery, 29, 509-514. Halpern J, Hinkle PM (1983). Binding and internalization of epidermal growth factor by rat pituitary tumor cells. Mol Cell Endocrinol, 33, 183-196. Harmar T, Lutz E (1994). Multiple receptors for PACAP and VIP. Trends Pharmacol Sci, 15, 97-99. Hartmann DP, Holoday JW, Bernton EW (1989). Inhibition of lymphocyte proliferation by antibodies to prolactin. FASEB, J3, 2194-2202.

Elsholtz HP, Lew AM, Albert PR, Sundmark VC (1991). Inhibitory control of prolactin and Pit-1 gene promoters by dopamine. J Biol Chem, 266, 22919-22925. Faccenda E, Melmed S, Bevan JS, Eidne KA (1996). Structure of the thyrotrophin-releasing hormone receptor in human pituitary adenomas. Clin Endocrinol, 44, 341-347.

Hashimoto K, Koga M, Motomura T, Kasayama S, Kouhara H, Ohnishi T, Arita N, Hayakama T, Sato B, Kishomoto T (1995). Identification of alternatively spliced messenger ribonucleic acid encoding truncated growth hormone-releasing hormone receptor in human pituitary adenomas. J Clin Endocrinol Metab, 80, 2933-2939.

Faglia G (1993). Epidemiology and pathogenesis of pituitary adenomas. Acta Endocrinol, 129 (suppl 1), 1-5. Felix R, Meza U, Cota G (1995). Induction of classical lactotropes by epidermal growth factor in rat pituitary cell cultures. Endocrinology, 136, 939-946.

Herbert DC, Ishikawa H, Rennels EG (1979). Evidence for the autoregulation of hormone secretion by prolactin. Endocrinology, 104, 97-100. Herman V, Drazin NZ, Gonsky R, Melmed S (1993). Molecular screening of pituitary adenomas for gene mutations and rearrangements. J Clin Endocrinol Metab, 77 (1), 50-55. Herman V, Fagin J, Gonsky R, Kovacs K, Melmed S (1990). Clonal origin of pituitary adenomas. J Clin Endocrinol Metab, 71, 1427-33. Hibi M, Murakami M, Saito M, Hirano T, Taga T, Kishimoto T (1990). Molecular cloning of an IL-6 signal transducer, gp 130. Cell, 63, 1149-1157.

Friedman E, Adams EF, Höög A, Gejman PV, Carson E, Larson C, De Marco L, Wermer S, Fahlsbusch R, Nordenskjöld M (1994). Normal structural dopamine type 2 receptor gene in prolactin-secreting and other pituitary tumors. J Clin Endocrinol Metab, 78, 568-574. Fuh G, Wells JA (1995). Prolactin receptor antagonists that inhibit the growth of breast cancer cell lines. J Biol Chem, 270, 13133-13137. Gershengorn MC, Thaw C (1985). Thyrotropin-releasing hormone (TRH) stimulates biphasic elevation of cytoplasmic

Hollstein M, Sidransky D, Vogelstein B, Harris CC (1991). P53 mutations in human cancers. Science, 253, 49-53.

381

Courvoisier: Development of prolactinomas Howard AD, Tan C, Shiao LL, Palyha OC, Mc Kee KK, Weinberg DH, Feighner SD, Cascieri MA, Smith RG, Van Der Ploeg LH, Sullivan KA (1997). Molecular cloning and characterization of a new receptor for galanin. FEBS Lett, 405, 285-290.

Landolt AM, Shibata T, Kleihues P, Tuncdogan E (1988). Proliferation of human pituitary adenomas: facts and speculations. Adv Biosci, 69, 53-62. Le Dafniet M, Grouselle D, Li JY, Kujas M, Bression D, Barret A, Tixier-Vidal A, Peillon F (1987). Evidence of thyrotropinreleasing hormone (TRH) and TRH-binding sites in human nonsecreting pituitary adenomas. J Clin Endocrinol Metab, 65, 1014-1019. Lev S, Moreno H, Martinez R, Canoll P, Peles E, Musacchio JM, Plowman GD, Rudy B, Schlessinger J (1995). Protein tyrosine kinase PYK2 involved in CA(2+)-induced regulation of ion channel and MAP kinase functions. Nature, 376, 737-745.

Hsu DW, Riskind PN, Hedley-Whyte ET (1989). Vasoactive intestinal peptide in the human pituitary gland and adenomas. An immunocytochemical study. Am J Pathol, 135, 329-338. Hyde JF, Moore JP, Cai A (1998). Galanin in normal and hyperplastic anterior pituitary cells. From pituitary tumor cell lines to transgenic mice. Ann NY Acad Sci, 863, 48-55. Ingraham HA, Chen R, Mangalam HJ, Elsholtz HP, Flynn SE, Lin CR, Simmons DM, Swanson L, Rosenfeld MG (1988). A tissue-specific transcription factor containing a homeodomain specifies a pituitary phenotype. Cell, 55, 519-529. Jackson DM, Westlind-Danielsson A (1994). Dopamine receptors: molecular biology, biochemistry and behavioural aspects. Pharmac Ther, 64, 291-369. Jiang Y, Yu VC, Buchholz F, O'Connell S, Rhodes SJ, Candeloro C, Xia YR, Lusis AJ, Rosenfeld MG (1996). A novel family of Cys-Cys, His-Cys zinc finger transcription factors expressed in developing nervous system and pituitary gland. J Biol Chem, 271 (18), 10723-10730.

Lew AM, Elsholtz HP (1995). A dopamine-responsive domain in the N-terminal sequence of Pit-1. J Biol Chem, 270 (13), 7156-7160. Li S, Crenshaw EB, Rawson EJ, Simmons DM, Swanson LW, Rosenfeld MG (1990). Dwarf locus mutants lacking three pituitary cell types result from mutations in the POU-domain gene pit-1. Nature, 347, 528-532. Li Y, Koga M, Kasayama S, Matsumoto K, Arita N, Hayakawa T, Sato B (1992). Identification and characterization of high molecular weight forms of basic fibroblast growth factor in human pituitary adenomas. J Clin Endocrinol Metab , 75, 14361441.

Jones TH, Daniels M, James RA, Justice SK, Mc Corkle R, Price A, Kendall-Taylor P, Weetman AP (1994). Production of bioactive and immunoreactive interleukin-6 (IL-6) and expression of IL-6messenger ribonucleic acid by human pituitary adenomas. J Clin Endocrinol Metab, 78, 180-187.

Lieberman ME, Slabaugh MB, Rutledge JJ, Gorski J (1983). 8. Steroids and differentiation. The role of estrogen in the differentiation of prolactin producing cells. J Steroid Biochem, 19, 275-281. Lloyd RV (1983). Estrogen-induced hyperplasia and neoplasia in the anterior pituitary gland. An immunohistochemical study. Am J Pathol, 113, 198-206.

Karga HJ, Alexander JM, Hedley-Whyte ET, Klibansky A, Jameson JL (1992). ras mutations in human pituitary tumors. J Clin Endocrinol Metab, 74, 914-919.

Lloyd RV, Jin L, Fields K, Kulig E (1991). Effects of estrogens on pituitary cell and pituitary tumor growth. Pathol Res Pract, 187, 584-586.

Katznelson L and Klibanski A (1997). Prolactinomas. Endocrine Neoplasms, 89, 41-55. Kelly MA, Rubinstein M, Asa SL, Zhang G, Saez C, Bunzow JR, Allen RG, Hnasko R, Ben-Jonathan N, Grandy DK, Low MJ (1997). Pituitary lactotroph hyperplasia and chronic hyperprolactinemia in dopamine D2 receptor-deficient mice. Neuron, 19, 103-113.

Lochrie MA, Hurley JB, Simon MI (1985). Sequence of the alpha subunit of photoreceptor G protein: homologies between transducin, ras, and elongation factors. Science, 228, 96-99. Lyons J, Landis CA, Harsh G, Vallar L, Grunewald K, Feichtinger H, Duh QY, Clark OH, Kawasaki E, Bourne HR, et al (1990). Two G protein oncogenes in human endocrine tumors. Science, 249, 655-659.

Kelly PA, Djiane J, Postel-Vinay MC, Edery M (1991). The prolactin/growth hormone receptor family. Endocr Rev, 12, 235-251. Knosp E, Kitz K, Perneczky A (1989). Proliferation activity in pituitary adenomas: measurement by monoclonal antibody. Neurosurgery, 25, 927-930.

Mason IJ (1994). The ins and outs of fibroblast growth factors. Cell, 78, 547-552. Massague J, Polyak K (1995). Mammalian antiproliferative signals and their targets. Curr Opin Genet Dev, 5, 91-96.

Kobrin MS, Asa SL, Samsoondar J, Kudlow JE (1987). Alpha transforming growth factor in bovine anterior pituitary gland: secretion by dispersed cells and immunohistochemical localisation. Endocrinology, 121, 1412-1416. Krown KA, Wang Y-F, Ho TWC, Kelly PA, Walker AM (1992). Prolactin isoform 2 as an autocrine growth factor for GH3 cells. Endocrinology, 131, 595-602.

Maurer RA, Notides AC (1987). Identification of an estrogenresponsive element from the 5â&#x20AC;&#x2DC;-flanking region of the rat prolactin gene. Mol Cell Biol, 7, 4247-4254. Mc Andrew J, Paterson AJ, Asa SL, Mc Carthy KJ, Kudlow JE (1995). Targeting of transforming growth factor-alpha expression to pituitary lactotrophs in transgenic mice results in selective lactotroph proliferation and adenomas. Endocrinology, 136, 4479-4488.

Kuo CB, Coss D, Walker AM (1998). Prolactin receptor antagonists. Endocrine, 9, 121-131. Landis CA, Masters SB, Spada A, Pace AM, Bourne HR, Vallar L (1989). GTPase inhibiting mutations activate the alpha chain of Gs and stimulate adenylyl cyclase in human pituitary tumors. Nature, 340, 692-696.

Mershon J, Sall W, Mitchner N, Ben-Jonathan N (1995). Prolactin is a local growth factor in rat mammary tumors. Endocrinology, 136, 3619-3623. Miller WL, Eberhardt NL (1983). Structure and evolution of the growth hormone gene family. Endocr Rev, 4, 97-130.

382

Gene Therapy and Molecular Biology Vol 4, page 383 Minowa T, Minowa MT, Mouradian MM (1992). Analysis of the promoter region of the D2 dopamine receptor gene. Biochemistry, 31, 8389-8396. Missale C, Boroni F, Frassine M, CarusoA, Spano P (1995a). Nerve growth factor promotes the differentiation of pituitary mammotroph cells in vitro. Endocrinology, 136, 1205-1213. Missale C, Boroni F, Sigala S, Buriani A, Fabris M, Leon A, Dal Toso R, Spano P (1996a). Nerve growth factor in the anterior pituitary: localization in mammotroph cells and co-secretion with prolactin by dopamine regulated mechanism. Proc Natl Acad Sci USA, 93, 4240-4250.

Onali P, Eva C, Olianas MC, Schwartz JP, Costa E (1983). In GH3 pituitary cells, acetylcholine and vasoactive intestinal peptide antagonistically modulate adenylate cyclase, cyclic AMP content, and prolactin secretion. Mol Pharmacol, 24, 189-194. Ormandy CJ, Lee CS, Ormandy HF, Fantl V, Shine J, Peters G, Sutherland RL (1998). Amplification, expression, and steroid regulation of the preprogalanin gene in human breast cancer. Cancer Res, 58, 1353-1357. Ouhtit A, Morel G, Kelly PA (1993). Visualization of gene expression of short and long forms of prolactin receptor in the rat. Endocrinology, 133, 135-144.

Missale C, Castelletti L, Boroni F, Memo M, Spano P (1991). Epidermal growth factor induces the functional expression of dopamine receptors in the GH3 cell lines. Endocrinology, 128, 13-20. Missale C, Losa M, Boroni F, Giovanelli M, Balsari A, Spano P (1995b). Nerve growth factor and bromocriptine: a segmented therapy for human bromocriptine resistant prolactinomas. Br J Cancer, 72, 1397-1399.

Pagesy P, Croissandeau G, Le Dafneit M, Peillon F, Li JY (1992). Detection of thyrotropin-releasing hormone (TRH) m RNA by the reverse transcription polymerase chain reaction in the human normal and tumoral anterior pituitary. Biochem Biophys Res Comm, 182, 182-187. Pastorcic M, De A, Boyadjieva N, Vale W, Sarkar DK (1995). Reduction in the expression and action of transforming growth factor beta-1 lactotrophs during estrogen-induced tumorigenesis in the anterior pituitary. Cancer Res, 55, 4892-4898. Pei L, Melmed S (1997). Isolation and characterization of a pituitary tumor-transforming gene (PTTG). Mol Endocrinol, 11, 433-441. Pei L, Melmed S, Scheithauer B, Kovacs K, Benedict WF, Prager D (1995). Frequent loss of heterozygosity at the retinoblastoma susceptibility gene (RB) locus in aggressive pituitary tumors: evidence for a chromosome 13 tumor suppressor geneother than RB. Cancer Res, 55, 1613-1616. Pei L, Melmed S, Scheithauer B, Kovacs K, Prager D (1994). H-ras mutations in human pituitary carcinoma metastases. J Clin Endocrinol Metab, 78, 842-846.

Missale C, Losa M, Sigala S, Balsari A, Giovanelli M, Spano P (1996b). Nerve growth factor controls proliferation and progression of human prolactinoma cell lines through an autocrine mechanism. Mol Endocrinol, 10, 272-285. Monden T, Yamada M, Konaka S, Satoh T, Ezawa H, Iwasaki T, Mori M (1995). Transcriptional down-regulation by epidermal growth factor of TRH receptor mRNA in rat pituitary cells. J Mol Endocrinol, 15, 73-79. Monsma FJ, Mc Vittie LD, Gerfen CR, Mahan LC, Sibley DR (1989). Multiple D2 dopamine receptors produced by alternative RNA splicing. Nature, 342, 926-929. Morel G, Ouhtit A, Kelly PA (1994). Prolactin receptor immunoreactivity in rat anterior pituitary. Neuroendocrinology, 59, 78-84.

Pellegrini I, Barlier A, Gunz G, Figarella-Branger D, Enjalbert A, Grisoli E, Jaquet P (1994). Pit-1 gene expression in the human pituitary and pituitary adenomas. J Clin Endocrinol Metab, 79, 189-196. Pellegrini-Bouiller I, Morange-Ramos I, Barlier A, Gunz G, Enjalbert A, Jaquet P (1997). Pit-1 gene expression in human pituitary adenomas. Horm Res, 47, 251-258.

Mouihate A, Lestage J (1995). Estrogen increases the release of epidermal growth factor from individual pituitary cells in female rats. J Endocrinol, 146 (3), 495-500. Mueller SG, Kobrin MS, Paterson AJ, Kudlow JE (1989). Transforming growth factor-alpha expression in the anterior pituitary gland: regulation by epidermal growth factor and phorbol ester in dispersed cells. Mol Endocrinol, 3, 967-983.

Pellegrini-Bouiller I, Morange-Ramos I, Barlier A, Gunz G, Figarella-Branger D, Cortet-Rudelli C, Grisoli F, Jaquet P, Enjalbert A (1996). Pit-1 gene expression in human lactotroph and sommatotroph pituitary adenomas is correlated to D2 receptor gene expression. J Clin Endocrinol Metab, 81, 33903396. Perez RL, Machiavelli GA, Romano MI, Burdman JA (1986). Prolactin release, oestrogens and proliferation of prolactinsecreting cells in the anterior pituitary gland of adult male rats. J Endocrinol, 108, 399-403.

Mueller SG, Kundlow JE (1991). Transforming growth factor beta (TGF beta) inhibits TGF alpha expression on bovine anterior pituitary-derived cells. Mol Endocrinol, 5, 14391446. Nagy E, Berczi I (1991). Hypophysectomized rats depend on residual prolactin for survival. Endocrinology, 128, 27762784.

PfĂ¤ffle RW, Dimattia GE, Parks JS, Brown MR, Wit JM, Jansen M, Van der Nat H, Van der Brande JL, Rosenfeld MG, Ingraham HA (1992). Mutation of the POU-specific domain of Pit-1 and hypopituitarism without pituitary hypoplasia. Science, 257, 11181121.

Namba H, Rubin SA, Fagin JA (1990). Point mutations of ras oncogenes are an early event in thyroid tumorigenesis. Mol Endocrinol, 4, 1474-1479. Oâ&#x20AC;&#x2DC;Halloran DJ, Jones PM, Steel JH, Gon G, Ghatei MA, Polak JM, Bloom SR (1990). Effect of endocrine manipulation on the anterior pituitary galanin in the rat. Endocrinology, 127, 467-475.

Picetti R, Saiardi A, Samad T A, Bozzi Y, Baik JH, Borrelli E (1997). Dopamine D2 receptors in signal transduction and behavior. Crit Rev Neurobiol, 11, 121-142. Radovick S, Nations M, Du Y, Berg LA, Weintraub BD, Wondisford FE (1992). A mutation in the POU-homeodomain of Pit-1 responsible for combined pituitary hormone deficiency. Science, 257, 1115-1118.

Ohmichi M, Hirota K, Koike K, Kurachi H, Ohtsuka S, Matsuzaki N, Yamaguchi M, Miyake A, Tanizawa O (1992). Binding sites for interleukin 6 in the anterior pituitary gland. Neuroendocrinology, 55, 199-203.

383

Courvoisier: Development of prolactinomas Rawlings SR, Hezareh M (1996). Pituitary adenylate cyclaseactivating polypeptide (PACAP) and PACAP/vasoactive intestinal polypeptide receptors: actions on the anterior pituitary gland. Endocr Rev, 17, 4-29. Ray D, Melmed S (1997). Pituitary cytokine and growth factor expression and action. Endocr Rev, 18 (2), 206-228. Raymond V, Beaulieu M, Labrie F, Boissier J (1978). Potent antidopaminergic activity of estradiol at the pituitary level on prolactin release. Science, 200 (4346), 1173-1175.Â°

Scully KM, Gleiberman AS, Lindzey J, Lubahn DB, Korach KS, Rosenfeld MG (1997). Role of estrogen receptor-alpha in the anterior pituitary gland. Mol Endocrinol, 11, 674-681. Seuwen K, Pouyssegur J (1990). Serotonin as a growth factor. Biochem Pharmacol, 39, 985-990. Sheng HZ, Westphal H (1999). Early steps organogenesis. Trends Genet, 15 (6), 236-240.

pituitary

Shimon I, Hinton DR, Weiss MH, Melmed S (1998). Prolactinomas express human heparin-binding secretory transforming gene (hst) protein product: marker of tumour invasiveness. Clin Endocrinol, 48, 23-29.

Rezai AR, Rezai A, Martinez-Maza O, Vander-Meyden M, Weiss MH (1994). Interleukin-6 and interleukin-6 receptor gene expression in pituitary tumors. J Neurooncol, 19, 131135.

Shimon I, Huttner, A, Said J, Spirina OM, Melmed S (1996). Heparin-binding secretory transforming gene (hst) facilitates rat lactotrope cell tumorigenesis and induces prolactin gene transcription. J Clin Invest, 97, 187-195. Shyr SW, Crowley WR, Grosvenor CE (1986). Effect of neonatal prolactin deficiency on prepubertal tuberoinfundibular and tuberohypophyseal dopaminergic neuronal activity. Endocrinology, 119, 1217-1221. Simmons DM, Voss JW, Ingraham HA (1990). Pituitary cell phenotypes involve cell-specific Pit-1 mRNA translation and synergic interactions with other clases of transcription factors. Genes Dev, 4, 695-711.

Rhodes SJ, Chen R, Dimattia GE, Scully KM, Kalla KA, Lin SC, Yu VC, Rosenfeld MG (1993). A tissue-specific enhancer confers Pit-1-dependant morphogen inducibility and autoregulation on the Pit-1 gene. Genes Dev, 7, 913-932. Riddle O, Bates RW, Dykshorn SW (1933). The preparation, identification and assay of prolactin â&#x20AC;&#x201C; a hormone of the anterior pituitary. Am J Physiol, 105, 191-216. Rosenthal A, Lindquist PB, Bringman TS, Goeddel DV, Derynck R (1986). Expression in rat fibroblasts of a transforminf growth factor alpha cDNA results in transformation. Cell, 46, 301-309. Saiardi A, Bozzi Y, Baik J-H, Borrelli E (1997). Antiproliferative role of dopamine: loss of D2 receptors causes hormonal dysfunction and pituitary hyperplasia. Neuron, 19, 115-128.

Sinha YN (1995). Structural variants of prolactin: occurence and physiological significance. Endocr Rev, 16, 354-369. Song JY, Jin L, Lloyd RV (1989). Effects of estradiol on prolactin and growth hormone messenger RNAs in cultured normal and neoplastic (MtT/W15 and GH3) rat pituitary cells. Cancer Res , 49, 1247-1253. Spangelo BL, Judd AM, Isakson PC, MacLeod RM (1989). Interleukin-6 stimulates anterior pituitary hormone release in vitro. Endocrinology, 125, 575-577. Steel JH, Gon G, O'Halloran DJ, Jones PM, Yanaihara N, Ishikawa H, Bloom SR, Polak JM (1989). Galanin and vasoactive intestinal peptide are co-localized with classical pituitary hormones and show plasticity of expression. Histochemistry, 93, 183-189. Steinfelder HJ, Hauser P, Nakayama Y, Radovick S, Mc Clasley JH, Taylor T, Weintraub BD, Wondisford FE (1991). Thyrotropinreleasing hormone regulation of human TSH expression: role of pituitary-specific transcription factor (pit-1/GHF-1) and potential interaction with a thyroid hormone-inhibitory element. Proc Natl Acad Sci USA, 88, 3130-3134.

Sakamoto H, Mori M, Taira M, Yoshida T, Matsukawa S, Shimizu K, Sekiguchi M, Terada M, Sugimura T (1986). Transforming growth from human stomach cancers and noncancerous portion of stomachmucosa. Proc Natl Acad Sci USA, 83, 3997-4001. Saltiel AR, Decker SJ (1994). Cellular mechanism of signal transduction for neurotrophins. Bioessays, 16, 405-407. Samad TA, Krezel W, Chambon P, Borrelli E (1997). Regulation of dopaminergic pathways by retinoids: activation of the D2 receptor promoter by members of the retinoic acid receptorretinoid X receptor family. Proc Natl Acad Sci USA, 94, 14349-14354. Sanchez-Pacheco A, Palomino T, Aranda A (1995). Retinoic acid induces expression of the transcription factor GHF-1/Pit-1 in pituitary prolactin- and growth hormone-producing cell lines. Endocrinology, 136, 5391-5398. Sarkar DK, Kim KH, Minami S (1992). Transforming growth factor beta 1 messenger RNA and protein expression in the pituitary gland: its action on prolactin secretion and lactotrophic growth. Mol Endocrinol, 6, 1825-1833. Schally AV, Guoth JG, Redding TW, Groot K, Rodriguez H, Szonyi E, Stults J, Nikolics K (1991). Isolation and characterization of two peptides with prolactin releaseinhibiting activity from porcine hypothalami. Proc Natl Acad Sci USA, 88, 3540-3544. Schwartz J, Cherny R (1992). Intercellular communication within the anterior pituitary influemcing the secretion of hypophysial hormones. Endocr Rev, 13, 453-475.

