haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Editor-in-Chief Jan Cools (Leuven)
Deputy Editor Luca Malcovati (Pavia)
Managing Director Antonio Majocchi (Pavia)
Associate Editors Hélène Cavé (Paris), Ross Levine (New York), Claire Harrison (London), Pavan Reddy (Ann Arbor), Andreas Rosenwald (Wuerzburg), Juerg Schwaller (Basel), Monika Engelhardt (Freiburg), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), Paolo Ghia (Milan), Swee Lay Thein (Bethesda), Pieter Sonneveld (Rotterdam)
Assistant Editors Anne Freckleton (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (English Editor), Kate O’Donohoe (English Editor)
Editorial Board Omar I. Abdel-Wahab (New York); Jeremy Abramson (Boston); Paolo Arosio (Brescia); Raphael Bejar (San Diego); Erik Berntorp (Malmö); Dominique Bonnet (London); Jean-Pierre Bourquin (Zurich); Suzanne Cannegieter (Leiden); Francisco Cervantes (Barcelona); Nicholas Chiorazzi (Manhasset); Oliver Cornely (Köln); Michel Delforge (Leuven); Ruud Delwel (Rotterdam); Meletios A. Dimopoulos (Athens); Inderjeet Dokal (London); Hervé Dombret (Paris); Peter Dreger (Hamburg); Martin Dreyling (München); Kieron Dunleavy (Bethesda); Dimitar Efremov (Rome); Sabine Eichinger (Vienna); Jean Feuillard (Limoges); Carlo Gambacorti-Passerini (Monza); Guillermo Garcia Manero (Houston); Christian Geisler (Copenhagen); Piero Giordano (Leiden); Christian Gisselbrecht (Paris); Andreas Greinacher (Greifswals); Hildegard Greinix (Vienna); Paolo Gresele (Perugia); Thomas M. Habermann (Rochester); Claudia Haferlach (München); Oliver Hantschel (Lausanne); Christine Harrison (Southampton); Brian Huntly (Cambridge); Ulrich Jaeger (Vienna); Elaine Jaffe (Bethesda); Arnon Kater (Amsterdam); Gregory Kato (Pittsburg); Christoph Klein (Munich); Steven Knapper (Cardiff); Seiji Kojima (Nagoya); John Koreth (Boston); Robert Kralovics (Vienna); Ralf Küppers (Essen); Ola Landgren (New York); Peter Lenting (Le Kremlin-Bicetre); Per Ljungman (Stockholm); Francesco Lo Coco (Rome); Henk M. Lokhorst (Utrecht); John Mascarenhas (New York); Maria-Victoria Mateos (Salamanca); Simon Mendez-Ferrer (Madrid); Giampaolo Merlini (Pavia); Anna Rita Migliaccio (New York); Mohamad Mohty (Nantes); Martina Muckenthaler (Heidelberg); Ann Mullally (Boston); Stephen Mulligan (Sydney); German Ott (Stuttgart); Jakob Passweg (Basel); Melanie Percy (Ireland); Rob Pieters (Rotterdam); Stefano Pileri (Milan); Miguel Piris (Madrid); Andreas Reiter (Mannheim); Jose-Maria Ribera (Barcelona); Stefano Rivella (New York); Francesco Rodeghiero (Vicenza); Richard Rosenquist (Uppsala); Simon Rule (Plymouth); Claudia Scholl (Heidelberg); Martin Schrappe (Kiel); Radek C. Skoda (Basel); Gérard Socié (Paris); Kostas Stamatopoulos (Thessaloniki); David P. Steensma (Rochester); Martin H. Steinberg (Boston); Ali Taher (Beirut); Evangelos Terpos (Athens); Takanori Teshima (Sapporo); Pieter Van Vlierberghe (Gent); Alessandro M. Vannucchi (Firenze); George Vassiliou (Cambridge); Edo Vellenga (Groningen); Umberto Vitolo (Torino); Guenter Weiss (Innsbruck).
Editorial Office Simona Giri (Production & Marketing Manager), Lorella Ripari (Peer Review Manager), Paola Cariati (Senior Graphic Designer), Igor Ebuli Poletti (Senior Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review)
Affiliated Scientific Societies SIE (Italian Society of Hematology, www.siematologia.it) SIES (Italian Society of Experimental Hematology, www.siesonline.it)
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies). Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the International edition for the year 2016 are as following: Print edition
Institutional Euro 500
Personal Euro 150
Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature. Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Tipografia PI-ME, via Vigentina 136, Pavia, Italy. Printed in September 2016.
haematologica calendar of events
Journal of the European Hematology Association Published by the Ferrata Storti Foundation
3rd International Conference on Multiple Myeloma European School of Haematology (ESH) Chairs: S Lonial, M Mohty, A Palumbo October 7-9, 2016 Milan, Italy
5th National Congress of Geriatric Hematology Society of Turkish Geriatric Hematology Chairs: O Ilhan, A Tunali October 7-9, 2016 Ankara, Turkey
XIV Congresso Nazionale SIES - Societa' Italiana Di Ematologia Sperimentale Societa' Italiana Di Ematologia Sperimentale (SIES) Chair: M Massaia October 19-21, 2016 Rimini, Italy
10th Hodgkin Symposium University hospital of Cologne Chairs: A Engert, B von Treskow, B Böll October 22-25, 2016 Cologne, Germany
2 MEGMA Conference on Thalassaemia and Other Haemoglobinopathies Thalassaemia International Federation (TIF) Chairs: A Taher, J Porter, A Piga, A Beshlawy November 11-12, 2016 Amman, Jordan nd
Highlights of Past EHA - HOPE Dubai 2016 Chairs: R Foà, M Qari November 24-26, 2016 Dubai, UAE
5th ESLHO educational meeting: New developments in the ESLHO networks European Scientific foundation for Laboratory Hemato Oncology (ESLHO) Chairs: JJM van Dongen, P Groenen, B Schäfer November 3, 2016 Prague, Czech Republic
EuroClonality Workshop: “Clonality assessment in Pathology” European Scientific foundation for Laboratory Hemato Oncology (ESLHO) Chairs: PJTA Groenen, JHJM van Krieken, AW Langerak February 13-15, 2016 Nijmegen, The Netherlands
EHA Hematology Tutorial on Lymphoid malignancies , Multiple myeloma and Bone Marrow Failure February 23-24, 2017 Colombo, Sri Lanka
EHA Hematology Tutorial on Lymphoid Malignancies March 17-18, 2017 Warsaw, Poland
EHA Scientific Meeting on Advances in Biology and Treatment of B Cell Malignancies with a Focus on Rare Lymphoma Subtypes Chairs: MJ Kersten and M Dreyling March 10-12, 2017 Barcelona, Spain
EHA Scientific Meeting on Aging and Hematology Chair: D Bron May 4-6, 2017 Location: TBC
22nd Congress of the European Hematology Association European Hematology Association June 22 - 25, 2017 Madrid, Spain
EHA Scientific Meeting on Challenges in the Diagnosis and Management of Myeloproliferative Neoplasms Chairs: JJ Kiladjian and C Harrison October 12-14, 2017 Location: TBC
EHA Scientific Meeting on Shaping the Future of Mesenchymal Stromal Cells Therapy Chair: W Fibbe November 23-25, 2017 Location: TBC
2nd MEGMA Conference on Thalassaemia and Other Haemoglobinopathies Thalassaemia International Federation (TIF) Chairs: A Taher, J Porter, A Piga, A Beshlawy November 11-12, 2016 Amman, Jordan
EHA Scientific Meeting on Anemia Diagnosis and Treatment in the Omics Era Chair: A Iolascon February 2-4, 2017 Barcelona, Spain
Calendar of Events updated on September 1, 2016
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
Table of Contents Volume 101, Issue 10: October 2016 Cover Figure Artistic representation of the mutant KIT receptor on the cell surface. Image accompanies the review article on mastocytosis on page 1133. (Image created by www.somersault1824.com)
Editorials 1129
The complex relationship of Tribbles pseudokinase 1, PML/RARA and C/EBPÎą in leukemia: two possible couples but not a trio Guillermo Velasco
1131
Matching inside and outside the HLA molecule in allogeneic hematopoietic stem cell transplantation J. Alejandro Madrigal and Linda D. Barber
Review Articles 1133
Advanced systemic mastocytosis: from molecular and genetic progress to clinical practice Celalettin Ustun, et al.
1144
Catching up with solid tumor oncology: what is the evidence for a prognostic role of programmed cell death-ligand 1/programmed cell death-1 expression in B-cell lymphomas? Fabienne McClanahan, et al.
Articles Coagulation & Its Disorders
1159
A population pharmacokinetic model for perioperative dosing of factor VIII in hemophilia A patients Hendrika Hazendonk, et al.
Platelet Biology & Its Disorders
1170
Whole exome sequencing identifies genetic variants in inherited thrombocytopenia with secondary qualitative function defects Ben Johnson, et al.
Bone Marrow Failure
1180
Marked overlap of four genetic syndromes with dyskeratosis congenita confounds clinical diagnosis Amanda J. Walne, et al.
Myeloproliferative Disorders
1190
Stat5 is critical for the development and maintenance of myeloproliferative neoplasm initiated by Nf1 deficiency Zohar Sachs, et al.
Chronic Myeloid Leukemia
1200
Nilotinib 300 mg twice daily: an academic single-arm study of newly diagnosed chronic phase chronic myeloid leukemia patients Fausto Castagnetti, et al.
Haematologica 2016; vol. 101 no. 10 - October 2016 http://www.haematologica.org/
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation Acute Myeloid Leukemia
1208
Ciprofloxacin versus colistin prophylaxis during neutropenia in acute myeloid leukemia: two parallel patient cohorts treated in a single center Michele Pohlen, et al.
1216
Acute myeloid leukemia cells polarize macrophages towards a leukemia supporting state in a Growth factor independence 1 dependent manner Yahya S. Al-Matary, et al.
1228
Co-operative leukemogenesis in acute myeloid leukemia and acute promyelocytic leukemia reveals C/EBPÎą as a common target of TRIB1 and PML/RARA Karen Keeshan, et al.
Hodgkin Lymphoma
1237
Definition of bulky disease in early stage Hodgkin lymphoma in computed tomography era: prognostic significance of measurements in the coronal and transverse planes Anita Kumar, et al.
Non-Hodgkin Lymphoma
1244
Non-Hodgkin lymphoma in the developing world: review of 4539 cases from the International Non-Hodgkin Lymphoma Classification Project Anamarija M. Perry, et al.
Cell Therapy & Immunotherapy
1251
A phase I study of CD25/regulatory T-cell-depleted donor lymphocyte infusion for relapse after allogeneic stem cell transplantation Sarah Nikiforow, et al.
1260
Donor and recipient sex in allogeneic stem cell transplantation: what really matters Haesook T. Kim, et al.
1267
Human leukocyte antigen supertype matching after myeloablative hematopoietic cell transplantation with 7/8 matched unrelated donor allografts: a report from the Center for International Blood and Marrow Transplant Research Aleksandr Lazaryan, et al.
Letters to the Editor Letters are available online only at www.haematologica.org/content/101/10.toc
e392
Specific antibody deficiency and autoinflammatory disease extend the clinical and immunological spectrum of heterozygous NFKB1 loss-of-function mutations in humans Cyrill Schipp, et al. http://www.haematologica.org/content/101/10/e392
e397
Genetic inactivation of calpain-1 attenuates pain sensitivity in a humanized mouse model of sickle cell disease Jennifer O. Nwankwo, et al. http://www.haematologica.org/content/101/10/e397
e401
Neutrophil to lymphocyte ratio and future risk of venous thromboembolism and mortality: the Tromsø Study Gro Grimnes, et al. http://www.haematologica.org/content/101/10/e401
e405
Discrepancies of applying primary myelofibrosis prognostic scores for patients with post polycythemia vera/essential thrombocytosis myelofibrosis Krisstina Gowin, et al. http://www.haematologica.org/content/101/10/e405
Haematologica 2016; vol. 101 no. 10 - October 2016 http://www.haematologica.org/
haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation
e407
European phase II study of mogamulizumab, an anti-CCR4 monoclonal antibody, in relapsed/refractory peripheral T-cell lymphoma Pier Luigi Zinzani, et al. http://www.haematologica.org/content/101/10/e407
e411
Maintenance rituximab following induction R-CHOP chemotherapy in patients with composite or discordant, indolent and aggressive, B-cell non-Hodgkin lymphomas Roopesh Kansara, et al. http://www.haematologica.org/content/101/10/e411
e415
Baseline bone involvement in multiple myeloma – a prospective comparison of conventional X-ray, low-dose computed tomography, and 18flourodeoxyglucose positron emission tomography in previously untreated patients Maja Hinge, et al. http://www.haematologica.org/content/101/10/e415
e419
Bone healing in multiple myeloma: a prospective evaluation of the impact of first-line anti-myeloma treatment Maja Hinge, et al. http://www.haematologica.org/content/101/10/e419
e423
Identification of a new potential mechanism responsible for severe bleeding in myeloma: immunoglobulins bind the heparin binding domain of antithrombin activating this endogenous anticoagulant Irene Martínez-Martínez, et al. http://www.haematologica.org/content/101/10/e423
e427
Predicting outcome of patients with myelodysplastic syndromes after failure of azacitidine: validation of the North American MDS consortium scoring system Thomas Prebet, et al. http://www.haematologica.org/content/101/10/e427
e429
The higher prevalence of missense mutations in hemophilia B compared to hemophilia A could be important in determining a milder clinical phenotype in patients with severe hemophilia B Daniela Melchiorre, et al. http://www.haematologica.org/content/101/10/e429
e430
A contribution to the debate about the possible different clinical severity between hemophilia A and B Daniela Melchiorre, et al. http://www.haematologica.org/content/101/10/e430
Haematologica 2016; vol. 101 no. 10 - October 2016 http://www.haematologica.org/
EDITORIALS The complex relationship of Tribbles pseudokinase 1, PML/RARA and C/EBPα in leukemia: two possible couples but not a trio Guillermo Velasco1,2 1
Department of Biochemistry and Molecular Biology I, School of Biology, Complutense University; and 2Instituto de Investigaciones Sanitarias San Carlos (IdISSC), Madrid, Spain E-mail: gvelasco@ucm.es doi:10.3324/haematol.2016.151654
T
ribbles gene was firstly identified in Drosophila where it was shown to be involved in cell cycle regulation.1-3 The name Tribbles was proposed by Seher and Leptin because the proliferating mesodermal cells observed in the tribbles-deficient drosophila embryos reminded them of the rapidly-dividing science fiction alien organism named “Tribbles” that appeared in the TV series “Star Trek” in December 29, 1967.3 There are three human orthologs of the drosophila Tribbles gene named Tribbles pseudokinase 1, 2 and 3 (TRIB1, TRIB2 and TRIB3).4 As indicated by their names, Tribbles genes encode “pseudokinases”: i.e. proteins that have a kinase domain lacking some of the amino acids that are essential for kinase enzymatic activity.5 In addition to the pseudokinase domain, Tribbles proteins also contain a C terminal E3 ligase-binding domain, that is involved in the regulation of the ubiquitination of several target proteins.4 Mammalian Tribbles play an important role on the control of different physiological functions including the regulation of lipid metabolism, inflammation and innate immunity.4 Likewise, Tribbles proteins have been implicated in both the regulation of tumorigenesis and the mechanism of action of several anticancer agents.6,7 TRIB1 and TRIB2 but not TRIB3 seem to play a relevant role in leukemia.6,8,9 Thus, TRIB1 (which was initially identified as a gene placed at a retroviral integration site10) has been shown to be over-expressed in patients with acute myeloid leukemia (AML).6,8,9 Moreover, the human TRIB1 gene is located in the same chromosome region as MYC, a gene that is also frequently amplified in AML.6,8,9 In line with these observations, MYC and TRIB1 have been shown to co-operate in AML.6,8,9 The oncogenic role of TRIB1 in AML relies at least in part on the ability of TRIB1 to regulate the stability of the transcription factor C/EBPα via the E3 ligase COP-1.11 C/EBPα plays an important role in the control of myeloid differentiation and its (TRIB1-COP1-dependent) degradation blocks this process thereby facilitating the transcription of leukemia promoting genes.6,8,9 Now, in an article included in this issue of Haematologica,12 Keeshan et al. further explore the role played by TRIB1, MYC, C/EBPα and PML/RARA in AML and acute promyelocytic leukemia (APL) shedding light (and adding also some layers of complexity) on the role played by this pseudokinase on these malignancies. The most frequent type of APL derives from a balanced chromosomal translocation [t(15;17)(q22;q12)] that leads to the fusion of the N-terminus of the promyelocytic leukemia protein (PML) with the C terminus of the retinoic acid receptor-alpha (RARA) transcription factor to produce the PML/RARA fusion protein.13 In the absence of its ligand (retinoic acid, RA) the nuclear receptor RARA represses the transcription of genes involved in myeloid differentiation whereas in the presence of physiological levels of RA, RARA promotes the expression of these genes. In contrast, PML/RARA does not respond to RA and haematologica | 2016; 101(10)
therefore cannot drive differentiation. In addition, PML/RARA also prevents the formation of the PML nuclear bodies (nuclear structures that are involved in the regulation of p53 and other important signaling mechanisms.13) The combination of the two events seems to be responsible for the accumulation of promyelocyte characteristic of the disease. In this context, Keeshan et al. investigated whether TRIB1 co-operates with PML/RARA and MYC in the development of AML and APL. Using an elegant approach, bone marrow cells derived from wild-type (WT) animals or from transgenic mice over-expressing PML/RARA were transduced with retroviral vectors encoding MYC and TRIB1. These cells were subsequently transplanted into lethally-irradiated animals and finally, the genotypes of the clones that produced leukemias in these animals analyzed. Interestingly, leukemias derived from WT-bone marrow cells expressed both TRIB1 and MYC whereas those derived from PML/RARA cells in most cases expressed MYC but not TRIB1. These observations suggest that TRIB1 co-operates with MYC (but not with PML/RARA) to induce leukemias. One of the reasons for this lack of co-operation between TRIB1 and PML/RARA could be that both proteins promote leukemia through the same mechanism and that therefore leukemia development in the mice is not facilitated by the common selection of the
Figure 1. C/EBPα is a common target of PML/RARA and TRIB1. Proposed model (based on the work by Keesham et al.12) by which PML/RARA and TRIB1 inhibit myeloid differentiation by targeting C/EBPα. PML/RARA leads to a decreased expression of C/EBPα whereas TRIB1 promotes its ubiquitination and subsequent proteasomal degradation. In the absence of TRIB1 overexpression, ATRA treatment promotes PML/RARA degradation and modifies PML/RARA transcriptional activity, which leads in turn to increased C/EBPα expression, an event that is required to restablish myeloid differentiation. In the presence of elevated levels of TRIB1, C/EBPα levels cannot be recovered which blocks the effect of ATRA on myeloid differentiation.
1129
Editorials
two genes. The authors hypothesized that the transcription factor C/EBPα (that plays a crucial role in the regulation of myeloid differentiation) could be the common target of TRIB1 and PML/RARA. To investigate this hypothesis, the authors used the well-established model of alltrans-RA (ATRA)-induced APL cells differentiation (ATRA treatment converts PML/RARA from a repressor to a transcriptional activator and also promotes PML/RARA degradation thereby triggering myeloid differentiation.)13 Using this model, the authors found that TRIB1 overexpression abrogates the response to ATRA of PML/RARA-expressing APL cells. Moreover, the authors also found that overexpression of TRIB1 (but not of a TRIB1 mutant that cannot promote C/EBPα degradation) prevented ATRAinduced C/EBPα upregulation. Furthermore, an additional in vivo experiment showed that increased expression of TRIB1 abolishes the effect of ATRA in PML/RARA and MYC-induced leukemias. Altogether, these findings support the idea that both TRIB1 and PML/RARA negatively regulate C/EBPα (although acting through different mechanisms) and that the regulation of C/EBPα plays a relevant role in the control of myeloid differentiation and leukemogenesis (Figure 1). In any case, several questions remain to be clarified in relation with the ideas presented in this work. For example, are TRIB1 (or TRIB2) endogenous levels down-regulated in APL patients? Initial analysis in published gene expression datasets suggest that this could be the case in APL. However, confirmation that TRIB1 protein levels, and not only mRNA levels, are affected would be important to understand the relevance in patients of the results obtained in animal models. Another point that still requires to be experimentally explored is whether PML/RARA negatively regulates TRIB1 (or TRIB2) expression. The existence of this specific regulatory mechanism is a conceivable (although still hypothetical) possibility that might explain why TRIB1 was not present in the PML/RARA leukemias investigated in this study. In addition, it would be important to understand the precise contribution of the increased expression of TRIB1 or TRIB2 to the development of AML or other leukemias. Experiments in animal models cogently support that a co-operation between MYC and TRIB1 exists. However, the authors did not find a correlation between the expression of these two genes in a human dataset obtained from patients with AML. Therapies based on the use of ATRA and/or arsenic trioxide (ATO, another treatment used in APL patients with PML/RARA that promotes degradation of this protein as well as of normal PML)13 have enormously improved the clinical outcome of APL patients.13,14 However, there is still a fraction of these patients that develop resistances to ATRA and ATO therapies.13,14 In this context, the results obtained by Keeshan and Kogan et al. may have interesting diagnostic/therapeutic implications. Further research should nevertheless clarify whether TRIB1 (or TRIB2) may play a role in
1130
the development of resistances in APL patients and also whether targeting TRIB1 or TRIB2 may be a therapeutic strategy to fight AML, APL or other leukemias. Research performed during the last decade has provided evidence that the Tribbles proteins are important regulators of cell function at many different levels. The recent development of new tools for the study of Tribbles biology, including the generation of Tribbles transgenic and conditional knockout mice, together with the recent establishment of novel Tribbles collaborative networks,15 should facilitate the development of additional studies and the acquisition of a more profound knowledge on the precise role played by these fascinating proteins in different physio-pathological conditions including cancer, and more specifically, leukemias. Acknowledgments Tribbles-related work at G Velasco laboratory is funded by the PI15/00339 grant, integrated into the State Plan for R & D + I2013-2016 and funded by the Instituto de salud Carlos III (ISCIII) and the European Regional Development Fund (ERDF)
References 1. Grosshans J, Wieschaus E. A genetic link between morphogenesis and cell division during formation of the ventral furrow in Drosophila. Cell. 2000;101(5):523-531. 2. Mata J, Curado S, Ephrussi A, Rorth P. Tribbles coordinates mitosis and morphogenesis in Drosophila by regulating string/CDC25 proteolysis. Cell. 2000;101(5):511-522. 3. Seher TC, Leptin M. Tribbles, a cell-cycle brake that coordinates proliferation and morphogenesis during Drosophila gastrulation. Curr Biol. 2000;10(11):623-629. 4. Kiss-Toth E. Tribbles: 'puzzling' regulators of cell signalling. Biochem Soc Trans. 2011;39(2):684-687. 5. Boudeau J, Miranda-Saavedra D, Barton GJ, Alessi DR. Emerging roles of pseudokinases. Trends Cell Biol. 2006;16(9):443-452. 6. Stein SJ, Mack EA, Rome KS, Pear WS. Tribbles in normal and malignant haematopoiesis. Biochem Soc Trans. 2015;43(5):1112-1115. 7. Salazar M, Lorente M, Orea-Soufi A, et al. Oncosuppressive functions of tribbles pseudokinase 3. Biochem Soc Trans. 2015;43(5):1122-1126. 8. Liang KL, Rishi L, Keeshan K. Tribbles in acute leukemia. Blood. 2013;121(21):4265-4270. 9. Dugast E, Kiss-Toth E, Soulillou JP, Brouard S, Ashton-Chess J. The Tribbles-1 protein in humans: roles and functions in health and disease. Curr Mol Med. 2012;13(1):80-85. 10. Jin G, Yamazaki Y, Takuwa M, et al. Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogenesis. Blood. 2007;109(9):39984005. 11. Dedhia PH, Keeshan K, Uljon S, et al. Differential ability of Tribbles family members to promote degradation of C/EBPalpha and induce acute myelogenous leukemia. Blood. 2010;116(8):1321-1328. 12. Keeshan K, Vieugué P, Chaudhury S, et al. Co-operative leukemogenesis in acute myeloid leukemia and acute promyelocytic leukemia reveal C/EBPα as a common target of TRIB1 and PML/RARA. Haematologica. 2016;101(10):1228-1236. 13. Dos Santos GA, Kats L, Pandolfi PP. Synergy against PML-RARa: targeting transcription, proteolysis, differentiation, and self-renewal in acute promyelocytic leukemia. J Exp Med. 2013;210(13):2793-2802. 14. Zeidan AM, Gore SD. New strategies in acute promyelocytic leukemia: moving to an entirely oral, chemotherapy-free upfront management approach. Clin Cancer Res. 2014;20(19):4985-4993. 15. Kiss-Toth E, Velasco G, Pear WS. Tribbles at the cross-roads. Biochem Soc Trans. 2015;43(5):1049-1050.