Supowit SC, Potter E, Evans RM, Rosenfeld MG (1984). Polypeptide hormone regulation of gene transcription: specific 5' genomic sequences are required for epidermal growth factor and phorbol ester regulation of prolactin gene expression. Proc Natl Acad Sci USA, 81, 2975-2979. Takemoto K, Rokaes A, Jornvall H, Mc Donald TJ, Matt V (1983). Galanin, a novel biological peptide from porcine intestine. FEBS Lett, 164, 124-128. Tata JR (1993). Gene expression during metamorphosis: an ideal model for post-embryonic development. Bioessays, 15, 239-248. Thapar K, Kalman K, Laws ER, Muller PJ (1993). Pituitary adenomas: current concepts in classification, histopathology and molecular biology. The Endocrinologist, 3, 39-57. Therrien M, Drouin J (1993). Cell-specific helix-loop-helix factor required for pituitary expression of the pro-opiomelanocortin gene. Mol Cell Biol, 13, 2342-2353.

Schweigerer L, Neufeld G, Gospodarowicz D (1987). Basic fibroblast growth factoras a growth inhibitor for cultured human tumor cells. J Clin Invest, 80, 1516-1520.

384

Gene Therapy and Molecular Biology Vol 4, page 385 Tordjman K, Stern N, Ouaknine G, Yossiphov Y, Razon N, Nordenskjold M, Friedman E (1993). Activating mutations of the Gs alpha-gene in non-functioning pituitary tumors. J Clin Endocrinol Metab, 77, 765-769. Treier M, Gleiberman AS, O'Connell SM, Szeto DP, Mc Mahon JA, Mc Mahon AP, Rosenfeld MG (1998). Multistep signaling requirements for pituitary organogenesis in vivo. Genes Dev, 12 (11), 1691-1704. Treier M, Rosenfeld MG (1996). The hypothalamic-pituitary axis: co-development of two organs. Curr Opinion Cell Biol, 8, 833-843.

Woloschak M, Yu A, Xiao J, Post KD (1996). Abundance and state of phosphorylation of the Rb gene product in human pituitary tumors. Int J Cancer, 67, 16-19. Wynick D, Hammond PJ, Akinsanya KO, Bloom SR (1993). Galanin regulates basal and estrogen-stimulated lactotroph function. Nature, 364, 529-532. Wynick D, Small CJ, Bloom SR, Pachnis V (1998). Targeted disruption of the murine galanin gene. Ann NY Acad Sci, 863, 22-47. Yajima Y, Saito T (1984). The effects of epidermal growth factor on cell proliferation and prolactin production by GH3 rat pituitary cells. J Cell Physiol, 120, 249-256.

Valdenaire O, Maus-Moatti M, Vincent JD, Mallet J, Vernier P (1998). Retinoic acid regulates the developmental expression of dopamine D2 receptor in rat striatal primary culture. J Neurochem, 71, 929-936. Valdenaire O, Vernier P, Maus M, Dumas Milne Edwards JB, Mallet J (1994). Transcription of the rat dopamine-D2receptor gene from two promoters. Eur J Biochem, 220, 577-584.

Yamada M, Hashimoto K, Satoh T, Shibusawa N, Kohga H, Ozawa Y, Yamada S, Mori M (1997). A novel transcript for the thyrotropin-releasing hormone receptor in human pituitary and pituitary tumors. J Clin Endocrinol Metab, 82, 4224-4228. Yamaguchi M, Matsuzaki N, Hirota K, Miyake A, Tanizawa O (1990). Interleukin-6 possibly induced by interleukin-1beta in the pituitary gland stimulates the release of gonadotrophins and prolactin. Acta Endocrinol, 122, 20551-20556. Zhang X, Horwitz GA, Heaney AP, Nakashima M, Prezant TR, Bronstein MD, Melmed S (1999). Pituitary tumor transforming gene (PTTG) expression in pituitary adenomas. J Clin Endocrinol Metab, 84, 761-767.

Vallar L, Spada, A, Giannattasio G (1987). Altered Gs and adenylate cyclase activity in human GH-secreting pituitary adenomas. Nature, 330, 566-568. Vonderhaar BK (1998). Prolactin: the forgotten hormone of human breast cancer. Pharmacol Ther, 79, 169-178. Voss JW, Rosenfeld MG (1992). Anterior pituitary development: short tales from dwarf mice. Cell, 70, 527-530.

Zhu J, Leon SP, Beggs AH, Busque L, Gilliliard DG, Black PM (1994). Human pituitary adenomas show no loss of heterozygosity at the retinoblastoma gene locus. J Clin Endocrinol Metab, 78, 922-927.

Voss JW, Wilson L, Rosenfeld MG (1991). POU-domain proteins Pit-1 and Oct-1 interact to form a heteromeric complex and can cooperate to induce expression of the prolactin promoter. Genes Dev, 5 (7), 1309-1320.

Zimering MB, Brandi ML, DeGrange DA, Marx SJ, Streeten E, Katsumata N, Murphy PR, Sato Y, Friesen HG, Aurbach GD (1990). Circulating fibroblast growth factor-like substance in familial multiple endocrine neoplasia type 1. J Clin Endocrinol Metab, 76, 1182-1187.

Vrontakis ME, Schroedter IC, Cosby H, Friesen HG (1992). Expression and secretion of galanin during pregnancy in the rat. Endocrinology, 130, 458-464. Wakabayashi I, Inokuchi K, Hasegawa O, Sugihara H, Minami S (1992). Expression of growth hormone (GH)-releasing factor gene in GH-producing pituitary adenoma. J Clin Endocrinol Metab, 74, 357-361. Wang S, Hashemi T, He C, Strader C, Bayne M (1997). Molecular cloning and pharmacological characterization of a new galanin receptor subtype. Mol Pharmacol, 52, 337-343. Waterman ML, Adler S, Nelson C, Greene GL, Evans RM, Rosenfeld MG (1988). A single domain of the estrogen receptor confers deoxyribonucleic acid binding and transcriptional activation of the rat prolactin gene. Mol Endocrinol, 2, 14-21. Watkins-Chow DE, Camper SA (1998). How many homeobox genes does it take to make a pituitary gland? Trends Genet, 14 (7), 284-290. Weiss ER, Maness P, Lauder JM (1998). Why do neurotransmitters act like growth factors? Perspect Dev Neurobiol, 5, 323-335. Williamson EA, Daniels M, Foster S, Kelly WK, Kendall-Taylor P, Harris PE (1994). Gs alpha and Gi2 alpha mutations in clinically non-functioning pituitary tumours. Clin Endocrinol, 41, 815-820. Woloschak M, Roberts JL, Post KD (1994). Loss of heterozygosity at the retinoblastoma locus in human pituitary tumors. Cancer, 74, 693-696.

385

Gene Therapy and Molecular Biology Vol 4, page 387 Gene Ther Mol Biol Vol 4, 387-395. December 1999.

Cks1 mediates the effects of mutant p53 proteins on the mitotic spindle cell cycle checkpoint Review Article

M. L. Hixon, and Antonio Gualberto Department of Physiology & Biophysics, Case Western Reserve University School of Medicine, 10900 Euclid Ave., Cleveland, Ohio 44106, USA __________________________________________________________________________________________________ Correspondence: Antonio Gualberto, M.D. Ph.D., Senior Research Investigator, Cardiovascular & Metabolic Diseases, Pfizer Central Research Division, Groton, CT 06340. Tel: (860) 715 4617; E-mail: antonio_gualberto@groton.pfizer.com Key words: p53, Cyclin B, Cks1, CksHs1, Cell Cycle Checkpoint Received: 24 June 1999; accepted 12 July 1999

Summary The p53 tumor suppressor is the most frequently mutated gene in human cancer. Alteration of the p53 locus predisposes human cells to chromosomal instability. This is due in part to interference of mutant p53 proteins with the activity of the mitotic spindle and postmitotic cell cycle checkpoints. Recent data indicates that mutant p53 proteins affects the control of the metaphase-to-anaphase transition by up-regulating the expression of Cks1, a protein that mediates the activatory phosphorylation of the anaphase promoting complex (APC) by Cdc2. Cells carrying mutant p53 proteins overexpress Cks1 and are unable to sustain APC inactivation and mitotic arrest. These data implicate Cks1 in the onset of chromosomal instability in cells carrying mutant p53 proteins.

I. Introduction It has been hypothesized that the genesis of cancer is a multistep process that requires a normal cell to undergo a series of changes in order to progress to a tumorigenic state (Bishop 1987). It has been also hypothesized that both, the generation of new cellular variants and their progression to neoplasia, are originated by the acquisition of genomic instability (Nowell 1976). The hypothesis that loss of genomic integrity fuels tumor progression is largely supported by karyotypic analyses of human tumors (Nowell 1986; Wolman 1983), studies of the loss of tumor suppressor function in cell culture (Livingstone et al., 1992; Yin et al., 1992; White et al., 1994; Xiong et al., 1996) and by numerous studies conducted using transgenic animal models to disrupt genes identified as playing key roles in tumor initiation and/or progression (Donehower et al., 1992; Harvey et al., 1992; Jacks et al., 1992; Moser et al.,1993; Moser et al., 1990; Nakayama et al., 1996; Jacks et al., 1994; Williams et al., 1994; Barlow et al., 1996; Elson A et al., 1996; Kinzler and Vogelstein 1996; Baross-Francis et al., 1998; Wang et al., 1998; Kamijo et al.,1999; Toft et al., 1999). However, despite intense investigation, the molecular mechanisms that determine why normal cells go awry

remain to be elucidated. Research has led to the hypothesis that in preneoplastic cells, loss of fidelity in processes which replicate, repair, and segregate the genome allow for the accumulation of the genetic alterations that eventually lead to a malignant phenotype (Hartwell and Kastan 1994). Research addressing the relationship between cell cycle control and the maintenance of genomic integrity in human cells is still in its infancy. Experimental evidence in both yeast and mammals demonstrates that the transition from one phase of the cell cycle to the next depends critically on completion of the previous phase. Superimposed upon these transitions are cell cycle checkpoint pathways (Elledge 1996). These checkpoints oversee cell cycle transitions thus playing a crucial role in the maintenance of genomic integrity by integrating the cell cycle regulatory machinery with DNA repair and cell death pathways (Tlsty et al., 1995). The first set of checkpoints occurs at the G1/S transition and results in a delay in the progression from G 1 into S phase preventing the replication of damaged DNA (Weinert 1998). A second set of checkpoints associated with DNA replication and repair activities may occur during S phase (Foiani et al., 1998; Boddy et al., 1998; Lindsay et al., 1998). Importantly, the dependence of mitosis on the completion of DNA replication in the cell cycle ensures that chromosome segregation takes place only after the genome has been fully replicated (Agarwal et al., 1998;

387

Hixon and Gualberto: Mutant p53 and mitosis Michael and Newport 1998). A third set of checkpoints is located at M phase and the G2/M transition and results in a delay in the progression into and out of mitosis (Minshull et al., 1994; Murray 1995). At mitosis the proper segregation of chromosomes requires the execution of a number of processes: assembly of a bipolar spindle; attachment of chromosomes to the spindle through the kinetochore; and alignment of chromosomes at the metaphase plate. The mitotic spindle checkpoint exerts its function by preventing the onset of anaphase, the actual segregation of chromosomes, until all these events have been properly completed (DeWald et al., 1994). In addition, cells with anomalous mitoses that escape the control of the mitotic checkpoint may be growth arrested (Stewart et al. 1999) or destroyed (Lanni and Jacks 1998) at a tetraploid G1 phase by the postmitotic checkpoint, a DNA damage-responsive pathway which is illustrated in Figure 1. Therefore, abrogation of the mitotic spindle cell cycle checkpoint results in cell death and/or aneuploidy, a hallmark of tumor progression (Cahill et al., 1998).

Fig. 1: Cells are protected from polyploidy by the activity of the mitotic spindle and postmitotic cell cycle checkpoints.

Commonly referred to as the â&#x20AC;&#x153;cellular gatekeeperâ&#x20AC;? (Levine 1997), the p53 gene is the most frequently mutated gene in human cancer (Hollstein et al., 1994) and has been extensively studied as a key cell cycle checkpoint gene at the G 1/S transition of the cell cycle (Levine 1997). Recent reports have also suggested a role for the p53 tumor suppressor gene product at the mitotic spindle cell cycle checkpoint. For example, primary fibroblasts isolated from Li-Fraumeni syndrome (LFS) individuals, who are born with heterozygous mutations in the p53 tumor suppressor gene (Malkin et al., 1990; Srivastava et al., 1990), have a marked tendency to become heteroploid in culture (Bischoff et al., 1990). Heteroploidy is also commonly observed in p53 null mouse cells in vitro (Harper et al., 1993), and in p53 knockout mice (Cross et al., 1995). Likewise the expression of mutant p53 proteins in human colon carcinoma cells and murine cell lines causes chromosomal abnormalities including increased ploidy levels during growth in culture (Agapova et al.,

1996). Intriguingly, the overexpression of a mutant p53 protein on a p53 null background accelerated the polyploidy in a myelomonocytic cell line (Peled et al., 1996). Moreover, LFS fibroblasts that carry structural dominant p53 mutant proteins progress quickly to polyploidy when incubated in the presence of mitotic spindle inhibitors such as colcemid (Gualberto et al., 1998).

II. Discussion The altered mitotic spindle cell cycle checkpoint status generated by structural p53 mutant proteins is a gain of function property that can not be explained by the loss of wild type p53 function. As stated, we have shown that LFS fibroblasts that carry structural dominant p53 mutant protein progress quickly to polyploidy when incubated in the presence of mitotic spindle inhibitors such as colcemid (Gualberto et al. 1998). In contrast, normal human fibroblasts, p53 null LFS fibroblasts, or NHFs carrying the human papilloma virus 16E6 (HPV16 E6) that binds to and promotes the degradation of wild type p53 exhibit growth arrest in response to mitotic inhibitors (Gualberto et al. 1998). Furthermore, Jacks and coworkers recently reported that p53 null mouse fibroblasts have a normal mitotic checkpoint (Lanni and Jacks, 1998). Contrary to these results, polyploidy has been reported by others in HPV16 E6-expressing NHFs and in p53 null mouse embryo fibroblasts (Cross et al., 199s; Di Leonardo et al., 1997). The reason for these apparently contradictory results may be in the inherent differences in the mechanisms that underlie the onset of aneuploidy in p53 null and mutant p53expressing cell types. Importantly, we and others have demonstrated that the mitotic cell cycle spindle checkpoint is transient in p53 null cells (Hixon et al., 1998; Lanni and Jacks et al., 1998). Therefore, even cells with active mitotic checkpoint status may become polyploid due to inactivation of a p53-dependent postmitotic checkpoint (Minn et al., 1996; Di Leonardo et al., 1997; and Lanni and Jacks 1998). This is illustrated in Figure 2. In response to an anomalous chromosomal segregation, normal cells (p53 +/+) delay transiently the exit from mitosis (mitotic spindle checkpoint). P53 +/+ cells with non-segregated chromosomes may eventually re-enter the cell cycle but are destroyed at the subsequent G1 phase (postmitotic checkpoint-dependent apoptosis). Alternatively, normal cells may be growth arrested at a tetraploid G1 stage (Stewart et al., 1999). Loss of p53 (p53 -/-) abrogates the postmitotic cell cycle checkpoint whereas gain of a dominant p53 mutant (p53 +/mut) abrogates both the mitotic and postmitotic checkpoints. Loss of p53 function, either by loss of p53 (p53 -/-) or mutant p53 dominance (p53 +/mut), may then originate aneuploidy due to centrosome amplification (Fukasawa et al. 1996) and multipolar division. In summary, loss of wt p53 may result in aneuploidy due to lack of a postmitotic checkpoint activity; whereas, a structural dominant mutant of the p53 protein may abrogate both the mitotic checkpoint (gain-of-function) and the postmitotic checkpoint (dominant negative).

388

Gene Therapy and Molecular Biology Vol 4, page 389

Fig. 2: In response to an anomalous chromosomal segregation, normal cells (p53 +/+) delay transiently the exit from mitosis (mitotic spindle checkpoint). P53 +/+ cells with non-segregated chromosomes may eventually re-enter the cell cycle but are destroyed at the subsequent G1 phase (postmitotic checkpoint, an apoptotic pathway). Loss of p53 (p53 -/-) abrogates the postmitotic cell cycle checkpoint whereas gain of a dominant p53 mutant (p53 +/mut) abrogates both the mitotic and postmitotic checkpoints. Loss of p53 function, either by loss of p53 (p53 -/-) or mutant p53 dominance (p53 +/mut), may then originate aneuploidy due to centrosome amplification (Fukasawa et al. 1996) and multipolar division.

In spite these results, questions still remain as to how p53 mutant proteins affect the control of mitosis. Entry into M phase requires MPF activation, a process that depends upon an increase in cyclin B expression and the dephosphorylation of cdc2. Progression through mitosis and cytokinesis requires the subsequent inactivation of MPF, which depends in part on cyclin B degradation. Experiments in yeast indicate that the mitotic spindle cell cycle checkpoint feeds into the cell cycle regulatory machinery at mitosis by a pathway that delays the degradation of cyclin B and maintains cdc2 kinase activity (Basi and Draetta 1995). Thus cyclin B is degraded and MPF inactivated only after certain aspects of mitosis related to spindle assembly and disassembly have been properly completed. Evidence indicates that mutant p53 proteins render human cells unable to control cyclin B

metabolism in response to mitotic spindle depolymerizing agents (Hixon et al., 1998). Important, loss of wild type p53 does not duplicate this effect (Hixon et al., 1998). An experiment illustrating the differences in the control of cyclin B metabolism in mutant p53 cells versus cells with inactivated wild type p53 is shown in Figure 3. Using the retroviral vectors LXSN or pBabe we expressed the structural p53 mutant 143A or the human papilloma virus protein 16 E6 (HPV16 E6) in primary human fibroblasts. As a control, cell were also coinfected with empty LXSN or pBabe. HPV16 E6 promoted the selective degradation of wild type p53 but not mutant p53. Therefore, NHF pBabe p53 143A/LXSN-E6 showed a dramatic decrease in wild type p53 levels but expressed similar levels of mutant p53 protein to those found in the NHF pBabe p53 143A/LXSN population (Fig. 3A,B). An 80-hour delay in cyclin B metabolism was observed in NHFs exposed to colcemid (Fig. 3C). Similar to the NHFs, 389

Hixon and Gualberto: Mutant p53 and mitosis the NHF16 E6-expressing cell population also delayed cyclin B metabolism in response to colcemid (Figure 3C and D). In contrast, the cell population expressing E6 and p53 143A showed a short mitotic pause, approximately 16 hours, in response to colcemid (Figure 3E). These results demonstrate that expression of a mutant p53 protein, but not wild type p53 inactivation, abrogates the ability of normal cells to regulate the metabolism of cyclin B protein in response to an anomalous chromosomal segregation. Unfortunately, little is known about the molecular events that constitute the mitotic spindle cell cycle checkpoint pathway in normal human cells. However, biochemical and genetic evidence demonstrates that mitotic checkpoint signals, originated in part by the association of MAD2 proteins to unattached kinetochores (Chen et al., 1996; Li and Benezra 1996; Pangilinan and Spencer 1996), are transduced to modulate cyclin B metabolism and MPF activity (Murray et al., 1996). MAD2 has been shown to be associated with the cdc27 component of the anaphase promoting complex (APC), a protease that targets Cyclin B and other mitotic targets, inhibiting its proteolytic activity (Fang et al., 1998; Hwang et al., 1998; Wasserman and Benezra 1998). More exactly, cdc27-APC is activated by MPF in a reaction that requires its association with the Cdc2-Cks1 complex (Patra and Dunphy 1998) (Figure 4). Numerous studies indicate that the level of Cks1 or homologous proteins is rate limiting in Cyclin B degradation (Moreno et al., 1989; Basi and Draetta 1995b; Patra and Dunphy 1996; Hixon et al., 1998; Patra and Dunphy 1998). Human Cks1, also called CksHs1 for Cdc2/Cdc28 kinase subunit Homo sapiens 1 (Richardson et al.,1990), was initially identified as a homologue of the Cdc28/Cdc2- associated proteins of S. cerevisiae, Cks1, and S. pombe, Suc 1 (Richardson et al., 1990). Importantly, in S. pombe, inactivation of the suc 1 gene causes mitotic arrest with high levels of cdc13 (Cyclin B homologue) and high MPF kinase activity (Moreno et al., 1989; Basi and Draetta 1995b). To further elucidate the effect(s) of mutant p53 protein on mitosis we investigated how the presence of mutant p53 proteins could affect Cks1 functions (Hixon et al., 1998). We found that normal human fibroblasts downregulate CksHs1 expression in response to mitotic inhibition (Figure 5A). In contrast, human fibroblasts carrying mutant p53 proteins expressed higher levels of CKsHs1 and failed to downregulate CksHs1 expression in response to mitotic spindle depolymerization (Figure 5B). More importantly, ectopic expression of CKsHs1 originated altered cell cycle checkpoint status (Figure 5C). These results identified CksHs1 as a mutant p53 gene target and provide a molecular mechanism to understand the origin of heteroploidy in mutant p53 expressing cells. However, many details remain to be elucidated. It appears that CksHs1 induces aneuploidy by interfering with the control of cyclin B metabolism by the mitotic spindle checkpoint

(Hixon et al., 1998). However, whether mutant p53 directly regulates Cks1 or affects its control through interaction with other p53-related proteins warrants further study. In addition, CKsHs1 could be one of many mutant p53 targets at mitosis. How could mutant p53 protein affect the expression of mitotic regulators that are not wild type p53 targets? Mutant p53 proteins may directly activate gene expression. Several reports have indicated gain of function transcriptional properties for p53 mutants (Suber et al., 1992; Dittmer et al., 1993; Gualberto et al., 1995; Lin et al., 1995; Frazier et al., 1998). Also, it has been postulated that mutant p53 proteins may mimic the biological function of a â&#x20AC;&#x153;proliferativeâ&#x20AC;? conformational stage of wild-type p53 (Milner and Watson 1990; Ullrich et al., 1992). This proliferative p53 may work in association with other transcription factors, such as Sp1 (Gualberto and Baldwin 1995). Alternatively, mutant p53 could affect the expression of CKsHs1 and other mitotic proteins indirectly, affecting the levels of its regulators. Finally, recent data (Di Como et al., 1999 Ruiz-Lozano et al., 1999) suggest an additional molecular mechanism that could explain some of the gain of function properties of mutant p53 proteins. These authors demonstrate that the dominant negative activity of mutant p53 can be exerted not only over wild type p53, but also over other p53-related proteins, such as p51 and p73. Thus, certain differences between p53 mutant and p53 null phenotypes could be due to the ability of p53 related proteins to compensate for the lack of p53 function in the latter. In conclusion, loss of fidelity in the processes that replicate, repair, and segregate the genome may allow for the accumulation of the genetic alterations that eventually lead to a malignant phenotype. These processes are integrated with the cell cycle regulatory machinery by cell cycle checkpoint pathways. Loss or inactivation of cell cycle checkpoint genes, such as the p53, results in genomic instability and tumor progression. In addition, mutant p53 proteins may contribute to genomic instability by mechanisms that do not necessarily imply the inactivation of wild type p53. At mitosis, structural p53 mutants, but not loss of wild type p53 function, cause altered mitotic spindle cell cycle checkpoint status. This altered checkpoint status is originated, at least in part, by the upregulation of Cks1, a cofactor of MPF that targets cdc2 to the APC. In normal cells, mitotic checkpoint signals delay the progression of mitosis by inhibiting the activity of the APC through a dual mechanism, MAD2 association to APC components and dowregulation of Cks1. Overexpression of Cks1 abrogates the mitotic spindle checkpoint reproducing the effects of mutant p53 proteins on cyclin B metabolism and the progression through mitosis. In addition, p53 dominant mutants may facilitate the survival of aneuploid cells due to inactivation of a postmitotic p53-dependent checkpoint. These data contribute to our understanding of the origin of aneuploidy in mutant p53 cells. Future experiments should focus on the upregulation of CksHs1 and other members of this protein family in human tumor initiation and progression.

390

Gene Therapy and Molecular Biology Vol 4, page 391

Fig. 3: (A) Western analysis of p53 and !-actin in NHF carrying the retroviral vectors pBabe (puromycin resistance, no insert) and LXSN (neomycin resistance, no insert) or LXSN-E6. For the induction of p53, the cells were incubated for 2 days in 10 ÂľM mycophenolic acid (a GMP biosynthesis inhibitor). (B) Immunoprecipitation of mutant p53 and !-actin in primary NHF following double infection with the retroviral vector pBabe and LXSN or LXSN-E6 or with the vector pBabe p53 143A and LXSN or LXSN-E6. The cells were metabolically labeled and the proteins were analyzed (C to E ) Western analysis of cyclin B and !-actin in NHF carrying the retroviral vectors pBabe and LXSN (C), pBabe and LXSN-E6 (D), or pBabe p53 143A and LXSN-E6 (E). The cells were incubated and western blotting was carried out. Data are representative of three independent experiments.

391

Hixon and Gualberto: Mutant p53 and mitosis

Fig. 4: Model of the mitotic spindle cell cycle checkpoint pathway. Cdc2 phosphorylates and activates the protease complex APC. Cks1 allows docking of Cdc27 onto Cdc2. Dowregulation of Cks1 expression by the mitotic spindle checkpoint pathway block APC activation and degradation of its mitotic targets such as Cyclin B. Bishop, JM (1987) The molecular genetics of cancer Science 235, 305-311. Cahill, DP, Lengauer, C, Yu, J, Riggins, GJ, Willson, JK, Markowitz, SD, Kinzler, KW, and Vogelstein, B (1998). Mutations of mitotic checkpoint genes in human cancers Nature 392, 300-303. Boddy, MN, Furnari, B, Mondesert, O, and Russell, P (1998) Replication checkpoint enforced by kinases Cds1 and Chk1 Science 280, 909-912. Chen, RH, Waters, JC, Salmon, ED, and Murray, AW (1996). Association of spindle assembly checkpoint component XMAD2 with unattached kinetochores Science 274, 242246. Cross, SM, Sanchez, CA, Morgan, CA, Schimke, MK, Ramel, S, Idzerda, RL, Raskind, WH and Reid, BJ (1995). A p53dependent mouse spindle checkpoint Science 267, 13531356. DeWald, MG, Sharma, RC, Kung, AL, Wong, HE, Sherwood, SW, and Schimke, RT (1994). Heterogeneity in the mitotic checkpoint control of BALB/3T3 cells and a correlation with gene amplification propensity Cancer Res 54, 5064-5070. Di Leonardo, A, Khan, SH, Linke, SP, Greco, V, Seidita, G, and Wahl, GM (1997). DNA rereplication in the presence of mitotic spindle inhibitors in human and mouse fibroblasts lacking either p53 or pRb function Cancer Res 57, 10131019. Dittmer, D, Pati, S, Zambetti, G, Chu, S, Teresky, AK, Moore, M, Finlay, C, and Levine, AJ (1993). Gain of function mutations in p53 Nat Genet 4, 42-46. Donehower, LA, Harvey, M, Slagle, BL, McArthur, MJ, Montgomery, CA, Jr, Butel, JS, and Bradley, A (1992) Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 356, 215-221. Elledge, SJ (1996). Cell cycle checkpoints: preventing an identity crisis Science 274, 1664-1672.

Acknowledgments This work was supported in part by grants NIH AR39750, ACS IRG91022, and AHA to A.G.

References Agapova, LS, Ilyinskaya, GV, Turovets, NA, Ivanov, AV, Chumakov, PM, and Kopnin, BP (1996) Chromosome changes caused by alterations of p53 expression Mutat Res 354, 129-138. Agarwal, ML, Agarwal, A, Taylor, WR, Chernova, O, Sharma, Y, and Stark, GR (1998) A p53-dependent S-phase checkpoint helps to protect cells from DNA damage in response to starvation for pyrimidine nucleotides Proc Natl Acad Sci U S A 95, 14775-14780. Barlow, C, Hirotsune, S, Paylor, R, Liyanage, M, Eckhaus, M, Collins, F, Shiloh, Y, Crawley, JN, Ried, T, Tagle, D, and Wynshaw-Boris, A (1996) Atm-deficient mice: a paradigm of ataxia telangiectasia Cell 86, 159-171. Baross-Francis, A, Andrew, SE, Penney, JE, and Jirik, FR (1998) Tumors of DNA mismatch repair-deficient hosts exhibit dramatic increases in genomic instability Proc Natl Acad Sci U S A 95, 8739-8743. Basi, G and Draetta, G (1995) The Cdc2 kinase: structure, activation, and its role at mitosis in vertebrate cells. In Cell Cycle Control Hutchison, G. and Glover, D. M. (eds). Oxford University Press, Oxford. pp 106-134. Basi, G and Draetta, G (1995b) p13suc1 of Schizosaccharomyces pombe regulates two distinct forms of the mitotic Cdc2 kinase Mol Cell Biol 15, 2028-2036. Bischoff, FZ, Yim, SO, Pathak, S, Grant, G, Siciliano, MJ, Giovanella, BC, Strong, LC, and Tainsky, MA (1990). Spontaneous abnormalities in normal fibroblasts from patients with Li-Fraumeni cancer syndrome: aneuploidy and immortalization Cancer Res 50, 7979-7984.

392

Gene Therapy and Molecular Biology Vol 4, page 393

Fig. 5: Northern analysis of CKsHs1 expression in NHF-pBabe (A) and NHF-pBabe p53 143A (B) cells incubated in the presence of increasing concentrations of colcemid. Cells were synchronized at G0 by incubation in low serum media and then incubated for 40 h in 10% FBS in the presence of 0 (a), 100 (b), 200 (c), or 1000 (d) ng/ml of colcemid. Colcemid was added at 12 h after cell passage. (C) Flow cytometry analysis of cell cycle distribution of DNA content in C2C12 cells ectopically expressing CKsHs1 and/or bcl-2. Cells were incubated in the absence or presence of 200 ng/ml of colcemid for 2 PDL incubation times. PDLs for C2C12 bcl2/pBabe and C2C12 bcl2/pBabe CKsHs1 cells were 38 and 32 h, respectively. Following incubations, cells were harvested and processed for flow cytometry of DNA content.

Elson, A, Wang, Y, Daugherty, CJ, Morton, CC, Zhou, F, Campos-Torres, J, and Leder, P (1996) Pleiotropic defects in ataxia-telangiectasia protein-deficient mice Proc Natl Acad Sci U S A 93, 13084-13089 Fang, G, Yu, H, and Kirschner, MW (1998) The checkpoint protein MAD2 and the mitotic regulator CDC20 form a ternary complex with the anaphase-promoting complex to control anaphase initiation Genes Dev 12, 1871-1883.

Foiani M, Ferrari M, Liberi G, Lopes M, Lucca C, Marini F, Pellicioli A, Falconi MM, Plevani P (1998) S-phase DNA damage checkpoint in budding yeast Biol Chem 1998 379, 10191023 Frazier, MW, He, X, Wang, J, Gu, Z, Cleveland, JL, and Zambetti, GP (1998) Activation of c-myc gene expression by tumorderived p53 mutants requires a discrete C-terminal domain Mol Cell Biol 18, 3735-3743.

393

Hixon and Gualberto: Mutant p53 and mitosis Fukasawa, K, Choi, T, Kuriyama, R, Rulong, S, and Vande Woude, GF (1996) Abnormal centrosome amplification in the absence of p53 Science 271, 1744-1747. Gualberto, A, Aldape, K, Kozakiewicz, K, and Tlsty, TD (1998) An oncogenic form of p53 confers a dominant, gain-offunction phenotype that disrupts spindle checkpoint control Proc Natl Acad Sci U S A 95, 5166-5171. Gualberto, A, Hixon, ML, Finco, TS, Perkins, ND, Nabel, GJ, and Baldwin, AS, Jr. (1995). A proliferative p53-responsive element mediates tumor necrosis factor alpha induction of the human immunodeficiency virus type 1 long terminal repeat Mol Cell Biol 15, 3450-3459. Harper, JW, Adami, GR, Wei, N, Keyomarsi, K, and Elledge, SJ (1993) The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases Cell 75, 805-816. Hartwell, LH and Kastan, MB (1994) Cell cycle control and cancer Science 266, 1821-1828. Harvey, M, McArthur, MJ, Montgomery, CA, Jr, Butel, JS, Bradley, A, and Donehower, LA (1993) Spontaneous and carcinogen-induced tumorigenesis in p53-deficient mice Nat Genet 5, 225-229. Hermeking, H, Lengauer, C, Polyak, K, He, T-C, Zhang, L, Thiagalingam, S, Kinzler, KW, and Vogelstein, B (1998) 143-3" is a p53-regulated inhibitor of G2/M progression Mol Cell 1, 3-11. Hixon, ML, Flores, A-I, Wagner, MW, and Gualberto, A (1998). Ectopic expression of Cdc2/Cdc28 kinase subunit homo sapiens 1 uncouples Cyclin B metabolism from the mitotic spindle cell cycle checkpoint Mol Cell Biol 18, 6224-6237. Hwang, LH, Lau, LF, Smith, DL, Mistrot, CA, Hardwick, KG, Hwang, ES, Amon, A, and Murray, AW (1998) Budding yeast Cdc20: a target of the spindle checkpoint Science 279, 1041-1044. Jacks, T, Fazeli, A, Schmitt, EM, Bronson, RT, Goodell, MA, and Weinberg, RA (1992) Effects of an Rb mutation in the mouse Nature 359, 295-300. Jacks, T, Shih, TS, Schmitt, EM, Bronson, RT, Bernards, A, and Weinberg, RA (1994) Tumour predisposition in mice heterozygous for a targeted mutation in Nf1. Nat Genet 7, 353-361 Kamijo, T, Bodner, S, van de Kamp, E, Randle, DH, and Sherr CJ (1999) Tumor spectrum in ARF-deficient mice. Cancer Res 59, 2217-2222. Kinzler, KW and Vogelstein, BW (1996) Lessons from Hereditary Colorectal Cancer Cell 87, 159-170. Lanni, JS and Jacks, T (1998) Characterization of the p53dependent postmitotic checkpoint following spindle disruption Mol Cell Biol 18, 1055-1064. Levine, AJ (1997) p53, the Cellular Gatekeeper for Growth and Division Cell 88, 323-331. Li, Y and Benezra, R (1996) Identification of a human mitotic checkpoint gene: hsMAD2 Science 274, 246-248. Lin, J, Teresky, AK, and Levine, AJ (1995) Two critical hydrophobic amino acids in the N-terminal domain of the p53 protein are required for the gain of function phenotypes of human p53 mutants Oncogene 10, 2387-2390. Lindsay, HD, Griffiths, DJ, Edwards, RJ, Christensen, PU, Murray, JM, Osman, F, Walworth, N, and Carr, AM (1998) S-phase-specific activation of Cds1 kinase defines a subpathway of the checkpoint response in Schizosaccharomyces pombe. Genes Dev 12, 382-395. Livingstone, LR, White, A, Sprouse, J, Livanos,E, Jacks, T, and Tlsty TD (1992) Altered cell cycle arrest and gene

amplification potential accompany loss of wild-type p53 Cell 70, 923-935. Malkin, D, Li, FP, Strong, LC, Fraumeni, JF, Jr, Nelson, CE, Kim, DH, Kassel, J, Gryka, MA, Bischoff, FZ, Tainsky, MA, and et al. (1990). Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms Science 250, 1233-1238. Michael, WM and Newport, J (1998) Coupling of mitosis to the completion of S phase through Cdc34-mediated degradation of Wee1 Science 282, 1886-1889. Minn, AJ, Boise, LH, and Thompson, CB (1996). Expression of BclxL and loss of p53 can cooperate to overcome a cell cycle checkpoint induced by mitotic spindle damage Genes Dev 10, 2621-2631. Minshull, J, Sun, H, Tonks, NK, and Murray, AW (1994) A MAP kinase-dependent spindle assembly checkpoint in Xenopus egg extracts Cell 79, 475-486. Moreno, S, Hayles, J, and Nurse, P (1989) Regulation of p34Cdc2 protein kinase during mitosis Cell 58, 361-372. Moser, AR, Mattes, EM, Dove, WF, Lindstrom, MJ, Haag, JD, and Gould, MN ( 1993) ApcMin, a mutation in the murine Apc gene, predisposes to mammary carcinomas and focal alveolar hyperplasias Proc Natl Acad Sci U S A 90, 8977-8981. Moser, AR, Pitot, HC, and Dove, WF (1990) A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 247, 322-324. Murphy, M, Hinman, A, and Levine, AJ (1996) Wild-type p53 negatively regulates the expression of a microtubule-associated protein Genes Dev 10, 2971-2980. Murray, AW (1995) The genetics of cell cycle checkpoints Curr Opin Genet Dev 5, 5-11. Murray, AW, Desai, AB, and Salmon, ED (1996) Real time observation of anaphase in vitro Proc Natl Acad Sci U S A 93, 12327-12332. Nakayama, K, Ishida, N, Shirane, M, Inomata, A, Inoue, T, Shishido, N, Horii, I, Loh, DY, and Nakayama, K (1996) Mice lacking p27(Kip1) display increased body size, multiple organ hyperplasia, retinal dysplasia, and pituitary tumors Cell 85, 707720. Niculescu, AB, Chen, X, Smeets, M, Hengst, L, Prives, C, and Reed, SI (1998) Effects of p21 (cip/Waf1) at both the G1/S and the G2/M cell cycle transitions: pRb is a critical determinent in blocking DNA replication and in preventing endoreduplication Mol Cell Biol 18, 629-643. Norbury, C and Nurse, P (1992) Animal cell cycles and their control Annu Rev Biochem 61, 441-470. Nowell, PC (1976) The clonal evolution of tumor cell populations Science 194, 23-28. Nowell PC (1986) Mechanisms of tumor progression Cancer Res 46, 2203-2207. Ookata, K, Hisanaga, S, Okumura, E, and Kishimoto, T (1993) Association of p34Cdc2/Cyclin B complex with microtubules in starfish oocytes J Cell Sci 105, 873-881. Pangilinan, F and Spencer, F (1996). Abnormal kinetochore structure activates the spindle assembly checkpoint in budding yeast Mol Biol Cell 7, 1195-1208. Patra, D and Dunphy, WG (1996) Xe-p9, a Xenopus Suc1/Cks homolog, has multiple essential roles in cell cycle control Genes Dev 10, 1503-1515. Patra, D and Dunphy, WG (1998). Xe-p9, a Xenopus Suc1/Cks protein, is essential for the Cdc2-dependent phosphorylation of the anaphase-promoting complex at mitosis Genes Dev 12, 2549-2559.

394

Gene Therapy and Molecular Biology Vol 4, page 395 Peled, A, Schwartz, D, Elkind, NB, Wolkowicz, R, Li, R, and Rotter, V (1996) The role of p53 in the induction of polyploidity of myelomonocytic leukemic M1/2 cells Oncogene 13, 1677-1685. Richardson, HE, Stueland, CS, Thomas, J, Russell, P, and Reed, SI (1990) Human cDNAs encoding homologs of the small p34Cdc28/Cdc2-associated protein of Saccharomyces cerevisiae and Schizosaccharomyces pombe Genes Dev 4, 1332-1344. Rudner, AD and Murray, AW (1996) The spindle assembly checkpoint Curr Opin Cell Biol 8, 773-780. Ruiz-Lozano, P, Hixon, ML, Wagner, MW, Flores, A, Chien, K, Baldwin, AS, and Gualberto, A (1999) p53 is a transcriptional activator of the muscle-specific phosphoglycerate mutase gene and contributes in vivo to the control of its cardiac expression Cell Growth & Diff. In press. Srivastava, S, Zou, ZQ, Pirollo, K, Blattner, W, and Chang, EH (1990) Germ-line transmission of a mutated p53 gene in a cancer-prone family with Li-Fraumeni syndrome Nature 348, 747-749. Stewart, ZA, Leach, SD, and Pietenpol, JA (1999) p21(Waf1/Cip1) inhibition of cyclin E/Cdk2 activity prevents endoreduplication after mitotic spindle disruption Mol Cell Biol 19, 205-215. Subler, MA, Martin, DW, and Deb, S (1992) Inhibition of viral and cellular promoters by human wild-type p53 J Virol 66, 4757-4762. Tlsty, TD, Briot, A, Gualberto, A, Hall, I, Hess, S, Hixon, M, Kuppuswamy, D, Romanov, S, Sage, M, and White, A (1995) Genomic instability and cancer Mutat Res 337, 17. Wang, XJ, Greenhalgh, DA, Jiang, A, He, D, Zhong, L, Brinkley BR, and Roop DR (1998) Analysis of centrosome abnormalities and angiogenesis in epidermal-targeted p53172H mutant and p53-knockout mice after chemical carcinogenesis: evidence for a gain of function Molecular Carcinogensis 23, 185-92. Toft, NJ, Winton, DJ, Kelly J, Howard, LA, Dekker, M, te Riele, H, Arends, MJ, Wyllie, AH, Margison, GP, and Clarke AR (1999) Msh2 status modulates both apoptosis and mutation frequency in the murine small intestine. Proc Natl Acad Sci U S A 96, 3911-3915. Wassmann, K and Benezra, R ( 1998) Mad2 transiently associates with an APC/p55Cdc complex during mitosis Proc Natl Acad Sci U S A 95, 11193-11198. Weinert, T (1998) DNA damage checkpoints update: getting molecular Curr Opin Genet Dev 8, 185-193. White, AE, Livanos, EM, and Tlsty, TD (1994) Differential Disruption of Genomic Integrity and Cell Cycle regulation in Normal Human Fibroblasts by the Human Papilloma Virus Oncoproteins Genes and Dev 8, 666-677. Williams, BO, Schmitt,, EM, Remington, L, Bronson, RT, Albert, DM, Weinberg, RA, and Jacks T (1994) Extensive contribution of Rb-deficient cells to adult chimeric mice with limited histopathological consequences EMBO J 13, 42514259. Wolman, SR (1983). Karyotypic progession in human tumors. Cancer Metastasis Rev 2, 257-293. Xiong, Y, Kuppuswamy, D, Li, Y, Livanos, EM, Hixon, M, White AE, Beach, D, and Tlsty, TD (1996) Alteration of cell cycle kinase complexes in HPV E6- and E7- expressing fibroblasts precedes neoplastic transfromation J Virol 70, 999-1008.

Yin, Y, Tainsky, MA, Bischoff, FZ, Strong, LC, and Wahl GM (1992) Wild-type p53 restores cell cycle control and inhibits gene amplification in cells with mutant p53 alleles Cell 70, 937-948.

Antonio Gualberto 395

Gene Therapy and Molecular Biology Vol 4, Page 397 Gene Ther Mol Biol Vol 4, 397-404. December 1999

BRCA1 function in transcription Review Article

Neelima Mondal and Jeffrey D. Parvin Department of Pathology, Harvard Medical School and, Brigham and Womenâ&#x20AC;&#x2122;s Hospital, Boston, MA ______________________________________________________________________________________________________ Correspondence: Jeffrey D. Parvin, M.D., Ph.D., Brigham and Womenâ&#x20AC;&#x2122;s Hospital, Department of Pathology, 75 Francis Street, Boston, MA 02115, USA. Tel: 617-278-0818; Fax: 617-732-7449; E-mail: jparvin@rics.bwh.harvard.edu Received: 30 April 1999; accepted: 7 May 1999

Summary The breast and ovarian specific tumor suppressor protein, BRCA1, i s associated with the transcriptional regulatory machine termed the RNA polymerase II (Pol II) holoenzyme. Experiments support a model in which molecules such as the BRCA1 protein and the CREB binding protein (CBP) function similarly as regulatory targets which link the enhancer-binding regulatory proteins to the basal transcriptional machinery. As an example, phosphorylated CREB protein interacts with the CBP molecule present in the holoenzyme complex and will activate transcription in vitro. We propose a similar model for the BRCA1 protein, but with a different subset of upstream activators functioning via BRCA1 to regulate the Pol II holoenzyme. We have established an in vitro transcriptional assay dependent upon the carboxy-terminal domain of BRCA1 fused to the DNA-binding domain of GAL4. We have also shown that both CBP and BRCA1 are linked to the Pol II holoenzyme complex via a common subunit -- the RNA Helicase A (RHA) protein. CBP binds to an amino terminal domain of RHA, and BRCA1 binds to a separate internal domain of RHA, and truncated RHA molecules have been shown to be dominant negative transcriptional repressors of either CBP or BRCA1. Since most t u m o r - a s s o c i a t e d B R C A 1 m u t a t i o n s r e s u l t i n t h e t r u n c a t i o n o f t h e B R C A 1 p r o t e i n a n d l o s s o f its carboxy-terminal holoenzyme interaction domain, it is suggested that an important component of the tumor suppressive function of BRCA1 occurs via the gene expression process.