haematologica | 2016; 101(10)
Editorials
Matching inside and outside the HLA molecule in allogeneic hematopoietic stem cell transplantation J. Alejandro Madrigal,1 and Linda D. Barber2 1
Anthony Nolan Research Institute, Royal Free Campus and UCL Cancer Institute, London; and 2Department of Haematological Medicine, King’s College London, UK E-mail: a.madrigal@ucl.ac.uk doi:10.3324/haematol.2016.150995
A
llogeneic hematopoietic stem cell transplantation (HSCT) remains the most effective cure for many patients suffering from hematologic disorders, and more than one million hematopoietic stem cell transplants have been performed worldwide. When a human leukocyte antigen (HLA)-matched sibling donor is not available, a search for an HLA-matched unrelated donor is initiated. Due to the international effort to establish registries of potential donors (currently almost 28 million), the success of unrelated donor HSCT has improved significantly.1 Matching for HLA is a critical factor in reducing the risk of the post-transplant complications of graft failure and graft-versus-host disease (GvHD). Ideally, HLA matching for all loci (12/12) should be the gold standard; however, the HLA system is highly polymorphic, with 7897 different HLA class I (HLA-A, -B and -C) proteins and 2768 different HLA class II (HLA-DRβ, -DQβ and -DPβ) proteins currently known (from http://www.ebi.ac.uk/ipd/imgt/hla/stats.html; accessed: July 2016). It is, therefore, often necessary to transplant patients using partially HLA-mismatched unrelated donors. The strategies adopted in an attempt to limit GvHD in the HLA-mismatched setting include the use of cord blood donor cells, where HLA mismatching is better tolerated,2 or haplo-identical family members as donors with post-transplant cyclophosphamide to selectively eliminate the alloreactive T cells that cause GvHD.3 Nonetheless, standard practice remains the use of adult unrelated donors, often mismatched for one or more HLA loci. Studies of HSCT survival have provided the basis for the current recommendations from the National Marrow Donor Program (NMDP) and the Center for International Blood and Marrow Transplant Research (CIBMTR) for allele level matching unrelated donors with patients at the HLA-A, -B, -C and -DRB1 loci (8/8 match);4,5 if unavoidable, a 7/8 match can be used. Given the uncertainty regarding the impact of HLA mismatches on HSCT outcomes, strategies are being sought to help guide donor selection when several potential options are available. An algorithm has been developed for selecting favorable HLA-DPB1 mismatches based on clinical outcomes indicating a survival advantage from Tcell epitope matching.6 HLA-A and HLA-B proteins can be segregated according to expression of shared antibody epitopes known as cross-reactive groups (CREG); however, a large retrospective study showed that an HLA allele mismatch within a CREG group does not result in better transplant outcomes than a mismatch outside GREG groups.7 Similarly, HLA matchmaker is an algorithm for assessing compatibility at the antibody epitope level, but it also fails to predict outcomes after HSCT.8 Antibody epitopes are typically located on the outer surface of proteins, but the polymorphic amino acids of HLA proteins are primarily concentrated at positions in the peptide-binding haematologica | 2016; 101(10)
site. The HistoCheck scoring system was developed in an attempt to rank HLA-A, -B or -C mismatches taking into account all amino acid differences between allele mismatched pairs; however, retrospective review again showed that the strategy does not predict clinical outcomes after HSCT.9,10 In this issue of Haematologica, Lazaryan et al.11 report a new approach to assessing the impact of HLA mismatches on the success of HSCT. They performed a retrospective analysis of outcomes of 1,934 patients after myeloablative HSCT for non-lymphoid malignancies using 7/8 HLAmatched unrelated donors. The single allele mismatches were grouped according to supertypes. The six HLA-A and six HLA-B groups were based on HLA class I peptidebinding motifs using the supertype classifications described by Sidney et al.12 The polymorphic HLA residues lining the peptide-binding site determine the shape and therefore types of peptides bound. Although an HLA molecule can bind a diverse range of peptide sequences for surveillance by T cells, those bound by each allelic protein product share common motifs dictated by the shape of the peptide-binding site. Analysis of the sequences of peptides bound by HLA molecules led to identification of peptidebinding motifs and the realization that HLA molecules can be clustered into groups that bind overlapping peptide repertoires reflecting similarities in the structure of their peptide-binding sites. The five HLA-DR supertypes used by Lazaryan et al. are based on sequence and structural similarities in the peptide-binding sites defined using an algorithm developed by Doytchinova and Flower,13 and the groupings agree well with known HLA-DR peptidebinding motifs. The classification of HLA-C into two groups was based on polymorphism at residue 77 in the peptide-binding site14 that influences killer Ig-like receptor (KIR) binding. HLA-C supertypes based on peptide preferences have not been defined because less is known about the peptide-binding specificities of HLA-C.15 Of the 694 single HLA-A allele mismatches, 38% were supertype matched; of the 322 single HLA-B allele mismatches, 71% were supertype matched; of the 714 HLAC single allele mismatches, 51% were supertype matched and of the 204 HLA-DRB1 single allele mismatches, 75% were supertype matched. Mismatching of HLA-B super-
Table 1. Number of protein polymorphisms at each HLA locus.
HLA-CLASS I HLA-A 2480
HLA-B 3221
HLA-C 2196
HLA-DQβ 647
HLA-DPβ 552
HLA-CLASS II HLA-DRβ 1569
1131
Editorials
types was found to be a significant independent risk factor for grade II-IV acute GvHD, with a cumulative incidence of 67% when HLA-B supertypes were mismatched compared to 47% when HLA-B supertypes were matched (P=0.007). HLA supertype mismatching was not significantly associated with any other major post-transplant outcome. Grouping HLA alleles according to similarities in their peptide-binding motifs already has proven utility in identifying epitopes recognized by pathogen- and tumor-specific T cells, and understanding HLA associations with disease and protective immunity.12 The association of HLA-B supertype mismatching with increased GvHD risk is the first evidence indicating that knowledge of peptide-binding motif supertypes might help guide prediction of the strength of allogeneic immune responses after transplantation. There has been uncertainty regarding the molecular basis of T-cell allorecognition. Some alloreactive T cells may recognize features on the outside surface of allogeneic HLA molecules independently of the peptide inside the binding site although most are specific for single peptides.16 It is perhaps surprising that an association between peptide-binding motifs and GvHD was only seen for HLA-B supertype mismatches. The authors speculate that this may be due to higher polymorphism at the HLA-B locus (Table 1) driven by evolutionary pressures from infectious pathogens. Of note, HLA-B has diverse peptidebinding motifs covering preferences for proline or amino acids with basic, acidic, small or aliphatic properties at peptide position 2.12 In contrast, HLA-A peptide-binding motifs have more limited preferences for peptides with small, aliphatic or aromatic amino acids at position 212 and HLA class II supertypes defined by peptide binding motifs have been shown to exhibit substantial repertoire overlap.17 The peptides presented by HLA-B supertype mismatches may look more different and promote stronger alloreactive T-cell responses. Despite the large size of the single-allele 7/8 HLA mismatched dataset used in this study (collated by the CIBMTR from multiple transplant centers), the extent of HLA diversity meant that numbers of individual HLA-B supertype mismatches were small. The capacity to detect specific combinations significantly associated with GvHD was limited to the HLA-B07-B44 supertype mismatch. Findings from this study indicate that HLA-B supertype matching is beneficial, but refinement to identification of specific mismatches to avoid was not achieved. Clustering HLA alleles into supertypes based on peptide-binding motifs is an encouraging beginning, but further development of reliable criteria for selecting optimal HLA mismatches in the unrelated donor HSCT setting will be challenging.
1132
Acknowledgment The authors would like to thank Dr Bronwen Shaw for her valuable input to this paper.
References 1. Gratwohl A, Pasquini MC, Aljurf M, et al. One million haemopoietic stem-cell transplants: a retrospective observational study. Lancet Haematol. 2015;2(3):e91-100. 2. Ballen KK, Gluckman E, Broxmeyer HE. Umbilical cord blood transplantation: the first 25 years and beyond. Blood. 2013;122(4):491-498. 3. Luznik L, Jalla S, Engstrom LW, Iannone R, Fuchs EJ. Durable engraftment of major histocompatibility complex-incompatible cells after nonmyeloablative conditioning with fludarabine, low-dose total body irradiation, and posttransplantation cyclophosphamide. Blood. 2001;98(12):3456-3464. 4. Spellman SR, Eapen M, Logan BR, et al. A perspective on the selection of unrelated donors and cord blood units for transplantation. Blood. 2012;120(2):259-265. 5. Bray RA, Hurley CK, Kamani NR, et al. National marrow donor program HLA matching guidelines for unrelated adult donor hematopoietic cell transplants. Biol Blood Marrow Transplant. 2008;14(9 Suppl):45-53. 6. Shaw BE, Robinson J, Fleischhauer K, Madrigal JA, Marsh SG. Translating the HLA-DPB1 T-cell epitope-matching algorithm into clinical practice. Bone Marrow Transplant. 2013;48(12):1510-1512. 7. Wade JA, Hurley CK, Takemoto SK, et al. HLA mismatching within or outside of cross-reactive groups (CREGs) is associated with similar outcomes after unrelated hematopoietic stem cell transplantation. Blood. 2007;109(9):4064-4070. 8. Duquesnoy R, Spellman S, Haagenson M, Wang T, Horowitz MM, Oudshoorn M. HLAMatchmaker-defined triplet matching is not associated with better survival rates of patients with class I HLA allele mismatched hematopoietic cell transplants from unrelated donors. Biol Blood Marrow Transplant. 2008;14(9):1064-1071. 9. Shaw BE, Barber LD, Madrigal JA, Cleaver S, Marsh SG. Scoring for HLA matching? A clinical test of HistoCheck. Bone Marrow Transplant. 2004;34(4):367-368; author reply 369. 10. Spellman S, Klein J, Haagenson M, et al. Scoring HLA class I mismatches by HistoCheck does not predict clinical outcome in unrelated hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2012;18(5):739-746. 11. Lazaryan A WT, et al. Human leukocyte antigen supertype matching after myeloablative hematopoietic cell transplantation with 7/8 matched unrelated donor allografts: a report from the Center for International Blood and Marrow Transplant Research. Haematologica. 2016;101(10): 000-000. 12. Sidney J, Peters B, Frahm N, Brander C, Sette A. HLA class I supertypes: a revised and updated classification. BMC Immunol. 2008;9:1. 13. Doytchinova IA, Flower DR. In silico identification of supertypes for class II MHCs. J Immunol. 2005;174(11):7085-7095. 14. Doytchinova IA, Guan P, Flower DR. Identifiying human MHC supertypes using bioinformatic methods. J Immunol. 2004;172(7):4314-4323. 15. Rasmussen M, Harndahl M, Stryhn A, et al. Uncovering the peptidebinding specificities of HLA-C: a general strategy to determine the specificity of any MHC class I molecule. J Immunol. 2014;193(10):4790-4802. 16. Amir AL, van der Steen DM, Hagedoorn RS, et al. Allo-HLA-reactive T cells inducing graft-versus-host disease are single peptide specific. Blood. 2011;118(26):6733-6742. 17. Greenbaum J, Sidney J, Chung J, Brander C, Peters B, Sette A. Functional classification of class II human leukocyte antigen (HLA) molecules reveals seven different supertypes and a surprising degree of repertoire sharing across supertypes. Immunogenetics. 2011;63(6):325-335.
haematologica | 2016; 101(10)
REVIEW ARTICLE
Advanced systemic mastocytosis: from molecular and genetic progress to clinical practice Celalettin Ustun,1 Michel Arock,2,3 Hanneke C. Kluin-Nelemans,4 Andreas Reiter,5 Wolfgang R. Sperr,6,7 Tracy George,8 Hans-Peter Horny,9 Karin Hartmann,10 Karl Sotlar,9 Gandhi Damaj,11 Olivier Hermine,12,13 Srdan Verstovsek,14 Dean D. Metcalfe,15 Jason Gotlib,16 Cem Akin,17 and Peter Valent6,7
Department of Medicine, Division of Hematology, Oncology and Transplantation, University of Minnesota, Minneapolis, MN, USA; 2Molecular and Cellular Oncology, LBPA CNRS UMR8113, Ecole Normale Supérieure de Cachan, France; 3Laboratory of Hematology, PitiéSalpêtrière Hospital, Pierre et Marie Curie Paris VI University, France; 4Department of Hematology, University Medical Center Groningen, University of Groningen, The Netherlands; 5 Department of Hematology and Oncology, University Medical Centre Mannheim, Germany; 6 Department of Internal Medicine I, Division of Hematology & Hemostaseology, Medical University of Vienna, Austria; 7Ludwig Boltzmann Cluster Oncology, Medical University of Vienna, Austria; 8Hematopathology Division, University of New Mexico and TriCore Reference Laboratories, Albuquerque, NM, USA; 9Institute of Pathology, Ludwig-Maximilians-University, Munich, Germany; 10Department of Dermatology, University of Cologne, Germany; Department of Dermatology, University of Luebeck, Germany; 11Department of Clinical Hematology, Caen University Hospital, France; 12Clinical Hematology Department, Faculty of Medicine and AP-HP Necker-Enfants Malades, Paris Descartes University, France; 13Faculty of Medicine and AP-HP Necker-Enfants Malades, Mastocytosis Reference Center, Paris, France; 14Department of Leukemia, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 15Laboratory of Allergic Diseases, National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD, USA; 16 Division of Hematology, Stanford Cancer Institute, CA, USA; and 17Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA
EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
1
Haematologica 2016 Volume 101(10):1133-1143
ABSTRACT
S
ystemic mastocytosis is a heterogeneous disease characterized by the accumulation of neoplastic mast cells in the bone marrow and other organ organs/tissues. Mutations in KIT, most frequently KIT D816V, are detected in over 80% of all systemic mastocytosis patients. While most systemic mastocytosis patients suffer from an indolent disease variant, some present with more aggressive variants, collectively called “advanced systemic mastocytosis”, which include aggressive systemic mastocytosis, systemic mastocytosis with an associated hematologic, clonal non mast cell-lineage disease, and mast cell leukemia. Whereas patients with indolent systemic mastocytosis have a near normal life expectancy, patients with advanced systemic mastocytosis have a reduced life expectancy. Although cladribine and interferon-alpha are of benefit in a group of patients with advanced systemic mastocytosis, no curative therapy is available for these patients except possible allogeneic hematopoietic stem cell transplantation. Recent studies have also revealed additional somatic defects (apart from mutations in KIT) in a majority of patients with advanced systemic mastocytosis. These include TET2, SRSF2, ASXL1, RUNX1, JAK2, and/or RAS mutations, which may adversely impact prognosis and survival in particular systemic mastocytosis with an associated hematological neoplasm. In addition, several additional signaling molecules involved in the abnormal proliferation of mast cells in systemic mastocytosis have been identified. These advances have led to a better understanding of the biology of advanced systemic mastocytosis and to the development of new targeted treatment concepts. Herein, we review the biology and pathogenesis of advanced systemic mastocytosis, with a special focus on novel molecular findings as well as current and evolving therapeutic options.
haematologica | 2016; 101(10)
Correspondence: custun@umn.edu
Received: March 31, 2016. Accepted: May 25, 2016. Pre-published: no prepublication. doi:10.3324/haematol.2016.146563
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1133
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1133
C. Ustun et al.
Introduction Mastocytosis comprises a pathomorphologically and clinically heterogeneous spectrum of localized or systemic disorders characterized by an abnormal accumulation of mast cells (MCs) in one or more organs.1 In children, the disease is mostly restricted to the skin (cutaneous mastocytosis: CM).2,3 By contrast, adult patients usually present with systemic mastocytosis (SM). In patients with SM, neoplastic MCs are almost always detectable in the bone marrow (BM), and usually also in other internal organs.1,4-7 The exact incidence of SM remains uncertain, but a prevalence of mastocytosis including all the subtypes is estimated to be approximately 1 in 10,000 people.8 A recent study from Denmark showed the incidence rate for all SM, including CM, was 0.89 per 100,000/year.9
matic site.29 These latter mutations induce stabilization of the activation loop in an active conformation and/or structural alteration at the ATP-binding site of KIT, resulting in a decreased affinity for type I TK inhibitors (TKI), such as imatinib, that recognize the active conformation of a kinase. The MCL-like cell line HMC-1 has developed two sub-clones: HMC-1.1 which harbors a juxtamembrane domain (JMD) regulatory type mutation, KITV560G, and HMC-1.2 expressing both KIT D816V and KITV560G.30 Imatinib inhibits only the regulatory type mutant affecting the juxtamembrane inhibitory helix, but does not significantly inhibit KIT D816V.31 However, even some JMDtype KIT mutations (e.g. KITV559I) can cause imatinib resistance by leading to structural changes of the JMD of KIT, which affects the structure of the kinase domain.32 Other TKIs, such as PKC412 (midostaurin) effectively suppress the activity of imatinib-resistant KIT mutants.33-35 Of
The World Health Organization (WHO) classification has defined major categories and variants of SM (Online Supplementary Table S1).1,6,10 Most adult patients present with indolent SM (ISM), which is mainly characterized by mediator-related symptoms, frequent skin involvement, no organ dysfunction and a nearly normal life expectancy.1 By contrast, in advanced variants of the disease (AdvSM), including SM with an associated clonal hematologic nonMC lineage disease (SM-AHNMD; recently updated to systemic mastocytosis with an associated hematological neoplasm (SM-AHN) by WHO),11 aggressive SM (ASM), and mast cell leukemia (MCL), the malignant expansion and accumulation of neoplastic MCs can lead to organ damage (“C-findings", Online Supplementary Table S2).6,7 No skin lesions are found in some patients.12 Depending on the subtype, the survival of patients with AdvSM ranges from a few months to several years,1,13,14 therefore cytoreductive therapy is indicated in most of these patients.15 Response criteria were developed (Online Supplementary Table S2),16 and updated and detailed17 for clinical trials by a consensus group.
Molecular defects found in advanced systemic mastocytosis KIT mutations and their sensitivity to tyrosine kinase inhibitors KIT is a type III tyrosine kinase (TK) transmembrane receptor for stem cell factor (SCF), which is the major growth factor of MCs in humans (Figure 1).18 Interestingly, in most cases of SM (overall >80%, in typical ISM >90%, and in AdvSM >70%), an acquired point mutation in the gene coding for KIT (CD117) is found. Although KIT D816V, an activation loop mutation, is the most common mutation found, more than 20 other mutations in KIT have been described in SM.19,20 The exact percentages vary, depending on disease subtypes (e.g. ISM vs. ASM) and cell source [e.g. BM vs. peripheral blood (PB)].18 The KIT D816V mutation is detected in AHN cells in the majority of cases, which reflects multilineage involvement.21-23 There are, however, cases in which two independent (sub)clones exist and this might depend on the type of AHN.24,25 KIT mutations often cause ligand-independent constitutive phosphorylation and activation of KIT, which transforms cell lines from factor-dependent growth to factor independence and tumorigenicity.26-28 Longley et al. proposed to divide activating mutations of KIT into two types: “regulatory type” mutations affecting regulation of the kinase molecule, and “enzymatic pocket type” mutations, which change the amino acid sequence of the enzy1134
Figure 1. Structure of the KIT receptor and position of the major mutation (KIT D816V) found in systemic mastocytosis. The KIT gene, located on chromosome 4q12 in humans, contains 21 exons transcribed/translated into a transmembrane receptor tyrosine kinase (RTK) of 145 kDa and 976 amino acids. The Figure shows the receptor under its monomeric form, comprising 5 immunoglobulin (Ig)-like subunits in the extracellular domain (ECD) with a ligand binding site (SCF for KIT) and a dimerization site, and a cytoplasmic region with a transmembrane domain (TMD) made by a single helix. The cytoplasmic region of KIT contains an autoinhibitory juxtamembrane domain (JMD) and a kinase domain (in blue) arranged in a proximal (N-) and a distal (C-) lobe linked by a hinge region. The C-lobe of RTKs type III includes a large Kinase Insert Domain (KID) of ~ 60-100 residues. In adults, depending on the category of mastocytosis, the KIT D81V located in the phosphotransferase domain mutant (in red) is found in at least 80% of all patients, while other mutations at position 816 (in black) are less frequent by far. TK: tyrosine kinase.
haematologica | 2016; 101(10)
Treatment in advanced SM
note, the allele burden of the KIT mutant, determined by highly sensitive techniques, such as allele specific quantitative PCR (ASO-qPCR), correlates with the burden of neoplastic MCs, and with survival and prognosis.18,36,37 Finally, although the KIT D816V mutant is recurrently found in SM patients, a recent report has pointed to the possibility that such patients may present with concurrent mutations in other codons of the KIT gene.38 Indeed, out of 21 patients analyzed, the authors found 3 (15%) patients with KIT D816V and a concurrent mutation.38 Overall, these data suggest an advantage for double mutations that might contribute to the aggressiveness of SM.
Tyrosine kinase inhibitors (TKI) Midostaurin (PKC412): Midostaurin (PKC412) is an oral multi-kinase inhibitor with activity against protein kinase C (PKC), FMS-related tyrosine kinase 3 (FLT3), PDGFRA/B, vascular endothelial growth factor receptor 2 (VEGFR-2), and KIT. Midostaurin was evaluated in a centrally adjudicated, phase II multi-center international
study in 116 patients with ASM, of which 89 were evaluable for efficacy.39 Overall, 73 patients (82%) had ASM, 16 (18%) had MCL, and 63/89 patients (71%) had an AHN. Seventy-seven patients (87%) were positive for a codon 816 KIT mutation. After a median follow-up of 26 months (range 12-54 months), the overall response rate (ORR) was 60%. Most responses were major (75%), including decreases of >50% in serum tryptase and BM mast cell levels. These responses were durable: the median duration of response and median OS were 24.1 and 28.7 months, respectively. Median OS was 9.4 months in patients with MCL; however, responders in the MCL group did not reach a median OS. Midostaurin was tolerated fairly well with grade 1-2 gastrointestinal side effects being the most common adverse events (Table 1). Patient-reported outcomes, including symptoms and quality of life, measured by the Memorial Symptom Assessment Scale and the Short Form-12 Health Survey, respectively, significantly improved with midostaurin therapy. These results indicate that the drug has a favorable efficacy and safety pro-
Table 1. Treatment and outcomes in advanced SM
Author, (Reference#)
Therapy
Patient#
Vega-Ruiz(40)
Imatinib
20 with ISM or AdvSM (n=9)
Verstovsek(50)
Dasatinib
33 with ISM or AdvSM (n=15)
Midostaurin
116 with AdvSM
Gotlib(39)
Study Type
Complications
Grade IV: Thrombocytopenia 5% Neutropenia 5% Symptomatic improvement, 30% Median OS was NR No Grade IV Prospective, Phase II Grade III: Pleural effusion 21% Thrombocytopenia 18% Nausea, headache Fatigue, pain, dyspnea Grade III/IV: Neutropenia 5% Prospective, Phase II Leukopenia 4% Anemia 3% Febrile neutropenia 3% Thrombocytopenia 3% Prospective, Phase II
Non-hematologic AEs: Nausea 6%, increased lipase 4%, fatigue 4%
Kluin-Nelemans(119)Cladribine
Outcomes
10 with ISM and AdvM (n=6)
Prospective
Cytopenia Grade III/IV Lymphopenia 82% Neutropenia 47% Infections 13% TRM at 6 months: 11%
Barete(123)
Cladribine
68 with ISM and AdvSM (n=32)
Retrospective, registry study with a long follow-up (>10 years)
Ustun(125)
Allo-HCT
57 with Adv SM
Retrospective
CR, 5%
ORR, 33% CR, 6.6% Median OS was NR
ORR, 60% MR, 75% IR, 36% PCR, 28% Unspecified, 11%. Good PR, 21% Minor PR, 4% Median OS All, 29 mos MCL 9.4 mos Median duration of response 24 mos All patients responded, no CR Median OS was NR ORR, 72% , No CR ORR in AdvSM 50% Median duration of response 44 mos OS at 3 years: 57% SM-AHN: 74% ASM: 43% MCL: 17% DFS at 3 years: 51% SM-AHN: 63% ASM: 43% MCL: 17%
AHN: associated hematological neoplasm; AdvSM: advanced systemic mastocytosis; AE: adverse event; Allo-HCT: allogeneic hematopoietic cell transplantation; CR: complete remission; DFS: disease-free survival; IR: incomplete remission; ISM: indolent systemic mastocytosis; MCL: mast cell leukemia; Mos, months; MR: major response; NR, not reported; ORR: overall response rate; OS: overall survival; PCR, pure clinical response; PR: partial response; SM: systemic mastocytosis; TRM: transplant-related mortality.