I. Genetics of BRCA1

The genetic predisposition to breast cancer clearly implicates BRCA1 protein as being important in the etiology of the disease, but these data yield no insights into how this protein functions. Data reviewed below will outline new developments from our laboratory on how BRCA1 functions to regulate the process of gene expression.

The BRCA1 gene was the first identified tumor susceptibility gene specific for breast and/or ovarian cancer (King, 1980). The gene encodes a 220 kd protein which has at its carboxy terminus an acidic domain consistent with a transcriptional regulator (Miki et al, 1994; Futreal et al, 1994). About 5% of breast cancer cases are attributable to the inheritance of a mutant BRCA1 allele. A mutant BRCA1 allele is present among about 50% of families with a predisposition to breast cancer and among about 80% of families with a predisposition to both breast and ovarian cancer (Szabo and King, 1995). Among these families, the genetics of cancer due to BRCA1 are similar to the retinoblastoma (RB) tumor suppressor. Predisposition to cancer is due to inheritance of a mutant allele, and disease occurs when the second allele becomes mutated (Neuhausen and Marshall, 1994; Smith et al, 1992).

II. Clues to BRCA1 function Murine knock-out studies were largely uninformative regarding the function of BRCA1 since nullizygous embryos died at very short gestational ages (Ludwig et al, 1997; Hakem et al, 1996; Hakem et al, 1997; Shen et al, 1998). An interaction with the function of the p53 tumor suppressor was suggested in several of these studies since animals doubly deficient in p53 and in BRCA1 remained alive two embryonic days longer, suggesting an antagonistic genetic interaction between these two tumor suppressors (Hakem et

397

Mondal and Parvin: BRCA1 function in transcription al, 1997; Shen et al, 1998). Since p53 exerts its tumor suppressing effects in large part by transcriptional regulation, it is possible that BRCA1 is also a transcriptional regulator. Transcription function of BRCA1 was first suggested in experiments in which the carboxy terminus of BRCA1 was fused to the DNA binding domain of the GAL4 transcription factor and the activation of transcription in transfected cells was observed dependent upon the presence of the GAL4 DNA element in the reporter gene (Chapman and Verma, 1996; Monteiro et al, 1996). Further transcription function of the BRCA1 protein was revealed by experiments in our laboratory in which the master transcriptional regulatory complex known as the RNA polymerase II holoenzyme (see below) was found to have BRCA1 associated with the complex when purified from HeLa or B cell lines (Scully et al, 1997a; Neish et al, 1998). BRCA1 protein transcriptional function has been associated with the regulation of transcription by p53. Overexpression of full-length BRCA1 was found to potentiate the transcriptional activation of reporter genes controlled by p53 response elements (Ouchi et al, 1998). The p53 protein was shown to bind directly to the amino acid residues 224-500 of the BRCA1 protein (Zhang et al., 1998). These data suggest that BRCA1 may somehow interact with p53 in order to regulate gene expression and thus function in tandem as tumor suppressors. BRCA1 has also been shown to function independent of p53 in regulating the expression of the p21 cell cycle progression inhibitor (Somasundarum et al, 1997). Taken together these data suggest that BRCA1 protein acts as a tumor suppressor protein in concert with the p53 protein. Mechanistically how this occurs is the subject of the model developed in this review. Why such an interaction should result in tumor suppression specifically in breast and ovarian cancer but have less importance in other cancers is unclear. Other functions of BRCA1 have been identified and will not be reviewed here. These include the association with the RAD51 protein in nuclear dots during the S phase of the cell cycle (Scully et al, 1997b) and BRCA1 regulation of the number of centrosomes and resultant effects on chromosome sorting during mitosis (Xu et al, 1999). It has also been shown that embryonic stem cells nullizygous for BRCA1 have defects in the process of transcription-coupled repair, although it was possible that those results could have arisen from defects in transcription alone (Gowen et al, 1998).

vitro by enhancer-binding proteins even more protein factors are required. The mRNA-synthesizing enzyme, RNA polymerase II (Pol II) exists in at least two forms: core polymerase and Pol II holoenzyme (Koleske et al, 1994). The core polymerase contains 12 subunits and has a mass of 500 kD. Several basal transcription factors associate with core Pol II in the absence of DNA (Flores et al, 1989; Sopta et al, 1989), possibly foreshadowing the existence of a super complex containing Pol II and other basal factors pre-assembled independent of DNA (Figure 1). The existence of the Pol II holoenzyme was first revealed in a yeast genetic screen for s uppressors of RNA polymerase B mutations which truncated the carboxy terminal domain of the largest RPB1 subunit (Nonet et al, 1989). Nine SRB genes resulted from this genetic screen and the SRB polypeptides were found exclusively in association with the Pol II in a complex termed the holoenzyme (Koleske et al, 1994). Antibodies directed against SRB proteins provide direct evidence for a holoenzyme complex that includes a subset of basal factors, as well as SRB Proteins. The yeast Pol II holoenzyme was independently discovered based upon a biochemical purification of a â&#x20AC;&#x153;mediatorâ&#x20AC;? fraction which allowed regulation of transcription in in vitro reactions (Kim et al, 1994). In these studies, the association of the yeast Pol II with mediator was required for transcriptional activation. The yeast RNA polymerase II holoenzyme consists of core Pol II, an SRB/mediator complex, the SWI/SNF chromatin-remodeling complex and subset of basal transcription factors (Wilson et al, 1996). The SRB proteins are considered hallmarks of yeast and mammalian holoenzyme as they are found almost exclusively in the holoenzyme complex and appear to be limiting for holoenzyme formation. The most highly defined yeast SRB/mediator complex consists of polypeptides: SRB2, 4-7. Med1, 2, 4, 6-8, GAL11, Sin4, Rgr1, Rox3 and Pgd1 (Myers et al, 1998). Yeast SRB10 is a cyclin-dependent kinase that is associated with SRB-8, -9 and -11. These polypeptides are components of the RNA polymerase II holoenzyme purified according to the method of Koleske and Young (1994). At promoters for genes involved in cell-type specificity, meiosis and sugar utilization, SRB10/11 is stimulated to phosphorylate the RNA polymerase II carboxy-terminal domain (CTD) prior to stable pre-initiation complex formation and, by doing so, inhibits transcription initiation (Hengartner et al, 1998). The Pol II holoenzyme is best characterized from yeast cells, but new studies from our laboratory and others have characterized the Pol II holoenzyme complexes found in human cells (reviewed in Parvin and Young, 1998). Just as in the yeast holoenzyme, the mammalian version includes

III. RNA polymerase II holoenzyme Transcription of all protein-encoding genes from an eukaryotic cell into mRNA occurs via the action of RNA polymerase II (Pol II). Pol II cannot function alone, but in vitro requires the basal transcription factors TFIID, TFIIB, TFIIF, TFIIE, and TFIIH (reviewed by Orphanides et al, 1996). In order to observe activation of transcription in 398

Gene Therapy and Molecular Biology Vol 4, Page 399 SRB/mediator components and basal transcription factors: SRB7 (hSRB7), Rgr1, Med6, Med7, SRB10/11 (CDK8/cyclin C), SWI2 Family member BRG1 (Brm/SWI2 related gene), TFIIF, TFIIE, TFIIH (Parvin and Young, 1998; Neish et al, 1998; Maldonado et al, 1996; Pan et al, 1997; Ossipow et al, 1995). Unlike the yeast holoenzyme, the human holocomplex contains other proteins which have been shown to have coactivator function such as CBP and p300 and also the BRCA1 protein which has been suggested to be a coactivator (Nakajima et al, 1997a; Neish et al, 1998; Ouchi et al, 1998).

ligand binding to the cell surface results in a transient increase in the concentration of the second messenger cAMP which leads to the activation of protein kinase A (PK-A) which stimulates the phosphorylation of specific nuclear factors. Genes which respond to cAMP were found to have specific DNA binding elements known as CREs, for cAMP response elements, and the factor CREB bound to these DNA sequences. Activated PK-A phosphorylates Ser-133 of CREB prebound on its DNA element and this leads to the activation of neighboring genes (Gonzalez and Montminy, 1989). Although phosphorylation has been shown to stimulate a number of nuclear factors via their DNA-binding or nuclear-targeting activities, CREB belongs to a group of activators whose transactivation potential is specifically affected (Gonzalez and Montminy, 1989; Hagiwara et al, 1992).

IV. Model for activation by phosphoCREB A number of hormones and growth factors induce the expression of certain genes via a cascade of events by which

F i g u r e 1 . Ordered assembly of individual basal transcription factors versus pre-assembled holoenzyme complex. The topmost left panel shows DNA strand with TATA box element. The small arrow indicates the start site and direction of transcription. TFIID first binds to the TATA box and facilitates the binding of TFIIA and TFIIB. This in turn gives the signal for binding of TFIIF and RNA Pol II (core). This complex then facilitates the binding of TFIIE and finally TFIIH to form the initiation complex for transcription. In contrast the pre-assembled holoenzyme containing TFIIE, TFIIF and TFIIH as well as other subunits not shown, is in fact the basal factor complex, which catalyzes most, or all, of the mRNA transcription in the eukaryotic cell.

399

Mondal and Parvin: BRCA1 function in transcription 1805-1890 (Nakajima et al, 1997a). This domain was used as a probe in far western gels to identify a 140 kd protein which copurified with the holoenzyme complex, and probing of an expression library revealed RNA helicase A (RHA) as a candidate for the subunit in the holoenzyme complex which binds CBP (Nakajima et al, 1997b). RHA (Lee and Hurwitz, 1993) is the homologue of the drosophila maleless protein which regulates the levels of transcription from the X chromosome in fruit flies (Bone et al, 1994). Experiments indicated that CBP amino acid residues 1805-1890 directly contact RHA residues 1-250, and that overexpression in tissue culture of RHA (1-250) resulted in an inhibition of CREB-CBP-dependent transcriptional activation (Nakajima et al, 1997b). Further, the E1A oncoprotein competed with RHA for binding to CBP, suggesting the mechanism by which the truncation mutant functioned as a dominant negative protein. These data support the model presented in Figure 2.

The CREB transactivation domain is bipartite, consisting of constitutive and inducible activators, termed Q2 and kinase-i nducible domain (KID), respectively, which function cooperatively in response to cAMP (Brindle et al, 1993; Quinn, 1993). The glutamine rich Q2 domain has been shown to involve the transcriptional apparatus via a constitutive interaction with the TBP-associated factor hTAFII130 (Ferreri et al, 1994; Nakajima et al, 1997a). In contrast, increased intracellular cAMP levels results in the phosphorylation of the KID domain on serine-133 which modulates association with CREB binding protein (CBP; Arias et al, 1994; Kwok et al, 1994). CBP and its related protein, p300, have since been shown to be coactivators for numerous transcriptional regulators in addition to CREB including c-Jun ATF-2, STAT-2, NF-!B, Mi, and p53 (Kamei et al, 1996; Arias et al, 1994; Hanstein et al, 1996; Gerritsen et al, 1997; Price et al, 1998; Gu et al, 1997; reviewed in Parvin and Young, 1998). For CREB and Mi, it has been shown that phosphorylation of the transcriptional regulator is the essential signal for the activator to interact with CBP/P300 (Chrivia et al, 1993; Price et al, 1998). For other factors, such as NF-!B, nuclear transport may be the regulated step for binding CBP. We recently showed that CBP was found associated with functional RNA Pol II holoenzyme complex, and the regulation of transcription in vitro was dependent upon the CBP-containing Pol II complex (Figure 2; Nakajima et al, 1997a). Stimulation of transcription required a phosphorylated Ser-133 CREB as well as interaction with TAF-130 of TFIID. The E1A oncoprotein inhibits transcriptional activation of CREB without interfering with complex formation between phospho (Ser-133)-CREB and CBP. Our in vitro transcription and binding data demonstrated that the mechanism by which E1A suppresses CREB-dependent transcriptional activation is by blocking the interaction of CBP with holoenzyme which causes the release of CBP/p300 and thus suppresses transcription (Nakajima et al, 1997a). CBP and p300 have demonstrated histone acetyltransferase (HAT) activity (Bannister and Kouzarides, 1996; Orgryzko et al, 1996). It is possible that the enhancer-binding regulator of transcription recruits the Pol II holoenzyme via CBP, or it is possible that the activator recruits the HAT activity to the chromatin in the region of the promoter. The repression of transcription by histone deacetylase supports the role of the HAT activity of CBP (Hassig et al, 1997; Pazin and Kadonaga, 1997). The activation of transcription in vitro via CBP has been observed in the absence of the key HAT substrate, histone (Nakajima et al, 1997a), suggesting that at least part of the activity of CBP/p300 is as a holoenzyme-bound target. The domain of CBP which interacts with the holoenzyme complex was mapped to amino acid residues

F i g u r e 2 . Model for regulation of transcription by phosphoCREB-CBP-holoenzyme. The CREB transcription factor bound to its cognate DNA enhancer element is phosphorylated on the KID domain as a result of elevated intercellular cAMP levels. Phospho-CREB then binds to CBP and recruits the holoenzyme via RHA (1-250) to the promoter. A second activation signal from the Q2 domain of CREB interacts with TFIID. Our data demonstrate that both activation signals are required in order to stimulate high levels of mRNA synthesis

V. BRCA1 is a component of RNA Pol II holoenzyme We have identified that a significant fraction of the BRCA1 protein in HeLa and B cells is associated with the Pol II holoenzyme complex (Scully et al, 1997a; Neish et al, 1998). The carboxy terminus of BRCA1, when fused to the DNA binding domain of the GAL4 protein, has been shown to activate transcription of GAL4-site dependent reporters in transfected cells (Chapman and Verma, 1996; Monteiro et al, 1996). We demonstrated that this carboxy terminal domain

400

Gene Therapy and Molecular Biology Vol 4, Page 401 (residues 1560-1863) could function as an affinity matrix for purification of the holoenzyme and that point mutations in this domain at sites which cause cancer resulted in BRCA1 polypeptide which no longer bound to the holoenzyme with high affinity (Neish et al, 1998). We found that this carboxy-terminal BRCA1 domain also binds to the RHA subunit of the holoenzyme complex. RHA (1-250) binds to CBP, and a separate RHA domain from residues 235-325 binds to BRCA1 (Anderson et al, 1998). The BRCA1 carboxy terminal domain containing a point mutation at a site associated with cancer binds to RHA with significantly lower affinity suggesting that this interaction is specific. The in vitro binding data are supported by directed yeast two-hybrid data which demonstrate the interaction between BRCA1 and RHA occurs in that setting. The BRCA1-RHA interaction is central to the activation of transcription of GAL4-BRCA1 fusions in transfection assays. This was determined using truncated RHA molecules which retain BRCA1 binding, and it was shown that expression of these peptides inhibited GAL4BRCA1 transcriptional activation (Anderson et al, 1998). Taken together with the CBP data, it is suggested that BRCA1 functions as an holoenzyme component similarly as does CBP (Figure 3). The similar interaction of holoenzyme component RHA with both CBP and BRCA1 suggest analogous function. In addition the findings that full length BRCA1 potentiates activation by the p53 transcriptional activator (Ouchi et al, 1998; Zhang et al, 1998) and that p53 directly contacts BRCA1 residues 224500 support this model whereby a transcriptional activator, such as p53, binds to the BRCA1 amino terminus and is thus linked with the basal transcriptional machinery in the holoenzyme. Further, other factors may interact in this fashion with BRCA1. For example, the p21 promoter was found to be regulated by BRCA1 via an as yet undetermined transcriptional activator which binds to specific promoter sequences (Somasundarum et al, 1997). The p53 protein could thus interact with the holoenzyme via CBP (Gu et al, 1997) or via BRCA1, raising the issue of why p53 would interact with both CBP and BRCA1. It is likely that interaction with BRCA1 would have a different outcome than activation by p53 via CBP.

VP16, the most powerful transcriptional activator known, the BRCA1 fusion was the more powerful transcriptional activator. This activation by BRCA1 was conditionspecific. When certain transcriptional coactivators were present, then the VP16 fusion was more potent. Different constellations of coactivators and TFIIA were checked for activation by GAL4-BRCA1 and GAL4-VP16 and the most striking difference was observed between Positive Component 4 (PC4) alone versus PC4 plus HMG2. When PC4 alone was used, BRCA1 about 5 fold more effective than VP16, whereas when PC4 plus HMG2 was used, VP16 was twice as effective than was BRCA1. Since PC4 binds to single-stranded bubbles in the DNA, the effect of negative superhelical turns of the template on the transcriptional stimulation was assayed. Activation by BRCA1 was highly dependent upon the supercoiled DNA, whereas the VP16 was more effective on a linear template. These data suggest that VP16 may recruit the TFIIH basal transcription factor to the promoter DNA, and the TFIIH would then generate the bubble of unwound DNA necessary for high levels of transcription. In contrast, BRCA1 would rely upon the supercoiling of the DNA to drive the formation of the transcription bubble. We hypothesize from these data that BRCA1 and VP16 function via different rate-limiting steps leading to the regulation of transcription (Haile and Parvin, 1999).

F i g u r e 3 . Model for regulation of transcription by BRCA1 as an holoenzyme component. Analogous to the CBP-dependent model outlined in Figure 2, activators which binds upstream of promoter to an enhancer element recruit Pol II holoenzyme via BRCA1. Just as CBP signals the holoenzyme via RHA subunit, BRCA1 is modeled to interact with the holoenzyme complex via a second RHA domain. As described in the text, an activator which potentially functions via BRCA1 is p53.

VI. Function of BRCA1 using in vitro transcription assay BRCA1 does not bind DNA, but its transcriptional function in transfected cells has been demonstrated when it is linked to a specific DNA binding domain in the GAL4 protein (cited above). The GAL4 DNA binding domain fused to BRCA1 (1560-1863) was purified in bacteria and tested for GAL4-site-dependent transcriptional activation using highly purified basal transcription factors in vitro (Haile and Parvin, 1999). When compared with GAL4-

Conclusions. Data are accumulating to suggest that BRCA1 plays an important role in transcriptional regulation. GAL4-BRCA1 fusions have been shown to activate transcription when expressed in cells or when assayed in vitro with purified transcription factors. Full length BRCA1 has been found to potentiate the regulation of transcription by p53 and possibly

401

Mondal and Parvin: BRCA1 function in transcription via other proteins which function on the p21 promoter. The model presented in this review suggests that BRCA1 functionally links certain upstream enhancer-binding factors to the basal transcription machinery in the holoenzyme. This is similar to a model for CBP function which we demonstrated applies for CREB-dependent transcriptional activation. Ongoing studies are aimed at identifying which factors are being directly contacted by BRCA1 and which amino acid residues of BRCA1 are involved in proteinprotein interactions. Further, it may be possible to identify which genes are directly regulated by BRCA1.

Gerritsen ME, Williams AJ, Neish AS, Moore S, Shi Y and Collins T (1 9 9 7 ). CREB-binding protein/p300 are transcriptional coactivators of p65. P r o c N a t l A c a d S c i U S A 94, 2927-2932.

References.

Hagiwara M, Alberts A, Brindle P, Meinkoth J, Feramisco J, Deng T, Karin M, Shenolikar S and Montminy M (1 9 9 2 ). Transcriptional attenuation following cAMP induction requires PP-1-mediated dephosphorylation of CREB. C e l l 70, 105-113.

Gonzalez GA, M.M. (1 9 8 9 ). Cyclic AMP stimulates somatostatin gene transcription by phosphorylation of CREB at serine 133. C e l l 59, 675-680. Gowen LC, Avrutskaya AV, Latour AM. Koller BH and Leadon SA (1 9 9 8 ). BRCA1 required for transcription-coupled repair of oxidative DNA damage. S c i e n c e 281, 1009-1012. Gu W, Shi XL and Roeder RG (1 9 9 7 ). Synergistic activation of transcription by CBP and p53. Nature 387, 819-823.

Anderson SF, Schlegel BP, Nakajima T, Wolpin ES and Parvin JD (1 9 9 8 ). BRCA1 protein is linked to the RNA polymerase II holoenzyme complex via RNA helicase A. Nat Genet 19, 254-256.

Haile DT and Parvin JD (1 9 9 9 ). Activation of transcription in vitro by the BRCA1 carboxyl-terminal domain. J . B i o l . C h e m . 274, 2113-2117.

Arany Z, Newsome D, Oldread E, Livingston DM and Eckner R (1 9 9 5 ). A family of transcriptional adaptor proteins targeted by the E1A oncoprotein. Nature 374, 81-84.

Hakem R, de la Pompa JL, Elia A, Potter J and Mak TW (1 9 9 7 ). Partial rescue of Brca1 (5-6) early embryonic lethality by p53 or p21 null mutation. Nat Genet 16, 298-302.

Arany Z, Sellers WR, Livingston DM and Eckner R (1 9 9 4 ). E1A-associated p300 and CREB-associated CBP belong to a conserved family of coactivators. C e l l 77, 799-800.

Hakem R, de la Pompa JL, Sirard C, Mo R, Woo M, Hakem A, Wakeham A, Potter J, Reitmair A, Billia F, Firpo E, Hui CC, Roberts J, Rossant J and Mak TW (1 9 9 6 ). The tumor suppressor gene Brca1 is required for embryonic cellular proliferation in the mouse. C e l l 85, 1009-1023.

Arias J, Alberts AS, Brindle P, Claret FX, Smeal T, Karin M, Feramisco J and Montminy M (1 9 9 4 ). Activation of cAMP and mitogen responsive genes relies on a common nuclear factor. Nature 370, 226-229. Bannister AJ and Kouzarides T (1 9 9 6 ). The CBP co-activator is a histone acetyltransferase. Nature 384, 641-643.

Hanstein B, Eckner R, DiRenzo J, Halachmi S, Liu H, Searcy B, Kurokawa R and Brown M (1 9 9 6 ). p300 is a component of an estrogen receptor coactivator complex. Proc Natl Acad Sci U S A 93, 11540-11545.

Bone JR, Lavender J, Richman R, Palmer MJ, Turner BM and Kuroda MI (1 9 9 4 ). Acetylated histone H4 on the male X chromosome is associated with dosage compensation in Drosophila. G e n e s D e v 8, 96-104.

Hassig CA, Fleischer TC, Billin AN, Schreiber S.L and Ayer DE (1 9 9 7 ). Histone deacetylase activity is required for full transcriptional repression by mSin3A. C e l l 89, 341-347.

Brindle P, Linke S and Montminy M (1 9 9 3 ). Protein-kinase-Adependent activator in transcription factor CREB reveals new role for CREM repressors. Nature 364, 821-824.

Hengartner CJ, Myer VE, Liao SM, Wilson CJ, Koh SS and Young RA (1 9 9 8 ). Temporal regulation of RNA polymerase II by Srb10 and Kin28 cyclin-dependent kinases. M o l . C e l l 2, 43-53.