haematologica | 2016; 101(10)
1135
C. Ustun et al.
file with activity in AdvSM regardless of KIT mutation status. Although midostaurin has not been approved by major drug authorities in either Europe or the USA, it is available for patients with AdvSM within a compassionate use program sponsored by the drug company. Imatinib: After the remarkable success of TKIs in chronic myeloid leukemia (CML), significant enthusiasm for TKI in the treatment of SM emerged in the early 2000s.19 However, imatinib is largely ineffective in patients with KIT D816V+ SM.40 On the other hand, some patients with SM may respond very well to imatinib, especially those with other KIT mutations such as K509I,41 F522C42 or KIT WT.43 In patients with FIP1L1-PDGFRA-positive myeloid neoplasms with eosinophilia, small doses of imatinib (100 mg/d) will effect durable hematologic and cytogenetic/molecular remission in almost all cases.44.45 Although some of these patients may exhibit scattered/interstitial distributions of increased abnormal CD25+ MCs in the BM, these cases are not considered a subtype of SM by the WHO because typical dense infiltrates of spindle-shaped mast cells are missing. In SM patients with KIT WT, imatinib may even induce CR with the disappearance of skin lesions and return of elevated serum tryptase levels to the normal range (<15 ng/mL).46 Imatinib (400 mg daily) is still the only TKI approved by the US Food and Drug Administration (FDA) for adult patients who have ASM either without the KIT D816V mutation or with unknown KIT mutational status. Dasatinib: Dasatinib, a multikinase inhibitor (e.g. BCRABL1, KIT, and PDGFRα),47,48 has proven to be effective in vitro against KIT D816V+ neoplastic MCs.49 However, the half-life of the drug is very short, and no durable and meaningful clinical responses were observed in clinical studies in AdvSM (Table 1).50 Masitinib: Masitinib, which inhibits KIT WT and LYN,51 is an effective drug for canine MC tumors.52 However, in humans, the KIT D816V mutation introduces resistance against masitinib. In one study, masitinib was administered daily (3-6 mg orally) for 12 weeks in 25 patients diagnosed as having SM or CM with a related “handicap” (i.e. disabilities associated with flushes, depression, pruritus and quality of life).53 ORR was 56% by AFIRMM response criteria.54 Severe toxicities occurred in <10% of all patients. Currently, a larger study is being performed in patients with CM and ISM with a “handicap” (AdvSM was excluded). Nilotinib: In a phase II trial of 61 patients with SM (37 with AdvSM), nilotinib (400 mg twice a day) induced overall responses of 21.6% (including a decrease in serum tryptase and BM mast cells) and of 21% in ASM.55 All responders had the KIT D816V mutation. Nine AdvSM patients died during 34.7 months of follow-up. No active study is currently being performed with nilotinib. Other targeted small-molecule inhibitors: Most of the data on these TKIs resulted from pre-clinical studies or case reports. Ponatinib, a multi-kinase blocker, inhibits the kinase activity of KITV560G and, less effectively, KIT D816V in HMC-1 cells.35,56 Ponatinib induced dose-dependent growth inhibition and apoptosis in primary neoplastic MCs, HMC-1.1 cells, and HMC-1.2 cells.56 Ponatinib and midostaurin were found to exert synergistic growthinhibitory effects against neoplastic MCs harboring the KIT D816V mutant.56 Other novel TKIs with potent TKI inhibiting properties (e.g. EXEL-0862)57 and thiazole amine 1136
derivatives inhibiting β-catenin signaling (e.g. semaxinib (SU5416) and compound 126332).58,59 BLU-285, a selective KIT D816V inhibitor with encouraging pre-clinical activity and a narrow target profile, is expected to enter clinical trial testing in AdvSM in the near future.60
Progress in somatic mutations other than KIT in SM Recent studies have reported the presence of additional, recurrent somatic mutations (apart from KIT mutations) in AdvSM, especially in SM-AHN, including mutations in TET2, SRSF2, ASXL1, RUNX1, JAK2, and/or RAS (Figure 2).61-65 Mutations in TET2, also detected in healthy individuals,66 cause loss of function (i.e. regulating gene expression at the cellular level),67 and are associated with increased self-renewal capacity of hematopoietic stem cells.68 Recently, several investigators have identified TET2 mutations scattered across several of its 12 exons in 1 or both TET2 alleles, as an early event during the development of various malignancies.69 Patients with mutant TET2+ myeloid disorders show a decreased level of 5-hmC with hypomethylation or hypermethylation of DNA.70 Altogether these data show that TET2 plays a role in various hematologic malignancies. In line with these recently published data, TET2 mutations have been reported in 2040% of KIT D816V-positive AdvSM patients.25,61,62,64 The cooperation between KIT D816V and loss of function of TET2 in MC results in transformation to a more aggressive disease phenotype in mice.71 It has also been suggested that TET2 mutations can occur before KIT D816V in ASMAHN patients.72 Thus, the acquisition of KIT D816V might act as a phenotype modifier of ASM in these cases.72 Patients carrying a combination of TET2 and DNMT3A (a DNA methyltransferase) mutations have a poor prognosis compared to those with wild-type genes.61 In vitro, a combination of dasatinib and decitabine (a hypomethylating agent) was more effective at inducing apoptosis and cell death in HMC-1.2 cells harboring a TET2 mutant compared to each compound alone.71 This combination also had less effect in TET2 wild-type cells due to a lower efficacy of decitabine. The impact of TET2 mutations on overall survival remains uncertain.61,62,73 The spliceosome machinery includes SRSF2, U2AF1, and SF3B1 proteins, and is involved in the removal of introns from a transcribed pre-mRNA.74 Mutations in the spliceosome machinery have recently been identified using whole exome/genome technologies in MDS and MPN.75 A mutation in the hotspot region of SRSF2 (codon P95) is found in approximately 1/3 of AdvSM patients64,65 but is usually not detectable in patients with ISM or SSM.25 It is more common in ASM-AHN25,64,65 and precedes KIT D816V in these patients.25 The frequency of SF3B1 mutations in AdvSM is low, ranging from 0 to 5%.64,65 U2AF1 mutations are less frequently reported in SM.64,65 The gene ASXL1 (additional sex combs–like 1) encodes for a protein of the polycomb group and trithorax complex family, which interacts with retinoic acid receptor and may be involved in chromatin remodeling.76 The presence of ASXL1 mutations has been reported in SM at various frequencies,25,61,64,73 and alone or with other mutations seems to be a poor prognostic factor for OS in patients.61,64,73 RUNX1, and less frequently, JAK2 mutations, haematologica | 2016; 101(10)
Treatment in advanced SM
are found in AdvSM, but not in ISM or SSM.64 The frequency of RAS mutations (e.g. NRAS, KRAS or HRAS) in SM has been investigated, with KRAS and NRAS mutants being found in AdvSM at a relatively low frequency, and not usually detectable in patients with ISM.63,64 The presence of additional genetic defects in KIT D816V+ AdvSM patients may confer adverse prognosis as compared with patients without such abnormalities.64,72 In a recent study, Jawhar et al. have analyzed the impact of several additional defects on 70 multi-mutated KIT D816V+ patients with an AHN.77 In this study, the most frequently identified mutated genes were TET2 (n=33 of 70 patients), SRSF2 (n=30), ASXL1 (n=20), RUNX1 (n=16) and JAK2 (n=11).77 In multivariate analysis, SRSF2 and ASXL1 remained the most predictive adverse indicators concerning OS. Furthermore, the authors found that inferior OS and adverse clinical characteristics were significantly influenced by the number of mutated genes in the SRSF2/ASXL1/RUNX1 (S/A/R) panel (P<0.0001).77 It appears that, based on these findings, the inclusion of molecular markers should be considered in upcoming prognostic scoring systems for patients with SM. This might be particularly important for patients with SMAHN given that most of these studies were done in patients with SM-AHN.73,78 Although it is arguable that these mutations could be detected due to the copresence of an AHN, there are recent studies in pure SM showing these mutations as well.72,79 In addition, it has been described in many previous reports that KIT mutations are not restricted to the mast cell disease components in SMAHN.80 Although we are at an early stage in the understanding of the clinical and biological importance of these mutations in SM, most likely these mutations affect hematopoietic stem and progenitor cells, and the rate of multilineage involvement increases with the aggressiveness of SM. In addition, recent investigations on mutational profiles of colonies grown from granulocyte-macrophage colonyforming progenitor cells (CFU-GM) and microdissected mature cells (tryptase or CD15 positive) revealed that these additional mutations develop prior to KIT D816V in almost all patients, indicating a multi-mutated stem cell disease with strong phenotype modification (i.e. the mastocytosis component) driven by KIT D816V.25
Critical intracellular pro-oncogenic pathways in neoplastic mast cells as novel potential therapeutic targets Several studies have reported that the ability of wildtype and oncogenic mutant forms of KIT to induce signal transduction differs not only quantitatively but also qualitatively. These altered pathways, which are presented in Figure 3 together with potential targeted drugs, may have an effect on several properties of neoplastic MCs by reducing apoptosis and/or by inducing alterations in the cell cycle. MCL-1, a BCL-2 family member with anti-apoptotic properties, is expressed in primary neoplastic MCs in SM as well as in the HMC-1.1 and HMC-1.2 cell lines.81 The targeting of MCL-1 by antisense oligonucleotides (ASOs) or MCL-1-specific siRNA resulted in reduced survival and haematologica | 2016; 101(10)
Figure 2. Synthesis of the frequency of the various molecular defects found in AdvSM, which sums up all the advanced (AdvSM) SM patients (n=122) reported in the studies by Tefferi et al.,62 Wilson et al.,63 Traina et al.,61 Schwaab et al.64 and Hanssens et al.65 The frequency (%) of cases found positive for each genetic defect is represented in red, whereas the frequency of patients for whom the corresponding defect was not tested is represented in blue.
increased apoptosis in these cell lines.81 Moreover, MCL-1 ASOs cooperated with various KIT-targeting TKIs in producing growth inhibition in neoplastic MC lines.81 BIM, a pro-apoptotic member of the BCL-2 family, has been identified as a tumor suppressor in neoplastic MCs.82 BIM is downregulated in neoplastic MCs by SCF as well as by KIT D816V.82 Midostaurin, bortezomib (a proteasome inhibitor), and obatoclax (a pan-BCL-2 family blocker) reportedly upregulate BIM expression in HMC-1 cells and may thereby promote apoptosis.82,83 Obatoclax also increased apoptosis in these cells.83 Activated LYN and BTK are expressed in neoplastic MCs in a KIT-independent manner in patients with ASM and MCL, and may thus contribute to malignant transformation.49 LYN is a member of the SRC family involved in cellular signaling processes regulating growth, differentiation, and apoptosis. Activated LYN regulates BTK function and may influence the process of degranulation and cytokine production in MCs.84,85 Dasatinib and bosutinib (SRC inhibitors) disrupt LYN and BTK activation and oncogenic signaling in neoplastic MCs.49 Bosutinib inhibits the growth of neoplastic MCs in vitro at relatively high concentrations, with no effect on KIT.49,86 Bosutinib acts synergistically with midostaurin on HMC-1 cell proliferation.49 However, bosutinib is unable to induce any response in patients with AdvSM.87 1137
C. Ustun et al.
Figure 3. Intracellular pathways involved in the accumulation/proliferation of neoplastic mast cells in SM and agents which could be potentially used to target one or the other of these molecules. That KIT D816V dimerizes spontaneously with itself or with KIT WT, or is capable of transmitting oncogenic signals as a single molecule, remains largely unexplored. However, it has been postulated whether the KIT D816V protein could activate substrates under a monomeric form and could even be located in the cell cytoplasm. The KIT D816V oncogenic mutation alters the substrate specificity of the mutant protein, which shows a substrate specificity resembling that of SRC and ABL TKs. In addition, FES TK is activated by mutant KIT protein and negatively regulates the STAT pathway, although it induced phosphorylation of mTOR. Furthermore, AKT activation has been identified as a key signaling molecule involved in KIT D816Vdependent differentiation and growth of neoplastic MCs. Also, STAT5 is believed to play a pivotal role in the growth of KIT D816V+ neoplastic MCs and is constitutively phosphorylated in such cells, probably because KIT D816V can promote direct STAT5 activation, thus diverting the canonical JAK-STAT pathway. A number of drugs (in red and in italics) can potentially selectively inhibit some of these critical pathways. Red arrows: inhibition; black arrows: induction of survival or functions; green arrows: activation of signaling pathways; dark blue arrow: induction of increased synthesis.
Phosphoinositide 3-kinase (PI3-K), a lipid kinase, is important for the function of intracellular signaling molecules, like BTK, AKT and PDK1, by inducing phosphatidylinositol 3,4,5-trisphosphate (PIP3) that provides membrane docking sites for these signaling molecules.88 In both HMC-1 subclones (HMC-1.1 and HMC-1.2), mutated KIT leads to constitutive activation of PI3-K.89 Once activated, the PI3-K subsequently activates AKT,89 a key signaling molecule involved in KIT-dependent differentiation and growth of neoplastic MCs harboring oncogenic KIT mutants.90 Indeed, AKT was found to be phosphorylated in neoplastic MCs in patients with KIT D816V+ SM and in the HMC-1.2 cell line.90,91 PI3-K and AKT are also important for the regulation of the mammalian target of rapamycin (mTOR), a serine/threonine kinase that interacts with 2 regulatory protein complexes called mTOR complex 1 (mTORC1) and complex 2 (mTORC2). PI3-K regulates the mTORC1 pathway via the activation of AKT which directly inactivates tuberin, the inhibitor of mTOR activation. Once activated, mTORC1 phosphorylates p70 ribosomal S6 kinase (p70S6K), resulting in increased gene transcription that regulates cell growth, survival, protein synthesis and 1138
metabolism. Smrz et al. showed that the expression and activation of mTORC1 and mTORC2 was increased in neoplastic human MC lines and in immature normal MCs, as compared with mature normal MCs.92 Interestingly, the authors demonstrated that mTORC1 might contribute to MC survival, while mTORC2 might only fulfill critical functions in the context of proliferating (dividing) neoplastic and immature MCs.92 Rapamycin, a specific inhibitor of mTORC1, has been shown to block FcÎľRI- and KITinduced mTORC1-dependent p70S6K phosphorylation in normal MCs.88 Furthermore, BEZ235, a dual PI3-K/mTOR blocker, exerted strong growth-inhibitory effects on neoplastic MCs in vitro.93 Of note, BEZ235 was also found to reduce the engraftment and growth of HMC-1 cells in a xenotransplanted mouse model employing NMR1Foxn1(nu) mice.93 Everolimus, another mTOR-blocker, was ineffective in patients with SM.94 Neoplastic MCs express cytoplasmic and nuclear phospho-STAT5 (pSTAT5).95 In an in vitro study,90 knockdown of STAT5 was followed by growth inhibition of neoplastic MCs. Furthermore, it has been shown that KIT D816V directly promotes STAT5-activation, and that pSTAT5 contributes to the growth of neoplastic MCs.95 This makes haematologica | 2016; 101(10)
Treatment in advanced SM
Figure 4. Proposed algorithm for the diagnosis and classification of the different categories of SM patients and for the preferable therapeutic options adapted to each category of the disease. AHN: associated hematological neoplasm; Allo-HSCT: allogeneic hematopoietic stem cell transplantation; ASM: aggressive systemic mastocytosis; BM: bone marrow; 2-CdA: cladribine; GC: glucocorticoids; HR: histamine receptors; IFN-Îą: interferon-alpha; IM: imatinib mesylate; ISM: indolent systemic mastocytosis; MCL: mast cell leukemia; PKC412: midostaurin; SM: systemic mastocytosis; WT: wild-type.
STAT5 an attractive target for therapy in AdvSM. However, until now, most drugs targeting STAT5 exert anti-neoplastic effects only at high, non-pharmacological concentrations in vitro. The inhibition of the JAK-STAT signaling pathway in vitro decreased KIT D816V-mediated cell growth.96 Ruxolitinib, a JAK1/2 inhibitor, has shown clinical benefit in patients with MPN regardless of JAK2 V617F-mutation.97 Ruxolitinib decreased spleen size and improved blood counts in a KIT-mutated but not JAK2mutated patient with SM-MPN primary myelofibrosis.98 Therefore, JAK1/2 blockers can be considered in studies of patients with SM-MPN. NF-kB, a dimeric transcription factor of the REL family, was found to be spontaneously activated in HMC-1 cells.99 IMD-0354 inhibited translocation of NF-kB to the nucleus, and thus led to decreased cyclin D3 expression and increased cell cycle arrest in HMC-1 cells in vitro.99 Another transcription factor of the REL family, nuclear factor of haematologica | 2016; 101(10)
activated T cells (NFAT), has also been found constitutively activated in KIT-mutated neoplastic MCs.100 The combination of a KIT inhibitor and of a calcineurin phosphatase inhibitor (a NFAT regulator) exhibited a synergistic inhibitory effect on cell viability and survival in KITmutated MC lines.100 One promising class of targets within chromatin regulatory molecules and related antigens are the bromodomain (BRD)-containing proteins.101-103 Indeed, inhibition of the epigenetic reader bromodomain-containing protein-4 (BRD4) by exposure to RNA interference or treatment with JQ1, a drug blocking the specific interactions between BRD4 and acetylated histones, resulted in major antileukemic effects in murine and human AML cells.102 More recently, BRD4 has been identified as a novel drug target in AdvSM.104 The authors showed that neoplastic MCs expressed substantial amounts of BRD4 in ASM and MCL, as assessed by immunohistochemistry and PCR.104 1139
C. Ustun et al.
They also reported that the human MCL lines HMC-1 and ROSA also expressed BRD4, and that a BRD4-specific short hairpin RNA or the BRD4-targeting drug JQ1 induced dose-dependent growth inhibition and apoptosis in HMC-1 and ROSA cells, regardless of the presence or absence of the KIT D816V mutant.104 Moreover, the authors demonstrated that JQ1 suppressed the proliferation of primary neoplastic MCs obtained from patients with ASM or MCL. Finally, in drug combination experiments, midostaurin (PKC412) and all-trans retinoic acids were found by the authors to cooperate with JQ1 in producing synergistic effects on survival in HMC-1 and ROSA cells.104 Taken together, these data identified BRD4 as a promising drug target in advanced SM. However, whether JQ1 or other BET bromodomain inhibitors are effective in vivo in patients with AdvSM remains to be elucidated.
Antibody-mediated therapeutic approach to target neoplastic mast cells and stem cells Based on recent knowledge on the phenotype of malignant MCs and their neoplastic progenitors, a number of cell surface antigens might be aberrantly expressed, including CD13, CD25, CD30, CD33, CD44, CD52, CD87, and CD117, and therefore might be considered also as potential targets of therapy in AdvSM.105-110 For example, neoplastic MCs and their progenitors have been shown to respond in vitro to gemtuzumab ozogamicin (a monoclonal antibody targeting CD33 combined to a cytostatic agent).111 The CD52-targeting antibody alemtuzumab induces cell death in neoplastic MCs in vitro and in mice xenotransplanted with HMC-1 cells.108 CD30 is expressed on the surface of neoplastic MCs in a proportion of patients with AdvSM, but not on normal/reactive MCs, making this antigen an attractive target of specific therapy in these patients.107,112,113 A single-arm, open-label clinical trial applying brentuximab vedotin (SGN-35) to patients with CD30-positive AdvSM (clinicaltrials.gov identifier: 01807598) is ongoing in the US. Neoplastic (leukemic) stem cells (LSCs) have recently been identified in AdvSM. These cells reside within a CD34+ cell fraction and coexpress aminopeptidase N (CD13), leukosialin (CD43), Pgp-1 (CD44), the IL-3R Îą-chain (CD123), AC133 (CD133), CXCR4 (CD184), CD33, CD52 and CD117.114,115 As observed in chronic myeloid leukemia, a part of these LSCs might be non-cycling and therefore probably resistant to treatment with TKIs. Thus, a combination of a TKI that targets KIT on neoplastic MCs and a mAb targeting a surface antigen, such as CD52 for instance, expressed on non-cycling LSCs, may help to achieve a minimal residual disease negative state in AdvSM.
Conventional therapies with anti-neoplastic drugs and allogeneic hematopoietic cell transplantation Cytarabine, fludarabine, hydroxyurea (a drug of choice in palliative care)15 and interferon-alpha (IFNâ&#x2C6;&#x2019;Îą),116-118 have been frequently used for cytoreduction in the treatment of AdvSM. Hydroxyurea is useful to control leukocyte counts in AdvSM, especially in SM-AHN (palliative therapy) and in patients with comorbidity. Cladribine (2-CDA) is the most effective and frequently used drug. KluinNelemans et al., used 2-CdA in 10 patients with SM, most of them suffering from AdvSM (Table 1).119 All patients responded concerning clinical symptoms and MC burden as reflected in declining serum tryptase values and urinary histamine metabolite excretion. Although no patient 1140
achieved a complete remission (CR), clinically meaningful and some durable responses were seen, suggesting that 2CdA may be a potentially effective treatment option for some patients with severe SM.119 These results have been supported by more recent studies.120-123 For instance, in a study on 44 SM patients, the median duration of response was 20 months; however, none of the patients with SMAHN responded.122 However, 2-CdA usually does not control the disease for prolonged periods of time in rapidly progressing ASM and MCL. For these patients, more intensive therapy, such as AML-like multi-agent chemotherapy, including fludarabine and cytarabine124 should be considered in induction therapy and then for allogeneic hematopoietic cell transplantation (HCT) for consolidation therapy.125-127 Allogeneic HCT remains the only potentially curative treatment option for patients with AdvSM. We have recently reported data on the effect of allo-HCT in patients with AdvSM (Table 1).125 Most patients (the median age was 46) received a graft from HLA-identical siblings (n=34) or unrelated donors (URD) (n=17). Overall survival (OS) and SM progression-free survival (PFS) at 3 years for all patients were 57% and 51%, respectively. They were significantly affected, however, by the type of advanced SM: 74% and 63%, respectively, for SM-AHN; 43% and 43%, respectively, for ASM; and 17% and 17%, respectively, for MCL. Although the data presented are very encouraging, future prospective studies, perhaps per recommended consensus opinion to homogenously collect data,128 are required to confirm the safety129 and efficacy of this treatment approach in AdvSM.
Miscellaneous aspects of management in AdvSM Patients with SM-AHN should be treated according to generally accepted guidelines: the SM component of the disease is treated as if no AHN was diagnosed, and the AHN component of the disease is treated as if no SM was diagnosed, with the recognition of potential drug interactions10.15 deciding whether the SM or AHN component is primarily contributing to organ damage or other related clinical, laboratory concerns. However, admittedly it is often not possible to clearly delineate whether one or the other component is responsible for the clinical issues/organ damage. As a supportive therapy, H1-receptor antagonists, such as the classical antihistamine hydroxyzine, or non-sedating antihistamines, such as loratadine or fexofenadine, can be administered for the alleviation of symptoms caused by the release of the mediators (e.g. pruritus and flushing).130-132
Conclusion and perspectives Advanced variants of SM share two major characteristics: i) the prognosis of the disease remains poor, and ii) other than allogeneic HCT no curative therapy is available. Only a few drugs have shown beneficial effects in AdvSM (2-CdA, interferon-alpha, and midostaurin). We propose a treatment algorithm with current therapy options (Figure 4). However, this is a subject to change in the future due to remarkable progress in the biology of AdvSM. Neoplastic cells in SM are usually driven by a canonical KIT-downstream pathway as well as by addihaematologica | 2016; 101(10)
Treatment in advanced SM
tional somatic mutations and KIT-independent pathways and molecules, including TET2, the spliceosome machinery, ASXL1, or RAS. We may better prognosticate AdvSM using these additional genetic defects. The PI3-kinase, AKT, STAT-5, BTK, FES, mTORC2, and BCL-2 family members as well as certain surface molecules and disease initiating (quiescent) neoplastic stem cells can be a target for therapies in the future. Potentially, studies will combine the most effective targeted drugs with one another and/or with conventional chemotherapy options to improve patient survival.
References 1. Valent P, Horny HP, Escribano L, et al. Diagnostic criteria and classification of mastocytosis: a consensus proposal. Leuk Res. 2001;25(7):603-625. 2. Hartmann K, Metcalfe DD. Pediatric mastocytosis. Hematol Oncol Clin North Am. 2000;14(3):625-640. 3. Wiechers T, Rabenhorst A, Schick T, et al. Large maculopapular cutaneous lesions are associated with favorable outcome in childhood-onset mastocytosis. J Allergy Clin Immunol. 2015;136(6):1581-1590. 4. Robyn J, Metcalfe DD. Systemic mastocytosis. Adv Immunol. 2006;89:169-243. 5. Escribano L, Alvarez-Twose I, SanchezMunoz L, et al. Prognosis in adult indolent systemic mastocytosis: a long-term study of the Spanish Network on Mastocytosis in a series of 145 patients. J Allergy Clin Immunol. 2009;124(3):514-521. 6. Horny HP AC, Metcalfe DD, Escribano L, Bennett JM, Valent P, Bain BJ. Mastocytosis (Mast cell disease). . World Health Organization (WHO) Classification of Tumours Pathology & Genetics Tumours of Haematopoietic and Lymphoid TissuesLyon: IARC Press. 2008:54-63. 7. Gotlib J, Pardanani A, Akin C, et al. International Working GroupMyeloproliferative Neoplasms Research and Treatment (IWG-MRT) & European Competence Network on Mastocytosis (ECNM) consensus response criteria in advanced systemic mastocytosis. Blood. 2013;121(13):2393-2401. 8. Brockow K. Epidemiology, prognosis, and risk factors in mastocytosis. Immunol Allergy Clin North Am. 2014;34(2):283-295. 9. Cohen SS, Skovbo S, Vestergaard H, et al. Epidemiology of systemic mastocytosis in Denmark. Br J Haematol. 2014;166(4):521528. 10. Valent P, Akin C, Escribano L, et al. Standards and standardization in mastocytosis: consensus statements on diagnostics, treatment recommendations and response criteria. Eur J Clin Invest. 2007;37(6):435453. 11. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia. Blood. 2016 Apr 11. pii: blood-2016-03643544. [Epub ahead of print]. 12. Hein MS, Hansen L. Aggressive systemic mastocytosis: a case report and brief review of the literature. S D J Med. 2005;58(3):95100. 13. Lim KH, Tefferi A, Lasho TL, et al. Systemic
haematologica | 2016; 101(10)
14.
15. 16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Acknowledgments The authors thank Sabrina Porter for her assistance in the manuscript preparation. M. Arock is supported by Fondation de France. J. Gotlib is supported by the Charles and Ann Johnson Foundation. O. Hermine is supported by Agence nationale pour la recherche (ANR) and by Fondation pour la recherche médicale (FRM). P. Valent is supported by Austrian Science Funds (FWF) Projects SFB F4611 and SFB F4704-B20. A. Reiter is supported by research grants from the ‘Deutsche José Carreras LeukämieStiftung e.V. (H 11/03 and R 13/05) ). D. D. Metcalfe is supported by the Division of Intramural Research, NIAID/NIH.
mastocytosis in 342 consecutive adults: survival studies and prognostic factors. Blood. 2009;113(23):5727-5736. Pardanani A, Lim KH, Lasho TL, et al. Prognostically relevant breakdown of 123 patients with systemic mastocytosis associated with other myeloid malignancies. Blood. 2009;114(18):3769-3772. Valent P, Sperr WR, Akin C. How I treat patients with advanced systemic mastocytosis. Blood. 2010;116(26):5812-5817. Valent P, Akin C, Sperr WR, et al. Aggressive systemic mastocytosis and related mast cell disorders: current treatment options and proposed response criteria. Leuk Res. 2003;27(7):635-641. Gotlib J, Pardanani A, Akin C, et al. International Working GroupMyeloproliferative Neoplasms Research and Treatment (IWG-MRT) & European Competence Network on Mastocytosis (ECNM) consensus response criteria in advanced systemic mastocytosis. Blood. 2013;121(13):2393-2401. Arock M, Sotlar K, Akin C, et al. KIT mutation analysis in mast cell neoplasms: recommendations of the European competence network on mastocytosis. Leukemia. 2015;29(6):1223-1232. Ustun C, DeRemer DL, Akin C. Tyrosine kinase inhibitors in the treatment of systemic mastocytosis. Leuk Res. 2011;35(9): 1143-1152. Bibi S, Langenfeld F, Jeanningros S, et al. Molecular defects in mastocytosis: KIT and beyond KIT. Immunol Allergy Clin North Am. 2014;34(2):239-262. Pullarkat VA, Bueso-Ramos C, Lai R, et al. Systemic mastocytosis with associated clonal hematological non-mast-cell lineage disease: analysis of clinicopathologic features and activating c-kit mutations. Am J Hematol. 2003;73(1):12-17. Nagai S, Ichikawa M, Takahashi T, et al. The origin of neoplastic mast cells in systemic mastocytosis with AML1/ETO-positive acute myeloid leukemia. Exp Hematol. 2007;35(11):1747-1752. Ustun C, Corless CL, Savage N, et al. Chemotherapy and dasatinib induce longterm hematologic and molecular remission in systemic mastocytosis with acute myeloid leukemia with KIT D816V. Leuk Res. 2009;33(5):735-741. Sotlar K, Colak S, Bache A, et al. Variable presence of KITD816V in clonal haematological non-mast cell lineage diseases associated with systemic mastocytosis (SMAHNMD). J Pathol. 2010;220(5):586-595. Jawhar M, Schwaab J, Schnittger S, et al. Molecular profiling of myeloid progenitor cells in multi-mutated advanced systemic
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
mastocytosis identifies KIT D816V as a distinct and late event. Leukemia. 2015;29(5): 1115-1122. Hashimoto K, Tsujimura T, Moriyama Y, et al. Transforming and differentiation-inducing potential of constitutively activated c-kit mutant genes in the IC-2 murine interleukin3-dependent mast cell line. Am J Pathol. 1996;148(1):189-200. Tsujimura T, Hashimoto K, Kitayama H, et al. Activating mutation in the catalytic domain of c-kit elicits hematopoietic transformation by receptor self-association not at the ligand-induced dimerization site. Blood. 1999;93(4):1319-1329. Saleh R, Wedeh G, Herrmann H, et al. A new human mast cell line expressing a functional IgE receptor converts to factor-independence and tumorigenicity by KIT D816V-transfection. Blood. 2014;124(1):111120. Longley BJ, Reguera MJ, Ma Y. Classes of cKIT activating mutations: proposed mechanisms of action and implications for disease classification and therapy. Leuk Res. 2001;25(7):571-576. Furitsu T, Tsujimura T, Tono T, et al. Identification of mutations in the coding sequence of the proto-oncogene c-kit in a human mast cell leukemia cell line causing ligand-independent activation of c-kit product. J Clin Invest. 1993;92(4):1736-1744. Ma YS, Zeng S, Metcalfe DD, et al. The cKIT mutation causing human mastocytosis is resistant to STI571 and other KIT kinase inhibitors; kinases with enzymatic site mutations show different inhibitor sensitivity profiles than wild-type kinases and those with regulatory-type mutations. Blood. 2002;99(5):1741-1744. Nakagomi N, Hirota S. Juxtamembrane-type c-kit gene mutation found in aggressive systemic mastocytosis induces imatinib-resistant constitutive KIT activation. Lab Invest. 2007;87(4):365-371. Gleixner KV, Mayerhofer M, Sonneck K, et al. Synergistic growth-inhibitory effects of two tyrosine kinase inhibitors, dasatinib and PKC412, on neoplastic mast cells expressing the D816V-mutated oncogenic variant of KIT. Blood. 2006;108(11):159a-159a. Gleixner KV, Mayerhofer M, Sonneck K, et al. Synergistic growth-inhibitory effects of two tyrosine kinase inhibitors, dasatinib and PKC412, on neoplastic mast cells expressing the D816V-mutated oncogenic variant of KIT. Haematologica. 2007;92(11):1451-1459. Jin B, Ding K, Pan J. Ponatinib induces apoptosis in imatinib-resistant human mast cells by dephosphorylating mutant D816V KIT and silencing beta-catenin signaling. Mol Cancer Ther. 2014;13(5):1217-1230.
1141
C. Ustun et al. 36. Erben P, Schwaab J, Metzgeroth G, et al. The KIT D816V expressed allele burden for diagnosis and disease monitoring of systemic mastocytosis. Ann Hematol. 2014;93(1):8188. 37. Hoermann G, Gleixner KV, Dinu GE, et al. The KIT D816V allele burden predicts survival in patients with mastocytosis and correlates with the WHO type of the disease. Allergy. 2014;69(6):810-813. 38. Lasho T, Finke C, Zblewski D, et al. Concurrent activating KIT mutations in systemic mastocytosis. Br J Haematol. 2016;173 (1):153-156. 39. Gotlib J, Kluin-Nelemans HC, George TI, et al. Midostaurin (PKC412) Demonstrates a High Rate of Durable Responses in Patients with Advanced Systemic Mastocytosis: Results from the Fully Accrued Global Phase 2 CPKC412D2201 Trial. Blood, 2014:636636. 40. Vega-Ruiz A, Cortes JE, Sever M, et al. Phase II study of imatinib mesylate as therapy for patients with systemic mastocytosis. Leuk Res. 2009;33(11):1481-1484. 41. de Melo Campos P, Machado-Neto JA, Scopim-Ribeiro R, et al. Familial systemic mastocytosis with germline KIT K509I mutation is sensitive to treatment with imatinib, dasatinib and PKC412. Leuk Res. 2014;38(10):1245-1251. 42. Akin C, Fumo G, Yavuz AS, et al. A novel form of mastocytosis associated with a transmembrane c-kit mutation and response to imatinib. Blood. 2004;103(8):3222-3225. 43. Alvarez-Twose I, Gonzalez P, Morgado JM, et al. Complete response after imatinib mesylate therapy in a patient with well-differentiated systemic mastocytosis. J Clin Oncol. 2012;30(12):e126-129. 44. Tefferi A, Lasho TL, Brockman SR, et al. FIP1L1-PDGFRA and c-kit D816V mutationbased clonality studies in systemic mast cell disease associated with eosinophilia. Haematologica. 2004;89(7):871-873. 45. Klion AD, Robyn J, Akin C, et al. Molecular remission and reversal of myelofibrosis in response to imatinib mesylate treatment in patients with the myeloproliferative variant of hypereosinophilic syndrome. Blood. 2004;103(2):473-478. 46. Horny HP, Sotlar K, Valent P. Mastocytosis: immunophenotypical features of the transformed mast cells are unique among hematopoietic cells. Immunol Allergy Clin North Am. 2014;34(2):315-321. 47. Shah NP, Lee FY, Luo R, et al. Dasatinib (BMS-354825) inhibits KITD816V, an imatinib-resistant activating mutation that triggers neoplastic growth in most patients with systemic mastocytosis. Blood. 2006;108(1):286-291. 48. Olivieri A, Manzione L. Dasatinib: a new step in molecular target therapy. Ann Oncol. 2007;18 Suppl 6:vi42-46. 49. Gleixner KV, Mayerhofer M, Cerny-Reiterer S, et al. KIT-D816V-independent oncogenic signaling in neoplastic cells in systemic mastocytosis: role of Lyn and Btk activation and disruption by dasatinib and bosutinib. Blood. 2011;118(7):1885-1898. 50. Verstovsek S, Tefferi A, Cortes J, et al. Phase II study of dasatinib in Philadelphia chromosome-negative acute and chronic myeloid diseases, including systemic mastocytosis. Clin Cancer Res. 2008;14(12):3906-3915. 51. Dubreuil P, Letard S, Ciufolini M, et al. Masitinib (AB1010), a Potent and Selective Tyrosine Kinase Inhibitor Targeting KIT. PloS One. 2009;4(9). 52. Hahn KA, Oglivie G, Rusk T, et al. Masitinib
1142
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69. 70.
is Safe and Effective for the Treatment of Canine Mast Cell Tumors. J Vet Intern Med. 2008;22(6):1301-1309. Paul C, Sans B, Suarez F, et al. Masitinib for the treatment of systemic and cutaneous mastocytosis with handicap: A phase 2a study. Am J Hematol. 2010;85(12):921-925. Hermine O, Lortholary O, Leventhal PS, et al. Case-control cohort study of patients' perceptions of disability in mastocytosis. PloS One. 2008;3(5):e2266. Hochhaus A, Baccarani M, Giles FJ, et al. Nilotinib in patients with systemic mastocytosis: analysis of the phase 2, open-label, single-arm nilotinib registration study. J Cancer Res Clin Oncol. 2015;141(11):2047-2060. Gleixner KV, Peter B, Blatt K, et al. Synergistic growth-inhibitory effects of ponatinib and midostaurin (PKC412) on neoplastic mast cells carrying KIT D816V. Haematologica. 2013;98(9):1450-1457. Pan J, Quintas-Cardama A, Kantarjian HM, et al. EXEL-0862, a novel tyrosine kinase inhibitor, induces apoptosis in vitro and ex vivo in human mast cells expressing the KIT D816V mutation. Blood. 2007;109(1):315322. Kosmider O, Denis N, Dubreuil P, MoreauGachelin F. Semaxinib (SU5416) as a therapeutic agent targeting oncogenic Kit mutants resistant to imatinib mesylate. Oncogene. 2007;26(26):3904-3908. Jin Y, Ding K, Wang D, Shen M, Pan J. Novel thiazole amine class tyrosine kinase inhibitors induce apoptosis in human mast cells expressing D816V KIT mutation. Cancer Lett. 2014;353(1):115-123. Evans EK, Hodous BL, Gardino A, et al. First Selective KIT D816V Inhibitor for Patients with Systemic Mastocytosis. Blood. 2014;124(21):3217. Traina F, Visconte V, Jankowska AM, et al. Single nucleotide polymorphism array lesions, TET2, DNMT3A, ASXL1 and CBL mutations are present in systemic mastocytosis. PloS One. 2012;7(8):e43090. Tefferi A, Levine RL, Lim KH, et al. Frequent TET2 mutations in systemic mastocytosis: clinical, KITD816V and FIP1L1-PDGFRA correlates. Leukemia. 2009;23(5):900-904. Wilson TM, Maric I, Simakova O, et al. Clonal analysis of NRAS activating mutations in KIT-D816V systemic mastocytosis. Haematologica. 2011;96(3):459-463. Schwaab J, Schnittger S, Sotlar K, et al. Comprehensive mutational profiling in advanced systemic mastocytosis. Blood. 2013;122(14):2460-2466. Hanssens K, Brenet F, Agopian J, et al. SRSF2-p95 hotspot mutation is highly associated with advanced forms of mastocytosis and mutations in epigenetic regulator genes. Haematologica. 2014;99(5):830-835. Busque L, Patel JP, Figueroa ME, et al. Recurrent somatic TET2 mutations in normal elderly individuals with clonal hematopoiesis. Nat Genet. 2012;44(11): 1179-1181. Ito S, Shen L, Dai Q, et al. Tet Proteins Can Convert 5-Methylcytosine to 5Formylcytosine and 5-Carboxylcytosine. Science. 2011;333(6047):1300-1303. Moran-Crusio K, Reavie L, Shih A, et al. Tet2 Loss Leads to Increased Hematopoietic Stem Cell Self-Renewal and Myeloid Transformation. Cancer Cell. 2011;20(1):1124. Holmfeldt L, Mullighan CG. The role of TET2 in hematologic neoplasms. Cancer Cell. 2011;20(1):1-2. Ko M, Huang Y, Jankowska AM, et al.
71.
72.
73.
74. 75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
Impaired hydroxylation of 5-methylcytosine in myeloid cancers with mutant TET2. Nature. 2010;468(7325):839-843. De Vita S, Schneider RK, Garcia M, et al. Loss of function of TET2 cooperates with constitutively active KIT in murine and human models of mastocytosis. PloS One. 2014;9(5):e96209. Jawhar M, Schwaab J, Schnittger S, et al. Molecular profiling of myeloid progenitor cells in multi-mutated advanced systemic mastocytosis identifies KIT D816V as a distinct and late event. Leukemia. 2015;29(5): 1115-1122. Damaj G, Joris M, Chandesris O, et al. ASXL1 but not TET2 mutations adversely impact overall survival of patients suffering systemic mastocytosis with associated clonal hematologic non-mast-cell diseases. PloS one. 2014;9(1):e85362. Will CL, LĂźhrmann R. Spliceosome Structure and Function. Cold Spring Harb Perspect Biol. 2011;3(7). Visconte V, Makishima H, Maciejewski JP, Tiu RV. Emerging roles of the spliceosomal machinery in myelodysplastic syndromes and other hematological disorders. Leukemia. 2012;26(12):2447-2454. Katoh M. Functional proteomics of the epigenetic regulators ASXL1, ASXL2 and ASXL3: a convergence of proteomics and epigenetics for translational medicine. Expert Rev Proteomics. 2015;12(3):317-328. Jawhar M, Schwaab J, Schnittger S, et al. Additional mutations in SRSF2, ASXL1 and/or RUNX1 identify a high risk group of patients with KIT D816V+ advanced systemic mastocytosis. Leukemia. 2016;30(1):136-143. Jawhar M, Schwaab J, Schnittger S, et al. Additional mutations in SRSF2, ASXL1 and/or RUNX1 identify a high-risk group of patients with KIT D816V(+) advanced systemic mastocytosis. Leukemia. 2016;30(1): 136-143. Soverini S, De Benedittis C, Mancini M, et al. Genome-Wide Molecular Portrait of Aggressive Systemic Mastocytosis and Mast Cell Leukemia Depicted By Whole Exome Sequencing and Copy Number Variation Analysis. Blood. 2015;126(23): 4085 abstr. Yavuz AS, Lipsky PE, Yavuz S, Metcalfe DD, Akin C. Evidence for the involvement of a hematopoietic progenitor cell in systemic mastocytosis from single-cell analysis of mutations in the c-kit gene. Blood. 2002;100 (2):661-665. Aichberger KJ, Mayerhofer M, Gleixner KV, et al. Identification of MCL1 as a novel target in neoplastic mast cells in systemic mastocytosis: inhibition of mast cell survival by MCL1 antisense oligonucleotides and synergism with PKC412. Blood. 2007;109 (7):3031-3041. Aichberger KJ, Gleixner KV, Mirkina I, et al. Identification of proapoptotic Bim as a tumor suppressor in neoplastic mast cells: role of KIT D816V and effects of various targeted drugs. Blood. 2009;114(26):5342-5351. Peter B, Cerny-Reiterer S, Hadzijusufovic E, et al. The pan-Bcl-2 blocker obatoclax promotes the expression of Puma, Noxa, and Bim mRNA and induces apoptosis in neoplastic mast cells. J Leukocyte Biol. 2014;95(1):95-104. Kawakami Y, Kitaura J, Satterthwaite AB, et al. Redundant and opposing functions of two tyrosine kinases, Btk and Lyn, in mast cell activation. J Immunol. 2000;165(3):12101219.
haematologica | 2016; 101(10)
Treatment in advanced SM 85. Alvarez-Errico D, Yamashita Y, Suzuki R, et al. Functional analysis of Lyn kinase A and B isoforms reveals redundant and distinct roles in Fc epsilon RI-dependent mast cell activation. J Immunol. 2010;184(9):5000-5008. 86. Gleixner KV, Mayerhofer M, Hormann G, et al. Bosutinib Blocks Lyn and Btk Activation and Synergizes with the KIT D816VTargeting Drug Midostaurin in Inducing Apoptosis in Neoplastic Human Mast Cells. ASH Annual Meeting Abstracts. 2009;114 (22):1717. 87. Randall N, Courville EL, Baughn L, Afrin L, Ustun C. Bosutinib, a Lyn/Btk inhibiting tyrosine kinase inhibitor, is ineffective in advanced systemic mastocytosis. Am J Hematol. 2015;90(4):E74-E74. 88. Kim MS, Kuehn HS, Metcalfe DD, Gilfillan AM. Activation and function of the mTORC1 pathway in mast cells. J Immunol. 2008;180(7):4586-4595. 89. Sundstrom M, Vliagoftis H, Karlberg P, et al. Functional and phenotypic studies of two variants of a human mast cell line with a distinct set of mutations in the c-kit protooncogene. Immunology. 2003;108(1):89-97. 90. Harir N, Boudot C, Friedbichler K, et al. Oncogenic Kit controls neoplastic mast cell growth through a Stat5/PI3-kinase signaling cascade. Blood. 2008;112(6):2463-2473. 91. Yang Y, Letard S, Borge L, et al. Pediatric mastocytosis-associated KIT extracellular domain mutations exhibit different functional and signaling properties compared with KIT-phosphotransferase domain mutations. Blood. 2010;116(7):1114-1123. 92. Smrz D, Kim MS, Zhang S, et al. mTORC1 and mTORC2 differentially regulate homeostasis of neoplastic and non-neoplastic human mast cells. Blood. 2011;118(26): 6803-6813. 93. Blatt K, Herrmann H, Mirkina I, et al. The PI3-kinase/mTOR-targeting drug NVPBEZ235 inhibits growth and IgE-dependent activation of human mast cells and basophils. PloS one. 2012;7(1):e29925. 94. Parikh SA, Kantarjian HM, Richie MA, Cortes JE, Verstovsek S. Experience with everolimus (RAD001), an oral mammalian target of rapamycin inhibitor, in patients with systemic mastocytosis. Leuk Lymphoma. 2010;51(2):269-274. 95. Baumgartner C, Cerny-Reiterer S, Sonneck K, et al. Expression of activated STAT5 in neoplastic mast cells in systemic mastocytosis: subcellular distribution and role of the transforming oncoprotein KIT D816V. Am J Pathol. 2009;175(6):2416-2429. 96. Lasho T, Tefferi A, Pardanani A. Inhibition of JAK-STAT signaling by TG101348: a novel mechanism for inhibition of KITD816V-dependent growth in mast cell leukemia cells. Leukemia. 2010;24(7):13781380. 97. Vannucchi AM, Kantarjian HM, Kiladjian JJ, et al. A pooled analysis of overall survival in COMFORT-I and COMFORT-II, 2 randomized phase III trials of ruxolitinib for the treatment of myelofibrosis. Haematologica. 2015;100(9):1139-1145. 98. Santos FPS, Helman R, Pereira WO, et al. Activity Of a JAK1/JAK2 Inhibitor In a Patient With KIT-Mutated Systemic Mastocytosis (SM) Associated With Myelofibrosis. Blood. 2013;122(21):5246. 99. Tanaka A, Konno M, Muto S, et al. A novel NF-kappa B inhibitor, IMD-0354, suppresses neoplastic proliferation of human mast cells with constitutively activated c-kit receptors.
haematologica | 2016; 101(10)
Blood. 2005;105(6):2324-2331. 100. Macleod AC, Klug LR, Patterson J, et al. Combination therapy for KIT-mutant mast cells: targeting constitutive NFAT and KIT activity. Mol Cancer Ther. 2014;13(12):28402851. 101. Belkina AC, Denis GV. BET domain co-regulators in obesity, inflammation and cancer. Nat Rev Cancer. 2012;12(7):465-477. 102. Zuber J, Shi J, Wang E, et al. RNAi screen identifies Brd4 as a therapeutic target in acute myeloid leukaemia. Nature. 2011;478(7370):524-528. 103. Godley LA, Le Beau MM. The histone code and treatments for acute myeloid leukemia. N Engl J Med. 2012;366(10):960-961. 104. Wedeh G, Cerny-Reiterer S, Eisenwort G, et al. Identification of bromodomain-containing protein-4 as a novel marker and epigenetic target in mast cell leukemia. Leukemia. 2015;29(11):2230-2237. 105. Sotlar K, Horny HP, Simonitsch I, et al. CD25 indicates the neoplastic phenotype of mast cells: a novel immunohistochemical marker for the diagnosis of systemic mastocytosis (SM) in routinely processed bone marrow biopsy specimens. Am J Surg Pathol. 2004;28(10):1319-1325. 106. Valent P, Cerny-Reiterer S, Herrmann H, et al. Phenotypic heterogeneity, novel diagnostic markers, and target expression profiles in normal and neoplastic human mast cells. Best Pract Res Clin Haematol. 2010;23(3): 369-378. 107. Valent P, Sotlar K, Horny HP. Aberrant expression of CD30 in aggressive systemic mastocytosis and mast cell leukemia: a differential diagnosis to consider in aggressive hematopoietic CD30-positive neoplasms. Leuk Lymphoma. 2011;52(5):740-744. 108. Hoermann G, Blatt K, Greiner G, et al. CD52 is a molecular target in advanced systemic mastocytosis. FASEB J. 2014;28(8):35403551. 109. Teodosio C, Mayado A, Sanchez-Munoz L, et al. The immunophenotype of mast cells and its utility in the diagnostic work-up of systemic mastocytosis. J Leukoc Biol. 2015;97(1):49-59. 110. Quintas-Cardama A, Kantarjian H, Verstovsek S. Treatment of systemic mastocytosis with denileukin diftitox. Am J Hematol. 2007;82(12):1124. 111. Krauth MT, Bohm A, Agis H, et al. Effects of the CD33-targeted drug gemtuzumab ozogamicin (Mylotarg) on growth and mediator secretion in human mast cells and blood basophils. Exp Hematol. 2007;35(1): 108-116. 112. Sotlar K, Cerny-Reiterer S, Petat-Dutter K, et al. Aberrant expression of CD30 in neoplastic mast cells in high-grade mastocytosis. Modern Pathol. 2011;24(4):585-595. 113. Blatt K, Cerny-Reiterer S, Schwaab J, et al. Identification of the Ki-1 antigen (CD30) as a novel therapeutic target in systemic mastocytosis. Blood. 2015;126(26):2832-2841. 114. Florian S, Sonneck K, Hauswirth AW, et al. Detection of molecular targets on the surface of CD34+/CD38-- stem cells in various myeloid malignancies. Leuk Lymphoma. 2006;47(2):207-222. 115. Eisenwort G, Peter B, Blatt K, et al. Identification of a Neoplastic Stem Cell in Human Mast Cell Leukemia. Blood. 2014;124(21):817. 116. Casassus P, Caillat-Vigneron N, Martin A, et al. Treatment of adult systemic mastocytosis with interferon-alpha: results of a multicen-
tre phase II trial on 20 patients. Br J Haematol. 2002;119(4):1090-1097. 117. Hauswirth AW, Simonitsch-Klupp I, Uffmann M, et al. Response to therapy with interferon alpha-2b and prednisolone in aggressive systemic mastocytosis: report of five cases and review of the literature. Leuk Res. 2004;28(3):249-257. 118. Kluin-Nelemans HC, Jansen JH, Breukelman H, et al. Response to interferon alfa-2b in a patient with systemic mastocytosis. N Engl J Med. 1992;326(9):619-623. 119. Kluin-Nelemans HC, Oldhoff JM, Van Doormaal JJ, et al. Cladribine therapy for systemic mastocytosis. Blood. 2003;102(13): 4270-4276. 120. Lim KH, Pardanani A, Butterfield JH, Li CY, Tefferi A. Cytoreductive therapy in 108 adults with systemic mastocytosis: Outcome analysis and response prediction during treatment with interferon-alpha, hydroxyurea, imatinib mesylate or 2chlorodeoxyadenosine. Am J Hematol. 2009;84(12):790-794. 121. Bohm A, Sonneck K, Gleixner KV, et al. In vitro and in vivo growth-inhibitory effects of cladribine on neoplastic mast cells exhibiting the imatinib-resistant KIT mutation D816V. Exp Hematol. 2010;38(9):744755. 122. Hermine O, Hirsh I, Damaj G, et al. Long Term Efficacy and Safety of Cladribine In Adult Systemic mastocytosis: a French Multicenter Study of 44 Patients. ASH Annual Meeting Abstracts. 2010 November 19, 2010;116(21):1982-. 123. Barete S, Lortholary O, Damaj G, et al. Longterm efficacy and safety of cladribine (2CdA) in adult patients with mastocytosis. Blood. 2015;126(8):1009-1016. 124. Valent P, Blatt K, Eisenwort G, et al. FLAGinduced remission in a patient with acute mast cell leukemia (MCL) exhibiting t(7;10)(q22;q26) and KIT D816H. Leuk Res Rep. 2014;3(1):8-13. 125. Ustun C, Reiter A, Scott BL, et al. Hematopoietic stem-cell transplantation for advanced systemic mastocytosis. J Clin Oncol. 2014;32(29):3264-3274. 126. Sperr WR, Drach J, Hauswirth AW, et al. Myelomastocytic leukemia: evidence for the origin of mast cells from the leukemic clone and eradication by allogeneic stem cell transplantation. Clin Cancer Res. 2005;11(19 Pt 1):6787-6792. 127. Ustun C, Courville E. Resolution of osteosclerosis after alloHCT in systemic mastocytosis. Blood. 2016;127(14):18361836. 128. Ustun C, Gotlib J, Popat U, et al. Consensus opinion on allogeneic hematopoietic cell transplantation in advanced systemic mastocytosis. Biol Blood Marrow Transplant. 2016 Apr 27. [Epub ahead of print]. 129. Ustun C, Smith A, Cayci Z, et al. Allogeneic hematopoietic cell transplantation in systemic mastocytosis: Is there a high risk for veno-occlusive disease. Eur J Haematol. 2015 Dec 17. [Epub ahead of print] 130. Akin C, Scott LM, Kocabas CN, et al. Demonstration of an aberrant mast-cell population with clonal markers in a subset of patients with "idiopathic" anaphylaxis. Blood. 2007;110(7):2331-2333. 131. Akin C, Metcalfe DD. Systemic mastocytosis. Annu Rev Med. 2004;55:419-432. 132. Theoharides TC, Valent P, Akin C. Mast Cells, Mastocytosis, and Related Disorders. N Engl J Med. 2015;373(19):1885-1886.