Chapman MS and Verma IM (1 9 9 6 ). Transcriptional activation by BRCA1. Nature 382, 678-679. Chrivia JC, Kwok RP, Lamb N, Hagiwara M, Montminy MR and Goodman RH (1 9 9 3 ). Phosphorylated CREB binds specifically to the nuclear protein CBP. Nature 365, 855859.

Kamei Y, Xu L, Heinzel T, Torchia J, Kurokawa R, Gloss B, Lin SC, Heyman RA, Rose DW, Glass CK and Rosenfeld MG (1 9 9 6 ). A CBP integrator complex mediates transcriptional activation and AP-1 inhibition by nuclear receptors. C e l l 85, 403-414.

Ferreri K, Gill G and Montminy M (1 9 9 4 ). The cAMP-regulated transcription factor CREB interacts with a component of the TFIID complex. P r o c N a t l A c a d S c i U S A 91, 12101213.

Kim YJ, Bjorklund S, Li Y, Sayre MH and Kornberg RD (1 9 9 4 ). A multiprotein mediator of transcriptional activation and its interaction with the C-terminal repeat domain of RNA polymerase II. C e l l 77, 599-608.

Flores O, Maldonado E and Reinberg D (1 9 8 9 ). Factors involved in specific transcription by mammalian RNA polymerase II. Factors IIE and IIF independently interact with RNA polymerase II. J B i o l C h e m 264, 8913-8921.

King MC, Go RC, Elston RC, Lynch HT and Petrakis NL (1 9 8 0 ). Allele increasing susceptibility to human breast cancer may be linked to the glutamate-pyruvate transaminase locus. S c i e n c e 208, 406-408.

Futreal PA, Liu Q, Shattuck-Eidens D, Cochran C, Harshman K, Tavtigian S, Bennett LM, Haugen-Strano A, Swensen J, Miki Y, et al (1994). BRCA1 mutations in primary breast and ovarian carcinomas. S c i e n c e 266, 120-122.

Koleske AJ and Young RA (1 9 9 4 ). An RNA polymerase II holoenzyme responsive to activators. Nature 368, 466469.

402

Gene Therapy and Molecular Biology Vol 4, Page 403 Kwok RP, Lundblad JR, Chrivia JC, Richards JP, Bachinger HP, Brennan RG, Roberts SG. Green MR and Goodman RH (1 9 9 4 ). Nuclear protein CBP is a coactivator for the transcription factor CREB. Nature 370, 223-226.

components required for promoter-specific transcription initiation. C e l l 83, 137-146. Ouchi T, Monteiro ANA, August A, Aaronson SA and Hanafusa H (1 9 9 8 ). BRCA1 regulates p53-dependent gene expression. Proc Natl Acad Sci USA 95, 2302-2306.

Lee CG and Hurwitz J (1 9 9 3 ). Human RNA helicase A is homologous to the maleless protein of Drosophila. J B i o l Chem 268, 16822-16830.

Pan G, Aso T and Greenblatt J (1 9 9 7 ). Interaction of elongation factors TFIIS and elongin A with a human RNA polymerase II holoenzyme capable of promoter-specific initiation and responsive to transcriptional activators. J B i o l Chem 272, 24563-24571.

Ludwig T, Chapman DL, Papaioannou VE and Efstratiadis A (1 9 9 7 ). Targeted mutations of breast cancer susceptibility gene homologs in mice: lethal phenotypes of Brca1, Brca2, Brca1/Brca2, Brca1/p53, and Brca2/p53 nullizygous embryos. G e n e s D e v 11, 1226-1241.

Parker D, Ferreri K, Nakajima T, LaMorte VJ, Evans R, Koerber SC, Hoeger C and Montminy MR (1 9 9 6 ) Feb;16(2):694703). Phosphorylation of CREB at Ser-133 induces complex formation with CREB-binding protein via a direct mechanism. M o l C e l l B i o l 16, 694-703.

Maldonado E, Shiekhattar R, Sheldon M, Cho H, Drapkin R, Rickert P, Lees E, Anderson CW, Linn S and Reinberg D (1 9 9 6 ). A human RNA polymerase II complex associated with SRB and DNA-repair proteins. Nature 381, 86-89.

Parvin JD and Young RA (1 9 9 8 ). Regulatory targets in the RNA polymerase II holoenzyme. Curr Opin Genet D e v 8, 565-570.

Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harshman K, Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding W, et al (1 9 9 4 ). A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. S c i e n c e 266, 66-71.

Pazin MJ and Kadonaga JT (1 9 9 7 ). What's up and down with histone deacetylation and transcription? C e l l 89, 325-328.

Monteiro AN, August A and Hanafusa H (1 9 9 6 ). Evidence for a transcriptional activation function of BRCA1 C-terminal region. Proc Natl Acad Sci U S A 93, 13595-13599.

Price ER, Ding HF, Badalian T, Bhattacharyal S, Takemoto C, Yao TP, Hemesath TJ and Fisher DE (1 9 9 8 ). Lineagespecific signaling in melanocytes: c-Kit stimulation recruits p300/CBP to microphthalmia. J B i o l C h e m 373, 1798317986.

Myers LC, Gustafsson CM, Bushnell DA, Lui M, ErdjumentBromage H, Tempst P and Kornberg RD (1 9 9 8 ). The Med proteins of yeast and their function through the RNA polymerase II carboxy-terminal domain. Genes D e v 12, 45-54.

Quinn PG (1 9 9 3 ). Distinct activation domains within cAMP response element-binding protein (CREB) mediate basal and cAMP-stimulated transcription. J B i o l C h e m 268, 1699917009.

Nakajima T, Uchida C, Anderson SF, Lee CG, Hurwitz J, Parvin JD and Montminy M (1 9 9 7 b ). RNA helicase A mediates association of CBP with RNA polymerase II. C e l l 90, 11071112.

Scully R, Anderson SF, Chao DM, Wei W, Ye L, Young RA, Livingston DM and Parvin JD (1 9 9 7 a ). BRCA1 is a component of the RNA polymerase II holoenzyme. Proc Natl Acad Sci USA 94, 5605-5610.

Nakajima T, Uchida C, Anderson SF, Parvin JD and Montminy M (1 9 9 7 a ). Analysis of a cAMP-responsive activator reveals a two-component mechanism for transcriptional induction via signal-dependent factors. G e n e s D e v 11, 738-747.

Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, Ashley T and Livingston DM (1 9 9 7 b ). Association of BRCA1 with Rad51 in mitotic and meiotic cells. C e l l 88, 265-275.

Neish AS, Anderson SF, Schlegel BP, Wei W and Parvin JD (1 9 9 8 ). Factors associated with the mammalian RNA polymerase II holoenzyme. Nuc. Acids R e s . 26, 847853.

Shen SX, Weaver Z, Xu X, Li C, Weinstein M, Chen L, Guan XY, Ried T and Deng CX (1 9 9 8 ). A targeted disruption of the murine Brca1 gene causes gamma-irradiation hypersensitivity and genetic instability. O n c o g e n e 17, 3115-3124.

Neuhausen SL and Marshall CJ (1 9 9 4 ). Loss of heterozygosity in familial tumors from three BRCA1-linked kindreds. Cancer Res 54, 6069-6072.

Smith SA, Easton DF, Evans DG and Ponder BA (1 9 9 2 ). Allele losses in the region 17q12-21 in familial breast and ovarian cancer involve the wild-type chromosome. N a t G e n e t 2, 128-131.

Nonet ML and Young RA (1 9 8 9 ). Intragenic and extragenic suppressors of mutations in the heptapeptide repeat domain of Saccharomyces cerevisiae RNA polymerase II. G e n e t i c s 123, 715-724.

Somasundaram K, Zhang H, Zeng YX, Houvras Y, Peng Y, Zhang H, Wu GS, Licht JD, Weber BL and El-Deiry WS (1 9 9 7 ). Arrest of the cell cycle by the tumour-suppressor BRCA1 requires the CDK-inhibitor p21WAF1/CiP1. Nature 389, 187-190.

Ogryzko VV, Schiltz RL, Russanova V, Howard BH and Nakatani Y (1 9 9 6 ). The transcriptional coactivators p300 and CBP are histone acetyltransferases. C e l l 87, 953-959. Orphanides G, Lagrange T and Reinberg D (1 9 9 6 ). The general transcription factors of RNA polymerase II. G e n e s D e v . 10, 2657-2683.

Sopta M, Burton ZF and Greenblatt J (1 9 8 9 ). Structure and associated DNA-helicase activity of a general transcription initiation factor that binds to RNA polymerase II. Nature 341, 410-411.

Ossipow V, Tassan JP, Nigg EA and Schibler U (1 9 9 5 ). A mammalian RNA polymerase II holoenzyme containing all

403

Mondal and Parvin: BRCA1 function in transcription Szabo CI and King MC (1 9 9 5 ). Inherited breast and ovarian cancer. Hum Mol Genet 4, 1811-1817. Wilson CJ, Chao DM, Imbalzano AN, Schnitzler GR, Kingston RE and Young RA (1 9 9 6 ). RNA polymerase II holoenzyme contains SWI/SNF regulators involved in chromatin remodeling. C e l l 84, 235-244. Xu X, Weaver Z, Linke SP, Li C, Gotay J, Wang XW, Harris CC, Ried T and Deng CX (1 9 9 9 ). Centrosome amplification and a defective G2-M cell cycle checkpoint induce genetic instability in BRCA1 exon 11 isoform-deficient cells. M o l . C e l l 3, 389-395. Zhang H, Somasundaram K, Peng Y, Tian H, Zhang H, Bi D, Weber BL and El-Deiry WS (1 9 9 8 ). BRCA1 physically associates with p53 and stimulates its transcriptional activity. O n c o g e n e 16, 1713-1721.

404

Gene Therapy and Molecular Biology Vol. 4, Page 405 Gene Ther Mol Biol Vol 4, 405-410. December 1999.

Coding end resolution in scid recombination-inducible cell lines Review Article

Matthew L. Brown, Sandra Lew and Yung Chang* Department of Microbiology, Program in Molecular and Cellular Biology, Arizona State University, Tempe, AZ 85287-2701 ______________________________________________________________________________________________________ * Correspondence: Yung Chang, Ph.D. Department of Microbiology, Program in Molecular and Cellular Biology, Arizona State University, Tempe, AZ 85287-2701, USA. Tel: 602-965-8672; Fax: 602-965-0098; E-mail:yung.chang@asu.edu Abbreviations: Ab-MLV, Abelson murine leukemia virus; DNA-PK, DNA-dependent protein kinase; DSB, double strand break; LM-PCR, ligation mediated polymerase chain reaction; scid, severe combined immune deficiency; V(D)J, variable (diversity) joining Key Words: RAG1, hairpin, scid, V(D)J recombination, chromosomal instability, Abelson murine leukemia virus, KU70/80, DNA-

dependent protein kinase Received: 17 April 1999; accepted: 17 May 1999

Summary VDJ recombination is the mechanism by which antigen receptor genes are assembled. The site-specific cleavage mediated by recombination activating gene (RAG1 and RAG2) proteins generates two types of broken DNA ends: blunt signal ends and hairpin coding ends. The standard joining of these ends to form signal joints and coding joints employs several proteins involved in double strand break (DSB) repair, including KU70/80, the catalytic subunit of DNAdependent protein kinase (DNA-PKcs), XRCC4 and ligase IV. The cells from severe combined immunodeficient (scid) mice are defective in resolving recombination coding ends due to a point mutation in the DNA-PKcs gene. To study the effect of the scid defect on coding end resolution, we have established recombination-inducible cell lines from scid mice. These cells, at the nonpermissive temperature, actively initiate recombination at the endogenous light chain loci and produce large amounts of hairpin coding ends. After returning to the permissive temperature, scid cells are capable of resolving these coding ends. However, unlike the coding end resolution in normal cells, which is a rapid and regulated process, the resolution of hairpin coding ends in scid cells is slow and error prone. The resulting coding joints contain extensive nucleotide deletions. In addition, the interlocus recombination products are found at much higher frequency in scid cells than in their normal cell counterparts. Our results suggest that functional DNA-PKcs may play an important role in facilitating effective V(D)J recombination and minimizing chromosomal instability.

lead to an arrest of lymphocyte development and ultimately to immunodeficiency. For example, the severe combined immunodeficient (scid) mouse bears a point mutation at the gene encoding the catalytic subunit of DNA-PK, and has a defect in resolving recombination coding ends (Danska et al., 1996; Araki et al., 1997). To understand the molecular mechanisms of the V(D)J recombination joining process, we made recombinationinducible cell lines from both scid and normal mice by transforming pre-B cells with the temperature-sensitive Abelson murine leukemia virus (ts-Ab-MLV), which could be manipulated in vitro (Chen et al.,1994; Chang and Brown,1999). The recombination activity can be induced by incubating cells at the nonpermissive temperature, which leads to site-specific cleavage at L-chain gene loci, both !

I. Introduction V(D)J recombination is a site-specific process that is unique to developing T and B lymphocytes. This process involves a site-specific cleavage and imprecise end joining (Gellert, 1992; Steen et al., 1996). The cleavage, mediated by recombination activation gene (RAG1 and RAG2) proteins, produce two types of recombination intermediates, blunt signal ends and hairpin coding ends (Schatz and Leu, 1996; van Gent et al., 1996). Joining of these ends is dependent on a general double-strand DNA break (DSB) repair system, which includes the Ku-heterodimer, DNAdependent protein kinase (DNA-PK), XRCC4 and ligase IV (Jackson and Jeggo, 1995; Chu, 1996; Jeggo et al., 1996; Frank et al., 1998). Genetic alteration of any of the proteins that participate in the V(D)J recombination process could 405

Brown et al: V(D)J recombination in scid mice and " (Klug et al., 1994; Liu et al., 1997). Thus, these recombination-inducible scid and normal counterparts provide us with a good model to study the regulation of recombination initiation, and to delineate various steps in the resolution of V(D)J recombination intermediates. In this study, we have investigated the resolution of coding ends in scid cells. Substantial amounts of recombination hairpin coding ends were found to accumulate in scid but not in scid heterozygous (s/+) cells after induction of recombination. By analyzing the kinetics of coding end resolution in both scid and s/+ cells, we found that such resolution is much delayed in scid cells relative to s/+ cells. The inability of scid cells to promptly resolve their recombination ends makes these ends vulnerable to genetic mutation. For example, extensive losses of nucleotides were found at the junction of coding joints. In addition, the aberrant end resolution in scid cells was also revealed by an increased level of interlocus recombination products. In contrast, the control s/+ cells resolve their coding ends rapidly and show very few aberrations in their joined products. These findings lead to the conclusion that functional DNA-PKcs plays an important role in facilitating effective V(D)J recombination and minimizing chromosomal instability.

significant amount of VJ!1 coding joints (Fig. 2). Extended incubation at 39°C or after shifting back to 33°C does not further increase the amount of coding joints (Fig. 2). Thus, the coding joint formation in s/+ cells is a very rapid process and is not dependent on incubation at 33°C. In contrast, no detectable coding joints were found in the scid-ts cells even after 48 hour-incubation at 39°C (Fig.1A). Interestingly, however, coding joints do gradually appear in scid-ts cells with an increased level upon shifting cells from the nonpermissive temperature to the permissive temperature for an increased time (Fig. 2). This study clearly demonstrates that incubation at 33°C following culture at 39°C is essential for scid-ts cells to join their coding ends. Therefore, the ability of scid- ts cells to join coding ends can be facilitated by culturing them at the permissive temperature after exposure to the higher levels of RAG at the nonpermissive temperature.

C. Temperature-dependent resolution of recombination hairpin coding ends The inability of scid-ts cells to form coding joints at the nonpermissive temperature could result from a dysfunction at several stages: hairpin nicking, end filling, or joining. To directly examine hairpin-structured coding ends, we carried out a two-dimensional (2-D) gel electrophoresis followed by Southern blot analysis as described by Roth et al (Roth et al., 1992). The 3.4-kb band was derived from restriction digestion (Xba-I) of the germline !-locus. Cleavage initiated at the J!1 locus should give rise to 1.7 kb J!1 coding ends (shown in Fig. 3A). On the 2-D gel electrophoresis, the open coding ends would be expected to run along the diagonal whereas the hairpin coding ends would run off the diagonal.

II. Results and Discussion A. Recombination initiation Great effort has been devoted to developing ts-AblMuLV transformed pre-B cell lines from scid and s/+ control mice. The scid and s/+ ts-Abl-MuLV transformants, referred to as scid-ts and s/+-ts cells, respectively, can be induced to express high levels of RAG1 and RAG2 by incubating the cells at the nonpermissive temperature of 39°C. It has been reported that ts-Ab-MLV transformants can be induced to rearrange both of their !- and "-chain genes (Chen, et al., 1994; Klug, et al., 1994; Liu, et al., 1997). To directly assess the recombination initiation at these gene loci in scid-ts and s/+-ts cells, we examined the recombination signal ends of J!, V "1 and J"1 using a ligation-mediated PCR (LM-PCR) assay as described previously (Roth et al., 1993). It is clear from Fig.1 that at 39°C; the level of signal ends generated from three specific gene loci is comparable between scid-ts and s/+-ts cells. This indicates that both cell types attempt similar levels of recombination cleavages. Inferred from this analysis, we expect that a comparable level of coding ends is generated in these two cell types, as well. Therefore, the resolution and various joining products of the newly generated coding ends can be directly compared between scid and s/+- ts cells.

B. Coding joint formation

Figure 1. Recombination signal ends (SE) in both ! and " gene loci detected by Ligation-mediated PCR. The recombination cleavages are initiated upon shifting cells from the nonpermissive temperature (39°C) to the permissive temperature (33°C).

We first analyzed and compared coding joint formation in scid-ts and s/+-ts cells by PCR assay. The cells were cultured at 39°C for 48 hours. The s/+ ts-cells contained a 406

Gene Therapy and Molecular Biology Vol. 4, Page 407 The scid sample incubated at 39°C for 48 hours showed a 1.7kb band off the diagonal (Fig. 3B, middle panel), a signature of hairpin ends. The same cells cultured at 39°C for 48 hours followed by 24 hours at 33°C showed the 1.7kb band right on the diagonal (Fig. 2B, lower right panel), i.e., open coding ends. Thus, incubation at 39°C followed by 33°C seems to allow opening of the scid hairpin coding ends. These data clearly show that under the nonpermissive temperature, scid-ts cells are defective in converting the newly generated hairpin-coding ends to intact open ends. This conversion might be accomplished after the cell culture is shifted to the permissive temperature, which is concurrent with the formation of coding joints. In contrast, under the same culture conditions, the DNA isolated from the s/+-ts cells did not show an obvious 1.7 kb band either on or off the diagonal, which may reflect a rapid joining of coding ends (Chang and Brown, 1999). Similar conclusion was also confirmed by a more sensitive assay, a modified LMPCR (described below). These data clearly indicates that the ability of scid-ts cells to resolve hairpin-coding ends is conditional

Figure 2: Analysis of coding Joint (CJ) formation in both scid/+ and scid/scid cells. (A) Time course analysis for CJ formation. After the temperature sensitive cells were cultured under their respective conditions (H indicates non-permissive temperature, L indicates permissive temperature), the DNA was isolated and subjected to a PCR assay using a V! primer and a J!2 primer, along with actin controls. (B) CJ formation with extended culture. The J!2 primer was used to follow the formation of CJs over a four day period under various conditions as described in the text.

Figure 3: Hairpin coding end detection by two-dimensional electrophoresis. (A) Diagrammatic representation of J!-gene rearrangement and predicted sizes after XbaI digestion. (B) scid DNA samples were digested with XbaI and electrophoresed in the native direction, then shifted 90° clockwise and continued under denaturing conditions. Expected sizes are 3.4 kb for the germ-line ! locus, 1.65 kb for the J!1 CE, and 1.3 for the J!2 CE. The denatured hairpin CE (H, 48 hrs) departed from the diagonal line. In contrast, the opened CE (H, 48 hrs ! L, 24 hrs) remained within the diagonal line.

407

Brown et al: V(D)J recombination in scid mice To further delineate the process in end resolution, we analyzed various structures of coding ends: closed hairpin ends, recessive and blunt opened ends by a modified LMPCR. The blunt open ends can be directly amplified by LMPCR while the staggered open ends are first processed by T4 DNA polymerase before being subjected to the ligation assay. The hairpin ends are pretreated with both Mung Bean nuclease (MBN) and T4 DNA polymerase then followed by LM-PCR. To minimize any DNA damage caused by the routine extraction procedure, DNA samples were prepared in an agarose plug, as described by Schlissel (1998). Fig. 4 shows that large amounts of blunt open ends are present in cells cultured at 39°C followed by incubation at 33°C (sample 2-1, Lane 6). Pretreatment with T4 DNA polymerase or together with MBN did not significantly increase the detection of these ends (Lane 2 and Lane 4). This indicates that majority of the coding ends are in the form of open blunt ends. However, the size of these ends seems smaller than the full length of artificially nicked hairpin ends (compare lane 6 and lane 2), which may reflect nucleotide loss at the ends. In contrast, very few blunt open coding ends were detected in cells that were cultured at 39°C for three days (sample 3-0, lane 5). Instead, the majority of the coding ends remained in the hairpin structure (Fig. 4, top panel, lane 1). The amount of PCR products was also increased if the DNA sample had been pre-treated with T4 DNA polymerase, indicating the presence of recessive open coding ends (Lane 3). These ends migrate much faster than the other ends, reflecting various deletions of nucleotides. Thus, even though some nicking of hairpin ends may occur at the nonpermissive temperature (3-0), this event is much less frequent and more aberrant than the one observed in the cells returning from 39°C to 33°C (compare lane 3 and lane 4). Taken together, our data indicate that the ability of scid-ts cells to resolve hairpin coding ends is conditional as well as aberrant. Extensive deletions were found at the blunt open ends and the recovered coding joints (data not shown), which is consistent with our previous finding of scid-like coding joints. Thus, the end resolution in scid-ts cells may be fundamentally different from the control s/+-ts cells. In normal cells, functional DNA-PKcs may act to protect the ends while recruiting some nucleases and other DSB repair proteins for limited nicking and rapid joining. In scid-ts cells, on the other hand, the temperature-dependent resolution of the ends may reflect different accessibility of the ends to the nucleases and repair proteins under different culture temperatures. For example, the ends may be blocked from nicking at 39°C and exposed to non-regulated nicking after returning the cells from 39°C to 33°C. Therefore, the resolution of coding ends in scid cells proceeds through a DNA-PKcs-independent default pathway, which is inefficient and prone to error.

Figure 4. Coding end resolution in scid (s/s) temperature sensitive (ts) Abl-MLV cell line. Scid-ts cells were shifted to 39°C for either 3 days (3-0) or 2 days followed by I day at 33°C (2-1). DNA was prepared in agarose plugs as described in materials and methods. Coding ends (CE) were analyzed by ligation mediatedPCR (LM-PCR) at the J!1 locus. Mung Bean nuclease (MBN) and T4 DNA Polymerase (T4 Pol) were used to flush the coding ends. Coding joint (CJ) formation was assayed by PCR using a V! primer and a J!2 primer. Amplification of the actin gene was used as a control for the amount of DNA.