1143
REVIEW ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
Catching up with solid tumor oncology: what is the evidence for a prognostic role of programmed cell death-ligand 1/programmed cell death-1 expression in B-cell lymphomas? Fabienne McClanahan,1,2 Thomas G. Sharp1, and John G. Gribben1
Haematologica 2016 Volume 101(10):1144-1158
Department of Haemato-Oncology, Barts Cancer Institute, Queen Mary University of London, UK; and 2Comprehensive Cancer Center, The Ohio State University, Columbus, OH, USA
1
ABSTRACT
T
Correspondence: j.gribben@qmul.ac.uk
Received: March 14, 2016. Accepted: May 27, 2016. Pre-published: no prepublication. doi:10.3324/haematol.2016.145904
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1144
Š2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1144
herapeutic strategies targeting the programmed cell death-ligand 1/programmed cell death-1 pathway have shown significant responses and good tolerability in solid malignancies. Although preclinical studies suggest that inhibiting programmed cell death-ligand 1/programmed cell death-1 interactions might also be highly effective in hematological malignancies, remarkably few clinical trials have been published. Determining patients who will benefit most from programmed cell death-ligand 1/programmed cell death-1-directed immunotherapy and whether programmed cell death-ligand 1/programmed cell death-1 are adequate prognostic markers becomes an increasingly important clinical question, especially as aberrant programmed cell death-ligand 1/programmed cell death-1 expression are key mediators of impaired anti-tumor immune responses in a range of B-cell lymphomas. Herein, we systematically review the published literature on the expression and prognostic value of programmed cell death-ligand 1/programmed cell death-1 in these patients and identify considerable differences in expression patterns, distribution and numbers of programmed cell death-ligand 1+/programmed cell death-1+cells, both between and within lymphoma subtypes, which is reflected in conflicting findings regarding the prognostic value of programmed cell death-ligand 1+/programmed cell death-1+ cells. This can be partly explained by differences in methodologies (techniques, protocols, cutoff values) and definitions of positivity. Moreover, lymphomagenesis, disease progression, and prognosis appear to be determined not only by the presence, numbers and distribution of specific subtypes of T cells, but also by other cells and additional immune checkpoints. Collectively, our findings indicate that programmed cell death-ligand 1/programmed cell death-1 interactions play an essential role in B-cell lymphoma biology and are of clinical importance, but that the overall outcome is determined by additional components. To categorize the exact prognostic value of programmed cell death-ligand 1/programmed cell death-1 expressing cells and cell types, efforts should be made to harmonize their assessment and interpretation, optimally within ongoing clinical immune checkpoint inhibitor trials, and to identify and validate novel high-throughput platforms.
Introduction The immune checkpoint programmed cell death protein 1 (PD-1, CD279) and its ligand PD-L1 (B7-H1, CD274) have rapidly taken center stage in tumor immunology. This is because antibodies targeting this pathway have shown significant responses and good tolerability across a variety of solid malignancies, both in initial phase 1/2 studies and in recently published randomized trials or in combination haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas
with other substances.1-12 Although a plethora of preclinical studies suggest that inhibiting PD-L1/PD-1 interactions might also be highly effective in hematological malignancies,13,14 only few PD-L1/PD-1 antibody based clinical trials have been published to date. An initial phase I trial demonstrated a clinical benefit of the PD-1 antibody pidilizumab in several advanced hematological malignancies.15 Encouraging results were also observed in recently published phase II trials in relapsed follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL),16,17 as well as in relapsed/refractory Hodgkin lymphoma (HL) patients treated with nivolumab.18 Determining which patients benefit most from PDL1/PD-1-directed immunotherapy is an important clinical question. Yet again, the solid oncology field appears to be one step ahead. Several retrospective and correlative studies examining the prognostic significance of tumor PD-L1 expression and PD-1 expression on tumor-infiltrating lymphocytes (TILs) have already been published, although the exact associations are somewhat controversial and appear to be dependent on tumor entity, treatment setting and the presence of other predictive factors or biomarkers.19-23 Similar studies have not been reported in hematological malignancies, even though most of these tumor types, and especially lymphomas, are increasingly understood to closely interact with their surrounding microenvironment.24 Importantly, we and others have shown that aberrant PD-L1 expression by lymphoma cells and increased expression of PD-1 on T cells are key mediators of impaired anti-tumor immune responses in a range of B-cell lymphomas, including DLBCL, FL and chronic lymphocytic leukemia (CLL),25-27 and that inhibiting their interaction restores immune function in preclinical models.28 However, PD-L1 is also expressed on other cell types and in peripheral tissues and is up-regulated during inflammation and in the tumor microenvironment.29-31 Similarly, PD-1 can be expressed on a variety of physiological immune cells, for example on CD4+ germinal center (GC) follicular helper T cells (TFH), which are required for GC development and high-affinity antibody production.32 As TFH cells also act as negative regulators of immune responses, their numbers and tissue distribution may shape the microenvironment in GC-type lymphomas.33 Indeed, across multiple solid cancer types, it was recently demonstrated that clinical responses were not only observed in patients with high tumor PD-L1 levels, but also when PDL1 was expressed by tumor-infiltrating immune cells and when T helper type 1 (TH1) gene signatures and CTLA-4 expression were detected in baseline specimens.23 Herein, we aimed to collate and review data from the literature on the prognostic value of PD-L1 or PD-1 expression in patients with the most frequent types of B-cell lymphomas. We hypothesized that increased PD-L1/PD-1 expression confers an adverse prognosis, but that differences exist between lymphoma subtypes and between lymphoma and tumor infiltrating lymphocytes (TIL) expression. Such a systematic comparison has several clinical implications. First, it allows the identification of entity- and cell-type-specific expression patterns and their association with prognosis and survival. Second, it elucidates the clinical importance of this pathway in specific lymphomas, contributing to identifying patient groups that might benefit most from blocking PD-L1/PD-1 interactions. Ultimately, these findings provide direct translational guidance in the implementation and interpretation haematologica | 2016; 101(10)
of assays and techniques assessing PD-L1 or PD-1 as biomarkers in future clinical trials of immune checkpoint inhibitors.
Methods and Materials Full-text publications were included if they met prospectively defined criteria: i) investigated DLBCL, FL, CLL/ small lymphocytic leukemia (SLL), Hodgkin lymphoma (HL) or primary mediastinal large B-cell lymphoma (PMBCL), ii) quantified PD-1/PD-L1 expression on tumor and/or microenvironmental components by immunohistochemistry (IHC) or flow cytometry, iii) described techniques and quantification methods, and iv) were written in English. Abstracts from conference proceedings were not reviewed, and less frequent B-cell lymphomas such as mantle cell, marginal zone and Burkitt lymphoma were not included. Suitable publications were retrieved from two independent MEDLINE database queries and information on study characteristics, methods/materials (examined tissues, techniques, quantification of PDL1/PD-1 expression, antigens/antibodies, controls, statistical analyses), patients and treatment characteristics and findings on PD-L1/PD-1 expression and prognostic significance were extracted. The majority of retrieved results were excluded because studies examined T-cell or cutaneous lymphomas. An overview of key information on included studies can be found in Table 1. Expression patterns on lymphoma and lymphoma-associated immune and/or surrounding cells are summarized according to lymphoma type in Table 2 (DLBCL), Table 3 (FL), Table 4 (CLL/SLL) and Table 5 (HL). The prognostic value of PDL1/PD-1 in all examined lymphoma types is depicted in Table 6.
Results DLBCL PD-L1/PD-1 expression on DLBCL cells One of the first studies to characterize PD-L1/PD-1 expression in a series of 161 B-cell non-Hodgkin lymphoma (NHL) tissues contained only 25 DLBCL specimens, of which 4 out of 14 examined samples were PDL1+ on 1-75% of tumor cells34 (Table 2). In a cohort comprising (Epstein-Barr virus) EBV+ and EBV- patients, the proportion of PD-L1+ malignant cells ranged from 1090%.35 All EBV+ DLBCLs showed strong PD-L1 expression, in contrast to 11% of EBV- DLBCL patients. Another study found at least 5% of PD-L1+ tumor cells in 55 out of 73 interpretable tissue microarrays (TMAs), which did however not correlate with plasma PD-L1 levels.36 Slight differences were observed in frozen versus paraffin specimens, where heterogeneous PD-L1 tumor expression was observed in 27% of frozen and 20% of paraffin samples.37 A more recent study detected tumor PD-L1 expression in 61% of DLBCL TMAs, with variable intensities and proportions.38 Using a threshold of â&#x2030;Ľ30% of PD-L1+ malignant cells among all malignant cells, another recent study of a total of 1,253 DLBCL TMAs reported a tumor PD-L1+ prevalence rate of 11%.39 This was significantly associated with non-germinal center B-cell (GCB) type and EBV positivity, and with chromosome 9 gain but not structural abnormalities in chromosome 9p. PD-1 expression was 1145
F. McClanahan et al. Table 1. Key information on included studies. Information on aim of study, patient/ sample numbers, techniques and examined tissues and PD-L1/ PD-1 scoring methods was extracted and is summarized according to B-NHL subtype.
Reference
Aim of study
Amé-Thomas 201243 Andorsky 201137
Functional characterization of intratumoral CD4+ T cells PD-L1 expression in cell lines and lymphoma specimens
Chen 201335
Examination of 237 primary tumors for expression of PD-L1 protein
Included studies examining several lymphoma types Patient/sample numbers Techniques and examined tissues DLBCL, FL numbers not specified Frozen specimens:9 HL, 33 DLBCL (11 GCB, 19 non-GCB), 3 PMBCL SSS: 16 FL, 2 SLL/CLL, 3 MZL, 1 MCL,1 BL paraffin specimens: 5 ALCL, 7 FL, 30 DLBCL 25 NSCHL, 8 MCCHL, 5 CHL-NOS, 15 NLPHL, 21 PMBCL, 11 TCHRBCL, 9 EBV+ DLBCL of the elderly, 7 EBV+ immunodeficiencyrelated DLBCL, 10 EBV+ PTLD, 7 EBV- PTLD, 66 DLBCL-NOS; 9 PMBCL, 4 PEL, 6 ENKTCL, 7 EBV+ BL, 18 NPC, 9 KS
PD-L1/PD-1 scoring methods
IHC on paraffin-embedded tissue sections Flow cytometry IHC on different sets of frozen or paraffinembedded DLBCL, HL, PMBCL, FL Fow cytometry on CLL/SLL, MZL, MCL, BL
Percent positive among CD4+ cells Not specified
IHC on paraffin-embedded tissue biopsies
Staining intensity: no staining: 0 weak: 1+ moderate: 2+ strong: 3+ Tumor PD-L1+ if ≥5% of tumor cells 2+/3+ membrane staining
Dorfman 200633
PD-1 expression in B and T-cell 42 B-LPD (25 HL, 4 CLL, 4 MCL, 6 FL, lymphoproliferative disorders 6 DLBCL, 3 MZL, 3 HCL, 7 BL, 3 LPL, 3 MM, 3 B-ALL), 23 T-LPD
IHC on paraffin-embedded tissues
Muenst 201040
Diagnostic potential and 8 BL, 184 DLBCL, 5 T-cell rich large BCL, prognostic importance of PD-1 7 DLBCL ex SLL/LPL/MZL, 11 DLBCL ex FL, in B-cell lymphomas 7 FL grade 3, 42 FL grade 1/2, 33 extranodal MZL, 19 extranodal DLBCL ex MZL, 10 MCL, 20 PMBCL, 58 SLL/CLL
IHC on total or paraffin-embedded sections
Microenvironment PD-L1+ if ≥20% of total tissue 2+/3+ membrane or cytoplasmic staining PD-1+ if ≥20% of neoplastic cells positive staining Staining specificity: comparison to isotype control Total number of PD-1+ TILs counted in one medium power field (1.33 mm2) at 200x magnification. % PD-1+TILs in relation to all cells
Ramsay 201227
Tonino 201263 Xerri 200834
Role of immune checkpoints in 68 CLL, 18 CLL median survival 38 mo, immune evasion mechanisms 17 CLL median survival >10 yrs, in lymphomas 6 untreated FL, 6 transformed FL, 34 diagnostic FL survival <5 yrs, 25 diagnostic FL survival > 15 yrs
IHC on TMAs
Changes in T cell compartment 29 CLL, 8 FL, 2 HCL, 3 MZL, in different B cell malignancies 2 low-grade lymphoma NOS, 13 aggressive lymphomas, 10 MM Expression profile of PD-1, 35 HL (5 LPHL, 22 NSCHL, 8 MCCHL), PD-L1 and PD-L2 in B-NHLs 11 MCL, 12 MZL, 3 BL,25 DLBCL, 43 FL, 11 T-NHL, 11 CLL
Flow cytometry of PB mononuclear cells
Reference
Aim of study
Ahearne 201442
Expression of PD-1 in combination with FoxP3 in DLBCL
Flow cytometry
IHC on total or paraffin-embedded sections Flow cytometry on CLL blood samples
Included studies examining DLBCL only or focus on DLBCL Patient/sample numbers Techniques and examined tissues 70
IHC on paraffin-embedded LN Flow cytometry to quantify T-cell subsets
Only absolute count of positive cells and not staining intensity were considered Staining on CD20+ cancer or reactive LN B cells and on CD3+ T cells evaluated for mean intensity expression using automated serial section overlay analysis Percent positive cells and median fluorescence intensity % cells positive
0:<1% of cells positive +: 1-50% of cells positive ++: 50-75% of cells positive +++: >75% of cells positive
PD-L1/PD-1 scoring methods Intensity threshold for definition of PD-1high cells by comparison to PD-1 expression within tonsil sections from normal subjects Continued on the next page
1146
haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas Continued from the previous page
Armand 201317
Correlative studies of lymphocyte subsets in phase II trial of pidilizumab in patients with DLBCL undergoing AHSCT
35 available patients
Flow cytometry on PB mononuclear cells from patients treated at least once with pidilizumab
Kiyasu 201539
Clinicopathological impact of PD-L1+ in newly diagnosed DLBCL
1,253 Among 273 pts with available clinical information: quantitative analysis of PD-1+ TILs
IHC on formalin-fixed paraffin-embedded tissues
41 prospectively specified leukocyte subsets evaluated for absolute (per μL) and relative numbers and median fluorescence intensity PD-L1+ DLBCL: ≥30% of lymphoma cells distinct membranous and/or cytoplasmic staining and nuclear staining of PAX5, regardless of PD-L1 positivity of nonmalignant stromal cells Microenvironmental PD-L1+ DLBCL: PD-L1– DLBCL cases in which PD-L1+ nonmalignant stromal cells represented ≥20% of total tissue
Ko 201141
Correlation between PD-1+ TILs and clinicopathologic prognostic factors in DLBCL
65
Kwon 201538
Expression patterns, 126 clinicopathological features and prognostic implications of PD-1 and PD-L1 in DLBCL tissues
73 interpretable TMAs
IHC on paraffin-embedded tumors
IHC on formalin-fixed paraffin-embedded tumor blocks
Rossille 201436
Clinical impact of soluble PD-L1 at diagnosis in DLBCL
IHC on paraffin-embedded blocks
Reference
Aim of study
Carreras 200950
Role of PD-1 in FL progression 100 diagnostic samples, 15 sequential biopsies IHC on paraffin-embedded whole tissue and outcome at relapse, 17 relapse samples only sections Flow cytometry in a subset of samples
Koch 201252
Prognostic significance of Treg and TFH in advanced-stage FL
Included studies examining FL only or focus on FL Patient/sample numbers Techniques and examined tissues
139 advanced stage, 125 early stage
IHC on paraffin-embedded tissue samples
Richendollar Prognostic relevance of 91 201153 numbers of PD-1+ T cells within the tumor microenvironment
IHC on paraffin-embedded tissue samples
Smeltzer 201454
IHC on paraffin embedded tissues
Cell subtypes associated with transformation in FL
58
Number of PD-1+ TILs Number of PD-1+ TILs, recorded as average value Positive if >20/hpf Negative if ≤20/hpf PD-L1 intensity and proportion of cells with membranous and/or cytoplasmic staining: 0: negative (no or any staining in<10% of cells) 1: weak 2: moderate 3: strong (>10% of cells) Numbers of PD-1+ cells: 0: no positive cells/hpf 1: <10 positive cells/hpf 2: 10–30 positive cells/hpf 3: >30 positive cells/hpf Protein expression recorded in 5% increments as percentage of positive tumor cells
PD-L1/PD-1 scoring methods Quantification using an automated scanning microscope and computerized image analysis system (under pathologist visual supervision) Number of positive cells among 100 cells/hpf (×400 magnification) Mean number of follicular PD-1+ cells/hpf (1000×, 3 follicles with 3 fields per follicle) Patterns of expression and 0–3 scale assessing quantity and intensity Follicular pattern: majority of cells in follicle/perifollicular area Continued on the next page
haematologica | 2016; 101(10)
1147
F. McClanahan et al. Continued from the previous page
Takahashi 201351
Prognostic implications of PD-1 82 in patients treated with R-CHOP
IHC on biopsy specimen 10 follicular areas quantified
Flow cytometry
using an automated scanning microscope and image analysis system Computerized image analysis, separating cells inside and outside the follicles Mean fluorescence intensity
IHC on paraffin embedded tissue Flow cytometry on SSS
Bright vs. dim Percent cells positive
Wahlin 201059
Prognostic significance of immune cell subsets
Westin 201416
Yang 201555
Correlative studies on available 25: 18 responders, 7 non-responders blood samples at baseline from FL patients treated with pidilizumab and rituximab Biological and clinical relevance 32 of PD-1 in FL
Reference
Aim of study
Brusa 201360
Expression and functional significance of PD-1/ PD-L1
117
Flow cytometry PB in all samples IHC on paraffin-embedded sections of LNs infiltrated by CLL cells (n=20)
Grzywnowicz 201261 Riches 201362
Characterization of PD-1 and PD-L1 expression Exhaustion in CD8+ T cells from CLL patients
45
Flow cytometry (n=45) PD-1 mRNA expression by qRT-PCR (n=43) Flow cytometry PB, in comparison to CMV-status matched controls
Reference
Aim of study
Ansell 201518
Correlative studies phase Pretreatment tumor specimens 1 trial assessing available from 10 patients PD-L1/PD-L2 loci and protein expression Characterization of CD4+ cells 18 cHL SSS, 122 cHL in the microenvironment of HL
IHC by automated staining system FISH to assess chromosome 9p24.1
Staining intensities and double-staining techniques
Flow cytometry SSS IHC on TMAs (n=122)
Koh 201569
Prognostic significance of and correlations between PD-1 and PD-L1 and PD-L2 expression in uniformly treated cHL
Diagnostic tissues from 109 cHL pts treated with ABVD
IHC on formalin-fixed, paraffin-embedded tumor samples
Muenst 200972
Distribution of PD-1+ lymphocytes in the HL microenvironment PD-1 expression on TFH cells in NLPHL and the entities involved in its differential diagnosis Clinical and prognostic importance of PD-1 and/or PD-L1 and association between EBV-encoded RNA
280 cHL (156 NSCHL, 93 MCCHL, 11 LRCHL, 7 LDCHL, 13 cHL-NOS), 3 nodular lymphocyte-predominant HL 43 NSCHL, 14 MCCHL, 13 LRCHL, 58 NLPHL, 7 NLPHL with diffuse areas, 12 T-cell rich BCL
IHC on TMAs (n=189 evaluable cases)
Percentage cells positive, median expression levels Median cell count/mm2 and expression levels based on automated image analysis ≥10 CD30+ HRS cells were read. PD-L1- or PD-L2positive if expression was detected in ≥20 % of HRS cells. PD-1-positive if PD-1 expression was detected in ≥20 % of the peritumoral microenvironment Absolute number of PD-1+ lymphocytes in relation to other lymphocyte populations Cells positive and forming rosettes around tumor cells
87 cases with newly diagnosed HL
IHC on formalin-fixed, paraffin-embedded tissue samples
Greaves 201373
Nam-Cha 200871
Paydas 201568
31 good and 33 bad prognosis patients
Diffuse pattern: majority of positive cells not confined to follicle Nucleated and PD-1+ cells of
IHC on paraffin-embedded TMAs Flow cytometry
Included studies examining CLL/ SLL only or focus on CLL/ SLL Patient/sample numbers Techniques and examined tissues
39
Included studies examining HL/ PMBCL only or focus on HL/ PMBCL Patient/sample numbers Techniques and examined tissues
IHC on paraffin-embedded tissues
PD-L1/PD-1 scoring methods Percent cells positive Percent positive area and patterns of expression in proliferation centers compared to other parts of same slide Percent cells positive Splicing variants of PD-1 gene Percent cells positive
PD-L1/PD-1 scoring methods
Staining intensity: no staining: 0 weak/ equivocal: 1+ moderate: 2+ Continued on the next page
1148
haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas Continued from the previous page
(EBER) and PD-1/PD-L1
Yamamoto 200870
Characterization of PD-L1 and PD-L2 expression
19 HL, 12 B-NHL
IHC (n=4) Flow cytometry LN SSS (n=3) and PB (n=10)
strong: 3+ Tumor PD-L1+ if ≥5% of tumor cells membrane staining Microenvironment positive if ≥20% of total tissue membrane or cytoplasmic staining HRS cells evaluated as positive or negative regardless of intensity Cells positive
ABVD: doxorubicin, bleomycin, vinblastine, and dacarbazine; AHSCT: autologous hematopoietic stem cell transplantation; ALCL: anaplastic large cell lymphoma; ALL: acute lymphoblastic leukemia; BCL: B-cell lymphoma; BL: Burkitt lymphoma; CHL: classical Hodgkin lymphoma; CLL: chronic lymphocytic leukemia; DLBCL: diffuse large B-cell lymphoma; EBV: Epstein–Barr virus; ENKTCL: extranodal NK/T cell lymphoma; FISH: fluorescence in situ hybridization; FL: follicular lymphoma; GCB: germinal center B cell; HCL: hairy cell leukemia; HL: Hodgkin lymphoma; hpf: high-power field; IHC: immunohistochemistry; KS: Kaposi sarcoma; LDCHL: lymphocyte-depleted classical Hodgkin lymphoma; LN: lymph node(s); LPD - lymphoproliferative disorder; LPL: lymphoplasmacytic lymphoma; LRCHL: lymphocyte-rich classical Hodgkin lymphoma; MCCHL: mixed cellularity classical Hodgkin lymphoma; MCL: mantle cell lymphoma; MM: multiple myeloma; mo: months; MZL: marginal zone lymphoma; NHL: Non-Hodgkin lymphoma; NLPHL: nodular lymphocyte-predominant Hodgkin lymphoma; NOS: not otherwise specified; NPC: nasopharyngeal carcinoma; NSCHL: nodular sclerosis CHL; PB: peripheral blood; PEL: primary effusion lymphoma; PMBCL: primary mediastinal large B-cell lymphoma; PTLD: post-transplant lymphoproliferative disorder; SLL: small lymphocytic lymphoma; SSS: single cell suspension(s); TCHRBCL: T-cell/histiocyte-rich large B-cell lymphoma; TIL: tumor infiltrating lymphocytes; TMA: tissue microarray; yrs: years. CHL: classical Hodgkin lymphoma; B-LPD: B cell lymphoproliferative disorder; TFH: T follicular helper; R-CHOP: Rituximab, cyclophosphamide, hydroxydaunorubicin, vincristine, and prednisolone; RNA: ribodeoxynucleic acid; NHL: non-Hodgkin lymphoma; LPHL: lymphocyte predominant Hodgkin lymphoma; qRT-PCR; quantitative real-time polymerase chain reaction; HRS: Hodgkin Reed-Sternberg cell; CMV: cytomegalovirus; mRNA: messenger RNA.
initally not detected on DLBCL cells,33 but heterogenous expression in a small number of patients was subsequently described.34,40
PD-L1/PD-1 expression on DLBCL-associated immune cells Initial studies described numerous PD-L1/PD-L2+ and variable, non-quantified amounts of PD-1+ reactive lymphocytes34 (Table 2). More recently, most DLBCL-infiltrating immune cells were characterized as PD-L1 expressing macrophages, with 30% of patients showing PD-L1 expression mainly in macrophages with little expression in tumor cells.38 Using a threshold of ≥20% PD-L1+ nonmalignant cells among the total tissue cellularity in PD-L1– patients, the study by Kiyasu et al. reported a microenvironment PD-L1+ prevalence rate of 15%.39 This was significantly associated with non-GCB type and EBV positivity, but not with gain of chromosome 9 nor structural abnormalities in chromosome 9p. Increased PD-1+ TILs were detected in 11% of 184 DLBCL, but numbers and percentages were lower compared with FL and PMBCL.40 Similarly variable and low numbers of PD-1+ TILs were described in a Korean cohort.41 More than half of the included patients were classified PD-1+, with no differences between GCB subtypes. PD-1+ cases had significantly higher clinical stage (P=0.025) and higher International Prognostic Index (IPI) (P=0.026) than PD-1- patients. Subsequent studies classified PD1+CD4+ TILs in DLBCL as TFH cells, and noted reduced TFH numbers in DLBCL and reactive lymph nodes (LNs) compared to tonsils.42,43 CD4+ T-cell numbers correlated with both PD-1+ and FoxP3+ numbers.42 More recently, PD-1 was detected on TILs in all but two cases, and their quantity correlated positively with the level of PD-L1 expression in tumor cells (P=0.042) or in tumor cells/macrophages (P=0.03).38 In the study by Kiyasu et al., the number of PD-1+ TILs was significantly lower in PDL1+ patients and in those with B symptoms (P=0.024), extranodal sites (P=0.042) and bulky disease (P=0.041), but higher in GCB-type DLBCL (P=0.034).39 haematologica | 2016; 101(10)
Prognostic relevance of PD-1 expression in DLBCL Distinct molecular subtypes determine biology and outcome in DLBCL,44,45 and molecular- and IHC-based algorithms have confirmed additional tumor-promoting roles of the microenvironment.46 However, findings regarding the prognostic relevance of TILs and tumor-associated macrophages (TAMs) are conflicting. Whereas infiltration with activated CD4+ cells generally correlates with better prognosis, the role of specific subtypes, such as FoxP3+ cells, has been largely contradictory.47-49 The same appears to be true for PD-1+ TILs in GC lymphomas (Table 6). While actual median values were not reported, the numbers of PD1+TFH (P=0.0007), FoxP3+ (P=0.0069), and total CD4+ cells (P=0.04) above the median were associated with improved overall survival (OS), and had independent prognostic significance in multivariate analyses.42 This was confirmed in more recent studies; although the quantity of PD-1+ TILs showed no significant association with clinicopathological variables, the presence of PD-1+ TILs (score 1–3) significantly prolonged OS (P=0.026) and progression-free survival (PFS) (P=0.005), and was an independent favorable prognostic factor in multivariate analyses.38 In contrast, in another study, patients with PD-1 expression >20/hpf had a trend to poorer OS (P=0.120).41 A similar trend was seen when groups were further refined to 1-10, 11-50, 51-100 and >100 PD-1+ cells/hpf, but numbers were too small to allow valid conclusions.
Prognostic relevance of PD-L1 expression in DLBCL The prognostic relevance of cellular PD-L1 has only recently been explored (Table 6). Strong tumor and tumor/macrophage PD-L1 expression were significantly associated with B symptoms (P=0.005 tumor only, P=0.011 tumor and/or macrophages) and EBV infection (P=0.015 tumor only, P=0.020 tumor and/or macrophages), and tended to be higher in activated B-cell (ABC) than GCB DLBCL.38 This however did not correlate with survival, which is somewhat inconsistent with another report showing that increased plasma PD-L1 lev1149
F. McClanahan et al. Table 2. Expression of PD-1 and PD-L1 on tumor infiltrating lymphocytes (TILs) and tumor cells in DLBCL.
DLBCL
Method of quantification
PD-1 expression on TILs Tumor cells
PD-L1 expression on TILs Tumor cells
Dorfman 200633 Xerri 200834 Muenst 201040
Positive cases/ all cases
nd
0/6
nd
nd
Proportion of positive cells#
Variable, not quantified
Numerous, not quantified
Mean number of positive cells/mm2 Mean % of positive cells/ all cells Pts with positive cells >mean % positive cells frozen specimen % positive cells paraffin specimen Mean number of PD-1+ TILs/hpf$ Pts with positive cells >mean % TFH cells/ all cells
27±93 (SD) 1.1 20/184 (11%) nd nd 21 (range 0-201) 33 (52.4%) Median 0.2% (0-20)
2/25 pts: + 20/184 pts nd nd nd nd nd nd nd
4/14 pts: + to +++ nd nd nd 9/33 pts: 27% 6/30 pts: 20% nd nd nd
% of positive cells
nd
nd
nd
% positive cells
nd
nd
nd
EBV-: in 7/66 pts on 10-90% of cells EBV+: present in all pts 55/73 pts: ≥5%
% positive cells/ all cells
0.1 - 1.5 %
nd
nd
nd
Prevalence rates of PD-L1+ DLBCL and microenvironment PD-L1+ DLBCL Median TILs/ mm2 N (%) pts positive
nd Reported according to various clinical features
nd nd
15.3% (172 of 1121) nd
10.5% (132 of 1253) nd
Andorsky 201137 Ko 201141 Amé Thomas 201243 Chen 201335 Rosille 201436 Ahearne 201442 Kiyasu 201539 Kwon 201538
In tumor cells and/ or 77 (61%) macrophages: 115 (91%) weak 55 (44%) weak 37 (29%) moderate 46 (37%) moderate 27(21%) strong 14 (11%) strong 13 (10%)
Staining intensities among positive cells Quantity of PD-1+ TIL/hpf
nd nd nd nd nd nd nd nd
0: 38 (31%) <10: 30 (25%) 10-30: 23 (19%) >30: 30 (25%)
hpf: high power field; nd: not done; pts: patients; SD: standard deviation. EBV: Epstein-Barr virus; DLBCL: diffuse large B-cell lymphoma. #+ 1-50%, ++ 50-75%, +++ >75% of cells positive; $classified as positive for >20/hpf, negative for ≤20/hpf.
els were associated with poorer prognosis in DLBCL patients.36 Inferior OS was also reported in patients with PD-L1+ DLBCL (P=0.0009), and the expression of PD-L1 maintained prognostic value for OS in multivariate analysis.39 Combining the median number of TILs with positive or negative PD-L1 expression patterns, the PD-L1+/TILlow group was significantly associated with poor prognosis compared to the PD-L1–/TILlow group, whereas no prognostic impact was observed in the other two groups (PDL1+/TILhigh and PD-L1–/TILhigh).
FL PD-L1/PD-1 expression on lymphoma cells The majority of published studies reported virtually PDL1 negative FL cells34,37,50 (Table 3). We found significantly increased PD-L1 on FL compared to healthy B cells, and on tumor cells from patients with <5-year (n=34) versus >15year (n=25) survival.27 PD-1 was heterogeneously expressed on 1-50% of tumor cells in a minority of FL specimens,34 whereas others excluded PD-1 expression on B cells.33
PD-L1/PD-1 expression on FL-associated immune cells PD-L1 expression was detected in some CD3+ cells in both reactive LN and FL samples50 (Table 3). A number of 1150
studies have characterized PD-1+ TFH cells, with similarly high proportions of TFH in tonsils and FL LNs (median 30% and 32%, respectively).43 At diagnosis (n=100), PD-1+ cells were mainly observed in follicular areas, but numbers were highly variable (mean 21.8%, range 0.12-73.6%) and similar to reactive tonsils.50 PD-1+ cells decreased with increasing histological grade (P=0.003), but correlated with the number of TRegs. PD-1+ cell numbers were also significantly lower in patients with poor performance status (P=0.014) and high serum lactate dehydrogenase (LDH, P=0.001). At relapse (n=32), the number of PD-1+ cells was similar to diagnosis for all grades. In transformed FL (n=10), PD-1+ numbers were significantly lower than either at diagnosis or relapse. Decreasing but numerous PD-1+ TILs with increasing grade (n=49) and transformation to DLBCL (n=11) were described by others.40 There might be an association between male gender and increased PD-1+ cells,51 but further confirmation is lacking. Several studies have focused on localization patterns of TFH cells. While PD-1 expression generally correlated with T-cell content in both interfollicular and follicular zones, it was mainly expressed within52 or restricted to follicles.53 FoxP3+ cells were predominantly found interfollicularly, but a high follicular content of FoxP3+ and PD-1+ haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas Table 3. Expression of PD-1 and PD-L1 on tumor infiltrating lymphocytes (TILs) and tumor cells in FL.
FL
Method of quantification
PD-1 expression on TILs Tumor cells
PD-L1 expression on TILs Tumor cells
Dorfman 200633 Xerri 200834 Carreras 200950
Positive cases/ all cases
nd
0/6
nd
nd
Proportion of positive cells#
nd
3/43 pts: +
nd
0/8 pts
Proportion of positive cells
Diagnosis vs. relapse (mean±SD): Gr1/2: 24.3±20% vs. 19.8±20%, Gr3: 13.2±17% vs. 20.6±18% Gr1/2: 287±228 Gr3: 128±105 tFL: 75±107 Gr1/2: 6.5, Gr3: 4.5, tFL: 2.3 Gr1/2: 7/42 (17%) Gr3: 2/7 (29%) tFL: 3/11 (27%) Total: 2.7 vs. 2.5 Follicular: 3.7 vs. 2.8 Interfoll.: 2.2 vs. 2.5 nd nd 35.6 cells/hpf (range 4.4-91.2) 45/91 (49%) Tonsils: 30% (5--57) FL LN: 32% (10- 57) Follicular: 12.7% Interfollicular: 3.3%
nd
Median 9% (2.4-29%)
Median 2.4% (0-4%)
nd
nd
nd
nd nd
nd nd
nd nd
nd
nd
nd
nd nd nd nd nd
nd nd nd nd nd
0/16 pts 0/7 nd nd nd
nd
nd
nd
nd nd nd
nd nd nd
CD20+ cells: ~90 vs. 150 CD20+ cells: ~135 vs. 175 nd
Muenst 201040
Mean number of positive cells/mm2 ±SD Mean % of positive cells/ all cells Pts with positive cells >mean
Wahlin 201059
Nmber of positive cells/ total area good vs. poor outcome pts
Andorsky % positive cells flow cytometry 201137 % positive cells paraffin specimen Richendollar Median number of positive cells/hpf 201153 Pts> median Amé Thomas Median % TFH cells/ all cells 201243 Koch Median % positive cells/ 100 cells/hpf 201252 Ramsay Mean intensity healthy vs. FL£ 27 2012 Mean intensity long vs. short survival£ Yang % positive cells 201555
CD3+ cells: ~105 vs. 150 CD3+ cells: ~140 vs. 175 + CD4 : PD-1high 26%, PD-1low 26.4% CD8+: PD-1high 4.8%, PD-1low 42.1%
Gr: grade; nd: not done; pts: patients; SD: standard deviation. FL: follicular lymphoma; TFH: T follicular helper; LN: lymph node. tFL: transformed follicular lymphoma; hpf: high-power field. #+ 150%, ++ 50-75%, +++ >75% of cells positive; £actual values not given, mean numbers estimated from graphs in figures.
cells was associated with high interfollicular content of the same cell type.52 Regardless of region, PD-1 content decreased with stage, and the interfollicular PD-1 content decreased in patients with a high Follicular Lymphoma International Prognostic Index (FLIPI) score. More recent evidence suggests that PD-1+CD4+ cells consist of several sub-populations and include conventional TFH cells and PD-1+TIM-3+ exhausted T cells, which primarily reside in the interfollicular space.54 Functionally exhausted TIM-3+ cells were PD-1low in another study, while the majority of CD4+PD-1high T cells were conventional TFH cells.55 Others identified distinct functional T-cell populations displaying specific gene expression profiles on the basis of CD25, namely CD25+ follicular regulatory T cells and CD25TFH.43 Changes in PD-1, PD-L1 and PD-L2 expression were analyzed with pidilizumab and rituximab treatment in relapsed FL patients.16 PD-L1 but not PD-1 or PD-L2 was significantly higher in blood T cells and monocytes of responders (n=18) than non-responders (n=7). Additional gene expression signature studies conducted in this trial suggested that T-effector cells had anti-tumor and TFH cells had pro-tumor effects, predicting tumor shrinkage and PFS. As this was not recapitulated in an external dataset of 191 patients largely treated with chemotherapy, the predictive power of the identified gene signature might only be relevant with PD-L1/PD-1 blockade. haematologica | 2016; 101(10)
Prognostic relevance of PD-1 expression in FL Gene expression profiling studies demonstrated that the cellular microenvironment plays an essential role in lymphomagenesis and outcome in FL, with enrichment in Tcell and monocyte-restricted genes conferring a favorable prognosis, and with activated macrophages/dendritic genes conferring a poor prognosis.56 However, it appears that both survival and transformation into DLBCL are influenced by the presence and perifollicular versus follicular localization of specific T-cell subtypes, including FOXP3+ TRegs57,58 (Table 6). Several studies have assessed the prognostic relevance of PD-1+ T cells, but findings are contradictory; increased levels of PD-1+ TILs were associated with improved 5-year OS (P=0.004) and PFS in one study (P=0.038), but patients were treated independently of the number of PD-1+ cells, and there was no correlation to the type of therapy and therapeutic response.50 In contrast, increased levels were associated with reduced survival in another study.53 PD-1 was an independent risk factor (RF) in a scoring system predicting 10-year survival rates of 80%, 60%, and 15% in the low (0 RFs, n=14), intermediate (1/2 RFs, n=64) and high-risk group (3/4 RFs, n=13). Using an extremes of survival approach, we detected increased PD-1 expression on follicular T cells in poor outcome versus long-term surviving patients, as well as on CD3+ T cells from patients compared to healthy controls.27 1151
F. McClanahan et al.
Two studies found no impact on time to treatment failure or OS.51,52 The numbers of CD4+ cells were associated with poor outcome, and CD8+ and PD-1+ cells with improved outcome, independently of FLIPI.59 In another study, increased numbers of CD4+PD-1high TFH cells had no impact on survival (P=0.411), while that of exhausted CD4+PD1low (P=0.007) and of CD8+PD-1+ (most likely also exhausted cytotoxic T cells) reduced survival (P=0.026).55 A potential prognostic role has also been attributed to patterns of PD-1+ TILs. The prognostic values of CD4+ and PD-1+ cells were accentuated when they were follicular, and that of CD8+ cells when they were interfollicular.59 Patients with PD-1+ in follicular patterns (i.e. TFH, n=38) also had prolonged time to transformation (TTT) and OS compared to patients with diffuse patterns (n=19), and transformation within one year occurred exclusively in patients with diffuse patterns.54 Multivariate analyses demonstrated that PD-1+ cells with diffuse patterns were associated with shorter TTT (HR 1.9, P=0.045) and inferior OS (HR 2.5, P=0.012), but that inferior outcome was also independently influenced by follicular dendritic cells (HR 3.0, P=0.004). In another study, transformation risk was significantly higher in patients (n=25) with less than 5% PD-1+ TILs compared to other patients.50
CLL/ SLL
sels in reactive LNs.60 Higher PD-L1 expression on CLL cells was also detected in blood in some27,60 but not all studies.61 PD-1 was strongly expressed on ≥50% of tumor cells in the majority of SLL LN specimens and on peripheral blood (PB) neoplastic cells from almost all CLL patients.34 Similar expression patterns were described in flow-cytometry-based studies.60,61 In contrast, the majority of examined SLL/CLL full tissue sections collected from three different institutions (n=58) were PD-1- in another study,40 similar to earlier findings of a lack of PD-1 expression on CLL cells.33
PD-L1/PD-1 expression on CLL/SLL-associated immune cells PD-1+ TILs are generally exceptionally low in CLL/SLL compared to other lymphomas40 (Table 4). We found significantly increased PD-1 expression on T cells from CLL patients compared to reactive LNs, and on PB CLL T cells compared to age-matched healthy donor T cells (both P<0.01).27 While percentages and numbers of CD4+ and CD8+ T cells are significantly increased in CLL patients,60,62,63 marked differences exist in the composition of both CD4+ and CD8+ T-cell subsets. This includes decreased naïve and relatively increased effector cells, with differential PD-1 expression compared to agematched controls, and in specific subpopulations such as BLIMP1HI CD4+ and CD8+ T cells and effector cells.60,62,63
PD-L1/PD-1 expression on tumor cells In initial IHC studies, neither PD-L1 nor PD-L2 were expressed on LN SLL/CLL cells34 (Table 4). Larger IHC studies later found significantly higher PD-L1 expression on CLL cells compared to control LN samples.27,60 Small vessels in CLL LNs also appear to express PD-L1 weakly, whereas this was confined to endothelial cells lining ves-
Prognostic relevance of PD-L1/ PD-1 expression in CLL Studies assessing the prognostic value of PD-L1/ PD-1 in CLL are lacking, and correlations between PD-L1/PD-1 and other conventional prognostic markers have not been identified.34,60,61 Using an extremes of survival approach and a limited number of patient samples, we found significant-
Table 4. Expression of PD-1 and PD-L1 on tumor infiltrating lymphocytes (TILs) and tumor cells in CLL/SLL.
CLL/SLL
Method of quantification TILs
Dorfman 200633 Xerri 200834 Muenst 201040 Grzywnowicz 201261 Ramsay 201227 Tonino 201263 Brusa 201360
Riches 201362
PD-1 expression on Tumor cells
PD-L1 expression on TILs Tumor cells
Positive cases/ all cases
nd
0/4
nd
nd
Proportion of positive cells (IHC)# Pts with positive cells (flow cytometry) Mean number of positive cells/mm2 ±SD Mean % of positive cells/ all cells Pts with positive cells >mean Median % positive
nd nd 13±37 0.2 15/58 (26%) nd
nd nd nd nd nd nd
MFI CLL vs. healthy B cells Mean intensity healthy vs. CLL LN (IHC)£ Mean intensity long vs. short survival LN (IHC)£ MFI healthy vs. CLL PB (flow cytometry)£ % positive effector cells CLL. vs. healthy controls£
nd ~120 vs. 150 nd ~10 vs. 25 CD4: ~20 vs.40 CD8: ~12.5 vs. 25 CD4: ~50 vs. 35 CD8: ~30 vs. 10 ~12 vs. 7
SLL: 12/13 pts ++ to +++ CLL: 10/11 nd Unequivocal in 8/66 pts (5%) nd CLL vs. healthy B cells: 47.2 vs. 14.81 nd nd nd nd nd
nd nd nd nd nd
SLL: 0/7 pts CLL: 0/11 nd nd nd CLL cells 52.52% (10.8–97.3) 9.96 vs. 7.93 ~80 vs. 150 ~120 vs. 150 ~12 vs. 20 nd
~18 vs. <5
nd
~35 vs. 20
nd
nd
~10 vs. 5
nd
nd
nd
Median ~25 vs. 18 CD8: median ~400 vs. 90
nd nd
nd nd
Diffuse: 9/20 pts patchy: 10/20 pts nd nd
% positive cells pts vs. healthy controls (flow cytometry)£ % positive areas in proliferation centers vs. other parts of same slide (IHC) Pattern of expression (IHC) % positive CLL vs. healthy controls£ AN positive cells/µl CLL vs. healthy controls£
IHC: immunohistochemistry; LN: lymph node(s); MFI: median fluorescence intensity; nd: not done; pts: patients; SD: standard deviation. SLL: small lymphocytic lymphoma; CLL: chronic lymphocytic leukemia; PB: peripheral blood. #+ 1-50%, ++ 50-75%, +++ >75% of cells positive; £actual values not given, mean numbers estimated from graphs in figures.
1152
haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas
ly increased expression of PD-L1 on CLL cells and of PD1 on CD3+ T cells in poor prognosis patients (median survival 38 months, n=18) compared with good prognosis patients (median survival >10 years, n=17)27 (Table 6). This, however, was based on a relatively small sample size and requires confirmation in independent patient cohorts. Others described an association between stage, need of therapy and molecular markers and levels of CD4+ and CD8+ subsets, but the exact role of PD-1 has not been established.60
HL/ PMBCL PD-L1/ PD-1 expression on HL and PMBCL cells An underlying molecular mechanism leading to elevat-
ed PD-L1/PD-L2 transcription is present in most patients with HL and PMBCL, as frequent cytogenetic alterations involve chromosome 9p, the coding region for PDL1/PD-L2.64-67 PD-L1 expression on malignant cells has been described by several studies for the majority of PMBCL patients and on Reed–Sternberg (RS) cells in patients with HL, mostly in conjunction with PD-L23335,37,68-70 (Table 5). Expression seems to differ with histological subtype, with strong tumor PD-L1 expression in the majority of patients with nodular sclerosis classical HL (cHL), mixed cellularity cHL and cHL-not otherwise specified (NOS), but only in a small fraction of nodular lymphocyte-predominant HL patients.35 Although tumor infiltration varied widely in this cohort, tumor PD-L1
Table 5. Expression of PD-1 and PD-L1 on tumor infiltrating lymphocytes (TILs) and tumor cells in HL/PMBCL.
HL/PMBCL
Method of quantification TILs
Dorfman 200633 Nam-Cha 200871 Yamamoto 200870 Xerri 200834 Muenst 200972
Greaves 201373 Ansell 201518 Koh 201569
Paydas 201568
TILs
PD-L1 expression on Tumor cells
Positive cases/ all cases
14/14
0/25
nd
RS positive but not quantified
Positive cases/ all cases (rosette formation)
NSCHL 0/43,MCCHL 0/14, LRCHL 10/13, NLPHL 57/58 CD4+: 54.3-76.8% CD8+: 53-66.6% ~5-15 vs. 5-53 Not quantified
nd
nd
nd
nd
nd
Increased, but not quantified
nd nd nd
nd cHL: 8/13 pts + to ++ LPHL: 4/4 pts + to ++ nd
nd
nd
nd
HL 8/9 pts 89% of cells, PMBCL 3/3 pts 100% NSCHL 5% (2-20), MCCHL 2% (2-10), CHL-NOS 50% (2-90), NLPHL 2% (2-5) NSCHL 21/25 (84%), MCCHL 7/8 (>2+ membranous staining)* (88%), CHL-NOS 5/5 (100%), NLPHL 2/15 (13%), PMBCL 15/21 (71%) NSCHL 19/25 (76%), MCCHL 7/8 (88%), CHL-NOS 5/5 (100%), NLPHL 1/15 (10%), PMBCL 19/21 (90%) nd
%positive cells SSS LN£ %positive cells PB healthy vs. HL£ Proportion of positive cells# Mean number of positive cells/mm2
Median number of positive cells/mm2 Andorsky 201137 Chen 201335
PD-1 expression on Tumor cells
% positive cells frozen specimens
nd cHL: 0/30 pts LPHL: 0/5 pts NSCHL 275 ± 493, MCCHL 129 ± 175, nd LRCHL 1044 ± 1116, LDCHL 202 ± 109, cHL-NOS 544 ± 794, NLPHL 296 ± 95 NSCHL 16, MCCHL 37, LRCHL 203, nd LDCHL 49, cHL-NOS 30, NLPHL 297 nd nd
Median and range percent of malignant cells
nd
nd
nd
N (%) cases with ≥5% malignant cells positive
nd
nd
nd
N (%) cases with ≥20% total cellularity positive (>2+ membranous and/or cytoplasmic staining)*
nd
nd
nd
Not detectable in 42%, <0.5% of all nucleated cells in another 40% Positive cells noted in all examined cases
nd
nd
nd
nd
13 pts (11%) membranous positivity
nd
nd
18 cases (20%)
nd
Pts with % positive cells (IHC) %positive cells N (%) pts with ≥20% malignant cells PD-L1 or PD-L2-positive N (%) pts with ≥20 % of microenvironment cells PD-1-positive N (%) cases with ≥5% malignant cells or ≥20% microenvironment cells positive staining intensity
Range 34-99% with staining intensity ++ to +++ 82 pts (75%) cytoplasmic and/or membranous positivity
18 cases (20%) Staining in HRS cells and in microenvironment
n=3: ++ n=15: +
LDCHL: lymphocyte-depleted classical HL; LRCHL: lymphocyte-rich classical HL; MCCHL: mixed cellularity classical HL; nd: not done; NLPHL: nodular lymphocyte-predominant HL; NSCHL: nodular sclerosis CHL; PB: peripheral blood; pts: patients; RS: Reed-Sternberg; SSS: single cell suspension. LN: lymph node; HL: Hodgkin lymphoma; NOS: not otherwise stated; PMBCL: primary mediastinal B-cell lymphoma; HRS: Hodgkin Reed-Sternberg cell. #+ 1-50%, ++ 50-75%, +++ >75% of cells positive; £actual values not given, mean numbers estimated from graphs in figures; *0 no staining, + weak or equivocal staining, ++ moderate staining, +++ strong staining.
haematologica | 2016; 101(10)
1153
F. McClanahan et al.
expression correlated with expression of PD-L1 on tumor-infiltrating macrophages. A more recent study reported PD-L1 positivity in only 20% of examined HL patients, with staining intensities and patterns not further specified.68 In contrast with findings in DLBCL, PDL1 expression is not increased in EBV+ patients.35,68,70 RS cells and variants appear to lack PD-1 expression, suggesting a potentially mutually exclusive expression pattern with PD-L1.33,35
Recently published studies assessing diagnostic cHL TMAs reported PD-1 positivity on microenvironment cells in 11%69 and 20%68 of patients. Interestingly, there were no clear correlations between PD-L1 and PD-1 expression in either study.68,69 PD-1+ cell numbers were lower in both cHL patients with 9p24 gains and with higher amounts of FOXP3+ cells, but correlated with Granzyme-B and T-cell restricted intracellular antigen (TIA-1) expression in another study.72
PD-L1/ PD-1 on HL/PMBCL-associated immune cells
Prognostic relevance of PD-1 expression in HL
Several early studies identified increased numbers of PD-1+ subsets, which frequently form rosettes around tumor cells, especially in lymphocyte-predominant Hodgkin lymphoma subtypes33,70-72 (Table 5). Elevated levels of PD-1+ TILs were also noted in blood T cells of HL patients (n=10) compared to healthy controls, and appeared to be higher in patients with active disease.70 In contrast, using both cHL-derived single-cell suspensions (n=18) and TMAs (n=122), our group found only little expression of PD-1 in TILs, with 40% of patients having less than 0.5% PD-1+ cells.73 In the phase I study on nivolumab, CD3+ TILs in available biopsy specimens largely expressed PD-1, albeit at similarly low levels.18
Associations between microenvironment PD-1 expression and PD-1+ cell numbers and clinical variables or other known phenotypic parameters have not yet been identified.69,72 Regardless, an increased amount of PD-1+ TILs above the prognostic cutoff score (23 cells/mm2) was a stage-independent negative prognostic factor of OS (P=0.005)72 (Table 6). In a prognostic score incorporating numbers of PD-1, Granzyme-B, and FOXP3 expressing cells, different age- and stage-independent outcomes were found between risk groups (FOXP3-PD-1+GrB+ median survival 91 months vs. FOXP3+PD-1-Gr-B- not reached, P<0.0001). Similar associations were noted by our group: albeit expressed at low levels; patients with
Table 6. Prognostic significance of PD-L1 and PD-1 in different types of B-NHL and HL. Orange color signifies reduced survival, green color improved survival, gray color no association between PD-L1/PD-1 and survival.