D. Chromosomal translocation: presence of interlocus recombination products It is conceivable that the unresolved recombination coding ends, due to an inefficient pathway for end resolution, could be vulnerable targets for gene mutations such as deletion, insertion and translocation. To directly examine the possibility of chromosomal translocations, we have analyzed the recombination events between two gene segments residing on different chromosomes, e.g., the ! and " gene loci. As shown in Fig. 1, both of these gene loci could be induced to undergo recombination cleavages. Thus, it is possible for a rearrangement to occur between these two gene loci. Fig. 5A is a schematic diagram that illustrates the detection of an interlocus rearrangement. Specifically, the coding joints between V"1 and J!1 gene segments were analyzed by PCR amplification, and revealed by hybridization with the V" probe. Fig. 5B is one representative example of the analyses of interlocus recombination. Although substantial amounts of V"J" coding joints are present in s/+-ts cells, no V"J! coding joints were detected. This finding is consistent with the previous report by Bailey and Rosenberg (1997) in which the V"J! joints were estimated to form at a frequency about 1,000-fold less than that of cis-rearrangement. 408

Gene Therapy and Molecular Biology Vol. 4, Page 409

Figure 5. Analysis of light chain interlocus rearrangements in scid and s/+ ts-Abl-MLV transformants. (A). Mechanism of interlocus as compared to intralocus rearrangements. Interlocus rearrangements consisted of V" joined to J! coding regions and signal joints made up of RSS from both " and ! loci. Primers are represented by small arrows. RSS are represented by triangles and coding regions by rectangles. (B) PCR and Southern blot analysis of standard V"J" coding joints, V"J! coding joints, and V"J! signal joints. Actin served as a control.

(unpublished observation), confirming the rearrangement between ! and ". The junctions of these joining products also contained large deletions (Table I), similar to the abnormality found in the coding joints made by scid-ts cells.

In contrast, V"J! coding joints were readily detected in the scid-ts cells that were cultured at 39째C followed by incubation at 33째C. The same PCR products detected by a V" probe were also revealed by a J!-specific probe

409

Brown et al: V(D)J recombination in scid mice Schatz, D. G., et al. (1996). rag-1 and rag-2: biochemistry and protein interactions. Curr Top Microbiol Immunol 217: 1130. Schlissel, M. (1998). Structure of nonhairpin coding-end DNA breaks in cells undergoing V(D)J recombination. Mol Cell Biol 18: 2029-2037. Steen, S. B., et al. (1996). Double-strand breaks, DNA hairpins, and the mechanism of V(D)J recombination. Curr Top Microbiol Immunol 217: 61-77. van Gent, D. C., et al. (1996). Initiation of V(D)J recomination in a cell-free system by RAG1 and RAG2 proteins. Curr Top Microbiol Immunol 217: 1-10.

These results indicate that recombination intermediates, including those made from different chromosomes, are in close proximity and accessible for joining to each other. It is possible that these ends are present in a common compartment or that they are free and flexible for engaging in various types of association. The lack of interlocus joining products in s/+-ts cells suggests that free association of coding ends is prohibited by the functional DNA-PKcs. Therefore, our finding suggests that DNA-PKcs plays an important role in facilitating the resolution of recombination intermediates, one which ultimately prevents the uncontrolled association of the ends.

References Araki, R., et al. (1997). Nonsense mutation at Tyr-4046 in the DNA-dependent protein kinase catalytic subunit of severe combined immune deficiency mice. Proc. Natl. Acad. Sci. USA 94: 2438-2443. Bailey, S. N and Rosenberg, N. (1997). Assessing the pathogenic potential of the V(D)J recombinase by interlocus immunoglobulin light-chain gene rearrangement. Mol Cell Biol. 17: 887-894. Chang, Y and Brown, M.L. (1999). Formation of coding joints in V(D)J recombination-inducible severe combined immune deficient pre-B cell lines. Proc. Natl. Acad. Sci. USA 96: 191196. Chen, Y. Y., et al. (1994). An active v-abl protein tryrosine kinase blocks immunoglobulin light-chain gene rearrangement. Genes Dev 8: 688-697. Chu, G. (1996). Role of the Ku autoantigen in V(D)J recombination and double-strand break repair. Curr Top Microbiol Immunol 217: 113-132. Danska, J. S., et al. (1996). Biochemical and genetic defects in the DNA-dependent protein kinase in murine scid lymphocytes. Mol Cell Biol. 16: 5507-5517. Frank, K., et al. (1998). Late embryonic lethality and impaired VDJ recombination in mice lacking DNA ligase IV. Nature 396: 173-177. Gellert, M. (1992). V(D)J recombination gets a break. Trends Genet 8: 408-412. Jackson, S. P., et al. (1995). DNA double-strand break repair and V(D)J recombination: involvement of DNA-PK. Trends Biochem Sci 20: 412-415. Jeggo, P. A., et al. (1996). Identification of the catalytic subunit of DNA dependent protein kinase as the product of the mouse scid gene. Curr Top Microbiol Immunol 217: 79-90. Klug, C. A., et al. (1994). The v-abl tyrosine kinase negatively regulates NF-!B/Rel factors and blocks ! gene transcription in pre-B lymphocytes. Genes Dev 8: 678-687. Liu, D., et al. (1997). ! and " rearrangement occur simultaneously in transformed pre-B cells. J Immunol 159: 6061-6069. Roth, D. B., et al. (1992). V(D)J recombination: Broken DNA molecules with covalently sealed (hairpin) coding ends in scid mouse thymocytes. Cell 70: 983-991. Roth, D. B., et al. (1993). Characterization of broken DNA molecules associated with V(D)J recombination. Proc. Natl. Acad. Sci. USA 90: 10788-10792.

410

Gene Therapy and Molecular Biology Vol 4, page 411 Gene Ther Mol Biol Vol 4, 411-416. December 1999.

Pak protein kinases as mediators of Ras signaling and cell transformation. A place for Pak on the Ras MAP: More than just another JNK bond Review Article

Yi Tang, Albert Chen, Ya Zhuo, Qi Wang, Albert Pahk and Jeffrey Field Department of Pharmacology, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, __________________________________________________________________________________________________ Correspondence: Jeffrey Field Ph.D., Tel: (215) 898-1912; Fax (215) 573-2236; E-mail: field@pharm.med.upenn.edu Key words: Raf, signal transduction, PI3-kinase, cancer Received: 1 November 1999; accepted: 8 November 1999

Summary Ras plays a key role in regulating cellular proliferation, differentiation, and transformation. Raf is the major effector o f Ras i n the Ras>Raf>Mek>Erk (MAPK) cascade. A second effector i s phosphoinositide 3-OH kinase (PI 3-kinase) which synthesizes several lipid second messengers that activate small G proteins such as Rac and Cdc42. Rac also has multiple effectors, one of which is the serine threonine kinase Pak (p65Pak). Here we review studies documenting a novel Ras signal through PI 3-kinase and Rac/Cdc42 t o Pak. The signal appears essential for maintaining cell transformation and Erk activation.

I. Introduction Ras is one of the most commonly mutated oncogenes and is found activated in 20-30% of tumors (Lowy and Willumsen, 1993). Ras encodes a small G protein that binds GTP and GDP and possesses an intrinsic GTPase activity. In its oncogenic form Ras acquires a point mutation that inactivates the GTPase activity and causes it to be locked into its activated GTP-bound state. Normally, Ras is activated by growth factor receptors through its guanine nucleotide exchange factors (Egan and Weinberg, 1993). The major oncogenic signal from Ras utilizes the serine threonine kinase, Raf, as the effector (Van Aelst et al., 1993; Vojtek et al., 1993). GTP bound Ras binds and activates Raf and simultaneously recruits it to the membrane. Upon activation Raf phosphorylates and activates another kinase Mek which in turn activates ERK (MAPK). This Ras>Raf>Mek>ERK signal is usually referred to as the MAP kinase cascade (Marshall, 1995). In recent years Ras has also been shown to bind other effectors, besides Raf, and activate other signaling pathways that cooperate with the Raf> ERK signal (White et al., 1995). The other pathways are not as well defined as

the Raf cascade. The three effectors that have been most widely studied are Rin, RalGDS and PI 3-kinase (Afar et al., 1997; Peterson et al., 1996; Rodriguez-Viciana et al., 1997). Each binds RasGTP and can, in some experimental systems, cooperate with partially activated Raf mutants to transform cells (Rin cooperates with Abl). Further evidence of the importance of these effectors in Ras signaling comes from new Ras point mutants, known as effector mutants, that bind and activate only subsets of Ras effectors (White et al., 1995). RasV12S35 binds and activates Raf, Ras V12G37 binds and activates RalGDS and Rin1 and RasV12C40 binds and activates PI 3-kinase (Joneson et al., 1996; Rodriguez-Viciana et al., 1997; White et al., 1995; White et al., 1996). These mutants are deficient in signaling when tested alone, but cooperate when introduced into cells together. Signals from Ras through the alternate effectors utilize other small G proteins. Ral GDS uses Ral, and PI 3-kinase uses the small G protein Rac (RodriguezViciana et al., 1997; White et al., 1996). Rho and two related proteins, Rac and Cdc42, are members of the Rho family of small G proteins. These proteins are about 50% identical to Ras and regulate the actin cytoskeleton. Rho induces stress fibers and focal adhesions (Ridley and Hall, 1992), Rac induces accumulation of actin rich ruffles or lamellipodia at the periphery of cells (Ridley et 411

Tang et al: Ras signaling through Pak al., 1992) and Cdc42 induces microspikes or filopodia (Nobes and Hall, 1995). Each Rho family member also activates a kinase cascade that leads to transcriptional activation similar to the MAP kinase cascade but not as well defined. Rho activates the ternary complex factors (TCF), and Rac and Cdc42 activate the Jun N-terminal kinase cascade JNK(SAPK) (Coso et al., 1995; Frost et al., 1996; Hill et al., 1995; Minden et al., 1995). Dominant negative mutants of Rac, Rho and Cdc42 each inhibit Ras transformation and activated mutants cooperate with Raf to transform cells (Khosravi-Far et al., 1995; Qiu et al., 1995; Qiu et al., 1997; Qiu et al., 1995). These observations suggest that the signals through the Rho family of small G proteins play essential roles in Ras transformation. The signals through Rac are directly connected to Ras (Bar-Sagi and Feramisco, 1986; Ridley et al., 1992). This is because Ras and Rac both cause membrane ruffling when microinjected into cells and a dominant negative Rac mutant inhibits Ras induced ruffling. The signal from Ras to Rac is likely to be mediated by PI 3-kinase since RasV12C40, which activates PI 3-kinase, and activated mutants of PI 3-kinase both induce ruffles (Joneson et al., 1996; Rodriguez-Viciana et al., 1997). The mechanism that PI 3-kinase uses to activate Rac probably involves stimulation of Rac GEFs by PI 3-kinase products such as phosphatidylinositol-3,4,5-triphosphate (Han et al., 1998; Nimnual et al., 1998). The immediate effector downstream of Rac in Ras signal transduction to both the JNK and actin pathways has remained elusive. The first candidate to be isolated was the serine threonine kinase p65Pak (Manser et al., 1994). Pak was isolated because it binds both Rac and Cdc42 in their GTP bound forms. Pak is homologous to Ste20, a protein kinase in the yeast S. cerevisiae regulated by Cdc42 (Lim et al., 1996; Sells and Chernoff, 1997). Some of the activities of Pak resemble those of Ras and Rac. For example, microinjection of Pak into some cells causes ruffling and breaks up stress fibers and membrane targeting of Pak in PC12 cells induces extension of neurites (Daniels et al., 1998; Manser et al., 1997; Sells et al., 1997). Although microinjection of Pak can cause membrane ruffling, ruffling does not require Pak kinase activity (Sells et al., 1997). In addition, RacV12H40, an effector mutant that does not bind to Pak, still cooperates with Raf to transform cells and causes membrane ruffling when it is microinjected (Joneson et al., 1996; Lamarche et al., 1996; Westwick et al., 1997). These studies suggest that there may not be a role for Pak in Ras transformation or signaling and even question the existence of any direct signals from Ras to Pak. This report reviews work linking Ras directly to Pak through a PI3-kinase dependent pathway and discusses the role of this link in cell transformation (Tang et al., 1997; Tang et al., 1998; Tang et al., 1999).

II. Evidence linking Pak to cell transformation Signal transduction studies routinely use reagents to activate or inhibit specific candidate components of the Ras signaling pathway. Such studies suffer from the limitation that high levels of expression may lead to non-physiological signals. To circumvent this problem, it is necessary to use multiple strategies to link signaling molecules to Ras. A general strategy has emerged in the field that is based on three types of observations. First, Ras must signal to the candidate. Second, Ras signals must be potentiated by activating or overexpressing the candidates. Third, Ras signals must be blocked by inhibiting the candidate molecule. Finally, the biological assays must be backed up with biochemical assays to known Ras signals. We shall now discuss how this has been applied to place Pak on the Ras signaling pathway. The observations are presented in reverse order for historical reasons.

III. Kinase deficient Pak inhibits Ras transformation The first observation linking Ras to Pak was the demonstration that expression of kinase deficient mutants of Pak inhibit Ras transformation in Rat-1 fibroblasts (Tang et al., 1997). This type of mutant is routinely used to study protein kinases because they behave as dominant negative mutants. The assumption is that the kinase defective mutants specifically bind to key substrates but fail to phosphorylate them while preventing the endogenous kinase from phosphorylating them. For Pak, a single amino acid substitution at amino acid number 299 (K299R), a key residue in the catalytic domain, inactivates the kinase activity (Sells et al., 1997; Zhang et al., 1995). Since Rac is required for Ras transformation, the trivial explanation that the mutants were sequestering Rac was addressed by mutating the Rac/Cdc42 binding site in the kinase deficient Pak. The mutant, Pak1L83,L86,R299, has substitutions of leucines for conserved histidines at positions 83 and 86, along with the original R299 in the kinase domain. This mutation is as effective at inhibiting Ras transformation as the original mutant, which rules out the sequestering model. Hence, kinase deficient Pak mutants are potent inhibitors of Ras signaling, even when they cannot bind Rac and Cdc42. Inhibition was specific for Ras since the Pak mutants did not inhibit v-Raf. Interestingly, almost no Ras inhibition was observed in NIH 3T3 cells, but was observed in other cells (Tang et al., 1998).

IV. Pak can be uncoupled from JNK but not MAPK signaling The major signaling pathway downstream of Ras is to the MAP kinase Erk, while the major signal downstream of

412

Gene Therapy and Molecular Biology Vol 4, page 413

Ras

PI 3 Kinase

Raf

Akt Kinase Rac & Cdc42

Mek

Survival Pak Actin

MAPK Erk JNK Proliferation

Figure 1 Model of Ras signaling through Pak to ERK. Although Pak can activate JNK, Ras does not appear to utilize Pak for this signal.

Rac is to the related kinase JNK. As expected, kinase deficient Pak, Pak1R299, inhibits JNK activation by Ras and by Rac (~75% ). However, no inhibition was observed with the Pak1 L83,L86,R299 construct which cannot sequester Rac or Cdc42. Since both of these mutants inhibit Ras transformation, JNK inhibition is not obligatory for Pak mutants to inhibit Ras transformation. However, when the analogous experiment was performed to measure Erk signaling, both of the kinase deficient Pak mutants inhibited Ras signals to Erk. These observations suggest that the dominant negative Pak1 mutants may inhibit Ras transformation by interfering with the MAP/ERK kinase cascade and not the JNK cascade. Thus, the third criteria is met--in both biochemical and biological assays, kinase deficient Pak mutants inhibit Ras.

V. Pak potentiates signals to Erk The studies described above suggest that Pak mediates signals to ERK to sustain cell transformation. Several laboratories have independently observed that Pak can stimulate Erk (Frost et al., 1997; King et al., 1998; Lu et al., 1997; Tang et al., 1999). The first was the study from Mayer and his colleagues who found that targeting Pak to

the membrane by fusion to a localization signal stimulated Erk (Lu et al., 1997). Two mechanisms have since emerged to trace the signals from Pak to Erk. (1) Pak was shown to phosphorylate Mek in vitro and in cell culture studies. In this study a kinase deficient Pak mutant that failed to bind Rac/Cdc42 also inhibited signals to Erk from several growth factors (Frost et al., 1997). (2) Pak was also shown to phosphorylate Raf at a novel residue, 338. Moreover, mutation of residue 338 prevented Raf from signaling to Erk and transforming cells (King et al., 1998). All four of the studies have in common the observation that Pak alone is not sufficient for Erk activation but can potentiate signals to Erk from components of the Ras signaling pathways. These data demonstrate that Pak may be necessary for Ras signaling. However, Pak is clearly not sufficient for Ras signaling since Pak1L83, L86 an hyperactive kinase mutant does not transform most cells nor does it cooperate with Ras or Raf to transform cells (Tang et al., 1997). The one exception is a cell line that is hypersensitive to Ras transformation because it has mutations that activate MEK. In this cell line, Pak cooperates with Raf to transform the cells (Tang et al., 1999). However, as discussed above evidence from many labs demonstrate that Pak can promote signals to Erk. This constitutes the second

413

Gene Therapy and Molecular Biology Vol 4, page 414 arm of the case for Pak as a key modulator of Ras signaling--potentiation signals to Ras targets.

VI. Ras activates Pak The third arm linking Ras to Pak comes from a study showing that Ras activates Pak (Tang et al., 1999). Ras activation was observed in transfection assays of Rat-1 fibroblasts. Activation was comparable to the levels observed by Rac. Activated PI 3-kinase, but not Raf also activated Pak suggesting that the signal was mediated by PI 3-kinase. Furthermore, the PI 3-kinase specific inhibitor LY294002 inhibited both Ras and PI3-kinase activation of Pak with similar dose responses. Dominant negative Rac and dominant negative Cdc42 both inhibited Ras and PI3kinase activation. Thus a Ras>PI3-kinase>Rac/Cdc42 signal was traced. The mechanism of action of Rac and Cdc42 dominant negative mutants both involve the sequestering of exchange factors. Many of these exchange factors can activate both Rac and Cdc42, so the dominant negative mutant results do not necessarily distinguish between them. Interestingly, in a remarkable synergy between the biological responses and biochemical signaling, Ras could not activate Pak in NIH 3T3 cells. The differences between the two cell lines and their signals from Ras to Pak lie downstream of PI3-kinase since neither Ras nor PI 3-kinase could activate Pak in NIH 3T3 cells.

VII. A convergence of signals from small G proteins on Pak The studies with Ras demonstrated that it can activate Pak directly. Further studies suggested that Pak mediates signals from other G proteins through indirect mechanisms (Tang et al., 1999). First, transformation by Rac, Rho, Ras and several Rac and Ras effector mutants were all inhibited by dominant negative Pak mutants. Second, cooperative ERK activation by all three GTPases was inhibited by the Pak dominant negative mutants. Third, all combinations of Ras, Rho and Rac mutants that yielded high efficiency transformations also activated Pak. Finally, in the presence of a partially activated Raf all small G proteins, including Rho, could activate Pak. It should be noted that these correlations suggest that Pak activation may be necessary for high efficiency transformation but clearly Pak activation is not sufficient for transformation since RasV12C40, RacV12, RacV12L37 and Pak1L83,L86, all of which activate the kinase activity of Pak, transform poorly when tested individually. This study suggested that multiple signals from several small G proteins utilize Pak even when they do not signal to Pak directly (Tang et al., 1999). A similar conclusion was reached by the study of Frost et. al. which found that Rho and Rac were signaling through Pak to Erk (Frost et al., 1997).

VIII. A Role for Pak in Ras signaling beyond Rat-1 fibroblasts While the studies with Rat-1 fibroblasts provide a useful experimental system to study the role of Pak in Ras signaling, it is not appropriate for tumor cells. To develop cell systems for studying Pak more appropriate for a specific tumor, Tang et. al. developed several cell systems relevant to Neurofibromatosis. Neurofibromatosis type 1 (NF1), a common autosomal dominant disorder caused by loss of the NF1 gene, is characterized clinically by neurofibromas, and more rarely by neurofibrosarcomas (Gutmann et al., 1997; Viskochil et al., 1990; Wallace et al., 1990). Abnormalities in Schwann cells are thought to be responsible for both types of tumors (Kim et al., 1997; Sheela et al., 1990). While the basis of the Schwann cell as the affected cell is not well understood, the underlying mechanism of transformation for both tumors is through the Ras signaling pathway. Neurofibromin, the protein encoded by NF1, possesses an intrinsic GTPase accelerating activity for the Ras protooncogene. Through this activity neurofibromin acts as a negative regulator of Ras. Hence, loss of neurofibromin causes elevated levels of activated Ras which leads to hyperactivation of downstream signals (Ballester et al., 1990; Basu et al., 1992; DeClue et al., 1992; Martin et al., 1990; Xu et al., 1990). First, transfection assays demonstrated the central role of Pak in Ras transformation and signaling to Erk in Rat Schwann cells. Then, kinase deficient Pak mutants were found to revert a transformed cell from a patient with NF1. Thus, in at least one tumor cell, Pak plays a central role in Ras signaling and transformation (Tang et al., 1998). The observation that Rho, Rac and Cdc42 are all required for Ras transformation has prompted several searches for key downstream effectors. Pak emerged as one candidate because it was the first protein kinase found to bind Rac/Cdc42 and it was homologous to Ste20, a Cdc42 effector in yeast (Manser et al., 1994). Subsequent studies with effector mutants suggested that Pak does not play a role in transformation. However, by the three criteria discussed earlier, Pak has been linked to Ras. First, Ras signals directly to Pak through PI-3 kinase. Second, several laboratories have found that Pak is required to sustain signals to Erk. Third, Ras transformation and Erk signaling can be inhibited by dominant negative Pak mutants. Hence, Pak becomes part of a growing list of Ras downstream signaling proteins, such as Raf, PI 3-kinase and Mek that may be targets for novel anti-neoplastic drugs.

Acknowledgments We thank members of the lab and Amita Sehgal for helpful discussions and for comments on the manuscript. Work in the lab is supported by grants to JF from the NIH (GM48241), the Lucille P. Markey Charitable Trust, the

414

Tang et al: Ras signaling through Pak American Cancer Society and the Neurofibromatosis Foundation. Note added in proof We have recently determined that the Akt protooncogene is a key intermediate between Ras and Pak suggesting that Pak may transduce cell survival signals from Ras [Tang Y, Zhou H, Chen A, Pittman RN, and Field J (2 0 0 0 ) The Akt ptoto-oncogene links Ras to Pak and cell survival signals. J B i o l C h e m 275, 9106].

References Afar, D. E. H., Han, L., McLaughlin, J., Wong, S., Dhaka, A., Parmar, K., Rosenberg, N., Witte, O. N., and Colicelli, J. (1 9 9 7 ). Regulation of the oncogenic activity of BCRABL by a tightly bound substrate protein RIN1. Immunity 6, 773-782. Ballester, R., Marchuk, D., Boguski, M., Saulino, A., Letcher, R., Wigler, M., and Collins, F. (1 9 9 0 ). The NF1 locus encodes a protein functionally related to mammalian GAP and yeast IRA proteins. C e l l 63 , 851-859. Bar-Sagi, D., and Feramisco, J. R. (1 9 8 6 ). Induction of Membrane Ruffling and Fluid-Phase Pinocytosis in Quiescent fibroblasts by ras Proteins. S c i e n c e 233, 1061-1068. Basu, T. N., Gutmann, D. H., Fletcher, J. A., Glover, T. W., Collins, F. S., and Downward, J. (1 9 9 2 ). Aberrant regulation of ras proteins in malignant tumor cells from type 1 neurofibromatosis patients. Nature 356, 713-715. Coso, O. A., Chiariello, M., Yu, J.-C., Teramoto, H., Crespo, P., Xu, N., Miki, T., and Gutkind, J. S. (1 9 9 5 ). The Small GTP-Binding Proteins Rac1 and Cdc42 Regulate the Activity of the JNK/SAPK Signaling Pathway. C e l l 81 , 1137-1146. Daniels, H. R., Hall, P. S., and Bokoch, G. M. (1 9 9 8 ). Membrane targeting of p21-activated kinase 1 (PAK1) induces neurite outgrowth from PC12 cells. EMBO J. 17 , 754-764. DeClue, J. E., Papageorge, A. G., Fletcher, J. A., Diehl, S. R., Ratner, N., Vass, W. C., and Lowy, D. R. (1 9 9 2 ). Abnormal regulation of mammalian p21ras contributes to malignant tumor growth in von Recklinghausen (type 1) neurofibromatosis. C e l l 69 , 265-273. Egan, S. E., and Weinberg, R. A. (1 9 9 3 ). The pathway to signal achievement. Nature 365, 781-783. Frost, J. A., Steen, H., Shapiro, P., Lewis, T., Ahn, N., Shaw, P. E., and Cobb, M. H. (1 9 9 7 ). Cross-cascade activation of ERKs and ternary complex factors by Rho family proteins. EMBO J. 16 , 6426-6438. Frost, J. A., Xu, S., Hutchison, M. R., Marcus, S., and Cobb, M. H. (1 9 9 6 ). Actions of Rho family small G proteins and p21 activated protein kinases on Mitogen-Activated Protein Kinase family members. M o l . C e l l . B i o l . 16 , 3707-3713. Gutmann, D. H., Aylsworth, A., Carey, J. C., Korf, B., Marks, J., Pyeritz, R. E., Rubenstein, A., and Viskochil, D. (1 9 9 7 ). The diagnostic evaluation and multidisciplinary management of Neurofibromatosis 1 and Neurofibromatosis 2. JAMA 278, 51-57.

Han, J., Luby-Pelps, K., Das, B., Shu, X., Xia, Y., Mosteller, R. D., Krishna, U. M., Falk, J. R., White, M. A., and Broek, D. (1 9 9 8 ). Role for substrates and products of PI3-kinase in regulating activation of Rac-related guanosine triphopsphatases by Vav. S c i e n c e 279, 558-560. Hill, C. S., Wynne, J., and Treisman, R. (1 9 9 5 ). The Rho family GTPases RhoA, Rac1, and CDC42Hs regulate transcriptional activation by SRF. C e l l 81 , 1159-1170. Joneson, T., McDonough, M., Bar-Sagi, D., and Van Aelst, L. (1 9 9 6 ). RAC regulation of actin polymerization and proliferation by a pathway distinct from Jun kinase. S c i e n c e 274, 1374-1376. Joneson, T., White, M. A., Wigler, M. H., and Bar-Sagi, D. (1 9 9 6 ). Stimulation of membrane ruffling and MAP kinase activation by distinct effectors of RAS. S c i e n c e 271, 810812. Khosravi-Far, R., Solski, P. A., Clark, G. J., Kinch, M. S., and Der, C. J. (1 9 9 5 ). Activation of Rac1, RhoA, and mitogenactivated protein kinases is required for Ras transformation. M o l . C e l l . B i o l . 15 , 6443-6453. Kim, H. A., Ling, B., and Ratner, N. (1 9 9 7 ). NF1-deficient mouse Schwann cells are angiogenic and invasive and can be induced to hyperproliferate: reversion of some phenotypes by an inhibitor of farnesyl protein transferase. M o l . C e l l . B i o l . 17 , 862-872. King, A. J., Sun, H., Diaz, B., Barnard, D., Miao, W., Bagrodia, S., and Marshall, M. S. (1 9 9 8 ). The protein kinase Pak3 positively regulates Raf-1 activity through phosphorylation of serine 338. Nature 396, 180-183. Lamarche, N., Tapon, N., Stowers, L., Burbelo, P. D., Aspenstrom, P., Bridges, T., Chant, J., and Hall, A. (1 9 9 6 ). Rac and Cdc42 induce actin polymerization and G1 cell cycle progression independently of p65 PAK and the JNK/SAPK MAP kinase cascade. C e l l 67 , 519-529. Lim, L., Manser, E., Leung, T., and Hall, C. (1 9 9 6 ). Regulation of phosphorylation by p21 GTPases. The p21 Ras-related Rho subfamily and its role in phosphorylation signalling pathways. Eur. J. Biochem . 242, 171-185. Lowy, D. R., and Willumsen, B. R. (1 9 9 3 ). Function and regulation of ras. A n n u . R e v . B i o c h e m . 62 , 851-891. Lu, W., Katz, S., Gupta, R., and Mayer, B. J. (1 9 9 7 ). Activation of Pak by membrane localization mediated by an SH3 domain from the adaptor protein Nck. Curr. Biol. 7, 85-94. Manser, E., Huang, H.-Y., Loo, T.-H., Chen, X.-Q., Dong, J.-M., Leung, T., and Lim, L. (1 9 9 7 ). Expression of constitutively active a-Pak reveals effects of the kinase on actin and focal complexes. M o l . C e l l . B i o l . 17 , 1129-1143. Manser, E., Leung, T., Salihuddin, H., Zhao, Z.-S., and Lim, L. (1 9 9 4 ). A brain serine/threonine protein kinase activated by Cdc42 and Rac1. Nature 367, 40-46. Marshall, C. J. (1 9 9 5 ). Specificity of receptor tyrosine kinase signaling: Transient versus sustained extracellular signalregulated kinase activation. C e l l 80 , 179-185. Martin, G. A., Viskochil, D., Bollag, G., McCabe, P. C., Crosier, W. J., Haubruck, H., L., C., Clark, R., O'Connell, P., Cawthon, R. M., Innis, M. A., and McCormick, F. (1 9 9 0 ). The GAP-related domain of the neurofibromatosis type 1 gene product interacts with ras p21. C e l l 63 , 843-849. Minden, A., Lin, A., Claret, F.-X., Abo, A., and Karin, M. (1 9 9 5 ). Selective activation of the JNK signaling cascade and c-Jun transcriptional activity by the small GTPases Rac and Cdc42Hs. C e l l 81 , 1147-1157.

415

Gene Therapy and Molecular Biology Vol 4, page 416 Nimnual, A. S., Yatsula, B. A., and Bar-Sagi, D. (1 9 9 8 ). Coupling of Ras and Rac guanosine triphosphatases through the Ras exchanger Sos. S c i e n c e 279, 560-563. Nobes, C. D., and Hall, A. (1 9 9 5 ). Rho, Rac, and Cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. C e l l 81 , 53-62. Peterson, S. N., Trabalzini, L., Brtva, T., Fischer, T., Altschuler, D. L., Martelli, P., Lapetina, E. G., Der, C. J., and White, G. C. (1 9 9 6 ). Identification of a novel RalGDS-related protein as a candidate effector for Ras and Rap1. J . B i o l . C h e m . 271, 29903-29908. Qiu, R., Chen, J., McCormick, F., and Symons, M. (1 9 9 5 ). A role for Rho in Ras transformation. P r o c . N a t l . A c a d . S c i . USA 92 , 11781-11785. Qiu, R.-G., Abo, A., McCormick, F., and Symons, M. (1 9 9 7 ). Cdc42 regulates anchorage-independent growth and is necessary for Ras transformation. M o l . C e l l . B i o l . 17 , 3449-3458. Qiu, R.-G., Chen, J., Kirn, D., McCormick, F., and Symons, M. (1 9 9 5 ). An essential role for Rac in Ras transformation. Nature 374, 457-459. Ridley, A. J., and Hall, A. (1 9 9 2 ). The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. C e l l 70 , 389-399. Ridley, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D., and Hall, A. (1 9 9 2 ). The small GTP-binding protein Rac regulates growth factor-induced membrane ruffling. C e l l 70 , 401-410. Rodriguez-Viciana, P., Warne, P. H., Khwaja, A., Marte, B. M., Pappin, D., Das, P., Waterfield, M. D., Ridley, A., and Downward, J. (1 9 9 7 ). Role of phosphoinositide 3-OH kinase in cell transformation and control of the actin cytoskeleton by Ras. C e l l 89 , 457-467. Sells, M. A., and Chernoff, J. (1 9 9 7 ). Emerging from the Pak: p21-activated protein kinase family. T r e n d s C e l l B i o l . 7, 162-167. Sells, M. A., Knaus, U. G., Bagrodia, S., Ambrose, D. M., Bokoch, G. M., and Chernoff, J. (1 9 9 7 ). Human p21activated kinase (Pak1) regulates actin organization in mammalian cells. Curr. Biol. 7, 202-210. Sheela, S., Riccardi, V. M., and Ratner, N. (1 9 9 0 ). Angiogenic and invasive properties of neurofibroma Schwann cells. J . C e l l . B i o l . 111, 645-653. Tang, Y., Chen, Z., Ambrose, D., Liu, J., Gibbs, J. B., Chernoff, J., and Field, J. (1 9 9 7 ). Kinase-deficient Pak1 mutants inhibit Ras transformation of Rat-1 fibroblasts. M o l . C e l l . B i o l . 17 , 4454-4464. Tang, Y., Marwaha, S., Rutkowski, J. L., Tennekoon, G. I., Phillips, P. C., and Field, J. (1 9 9 8 ). A role for Pak protein kinases in Schwann cell transformation. Proc. N a t l . A c a d . S c i . U S A 95 , 5139-5144. Tang, Y., Yu, J., and Field, J. (1 9 9 9 ). Signals from the Ras, Rac and Rho GTPases converge on the Pak protein kinase in Rat-1 fibroblasts. M o l . C e l l . B i o l . 19 , 1881-1891. Van Aelst, L., Barr, M., Marcus, S., Polverino, A., and Wigler, M. H. (1 9 9 3 ). Complex formation between RAS and RAF and other protein kinases. P r o c . N a t l . Acad. S c i . USA 90 , 6213-6217. Viskochil, D., Buchberg, A. M., Xu, G., Cawthon, R. M., Stevens, J., Wolff, R. K., Culver, M., Carey, J. C., Copeland, N. G., Jenkins, N. A., White, R., and

O'Connell, P. (1 9 9 0 ). Deletions and a translocation interrupt a cloned gene at the neurofibromatosis type 1 locus. C e l l 62 , 187-192. Vojtek, A. B., Hollenberg, S. M., and Cooper, J. A. (1 9 9 3 ). Mammalian Ras interacts directly with the serine/threonine kinase Raf. C e l l 74 , 205-214. Wallace, M. R., Marchuk, D. A., Anderson, L. B., Letcher, R., Odeh, H. M., Saulino, A. M., Fountain, J. W., Brereton, A., Nicholson, J., Mitchell, A. L., Brownstein, B. H., and Collins, F. S. (1 9 9 0 ). Type 1 Neurofibromatosis gene: Identification of a large transcript disrupted in three NF1 patients. S c i e n c e 249, 181-186. Westwick, J. K., Lambert, Q. T., Clark, G. J., Symons, M., Van Aelst, L., Pestell, R. G., and Der, C. J. (1 9 9 7 ). Rac regulation of transformation, gene expression, and actin organization by multiple, PAK-independent pathways. M o l . C e l l . B i o l . 17 , 1324-1335. White, M. A., Nicolette, C., Minden, A., Polverino, A., Van Aelst, L., Karin, M., and Wigler, M. H. (1 9 9 5 ). Multiple Ras functions can contribute to mammalian cell transformation. C e l l 80 , 533-541. White, M. A., Vale, T., Camonis, J. H., Schaefer, E., and Wigler, M. H. (1 9 9 6 ). A role for the Ral guanine nucleotide dissociation stimulator in mediating Ras-induced transformation. J . B i o l . C h e m . 271, 16439-16442. Xu, G. F., Lin, B., Tanaka, K., Dunn, D., Wood, D., Gesteland, R., White, R., Weiss, R., and Tamanoi, F. (1 9 9 0 ). The catalytic domain of the neurofibromatosis type 1 gene product stimulates ras GTPase and complements ira mutants of S. cerevisiae. C e l l 63 , 835-841. Zhang, S., Han, J., Sells, M. A., Chernoff, J., Knaus, U. G., Ulevitch, R. J., and Bokoch, G. M. (1 9 9 5 ). Rho family GTPases regulate p38 mitogen-activated protein kinase through the downstream mediator Pak1. J . B i o l . Chem 270, 23934-23936.

Jeffrey Field

416

Gene Therapy and Molecular Biology Vol 4, page 6 .

Gene Therapy and Molecular Biology Vol 4, page 417 Gene Ther Mol Biol Vol 4, 417-424. December 1999.

Negative regulation of cytoplasmic protein tyrosine kinase activity by adaptor proteins Review Article

Gael Manes, Paul Bello and Serge Roche CRBM, CNRS UPR1086, 1919 route de Mende, 34293 Montpellier, France __________________________________________________________________________________________________ Correspondence: Serge Roche, Ph.D., CRBM, CNRS UPR1086,1919 route de Mende, 34293 Montpellier, France. Tel: (33)-467-61-3373 ; Fax: (33) 467-52-15-59 ; E-mail: roche@crbm.cnrs-mop.fr Key Words: cytoplasmic protein, tyrosine, kinase, adaptor proteins, Jak, Syk, Fak, Src, Received: 10 August 1999; accepted: 16 August 1999

Summary Adaptor proteins are cytoplasmic signaling molecules that lack intrinsic catalytic activity but they are nevertheless crucial for signal transduction. They bear homology domains (SH2, SH3, PH, PTB, â&#x20AC;Ś) important for proteinprotein interactions and for function. The first adaptors identified were positive regulators of cell responses, with some even having oncogenic activities. More recently, a new subfamily has emerged that negatively regulates signaling responses. This review will focus on adaptors of this genre and specifically, those that inhibit cytoplasmic tyrosine kinase function and which define a new mechanism for in vivo kinase regulation. They include the inhibitors of the Jak, Syk, Fak and Src kinase family and their mechanism for inhibition as well as their possible function in cellular regulation will be discussed.

I. Cytoplasmic protein tyrosine kinases Tyrosine phospshorylation corresponds to less than 0.1% of the protein phosphorylation content in mammalian cells and yet protein tyrosine kinases play a pivotal role in cell regulation. These enzymes are classified in two distinct families: receptor and nonreceptor tyrosine kinases. The first includes receptors for growth factors and ligands involved in neuronal axon guidance. The later comprises cytoplasmic proteins grouped into 8 subfamilies, Src, Fak, Jak, Btk, Syk, Csk, Abl and Fps tyrosine kinases (Courtneidge, 1994). While having a striking homology in their catalytic sequences, they diverge greatly in their regulatory, non-enzymatic sequences, attributing to distinct kinase regulation, substrate phosphorylation and function. These kinases play important roles in cellular signaling and are activated by a large number of stimuli, including hormones, neurotransmitters, growth factors, cytokines, activation of T and B cells, cellular stress, adhesion and migration. In fact, most stimuli that use protein tyrosine phosphorylation for signaling activate two types of tyrosine kinases, one invariably being a member of the Src family. It is proposed that Src functions to

phosphorylate and activate a further enzyme responsible for substrate specific phosphorylation. For example, during T cell activation, the Src family member Lck, phosphorylates and activates Zap-70 which in turn phosphorylates downstream effectors for interleukin 2 (IL-2) gene expression (Courtneidge, 1994). Further evidence implicating cytoplasmic tyrosine kinases are genetic studies in mice: gene knockouts of fak (Ilic et al., 1995), csk (Imamoto and Soriano, 1993) or triple disruption of src, fyn and yes genes (Klinghoffer et al., 1999) all display an embryonic lethal phenotype. Finally, deregulation of these kinases has been linked to several human diseases: for example, oncogenic forms of Abl, Jak, and Src kinases have been identified in human cancers and have also been involved in leukemia or carcinoma development (Roche and Courtneidge, 1997). Moreover, a btk gene deficiency causes human X-linked agammaglobulinemia and murine X-linked immunodeficiency (Wen and Van Etten, 1997). Not surprisingly, their kinase activity is tightly regulated in vivo and much effort has gone towards unravelling the molecular mechanism(s) involved. Kinase activation generally begins with the phosphorylation of a tyrosine residue present in an activation loop (Courtneidge, 1994). However, less in known about the additional

417

Manes et al: Tyrosine kinase inhibitors mechanisms required for full kinase activation. The best examples studied to date are the Src family members. These enzymes are negatively regulated by phosphorylation of another regulatory tyrosine present at the C-terminus (refered to as pY527 in the chicken Src sequence). The importance of this conserved residue is confirmed by frequent residue deletions observed in oncogenic alleles (Roche and Courtneidge, 1997). Recent structural and mutagenic studies have revealed an even more elaborate model for regulation that in fact involves two intramolecular interactions: the SH2 domain associates with the pY527 and the SH3 complexes with a sequence between the SH2 domain and the kinase domain called the «SH2-linker» and the small lobe of the kinase domain itself. Regulated Src is in a closed or «off» conformation which when opened or turned «on», renders the kinase active (Superti-Furga and Gonfloni, 1997). Several mechanisms for activation can be proposed based on this model, for example, by dephosphorylating the C-terminal tyrosine and/or by interactions with ligands that disrupt either interaction (Superti-Furga and Gonfloni, 1997). While the 3dimentional structure for Abl is not yet known, mutagenic studies already hint to a similar intramolecular regulation involving the SH3 domain, implying a general mechanism (Barila and SupertiFurga, 1998). In this review, we propose a further mechanism for in vivo kinase regulation involving adaptor molecules which can prevent enzymatic activation and/or subtrate accessibility.

homology with the transcription factor Stat, and when overexpressed the cytokine response is inhibited. Since then several adaptors with similar properties have been identified. These include TGF! signaling proteins of the Smad family (Heldin et al., 1997), gene products related to Cis (the SOCS or SSI family) (Yoshimura, 1998), Cbl (Ota and Samelson, 1997), APS (Yokouchi et al., 1999), FRNK (Richardson and Parsons, 1996), Grb14 (Kasus-Jacobi et al., 1998) and Slap (Roche et al., 1998) (see Table 1). Smad (Heldin et al., 1997) and APS (Yokouchi et al., 1999) act by associating with other adaptors to inhibit cellular function, Grb14 (Kasus-Jacobi et al., 1998) appears to inhibit signaling by at least repressing the activity of targeted tyrosine kinase receptors and the others function by inhibiting cytoplasmic tyrosine kinase activities, in vivo. The molecular mechanism(s) and the function of the last group is discussed below.

II. Adaptors Adaptors are cytoplasmic proteins that do not possess any intrinsic catalytic activity but do contain known homology domains (SH2, SH3 , PTB, PH,…) important for protein interactions and signaling. For example SH2 domains interact with proteins through the recognition of phosphotyrosine residues and SH3 domains through association with proline rich sequences (van der Geer et al., 1994). The first members identified were either novel regulators of cell growth as induced by growth factors or oncogenes (Grb2 and Shc), or were transduced by transforming retroviruses (vCrk or vCbl) (see Table 1). They are thought to signal via complexing with their cognate effectors with subsequent targeting to the plasma membrane for activation. For example, the adaptor Grb2 is constitutively associated with the activator of the small GTP-binding protein Ras in the cytoplasm and upon growth factor stimulation, the complex is directed to the membrane enabling signaling (van der Geer et al., 1994). More recently, a new family of protein adaptors has emerged that inhibit cell responses. One of the first examples was Cis (Yoshimura et al., 1995) which was discovered by searching for new genes activated by cytokines. Cis is a small molecule that contains an SH2 domain which bears some

Adapter

Cell response

Grb2 Shc Crk Nck Grb10

+ + + + +

SOCS/CIS Slap FRNK Cbl Grb14 Smad APS

Cytoplasmic tyrosine kinase regulated by adapters

Jak Src Fak Syk

Table 1. Adapter proteins that activate (+) or inhibit (-) cell response and the cytoplasmic tyrosine kinases regulated

A. Jak inhibition by the SOCS adaptors The cytoplasmic tyrosine kinases of the Janus kinase family (Jak) are important regulators of the biological effects exerted by cytokines in haematopoietic and immune cells. Jak are constitutively associated with cytokine receptors and undergo tyrosine phosphorylation and activation following ligand binding. Important effectors of these kinases are the transcription factors of the Stat family. Upon cytokine stimulation, Stat associates with the cytoplasmic tail of the receptor through its SH2 domain and becomes phosphorylated by Jak. As a consequence, phosphorylated Stat dimerizes and translocates to the nucleus for target gene activation (Starr and Hilton, 1999). A functional approach was utilized in order to isolate inhibitors of cytokine signaling and led to the identification of the Supressor Of Cytokine Signaling family (SOCS) 418

Gene Therapy and Molecular Biology Vol 4, page 419 (Starr and Hilton, 1999). Additional members were further identified while searching for Jak interactors and new gene products with Stat SH2 homology. The SOCS family comprizes at least 8 members (SOCS1â&#x20AC;&#x201C;7 and Cis) (Starr and Hilton, 1999) and are composed of a variable N-terminus, a central SH2 domain with Stat SH2 homology and a conserved sequence of about 40 amino acids called the "SOCS box" (Figure 1). This new domain has also been found in several unrelated proteins and is thought to participate in protein degradation. SOCS adaptors inhibit Jak function in two ways: either by inhibiting the kinase activity (SOCS-1/JAB/SS-1) or by preventing association with and phosphorylation of the Stat proteins (Cis) (see Figure 2A). Mechanistically, SOCS-1 interacts with the phosphotyrosine in the activation loop of the Jak2 catalytic core preventing kinase activation both in vitro and in vivo (Figure 2A). While association involves the SOCS-1 SH2 domain, efficient inhibition also requires an additional 12 amino acid stretch present in the N-terminus (the "kinase inhibitory region"). SOCS-1 is activated at the transcriptional level by Stat and may define a negative regulatory loop as it is activated within the first minutes

of cytokine stimulation (Yasukawa et al., 1999). Recently, an additional mechanism has been proposed in SOCS regulation, namely protein degradation. Elongin B and C were identified as interactors of the "SOCS box" domain and which potentiated Jak inhibition and increased SOCS protein stability (Kamura et al., 1998; Yasukawa et al., 1999). Interestingly, elongin B displays some structural homology with Skp1, a protein involved in the proteasomal degradation pathway and so SOCS may also target elongin to Jak for subsequent degradation. The physiological function of SOCS-1 is not known and still awaits gene disruption in the whole animal, however when overexpressed, SOCS-1 is able to inhibit a number of cytokine responses suggesting that it plays a central role in the regulation of the Jak/Stat pathway. Thus far, the cytokine signaling pathways affected include oncostatin M, thrombopoietin, growth hormone, leukaemia inhibitory factor and IL-6 (Starr and Hilton, 1999). Finally, it should be noted that SOCS-1 function may not be solely restricted to the Jak tyrosine kinases since it was recently shown to reduce tyrosine kinase activity of the Tec family (Starr and Hilton, 1999). However, the mechanism and role of SOCS in Tec signaling is currently not known.