Reference DLBCL
Cell type analyzed
Cutoff value(s)
PD1+TFH
>median
R-CHOP
OS
PD1+ TIL
No positive cells/hpf vs. presence of positive cells
R-CHOP
OS
PD1+TFH
>20/hpf
Not reported
OS
PD1+ TIL in combination with PD-L1 expression patterns
Median number of PD-1+ TILs
Newly diagnosed and untreated
OS
PD-L1+ tumor cells and/or macrophages PD-L1+ tumor cells
No staining vs. staining ≥30% of lymphoma cells positive ≥20% of total tissue positive
R-CHOP
OS
In combination with PD-L1 expression patterns: improved prognosis in PD-L1–/ TILlow group (n=92) vs. PD-L1+/ TILlow group (n=25), P=0.0086 No prognostic difference between PD-L1+/ TILhigh group (n=3) and PD-L1– /TILhigh group (n=116). No impact
Newly diagnosed and untreated
OS
Decreased survival compared to PD-L1- DLBCL (P=0.009)
OS
Decreased survival compared to microenvironmental PD-L1- DLBCL but not significant
Ahearne 201442 Kwon 201438 Ko 201141 Kiyasu 201539
Kwon 201438 Kiyasu 201539
PD-L1+ microenvironment in patients without tumor PD-L1 expression
Treatment outcome measurement
Prognostic significance
Improved survival Independent prognostic significance in MV analysis Improved survival with increasing numbers of PD1+ TILs/ hpf Independent prognostic significance in MV analysis Decreased survival but not significant
FL Carreras 200950
PD1+TFH
<5% vs. 6-33% vs. >33%
n=80 fludarabine-based 5-year PFS regimes, n=6 alkylating 5-year OS monotherapy, n=3 RT, n=11 w&w
<5% Richendollar 201153 Ramsay 201227
PD1+TFH
>35.6 cells/hpf
CD3+
Extremes of survival
5-year risk of transformation n=23 w&w, n=8 RT, OS n=12 rituximab, n=48 immunochemotherapy Untreated Median OS
Increased survival with increasing PD-1+ TILs: 20% (95%CI 2-38), 46% (30-64), 48% (26-70) 50% (30-70), 77% (64-90), 95% (85-100) independent prognostic factor in MV analysis Increased risk of transformation: 29%, 95% CI 7-51% vs. 7%, 95% CI 1-13%, P<0.05 Decreased survival in MV analysis: HR 1.98, 95% CI 1.09-3.60, P=0.03 PD-1 independent risk factor in scoring system Increased PD-1 expression in poor survival group Increased PD-L1 expression in poor survival group Continued on the next page
1154
haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas
PD-1 expression in >15 cells/hpf had poorer 5-year disease specific survival, while OS was not affected.73 Multivariate analyses demonstrated that high PD-1 (P=0.007) and low FOXP3 expression (P=0.029) were predictors of adverse OS. Significantly reduced OS among PD-1+ patients was also reported in a recently published study, and multivariate analysis identified PD1 expression as an independent prognostic marker for OS (P=0.019) along with high-risk IPS ≥3.69 This was, however, dependent on Ann Arbor clinical stage; in limitedstage cHL, PD-1-positive patients had a worse OS compared with PD-1-negative patients (P=0.048), whereas in advanced stage cHL PD-1-positive status was not associated with OS (P=0.13). Another study found median OS and disease-free survival (DFS) to be shorter in patients with PD-1 compared to those without PD-1 expression, as well as in patients with PD-L1 expression compared to those without, but none of these differences were statistically significant.68 Interestingly, co-expression of PD-1 and PD-L1 emerged as an independent risk factor for prognosis (OR 6.9, 95 % CI 1.9–24.3), and both OS and DFS were significantly reduced among patients with PD1/PD-L1 coexpression compared to both PD-1 and PD-L1 negative patients.
Prognostic relevance of PD-L1 expression in HL Koh et al. reported that patients with tumor PD-L1 expression were more likely to have a low level of lactate dehydrogenase (P=0.024) than PD-L1-negative patients, but neither PD-L1 nor PD-L2 expression were significantly associated with OS (P=0.477 and P=0.676)69 (Table 6).
Discussion Preclinical studies suggest that PD-L1/PD-1 are key mediators of impaired anti-tumor immune responses in lymphomas.13,14 It is therefore reasonable to hypothesize that increased PD-L1/PD-1 expression confers an adverse prognosis, and that such patients might be prime candidates for therapeutic strategies targeting this axis. As prospective studies are currently lacking, we systematically reviewed published data on PD-L1/PD-1 expression and association with prognosis on B-cell lymphoma and lymphoma-associated cells. We found that PD-L1 expression on DLBCL cells is very heterogeneous and present in only a small number of examined samples, while being affected by EBV status and potentially molecular subtype. On FL cells, PD-L1 is absent
Continued from the previous page
PD1+TFH
Continuous variable
CHOP, MCP
OS, TTF
No impact
PD1+TFH
<7.5% vs. 7.5-24.4%vs. >24.4% >25% >26% >45% Extremes of survival
R-CHOP
OS
No impact
Untreated
OS
No impact Poorer survival Poorer survival Poorer survival, especially when follicular Increased survival, especially when interfollicular Increased survival, especially when follicular Follicular pattern prognostically favorable TTT 6.1 vs. 3.6yrs, P=0.033 OS 9.7 vs. 4.6yrs, P=0.009
CD3+ CD20+
Extremes of survival
Untreated
Median OS
Increased PD-1 expression in decreased survival group Increased PD-L1 expression in decreased survival group
Muenst 200972
PD-1+ TILs
>23 cells/mm2
Not specified
Mean OS
Greaves 201373
PD-1+ TILs
>15 cells/hpf
Increased PD-1+ TILs reduce survival: 198 (range 164-234) vs. 283 mo (247-318), P=0.005 PD-1+ counts risk factor in prognostic score Increased PD-1+ TILs reduce DSS but not OS DSS 63% vs. 86%, P =0.012 OS 63% vs. 84%, P=0.18 Predictor of adverse OS in MV analysis OS significantly worse in PD-1+ pts Adverse predictor of OS in MV analysis in limited-stage cHL (P=0.048). OS and DFS significantly worse in PD-1+ and PD-L1+ pts: OS 24 vs.135mo, P=0.002 DFS 20 vs.,107mo, P=0.003
Koch 201252 Takahashi 201351 Yang 201555 Wahlin 201059 Smeltzer 201454
CD4+PD-1high CD4+PD-1low CD8+PD-1low CD4+ CD8+ PD1+ PD1+TFH
Elaborate criteria OS for good versus bad risk pts but treatments not specified Follicular vs. n=42 w&w, Median TTT diffuse pattern n=9 CHOP, n= 5 Median OS anthracycline-combination
CLL Ramsay 201227
HL
Koh 201569
PD-1+ microenvironment
Paydas 201568
PD-L1+ tumor cells PD-1+ microenvironment
n=56 anthracyclines, 5-year DSS n=52 alkylator-based, 5-year OS n=14 RT, n=48 combined modality ≥20% of cells positive ABVD Cumulative OS ≥5% of tumor cells First-line ABVD positive vs. negative Second-line DHAP ≥20% of total tissue positive vs. negative
OS DFS
ABVD: doxorubicin, bleomycin, vinblastine, dacarbazine; CI: confidence interval; DFS: disease free survival; DHAP: dexamethasone, cytarabine, cisplatin; DSS: disease specific survival; HR: hazard ratio; MCP: melphalan, chlorambucil, prednisone; mo: months; MV: multivariate; OS: overall survival; pts: patients; R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone; RT: radiotherapy; TTT: time to transformation; w&w: watch and wait; yrs: years. CHOP: cyclophosphamide, hydroxydaunorubicin, vincristine, and prednisolone; TTF: time to treatment failure; TFH: T follicular helper; TIL: tumor infiltrating lymphoma; hpf: high power field; DLBCL: diffuse large B-cell lymphoma; PFS: progression-free survival.
haematologica | 2016; 101(10)
1155
F. McClanahan et al.
except in an extremes of survival approach. PD-L1 expression on CLL/SLL cells is increased on both LN and PB cells and in patients experiencing short-term survival. Malignant PMBCL and RS cells strongly express PD-L1 and PD-L2, especially in cHL subtypes, while being less affected by EBV serostatus. PD-1 expression was scarce on DLBCL and FL cells and absent on RS cells and variants, whereas highly conflicting findings exist in CLL. PD-1+ TILs in DLBCL are predominantly TFH cells, and numbers are reduced compared to tonsils and other lymphomas. This appears unaffected by molecular subtype, but numbers increase with advanced disease. In FL, PD-1+ cells mainly reside in follicles. Their numbers are comparable to tonsils, but decrease with increasing histological grade, advanced stage and transformation. Several sub-populations of PD-1+CD4+ cells with distinct localization preferences and functions have been identified, including conventional TFH, exhausted and follicular regulatory T cells. Compared to other lymphomas, PD-1+ TILs numbers appears to be low in SLL/CLL LNs, but increased relative and absolute T cell numbers and functionally distinct subsets are present in blood. In HL, conflicting findings exist regarding the architectural structure of PD-1+ T-cell subsets and levels of PD-1+ TILs, potentially due to differences in examined histological subtypes and disease activity. This heterogeneity within and across lymphoma entities is reflected by contradictory findings on the prognostic role of PD-1+ TILs, especially in DLBCL. On first sight, the same seems to be true for FL. However, both prognosis and transformation appear to be determined by follicular versus interfollicular localizations of exhausted versus functional or regulatory CD4+ and CD8+ cells. A more defined role exists in HL, where despite low and/ or variable overall numbers, elevated numbers of PD-1+ TILs confer a poor prognosis. PD-L1 expression was generally found to be an adverse prognostic marker across examined lymphoma types. Such heterogeneous findings can partly be explained by differences in the nature and composition of the examined cohorts (sample sizes, patient characteristics, treatment, etc.). Another explanation are differing methodologies, including the choice of reagents, analysis systems and definition of positivity and cutoff values. A validation study from a lymphoma consortium on the FL microenvironment reported considerable differences between manual scoring and automated microscopy systems and flow cytometry, which was also dependent on the investigating laboratory.74 Within semi-automated image analysis systems, a high concordance seems to exist.49 Among the included studies, expression was predominantly assessed by IHC. However, methods of quantifying positive cells and the definition of staining intensity and positivity varied widely. In selected studies, different counting methods were compared or verified with flow-cytometry results. Several studies have also accounted for intra- and inter-observer bias, showing good reproducibility especially in areas with fewer PD-1+ cells. Similar issues have been observed in solid malignancies, where the use of PD-L1 as a biomarker is confounded by detection antibodies, differing cutoffs and differences in tissue preparation and processing variability.75 It is also likely that biological behavior and prognosis are determined not only by overall PD-1+ TILs and tumor cells, but by functionally distinct subsets. PD-1+ numbers correlated with CD4+ T-cell and FoxP3+ numbers and GrB and TIA-1+ cells in several studies,42,52,72 and similar associations 1156
were found between distribution patterns of FoxP3+ and PD-1+ cells. Modulating effects might also be exerted by other microenvironment components such as TAMs,35 tumor-associated histiocytes,37 and small vessels.60 Studies in CLL, for example, suggest that monocyte-derived suppressor cells with high PD-L1 expression and/ or skewed monocyte subpopulations are increased in patients and preclinical models and modulate T-cell responses.76,77 In multiple solid cancer types, clinical responses were observed in patients with high PD-L1 expression on tumor-infiltrating immune cells and in those with TH1 gene signatures and Tcell CTLA-4 expression at baseline.23 Immune dysfunction might also be mediated by other (potentially inducible) immune checkpoint receptor-ligand interactions, for example, by the binding of PD-1 to PD-L278 or by signaling via CD200, CD270 and CD276,27 or by additional tumor-associated and/ or genetic determinants.22 Upregulation of TIM3 was recently reported in preclinical models of lung adenocarcinoma, where tumors progressed following response to anti-PD-1 therapy.79 Optimally, the importance of these components should be assessed within one analysis and in conjunction with established clinic-pathological features. Regardless of the expression and functions of PD-L1/PD1 expressing cell subsets, blocking PD-L1/PD-1 interactions is safe and effective in patients with relapsed/refractory FL, DLBCL and HL.16-18 This indicates that PD-L1/PD-1 expression on tumor cells or TILs cannot be used in isolation to predict outcome of treatment for individual patients. This is further supported by observations that the numbers of PDL1+ Tregs, CD4+ and CD8+ central memory cells, and PD-L1+ monocytes increased during treatment.17 PD-1/PD-L1/PDL2 expression changes could also be noted in responding versus non-responding patients.16 Altogether, this work highlights that PD-L1/PD-1 expression on tumor cells and the microenvironment is only one aspect, albeit an essential one, determining the biology of lymphomas, and that the inclusion of additional components will be required to form prognostic models. Therefore, attempts should be made to harmonize quantification methods and reporting of PD-L1/PD-1, optimally in the context of clinical studies on immune checkpoint inhibitors. Clinical study strategies should also include the identification of additional potential biomarkers using highthroughput technologies such as whole-exome sequencing, gene expression signatures/ patterns, epigenetic modifications, protein microarrays and flow and mass cytometry. To address the challenges of assay comparability, performance standardization, interpretation of test results and safe translation into patient care, the US Food and Drug Administration (FDA), the American Association for Cancer Research (AACR) and the American Society of Clinical Oncology (ASCO) recently convened a workshop entitled â&#x20AC;&#x153;Complexities in Personalized Medicine: Harmonizing Companion Diagnostics Across a Class of Targeted Therapiesâ&#x20AC;?. As a collaboration between several companies, a blueprint proposal was developed with the goal to agree on and deliver a package of information /data upon which analytic comparison of various diagnostic assays may be conducted in non-small cell lung cancer treated with PD1/PD-L1 inhibitors.80 It is anticipated that the proposed study will build the pre-clinical evidence for PDL1/PD-1 diagnostic characterization and lead to postapproval studies that will help inform personalized treatment decisions, and ultimately be applied to other tumor entities as well. haematologica | 2016; 101(10)
Prognostic value of PD-L1/PD-1 in B-cell lymphomas
References 1. Garon EB, Rizvi NA, Hui R, et al. Pembrolizumab for the treatment of nonsmall-cell lung cancer. N Engl J Med. 2015;372(21):2018-2028. 2. Robert C, Schachter J, Long GV, et al. Pembrolizumab versus Ipilimumab in Advanced Melanoma. N Engl J Med. 2015;372(26):2521-2532. 3. Robert C, Long GV, Brady B, et al. Nivolumab in Previously Untreated Melanoma without BRAF Mutation. N Engl J Med. 2015;372(4):320-330. 4. Larkin J, Chiarion-Sileni V, Gonzalez R, et al. Combined Nivolumab and Ipilimumab or Monotherapy in Untreated Melanoma. N Engl J Med. 2015;373(1):23-34. 5. Gettinger SN, Horn L, Gandhi L, et al. Overall Survival and Long-Term Safety of Nivolumab (Anti–Programmed Death 1 Antibody, BMS-936558, ONO-4538) in Patients With Previously Treated Advanced Non–Small-Cell Lung Cancer. J Clin Oncol. 2015;33(18):2004-2012. 6. Motzer RJ, Escudier B, McDermott DF, et al. Nivolumab versus Everolimus in Advanced Renal-Cell Carcinoma. N Engl J Med. 2015;373(19):1803-1813. 7. Powles T, Eder JP, Fine GD, et al. MPDL3280A (anti-PD-L1) treatment leads to clinical activity in metastatic bladder cancer. Nature. 2014;515(7528):558-562. 8. Robert C, Ribas A, Wolchok JD, et al. Antiprogrammed-death-receptor-1 treatment with pembrolizumab in ipilimumab-refractory advanced melanoma: a randomised dose-comparison cohort of a phase 1 trial. Lancet. 2014;384(9948):1109-1117. 9. Topalian SL, Sznol M, McDermott DF, et al. Survival, Durable Tumor Remission, and Long-Term Safety in Patients With Advanced Melanoma Receiving Nivolumab. J Clin Oncol. 2014;32(10):1020-1030. 10. Hamid O, Robert C, Daud A, et al. Safety and Tumor Responses with Lambrolizumab (Anti–PD-1) in Melanoma. N Engl J Med. 2013;369(2):134-144. 11. Brahmer JR, Tykodi SS, Chow LQM, et al. Safety and Activity of Anti–PD-L1 Antibody in Patients with Advanced Cancer. N Engl J Med. 2012;366(26):2455-2465. 12. Topalian SL, Hodi FS, Brahmer JR, et al. Safety, Activity, and Immune Correlates of Anti-PD-1 Antibody in Cancer. N Engl J Med. 2012;366(26):2443-2454. 13. Greaves P, Gribben JG. The role of B7 family molecules in hematologic malignancy. Blood. 2013;121(5):734-744. 14. Shi L, Chen S, Yang L, Li Y. The role of PD1 and PD-L1 in T-cell immune suppression in patients with hematological malignancies. J Hematol Oncol. 2013;6(1):74. 15. Berger R, Rotem-Yehudar R, Slama G, et al. Phase I Safety and Pharmacokinetic Study of CT-011, a Humanized Antibody Interacting with PD-1, in Patients with Advanced Hematologic Malignancies. Clin Cancer Res. 2008;14(10):3044-3051. 16. Westin JR, Chu F, Zhang M, et al. Safety and activity of PD1 blockade by pidilizumab in combination with rituximab in patients with relapsed follicular lymphoma: a single group, open-label, phase 2 trial. Lancet Oncol. 2014;15(1):69-77. 17. Armand P, Nagler A, Weller EA, et al. Disabling Immune Tolerance by Programmed Death-1 Blockade With Pidilizumab After Autologous Hematopoietic Stem-Cell Transplantation
haematologica | 2016; 101(10)
18.
19.
20.
21. 22.
23.
24.
25.
26.
27.
28.
29.
30. 31. 32. 33.
34.
35.
for Diffuse Large B-Cell Lymphoma: Results of an International Phase II Trial. J Clin Oncol. 2013;31(33):4199-4206. Ansell SM, Lesokhin AM, Borrello I, et al. PD-1 Blockade with Nivolumab in Relapsed or Refractory Hodgkin's Lymphoma. N Engl J Med. 2015;372(4):311-319. Wang A, Wang HY, Liu Y, et al. The prognostic value of PD-L1 expression for nonsmall cell lung cancer patients: a meta-analysis. Eur J Surg Oncol. 2015;41(4):450-456. Zhang Y, Kang S, Shen J, et al. Prognostic significance of programmed cell death 1 (PD-1) or PD-1 ligand 1 (PD-L1) Expression in epithelial-originated cancer: a meta-analysis. Medicine (Baltimore). 2015;94(6):e515. Wu P, Wu D, Li L, Chai Y, Huang J. PD-L1 and Survival in Solid Tumors: A MetaAnalysis. PLoS One. 2015;10(6):e0131403. Sui X, Ma J, Han W, et al. The anticancer immune response of anti-PD-1/PD-L1 and the genetic determinants of response to antiPD-1/PD-L1 antibodies in cancer patients. Oncotarget. 2015;6(23):19393-19404. Herbst RS, Soria JC, Kowanetz M, et al. Predictive correlates of response to the antiPD-L1 antibody MPDL3280A in cancer patients. Nature. 2014;515(7528):563-567. Burger JA, Gribben JG. The microenvironment in chronic lymphocytic leukemia (CLL) and other B cell malignancies: Insight into disease biology and new targeted therapies. Semin Cancer Biol. 2014;24:71-81. Gribben JG, Riches JC. Immunotherapeutic strategies including transplantation: eradication of disease. Hematology Am Soc hematol Educ Program. 2013;2013(1):151-157. Kiaii S, Clear AJ, Ramsay AG, et al. Follicular Lymphoma Cells Induce Changes in T-Cell Gene Expression and Function: Potential Impact on Survival and Risk of Transformation. J Clin Oncol. 2013;31(21): 2654-2661. Ramsay AG, Clear AJ, Fatah R, Gribben JG. Multiple inhibitory ligands induce impaired T-cell immunologic synapse function in chronic lymphocytic leukemia that can be blocked with lenalidomide: establishing a reversible immune evasion mechanism in human cancer. Blood. 2012;120(7):1412-1421. McClanahan F, Hanna B, Miller S, et al. PDL1 Checkpoint Blockade Prevents Immune Dysfunction and Leukemia Development in a Mouse Model of Chronic Lymphocytic Leukemia. Blood. 2015;126(2):203-211. Fife BT, Pauken KE. The role of the PD-1 pathway in autoimmunity and peripheral tolerance. Ann N Y Acad Sci. 2011;1217:4559. Zou W, Chen L. Inhibitory B7-family molecules in the tumour microenvironment. Nature Rev Immunol. 2008;8(6):467-477. Keir ME, Butte MJ, Freeman GJ, Sharpe AH. PD-1 and its ligands in tolerance and immunity. Annu Rev Immunol. 2008;26:677-704. Crotty S. Follicular helper CD4 T cells (TFH). Annu Rev Immunol. 2011;29:621663. Dorfman DM, Brown JA, Shahsafaei A, Freeman GJ. Programmed death-1 (PD-1) is a marker of germinal center-associated T cells and angioimmunoblastic T-cell lymphoma. Am J Surg Pathol. 2006;30(7):802-810. Xerri L, Chetaille B, Seriari N, et al. Programmed death 1 is a marker of angioimmunoblastic T-cell lymphoma and B-cell small lymphocytic lymphoma/chronic lymphocytic leukemia. Hum Pathol. 2008;39(7): 1050-1058. Chen BJ, Chapuy B, Ouyang J, et al. PD-L1 expression is characteristic of a subset of
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
aggressive B-cell lymphomas and virus-associated malignancies. Clin Cancer Res. 2013;19(13):3462-3473. Rossille D, Gressier M, Damotte D, et al. High level of soluble programmed cell death ligand 1 in blood impacts overall survival in aggressive diffuse large B-Cell lymphoma: results from a French multicenter clinical trial. Leukemia. 2014;28(12):2367-2375. Andorsky DJ, Yamada RE, Said J, Pinkus GS, Betting DJ, Timmerman JM. Programmed Death Ligand 1 Is Expressed by Non– Hodgkin Lymphomas and Inhibits the Activity of Tumor-Associated T Cells. Clin Cancer Res. 2011;17(13):4232-4244. Kwon D, Kim S, Kim PJ, et al. Clinicopathological analysis of programmed cell death-1 and programmed cell death-ligand 1 expression in the tumor microenvironments of diffuse large B-cell lymphomas. Histopathology. 2016;68(7):1079-1089. Kiyasu J, Miyoshi H, Hirata A, et al. Expression of programmed cell death ligand 1 is associated with poor overall survival in patients with diffuse large B-cell lymphoma. Blood. 2015;126(19):2193-2201. Muenst S, Hoeller S, Willi N, Dirnhofera S, Tzankov A. Diagnostic and prognostic utility of PD-1 in B cell lymphomas. Dis Markers. 2010;29(1):47-53. Ko YS, Oh YH, Park CK, et al. Prognostic Implication of Programmed Death-1Positive Tumor-infiltrating Lymphocytes in Diffuse Large B-Cell Lymphoma. Korean J Pathol. 2011;45(6):573-581. Ahearne MJ, Bhuller K, Hew R, Ibrahim H, Naresh K, Wagner SD. Expression of PD-1 (CD279) and FoxP3 in diffuse large B-cell lymphoma. Virchows Arch. 2014;465(3): 351-358. Ame-Thomas P, Le Priol J, Yssel H, et al. Characterization of intratumoral follicular helper T cells in follicular lymphoma: role in the survival of malignant B cells. Leukemia. 2012;26(5):1053-1063. Rosenwald A, Wright G, Chan WC, et al. The use of molecular profiling to predict survival after chemotherapy for diffuse large-Bcell lymphoma. N Engl J Med. 2002;346 (25):1937-1947. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature. 2000;403(6769):503-511. Scott DW. Cell-of-Origin in Diffuse Large BCell Lymphoma: Are the Assays Ready for the Clinic? Am Soc Clin oncol Educ Book. 2015;35:e458-466. Hasselblom S, Sigurdadottir M, Hansson U, Nilsson-Ehle H, Ridell B, Andersson PO. The number of tumour-infiltrating TIA-1+ cytotoxic T cells but not FOXP3+ regulatory T cells predicts outcome in diffuse large B-cell lymphoma. Br J Haematol. 2007;137(4):364373. Tzankov A, Meier C, Hirschmann P, Went P, Pileri SA, Dirnhofer S. Correlation of high numbers of intratumoral FOXP3+ regulatory T cells with improved survival in germinal center-like diffuse large B-cell lymphoma, follicular lymphoma and classical Hodgkin's lymphoma. Haematologica. 2008;93(2):193-200. Coutinho R, Clear AJ, Mazzola E, et al. Revisiting the immune microenvironment of diffuse large B-cell lymphoma using a tissue microarray and immunohistochemistry: robust semi-automated analysis reveals CD3 and FoxP3 as potential predictors of response to R-CHOP. Haematologica. 2015;100(3):363-369.