Figure 1. Structure of various cytoplasmic protein-tyrosine kinases, transcription factor and adapter proteins involved in negative regulation. The presence of the kinase region, the SH2, SH3 and unique (U) domains, as well as the DNA binding domain, the transactivation domain (TAD), the SOCS box domain (SB), the leucine zipper domain (L) and the proline rich region are indicated. (!) indicates the presence of a myristyl group.

419

Manes et al: Tyrosine kinase inhibitors

Figure 2. Mechanisms for negative regulation of the Jak/STAT (A) and Src (B) pathways. A: Inhibition of the Jak activity by SOCS-1 and inhibition of STAT receptor association and phosphorylation by CIS. B: Inhibition of Src activation by growth factor receptor and substrate phosphorylation by Slap.

In contrast to SOCS-1, Cis does not affect Jak kinase activity but rather acts by preventing substrate phosphorylation in vivo (Figure 2A). The Cis SH2 domain bears strong homology with the Stat SH2 domain and competes with Stat for cytokine receptor association, thus preventing Stat membrane translocation and Jak2 association. Like SOCS-1, Cis is activated at the transcriptional level by IL-3 and Epo. However, it again differs from SOCS-1 in the time course of activation since cis expression persists after 24h of ligand exposure (Yoshimura, 1998). Therefore, various

SOCS may regulate different aspects of cellular responses induced by cytokines and a combination of adaptors may define a fine tuning mechanism for the strength and length of signaling. This aspect of signal transduction can have dramatic consequences on cell response; for example, it has been shown that a prolonged activation of the serine/threonine Map kinase switches the cell from a growth response to a differentation process in the neuronal precursor cell line, PC12 (Dikic et al., 1994; Traverse et al., 1994).

420

Gene Therapy and Molecular Biology Vol 4, page 421

B. Syk inhibition by the adaptor Cbl Cbl is the second example of an adaptor that negatively regulates cytoplasmic tyrosine kinases in vivo. Cbl is a cellular proto-oncogene with a viral homolog v-Cbl originally found in a rodent leukemia retrovirus (Langdon et al., 1989). In contrast to the later, Cbl inhibits cell growth and transformation (Liu and Altman, 1998) and targets the tyrosine kinase Syk of the ZAP-70 family, which is important for lymphoid T and B cell activity and mast cell degranulation. Syk is activated by translocating to the membrane and associating with the T (or B) cell antigen receptor. This complex formation involves tyrosine phosphorylated sequences (ITAMs) in the receptor and the two Syk SH2 domains. Membrane associated Syk is then phosphorylated by Lck leading to maximal kinase activation. Moreover, Syk can be inactivated by phosphorylation of tyrosine 323 present in the linker region between the two SH2 domains and which prevents receptor association (Meng et al., 1999). When overexpressed, Cbl associates with phosphorylated Syk and is able to inhibit mast cell activation (Ota and Samelson, 1997). This adaptor also contains an SH2-like (SHL) domain at the N-terminus (see Figure 1) that has recently been shown to function as a phosphotyrosine binding domain (Meng et al., 1999). Cbl can inhibit Syk by association with pY323 through its SHL, leaving the kinase in a repressed form. Therefore, Cbl may regulate Syk signaling by preventing kinase activation rather than inhibiting the catalytic activity as proposed for SOCS-1. Just how the Cbl-Syk complex formation is regulated is not known. In addition to Syk, Cbl also regulates receptor tyrosine kinase function including the receptors for growth factors (Liu and Altman, 1998). However, while its inhibitory mechanisms are still unclear, several reports suggest that it induces receptor down regulation and degradation (Lee et al., 1999; Levkowitz et al., 1998). In contrast, microinjection data suggests that Cbl may specifically affect the DNA synthesis signaling pathway activated by cytoplasmic tyrosine kinases of the Src family (Broome et al., 1999). Finally, this adaptor becomes tyrosine phosphorylated in vivo creating binding sites for various signaling proteins including phosphoinositide 3-kinase and the adaptor Grb2, which all may influence signaling (Liu and Altman, 1998). Overall, the data suggests that Cbl may be a general inhibitor of cell activation and may affect several intracellular events. A physiological role for Cbl was first demonstrated in C. elegans where the orthologue, SLI-1, reduces Let23 (EGF receptor) signaling (Jongeward et al., 1995). In mammals, its function has also been addressed by gene disruption in mice ; while a number of negative functions have been attributed to Cbl using tissue culture models, cbl-/- mice are viable suggesting that this adaptor is not mandatory for

development. Rather, a marked change in haematopoietic profile was observed including lymphoid hyperplasia and enhanced T cell signaling (Murphy et al., 1998). This led to the postulation that adaptors with negative function may define a subtle mechanism for cell regulation rather than an ÂŤ on/off Âť process. Finally, an in vitro study attributed a positive function for Cbl during bone resorption (Tanaka et al., 1996) indicating that these adaptors could also have positive functions in some circumstances.

C. FRNK as a negative regulator of Fak The Focal adhesion kinase, Fak, is an important regulator of cell adhesion and migration. In contrast to the other subfamilies, it does not bear any known homology sequences like SH2 or SH3 domains. Instead, it has two proline-rich regions important for binding SH3-containing proteins and a C-terminal sequence called FAT, which mediates cellular focal contact association (Parsons, 1996) (Figure 1). While the mechanism for kinase activation is not fully understood, Fak tyrosine phosphorylation creates an SH2 binding site for Src (or Fyn) leading to enzymatic stimulation that in turns phosphorylates Fak further for maximal stimulation (Parsons, 1996). The product of an alternative spliced variant of the fak gene has been isolated which encodes the pp125FAK-related non-kinase, FRNK. This adaptor is composed of the C-terminal, non catalytic sequence of Fak and includes both the proline-rich and FAT sequences (Figure 1). When overexpressed, FRNK was shown to inhibit Fak in vivo preventing focal adhesion formation on fibronectin and substrate phosphorylation (Richardson and Parsons, 1996). Like Cis, FRNK may act by a competitive mechanism since Fak over expression reverses the FRNK inhibitory effect (Richardson et al., 1997). However, neither the importance of this spliced variant in vivo nor how its activity is regulated are known. Endogeneous protein has been detected in non transformed rodent embryo fibroblasts (Richardson and Parsons, 1996) suggesting that FRNK could define a signaling threshold but the relative protein levels of FRNK and Fak need to be investigated.

D. Negative regulation of Src by the adaptor Slap The Src family tyrosine kinases have important functions in a number of cell responses including cell growth, differentiation and migration. Moreover, when deregulated, they display oncogenic activity (Roche and Courtneidge, 1997). They are anchored at the membrane via myrsitoylation at the N-terminus and include in addition to a catalytic sequence an SH3 and SH2 domain important for kinase regulation (Figure 1). In non stimulated cells, endogenous Src kinases are inactive but are transiently activated by a number of extracellular stimuli including growth factors (Roche and Courtneidge, 1997). In the later case, activated receptors associate with Src members 421

Manes et al: Tyrosine kinase inhibitors through their SH2 domains leading to kinase activation. While substrates important for growth factor response are ill-defined, several groups including our own suggest a requirement for these kinases during mitogenesis (Broome and Hunter, 1996; Roche et al., 1995). A microinjection approach has also indicated that they act by upregulating the transcription factor, c-myc (Barone and Courtneidge, 1995). In contrast to most cytoplasmic kinases, Src members show a broad range of substrate phosphorylation. Nevertheless, they require intact SH3 and SH2 domains for signaling, probably for specific susbtrate association and phosphorylation. For example, despite a high kinase activity, deregulated Src lacking an SH3 domain fails to be transforming (Erpel et al., 1995) but is able to inhibit the mitogenic response induced by PDGF in fibroblasts (Erpel et al., 1996). The Src-Like Adaptor Protein (Slap) was originally identified in a 2-hybrid screen while searching for new interactors of the Eck tyrosine kinase receptor (Pandey et al., 1995) and has an SH2 and SH3 domain with high identity to those of Src. In addition, Slap is myristoylated and largely colocalizes with Src in vivo (Manes et al., 2000) and also bears a unique C-terminus sequence of about 100 amino acids that is involved in protein-protein interaction (Figure 1). We have previously shown that Slap inhibits growth factor-induced DNA synthesis when overexpressed (Roche et al., 1998) and mutagenesis analyses together with the microinjection studies have allowed us to conclude that Slap inhibits Src function probably by associating with and competing for Src signaling substrates (Manes et al., 2000). The Src and Slap SH2 domains show strong functional homology and are essential for their respective functions. As a consequence, Slap associates with growth factor receptors through the same phosphotyrosine binding site used by Src. Therefore it may act by preventing association with and activation of Src by the receptor. However, Slap does not act solely by a competition mechanism as described for FRNK ; indeed, Src overexpression does not reverse the Slap inhibitory effect and further analyses pointed to an involvement of the C-terminus (Manes et al., 2000). We therefore propose a model where Slap utilizes its SH2 domain primarily for proper cell localisation (binding to the PDGF receptor for example) whereas the whole Slap molecule would additionally titrate Src effectors implying the C-terminus sequence, thus preventing signaling despite Src overexpression (Figure 2B). From this we initially predicted two mechanisms of Src inhibition: a competitive (via the SH2 domain) or a non competitive (via both the SH2 and the C-terminus) mechanism. However, another important difference between Src and Slap lies in the SH3 domain. Despite the strong homology, three critical residues important for ligand binding specificity in Src SH3 are replaced in Slap SH3. As a consequence, they display distinct binding specificities suggesting that Slap will not inihibit

Src SH3-dependent signaling. Indeed, Slap over expression is unable to reverse cell transformation induced by oncogenic Src (Manes et al., 2000), probably by its inability to interfere with Src SH3 effectors required for oncogenic events such as cytoskeletal rearrengement. These latter observations predict that Slap SH3 may also have specific functions and binding partners over and above being a simple Src competitor. How Slap itself is regulated is not known although it should be noted that a proline-rich ÂŤSH2 linker Âť is present int the Slap C-terminus. It is enticing to speculate that a similar intramolecular regulation exists for Slap as does for Src. In contrast to SOCS inhibitors, its protein level does not change dramatically during the cell cycle, suggesting that it does not define a negative regulatory loop. Also, Slap is not tyrosine phosphorylated like the adaptor Cbl. We feel that Slap could also define a signaling threshold for cell response modulation and to support this hypothesis, inhibition of endogeneous Slap function by antibody microinjection led to an increased cell response toward growth factors in fibroblasts (Roche et al., 1998). Furthermore, SLAP may transduce a signal for cell differentiation as its gene was found upregulated in cells treated by a differentiation agent.

III. Conclusion The identification of this new family of adaptors with negative function in cellular processes provides a new mechanism for cytoplasmic tyrosine kinase regulation in vivo. These adaptors clearly involve several mechanisms including inhibition of the catalytic activity, prevention of activation by competitive association with receptors and inhibition of phosphorylation by substrate association. The use of adaptors for tyrosine kinase regulation may be a general phenomenon and additional members are likely to be discovered. For example, kinase inhibitors of Btk (Yamadori et al., 1999) and Abl (Wen and Van Etten, 1997) have recently been identified but have not been discussed in this review since they have no obvious homology domains. Regulation of their activity is not clearly established. Some, such as SOCS, define a negative regulatory loop and are activated by gene expression while others like Slap may orchestrate a signaling threshold response and could also be involved in cell differentiation processes. Regulation of their activity is an important issue that needs to be addressed, as it may have important consequences on cell responses and human diseases linked to cytoplasmic tyrosine kinases. For the most part, their physiological function in vivo are not known and this is another important issue that needs to be addressed. However, given the fact that they are involved in the fine tuning of signaling for optimal cell responses, one can comfortably predict that they will not be mandatory for embryogenesis.

422

Gene Therapy and Molecular Biology Vol 4, page 423

References Barila, D., and Superti-Furga, G. (1998). An intramolecular SH3-domain interaction regulates c-Abl activity. Nat Genet 18, 280-2. Barone, M. V., and Courtneidge, S. A. (1995). Myc but not Fos rescue of PDGF signalling block caused by kinaseinactive Src. Nature 378, 509-12. Broome, M. A., Galisteo, M. L., Schlessinger, J., and Courtneidge, S. A. (1999). The proto-oncogene c-Cbl is a negative regulator of DNA synthesis initiated by both receptor and cytoplasmic tyrosine kinases. Oncogene 18, 2908-12. Broome, M. A., and Hunter, T. (1996). Requirement for c-Src catalytic activity and the SH3 domain in platelet- derived growth factor BB and epidermal growth factor mitogenic signaling. J Biol Chem 271, 16798-806. Courtneidge, S. A. (1994). Non-receptor protein tyrosine kinases, J. Woodgett, ed.: (Oxford University Press).p212p240 Dikic, I., Schlessinger, J., and Lax, I. (1994). PC12 cells overexpressing the insulin receptor undergo insulindependent neuronal differentiation. Curr Biol 4, 702-8. Erpel, T., Alonso, G., Roche, S., and Courtneidge, S. A. (1996). The Src SH3 domain is required for DNA synthesis induced by platelet- derived growth factor and epidermal growth factor. J Biol Chem 271, 16807-12. Erpel, T., Superti-Furga, G., and Courtneidge, S. A. (1995). Mutational analysis of the Src SH3 domain: the same residues of the ligand binding surface are important for intra- and intermolecular interactions. EMBO J 14, 96375. Heldin, C. H., Miyazono, K., and ten Dijke, P. (1997). TGFbeta signalling from cell membrane to nucleus through SMAD proteins. Nature 390, 465-71. Ilic, D., Furuta, Y., Kanazawa, S., Takeda, N., Sobue, K., Nakatsuji, N., Nomura, S., Fujimoto, J., Okada, M., and Yamamoto, T. (1995). Reduced cell motility and enhanced focal adhesion contact formation in cells from FAKdeficient mice. Nature 377, 539-44. Imamoto, A., and Soriano, P. (1993). Disruption of the csk gene, encoding a negative regulator of Src family tyrosine kinases, leads to neural tube defects and embryonic lethality in mice. Cell 73, 1117-24. Jongeward, G. D., Clandinin, T. R., and Sternberg, P. W. (1995). sli-1, a negative regulator of let-23-mediated signaling in C. elegans. Genetics 139, 1553-66. Kamura, T., Sato, S., Haque, D., Liu, L., Kaelin, W. G., Jr., Conaway, R. C., and Conaway, J. W. (1998). The Elongin BC complex interacts with the conserved SOCS-box motif present in members of the SOCS, ras, WD-40 repeat, and ankyrin repeat families. Genes Dev 12, 3872-81. Kasus-Jacobi, A., Perdereau, D., Auzan, C., Clauser, E., Van Obberghen, E., Mauvais-Jarvis, F., Girard, J., and Burnol, A. F. (1998). Identification of the rat adapter Grb14 as an inhibitor of insulin actions. J Biol Chem 273, 26026-35. Klinghoffer, R. A., Sachsenmaier, C., Cooper, J. A., and Soriano, P. (1999). Src family kinases are required for integrin but not PDGFR signal transduction. EMBO J 18, 2459-71. Langdon, W. Y., Hartley, J. W., Klinken, S. P., Ruscetti, S. K., and Morse, H. C. d. (1989). v-cbl, an oncogene from a dual-recombinant murine retrovirus that induces early Blineage lymphomas. Proc Natl Acad Sci USA 86, 116872. Lee, P. S., Wang, Y., Dominguez, M. G., Yeung, Y. G., Murphy, M. A., Bowtell, D. D., and Stanley, E. R. (1999). The Cbl protooncoprotein stimulates CSF-1 receptor multiubiquitination and endocytosis, and attenuates macrophage proliferation. EMBO J 18, 3616-3628.

Levkowitz, G., Waterman, H., Zamir, E., Kam, Z., Oved, S., Langdon, W. Y., Beguinot, L., Geiger, B., and Yarden, Y. (1998). c-Cbl/Sli-1 regulates endocytic sorting and ubiquitination of the epidermal growth factor receptor. Genes Dev 12, 3663-74. Liu, Y. C., and Altman, A. (1998). Cbl: complex formation and functional implications. Cell Signal 10, 377-85. Manes, G., Bello, P., and Roche, S. (2000). Slap negatively regulates Src mitogenenic function but does not revert Srcinduced cell morphology change. Mol Cell Biol In press. Meng, W., Sawasdikosol, S., Burakoff, S. J., and Eck, M. J. (1999). Structure of the amino-terminal domain of Cbl complexed to its binding site on ZAP-70 kinase. Nature 398, 84-90. Murphy, M. A., Schnall, R. G., Venter, D. J., Barnett, L., Bertoncello, I., Thien, C. B., Langdon, W. Y., and Bowtell, D. D. (1998). Tissue hyperplasia and enhanced T-cell signalling via ZAP-70 in c-Cbl- deficient mice. Mol Cell Biol 18, 487282. Ota, Y., and Samelson, L. E. (1997). The product of the protooncogene c-cbl: a negative regulator of the Syk tyrosine kinase. Science 276, 418-20. Pandey, A., Duan, H., and Dixit, V. M. (1995). Characterization of a novel Src-like adapter protein that associates with the Eck receptor tyrosine kinase. J Biol Chem 270, 19201-4. Parsons, J. T. (1996). Integrin-mediated signalling: regulation by protein tyrosine kinases and small GTP-binding proteins. Curr Opin Cell Biol 8, 146-52. Richardson, A., Malik, R. K., Hildebrand, J. D., and Parsons, J. T. (1997). Inhibition of cell spreading by expression of the Cterminal domain of focal adhesion kinase (FAK) is rescued by coexpression of Src or catalytically inactive FAK: a role for paxillin tyrosine phosphorylation. Mol Cell Biol 17, 6906-14. Richardson, A., and Parsons, T. (1996). A mechanism for regulation of the adhesion-associated proteintyrosine kinase pp125FAK [published erratum appears in Nature 1996 Jun 27;381(6585):810]. Nature 380, 538-40. Roche, S., Alonso, G., Kazlauskas, A., Dixit, V. M., Courtneidge, S. A., and Pandey, A. (1998). Src-like adaptor protein (Slap) is a negative regulator of mitogenesis. Curr Biol 8, 975-8. Roche, S., and Courtneidge, S. A. (1997). oncogenic cytoplasmic tyrosine kinases, Volume 19, G. Peters and K. H. Vousden, eds. (Oxford New York Tokyo: Oxford University Press).p87p129 Roche, S., Koegl, M., Barone, M. V., Roussel, M. F., and Courtneidge, S. A. (1995). DNA synthesis induced by some but not all growth factors requires Src family protein tyrosine kinases. Mol Cell Biol 15, 1102-9. Starr, R., and Hilton, D. J. (1999). Negative regulation of the JAK/STAT pathway. Bioessays 21, 47-52. Superti-Furga, G., and Gonfloni, S. (1997). A crystal milestone: the structure of regulated Src. Bioessays 19, 447-50. Tanaka, S., Amling, M., Neff, L., Peyman, A., Uhlmann, E., Levy, J. B., and Baron, R. (1996). c-Cbl is downstream of c-Src in a signalling pathway necessary for bone resorption. Nature 383, 528-31. Traverse, S., Seedorf, K., Paterson, H., Marshall, C. J., Cohen, P., and Ullrich, A. (1994). EGF triggers neuronal differentiation of PC12 cells that overexpress the EGF receptor. Curr Biol 4, 694-701. van der Geer, P., Hunter, T., and Lindberg, R. A. (1994). Receptor protein-tyrosine kinases and their signal transduction pathways. Annu Rev Cell Biol 10, 251-337. Wen, S. T., and Van Etten, R. A. (1997). The PAG gene product, a stress-induced protein with antioxidant properties, is an Abl SH3-binding protein and a physiological inhibitor of c-Abl tyrosine kinase activity. Genes Dev 11, 2456-67. Yamadori, T., Baba, Y., Matsushita, M., Hashimoto, S., Kurosaki, M., Kurosaki, T., Kishimoto, T., and Tsukada, S. (1999).

423

Manes et al: Tyrosine kinase inhibitors Bruton's tyrosine kinase activity is negatively regulated by Sab, the Btk-SH3 domain-binding protein. Proc Natl Acad Sci USA 96, 6341-6. Yasukawa, H., Misawa, H., Sakamoto, H., Masuhara, M., Sasaki, A., Wakioka, T., Ohtsuka, S., Imaizumi, T., Matsuda, T., Ihle, J. N., and Yoshimura, A. (1999). The JAK-binding protein JAB inhibits Janus tyrosine kinase activity through binding in the activation loop. EMBO J J 18, 1309-20. Yokouchi, M., Wakioka, T., Sakamoto, H., Yasukawa, H., Ohtsuka, S., Sasaki, A., Ohtsubo, M., Valius, M., Inoue, A., Komiya, S., and Yoshimura, A. (1999). APS, an adaptor protein containing PH and SH2 domains, is associated with the PDGF receptor and c-Cbl and inhibits PDGF-induced mitogenesis. Oncogene 18, 759-67. Yoshimura, A. (1998). The CIS family: negative regulators of JAK-STAT signaling. Cytokine Growth Factor Rev 9, 197-204. Yoshimura, A., Ohkubo, T., Kiguchi, T., Jenkins, N. A., Gilbert, D. J., Copeland, N. G., Hara, T., and Miyajima, A. (1995). A novel cytokine-inducible gene CIS encodes an SH2-containing protein that binds to tyrosinephosphorylated interleukin 3 and erythropoietin receptors. EMBO J 14, 2816-26.

424