1157
F. McClanahan et al. 50. Carreras J, Lopez-Guillermo A, Roncador G, et al. High numbers of tumor-infiltrating programmed cell death 1-positive regulatory lymphocytes are associated with improved overall survival in follicular lymphoma. J Clin Oncol. 2009;27(9):1470-1476. 51. Takahashi H, Tomita N, Sakata S, et al. Prognostic significance of programmed cell death-1-positive cells in follicular lymphoma patients may alter in the rituximab era. Eur J Haematol. 2013;90(4):286-290. 52. Koch K, Hoster E, Unterhalt M, et al. The composition of the microenvironment in follicular lymphoma is associated with the stage of the disease. Hum Pathol. 2012;43(12):2274-2281. 53. Richendollar BG, Pohlman B, Elson P, Hsi ED. Follicular programmed death 1-positive lymphocytes in the tumor microenvironment are an independent prognostic factor in follicular lymphoma. Hum Pathol. 2011;42(4):552-557. 54. Smeltzer JP, Jones JM, Ziesmer SC, et al. Pattern of CD14+ follicular dendritic cells and PD1+ T cells independently predicts time to transformation in follicular lymphoma. Clin Cancer Res. 2014;20(11):2862-2872. 55. Yang ZZ, Grote DM, Ziesmer SC, Xiu B, Novak AJ, Ansell SM. PD-1 expression defines two distinct T-cell sub-populations in follicular lymphoma that differentially impact patient survival. Blood Cancer J. 2015;5:e281. 56. Dave SS, Wright G, Tan B, et al. Prediction of survival in follicular lymphoma based on molecular features of tumor-infiltrating immune cells. N Engl J Med. 2004;351(21): 2159-2169. 57. Lee AM, Clear AJ, Calaminici M, et al. Number of CD4+ Cells and Location of Forkhead Box Protein P3â&#x20AC;&#x201C;Positive Cells in Diagnostic Follicular Lymphoma Tissue Microarrays Correlates With Outcome. J Clin Oncol. 2006;24(31):5052-5059. 58. Farinha P, Al-Tourah A, Gill K, Klasa R, Connors JM, Gascoyne RD. The architectural pattern of FOXP3-positive T cells in follicular lymphoma is an independent predictor of survival and histologic transformation. Blood. 2010;115(2):289-295. 59. Wahlin BE, Aggarwal M, Montes-Moreno S, et al. A unifying microenvironment model in follicular lymphoma: outcome is predicted by programmed death-1--positive, regulatory, cytotoxic, and helper T cells and macrophages. Clin Cancer Res. 2010;16(2): 637-650.
1158
60. Brusa D, Serra S, Coscia M, et al. The PD1/PD-L1 axis contributes to T-cell dysfunction in chronic lymphocytic leukemia. Haematologica. 2013;98(6):953-963. 61. Grzywnowicz M, Zaleska J, Mertens D, et al. Programmed death-1 and its ligand are novel immunotolerant molecules expressed on leukemic B cells in chronic lymphocytic leukemia. PLoS One. 2012;7(4):e35178. 62. Riches JC, Davies JK, McClanahan F, et al. T cells from CLL patients exhibit features of Tcell exhaustion but retain capacity for cytokine production. Blood. 2013;121(9): 1612-1621. 63. Tonino SH, van de Berg PJ, Yong SL, et al. Expansion of effector T cells associated with decreased PD-1 expression in patients with indolent B cell lymphomas and chronic lymphocytic leukemia. Leuk Lymphoma. 2012;53(9):1785-1794. 64. Twa DD, Chan FC, Ben-Neriah S, et al. Genomic rearrangements involving programmed death ligands are recurrent in primary mediastinal large B-cell lymphoma. Blood. 2014;123(13):2062-2065. 65. Green MR, Monti S, Rodig SJ, et al. Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood. 2010;116(17):3268-3277. 66. Rosenwald A, Wright G, Leroy K, et al. Molecular diagnosis of primary mediastinal B cell lymphoma identifies a clinically favorable subgroup of diffuse large B cell lymphoma related to Hodgkin lymphoma. J Exp Med. 2003;198(6):851-862. 67. Joos S, Otano-Joos M, Ziegler S, et al. Primary mediastinal (thymic) B-cell lymphoma is characterized by gains of chromosomal material including 9p and amplification of the REL gene. Blood. 1996;87(4): 1571-1578. 68. Paydas S, Ba Äąr E, Seydaoglu G, Ercolak V, Ergin M. Programmed death-1 (PD-1), programmed death-ligand 1 (PD-L1), and EBVencoded RNA (EBER) expression in Hodgkin lymphoma. Ann Hematol. 2015;94(9):15451552. 69. Koh YW, Jeon YK, Yoon DH, Suh C, Huh J. Programmed death 1 expression in the peritumoral microenvironment is associated with a poorer prognosis in classical Hodgkin lymphoma. Tumour Biol. 2016;37(6):75077514.
70. Yamamoto R, Nishikori M, Kitawaki T, et al. PD-1/PD-1 ligand interaction contributes to immunosuppressive microenvironment of Hodgkin lymphoma. Blood. 2008;111 (6):3220-3224. 71. Nam-Cha SH, Roncador G, Sanchez-Verde L, et al. PD-1, a follicular T-cell marker useful for recognizing nodular lymphocyte-predominant Hodgkin lymphoma. Am J Surg Pathol. 2008;32(8):1252-1257. 72. Muenst S, Hoeller S, Dirnhofer S, Tzankov A. Increased programmed death-1+ tumorinfiltrating lymphocytes in classical Hodgkin lymphoma substantiate reduced overall survival. Hum Pathol. 2009;40(12):1715-1722. 73. Greaves P, Clear A, Owen A, et al. Defining characteristics of classical Hodgkin lymphoma microenvironment T-helper cells. Blood. 2013;122(16):2856-2863. 74. Sander B, de Jong D, Rosenwald A, et al. The reliability of immunohistochemical analysis of the tumor microenvironment in follicular lymphoma: a validation study from the Lunenburg Lymphoma Biomarker Consortium. Haematologica. 2014;99(4): 715-725. 75. Patel SP, Kurzrock R. PD-L1 Expression as a Predictive Biomarker in Cancer Immunotherapy. Mol Cancer Ther. 2015;14 (4):847-856. 76. Jitschin R, Braun M, Buettner M, et al. CLLcells induce IDOhi CD14+HLA-DRlo myeloid-derived suppressor cells that inhibit T-cell responses and promote TRegs. Blood. 2014;124(5):750-760. 77. Hanna BS, McClanahan F, Yazdanparast H, et al. Depletion of CLL-associated patrolling monocytes and macrophages controls disease development and repairs immune dysfunction in vivo. Leukemia. 2016;30(3):570-579. 78. Latchman Y, Wood CR, Chernova T, et al. PD-L2 is a second ligand for PD-1 and inhibits T cell activation. Nat Immunol. 2001;2(3):261-268. 79. Koyama S, Akbay EA, Li YY, et al. Adaptive resistance to therapeutic PD-1 blockade is associated with upregulation of alternative immune checkpoints. Nature communications. 2016;7:10501. 80. A Blueprint Proposal for Companion Diagnostic Comparability. Available at: http://www.aacr.org/AdvocacyPolicy/Gove rnmentAffairs/Pages/industry-workinggroup-blueprint-proposal.aspx#. VyEupkfVt5A [Last accessed: April 27th 2016.]
haematologica | 2016; 101(10)
ARTICLE
Coagulation & Its Disorders
A population pharmacokinetic model for perioperative dosing of factor VIII in hemophilia A patients
EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
Hendrika Hazendonk,1 Karin Fijnvandraat,2 Janske Lock,1 Mariëtte Driessens,3 Felix van der Meer,4 Karina Meijer,5 Marieke Kruip,6 Britta Laros-van Gorkom,7 Marjolein Peters,2 Saskia de Wildt,8,9 Frank Leebeek,6 Marjon Cnossen,1* and Ron Mathôt;10* for the “OPTI-CLOT” study group*
Department of Pediatric Hematology, Erasmus University Medical Center - Sophia Children’s Hospital Rotterdam; 2Department of Pediatric Hematology, Academic Medical Center, Amsterdam; 3Netherlands Hemophilia Patient Organization (NVHP), Nijkerk; 4 Department of Thrombosis and Hemostasis, Leiden University Medical Center; 5 University of Groningen, Department of Hematology, University Medical Center Groningen; 6Department of Hematology, Erasmus University Medical Center, Rotterdam; 7 Department of Hematology, Radboud university medical center; 8Intensive Care and Department of Pediatric Intensive Care, Erasmus University Medical Center - Sophia Children’s Hospital Rotterdam; 9Department of Pharmacology, Radboud university medical center; and 10Hospital Pharmacy-Clinical Pharmacology, Academic Medical Center Amsterdam, The Netherlands 1
*Both contributed equally to this work.
Haematologica 2016 Volume 101(10):1159-1169
ABSTRACT
T
he role of pharmacokinetic-guided dosing of factor concentrates in hemophilia is currently a subject of debate and focuses on longterm prophylactic treatment. Few data are available on its impact in the perioperative period. In this study, a population pharmacokinetic model for currently registered factor VIII concentrates was developed for severe and moderate adult and pediatric hemophilia A patients (FVIII levels <0.05 IUmL-1) undergoing elective, minor or major surgery. Retrospective data were collected on FVIII treatment, including timing and dosing, time point of FVIII sampling and all FVIII plasma concentrations achieved (trough, peak and steady state), brand of concentrate, as well as patients' and surgical characteristics. Population pharmacokinetic modeling was performed using non-linear mixed-effects modeling. Population pharmacokinetic parameters were estimated in 75 adults undergoing 140 surgeries (median age: 48 years; median weight: 80 kg) and 44 children undergoing 58 surgeries (median age: 4.3 years; median weight: 18.5 kg). Pharmacokinetic profiles were best described by a two-compartment model. Typical values for clearance, inter-compartment clearance, central and peripheral volume were 0.15 L/h/68 kg, 0.16 L/h/68 kg, 2.81 L/68 kg and 1.90 L/68 kg. Interpatient variability in clearance and central volume was 37% and 27%. Clearance decreased with increasing age (P<0.01) and increased in cases with blood group O (26%; P<0.01). In addition, a minor decrease in clearance was observed when a major surgical procedure was performed (7%; P<0.01). The developed population model describes the perioperative pharmacokinetics of various FVIII concentrates, allowing individualization of perioperative FVIII therapy for severe and moderate hemophilia A patients by Bayesian adaptive dosing. haematologica | 2016; 101(10)
Correspondence: r.mathot@amc.uva.nl
Received: September 6, 2015. Accepted: July 1, 2016. Pre-published: July 6, 2016. doi:10.3324/haematol.2015.136275
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1159
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is required for any other use.
1159
H. Hazendonk et al.
Introduction Hemophilia A is an X-linked hereditary bleeding disorder characterized by a deficiency of coagulation factor VIII (FVIII). Current management of hemophilia patients consists of replacement therapy with plasma derived or recombinant factor concentrates in case of acute bleeding (“on demand”) or to prevent spontaneous or perioperative bleeding (“prophylaxis”). The aim of long-term prophylactic treatment is to prevent severe joint damage and subsequent long-term invalidity by raising FVIII trough plasma concentrations to at least 0.01 IUml-1.1,2 To acquire adequate hemostasis in the surgical setting, normalization of coagulation factor levels is advocated for 7-14 days after surgery in most perioperative protocols.3 Treatment with factor concentrates is costly. In the Netherlands, total annual costs of replacement therapy are estimated at more than 130 million euro and include costs for prophylactic and “on demand” treatment.4-7 In the Canadian Hemophilia Registry, perioperative consumption amounts to 1%-3% of the total annual amount administered.8 As we have reported earlier, coagulation factor plasma concentrations as recommended by National and International Guidelines are often exceeded in the perioperative setting to avoid lower plasma concentrations and a possibly higher bleeding risk, with additional costs.9,10 In a retrospective analysis of hemophilia A patients undergoing surgery, 45% of FVIII plasma concentrations were below the target range during the first 24 hours after surgery and 75% of the plasma concentration were above the target range after six days of hospitalization. In addition, a reduction of 44% in factor concentrates could have been reached if plasma concentrations had been maintained within target levels in the perioperative setting.9 In the prophylactic setting, Carlsson et al. have shown that FVIII consumption can be significantly reduced by application of pharmacokinetic (PK) modeling to individualize dosing regimens.11-14 In the perioperative setting, Longo et al. have reported excessive FVIII consumption and clearance in 50% of surgical hemophilia patients due to unidentified factors.15 This suggests mechanisms of increased clearance due to hemostatic challenges during surgery. Although an initial preoperative factor concentrate bolus dose may be individualized by individual PK parameters obtained after an individual PK profile based on a prophylactic population PK model, this may not be applicable as soon as a surgical procedure is initiated. A perioperative population PK model, however, would make PK-guided iterative adaptive Bayesian dosing with a potential concomitant decrease of factor concentrate consumption possible. During this procedure individual PK parameters are iteratively up-dated by combining PK information (e.g. dose, concentration, time) from the indi-
vidual patient with a priori PK information (e.g. average clearance, variability) from the population. But this information is not currently available and has, therefore, never been performed. In order to construct a perioperative population PK model of this kind, to facilitate Bayesian adaptive dosing in severe and moderate hemophilia A, we collected detailed retrospective FVIII infusion data in patients who had undergone surgery under replacement therapy with various similar FVIII concentrates from five hemophilia treatment centers.
Methods Patients’ characteristics and data collection Severe and moderate hemophilia A patients of all ages with FVIII plasma concentration less than 0.05 IUml-1 who had undergone elective, minor or major surgical procedures between 2000 and 2013 from five Academic Hemophilia Treatment Centers in the Netherlands were included.9 Patients received replacement therapy consisting of various recombinant factor concentrates (Kogenate FS: Bayer, Berkely, Ca, USA; Helixate FS: CSL Behring, Marburg, Germany; Advate and Recombinate: Baxter Bioscience, Thousand Oaks, CA, USA; Refacto AF: Pfizer, New York, NY USA) or plasma derived factor concentrates (Aafact: Blood Transfusion council of the Netherlands Red Cross; Hemofil M: Baxter Bioscience, Thousand Oaks, CA, USA) to achieve target FVIII plasma concentrations as set by the National Hemophilia Consensus. This guideline recommends peak and trough FVIII plasma concentrations on consecutive postoperative days (Table 1): 0-24 hours 0.80-1.00 IUml-1; 24-120 hours 0.50-0.80 IUml-1 and more than 120 hours 0.30-0.50 IUml-1.3 The following retrospective data were collected: FVIII dosages, detailed timing of administration and timing of FVIII blood sampling, mode of infusion (continuous or bolus infusion), all achieved FVIII plasma concentrations (both trough, peak and steady state plasma concentrations), patients' and surgical characteristics, and concomitant medication with a possible effect on hemostasis (i.e. tranexamic acid, heparin, desmopressin and non-steroidal anti-inflammatory drugs). Patients' characteristics included: body weight, length, lean body mass,16,17 body mass index (BMI),18 blood group, von Willebrand Factor (VWF) antigen and VWF activity (historically measured), liver and renal function, clinical bleeding phenotype, history of FVIII inhibiting antibodies, intensity of prophylactic dosing regimen, brand of concentrate, and treatment center. Surgical characteristics included type and severity of surgical procedure categorized into minor, major and high risk according to Koshy et al.19 In all centers, FVIII plasma concentrations were measured by one-stage clotting assays. The study was not subject to the conditions of the Medical Research Involving Human Subjects Act, as patient data were analyzed anonymously. The study was approved by all local Medical Ethics Committees; one center required prior patient informed consent.
Table 1. Prevalence of under- and overdosing in the perioperative period.*
Time (hours) Consensus % above % below
0-24
24-120 -1
0.80-1.00 IUml 33% (>1.00 IUml-1) 45% (<0.80 IUml-1)
>120 -1
0.50-0.80 IUml 59% (>0.80 IUml-1) 7% (<0.50 IUml-1)
0.30-0.50 IUml-1 75% (>0.50 IUml-1) 9% (<0.30 IUml-1)
*According to the National Hemophilia Consensus.
1160
haematologica | 2016; 101(10)
A perioperative population pharmacokinetic model
Pharmacokinetic modeling Population PK is defined as the study of sources of variability in drug concentrations after dosing that occurs both in individual patients and between patients.20 In the present population analy-
sis, all plasma concentration time points were analyzed simultaneously using non-linear mixed-effects modeling software (NONMEM v.7.2.0; Globomax LLC, Ellicott City, Maryland, USA).21 All PK-related abbreviations and terminology are described in Online
Table 2. Characteristics of the study population.
Patients' characteristics
Total cohort
N. of patients 119 Age (years) 40 [0.2-78] Weight (kg) 75 [5-111] Severe hemophilia (FVIII levels <0.01 IUml-1) 83 (69.7) On prophylaxis 84 (70.6) Blood group O* 51 (50.5) Historical VWF levels (mmoll) Antigen 1.1 [0.3-2.5] Activity 1.1 [0.2-2.7] Surgical characteristics Total n. of surgical procedures 198 N. of patients undergoing: 1 procedure 75 (63.0) 2 procedures 26 (21.8) 3 procedures 9 (7.6) >4 procedures 9 (7.6) Major surgical procedure 97 (49.0) Type of surgical procedure General 6 (3.0) Colo-rectal 5 (2.5) Vascular 1 (0.5) Cardio-thoracic 1 (0.5) Orthopedic 94 (47.5) Urology 12 (6.1) Maxillofacial 2 (1.0) Ear-nose-throat 11 (5.6) Eye 3 (1.5) (Re)placement of central 32 (16.2) intravenous catheters Miscellaneous 31 (15.7) Replacement therapy with factor concentrate, hospitalization and blood loss Mode of infusion Continuous 115 (58.1) Bolus 83 (41.9) Product type Recombinant 152 (76.8) Plasma derived 46 (23.2) Duration of hospitalization (days) 9 [1-50] Complications during the perioperative period N. of patients with a complication Bleeding 48 (24.2) Re-operation 6 (3.0) Hemoglobin drop >20 gL-1 38 (19.2) and/or erythrocyte transfusion Bleeding with prolonged hospitalization 5 (2.5) Thrombosis 0 0 FVIII data FVIII measurements (trough, peak and SS) 1389 Prior to surgery 158 (11.4) Day 1 (0 - 24 hours) 323 (23.2) Day 2 - 5 (24 - 120 hours) 473 (34.0) Day > 6 (>120 hours) 436 (31.4)
Adults N. (%); or median [minimum; maximum]
Children
75 48 80
[19-78] [45-111]
44 4 19
[0.2-17.3] [5-85]
49 51 34
(65.3) (68.0) (50.0)
34 33 17
(77.3) (75.0) (51.5)
1.2 1.4
[0.3-2.5] [0.2-2.7]
0.9 0.9
[0.5-2.3] [0.4-1.7]
140
58
43 15 9 8 86
(57.3) (20.0) (12.0) (10.7) (61.4)
32 11 0 1 11
(72.7) (25.0) (0.0) (2.3) (19.0)
6 4 1 1 91 4 2 6 3 1
(4.3) (2.9) (0.7) (0.7) (65.0) (2.9) (1.4) (4.3) (2.1) (0.7)
0 1 0 0 3 8 0 5 0 31
0 (1.7) 0 0 (5.2) (13.8) 0 (8.6) 0 (53.4)
21
(15.0)
10
(17.2)
88 52
(62.9) (37.1)
27 31
(46.6) (53.4)
99 41 9
(70.7) (29.3) [1-50]
53 5 7
(91.4) (8.6) [1-16]
45 6 36
(32.1) (4.3) (25.7)
3 0 2
(5.2) 0 (3.4)
4 0
(2.9) 0
1 0
(1.7) 0
(10.1) (21.9) (32.3) (35.7)
265 44 76 110 35
(16.6) (28.7) (41.5) (13.2)
1124 114 246 363 401
N.: number; %: percentages; kg: kilogram; FVIII: coagulation factor VIII; IUml-1: international units per milliliter; BU: Bethesda Units; VWF: von Willebrand factor; mmoll-1: millimolar per liter; gL-1: gram per liter; SS: steady state; *blood group known for 101 patients.
haematologica | 2016; 101(10)
1161
H. Hazendonk et al.
Supplementary Table S1. More specifically, first-order conditional estimation (FOCE) method with interaction was applied, allowing interaction between structural and residual variance components. The statistical package R v.2.14.2 (The R Foundation for Statistical Computing) and Xpose version 422 were used for data set checkout, exploration and model diagnostics. Pirana software was used as an interface between NONMEM, R and Xpose.23 Model diagnostics included the evaluation of the goodness of fit plots, the objective function value (OFV), the precision of the parameter estimates and the shrinkage of estimated random parameters. The OFV is a measurement of goodness of fit of the model and is proportional to minus two times the logarithm of the likelihood (-2log likelihood) of the data. Competing hierarchical models were compared by calculating the difference between their OFV. This ratio is assumed to be χ2 distributed. Therefore, if models differ by one parameter, a decrease in OFV of 3.84 corresponds to P=0.05 (1 degree of freedom) and OFV decreases of 6.63 and 10.8 correspond to P=0.01 and 0.001, respectively.
Structural model development FVIII plasma concentrations were described by a two-compartment PK model. Estimated (fixed) parameters were clearance (CL), volume of distribution of the central compartment (V1), intercompartment clearance (Q), and volume of distribution of the peripheral compartment (V2). The structural model also accounted for the individual endogenous baseline FVIII plasma concentration. PK parameters were allometrically scaled to account for the wide range of body weights of both adult and pediatric patients. An allometric power model was used with power exponents fixed at 0.75 for clearances and 1.0 for volumes of distribution,24 as described in the following equations: CLi = θCL×(BWi /68)0.75 VLi=θV1×(BWi /68) In this expression, CLi and Vi are the typical clearance and central volume of distribution for an individual i with body weight BWi while θCL and θV1 are the respective parameter values for a subject with a body weight of 68 kilogram.
Figure 1. Perioperative FVIII plasma concentrations and visual predictive check for observed FVIII plasma concentrations. Perioperative FVIII plasma concentrations consist of trough, peak and steady state concentrations for both modes of therapy (continuous infusion and bolus infusion therapy). Visual predictive check for the observed FVIII plasma concentrations, given the final model. Observed FVIII plasma concentrations and mean, 5th percentile and 95th percentile observed and simulated FVIII plasma concentrations.
Table 3. Model-building steps resulting in significant decreases in objective function value (OFV).
Model Structural model* 1 One compartment with IIV on V1 and CL 2 Two compartments with IIV on V1 and CL 3 Inclusion of individual endogenous baseline FVIII plasma concentrations Covariates on CL (added to model 3) 4 Age 5 Age, blood group 6 Age, blood group, bleeding complication 7 Age, blood group, bleeding complication, severity of surgical procedure Covariates on V1 (added to model 7) 8 Age Error model (added to model 8) 9 Center (two categories)
NOP
OFV
7 9 9
-2604.5 -2799.3 -2816.1
10 11 12 13
-2851.8 -2862.3 -2886.7 -2895.2
14
-2911.8
16
-2930.6
*Allometric scaling based on body weight was applied with an allometric exponent of 0.75 for the clearance parameters and 1 for the volume terms; under prediction of FVIII plasma concentrations of a B-domain deleted product was implemented. NOP: number of estimated parameters; OFV: objective function value; IIV: inter-individual variability; V1: volume of the central compartment; CL: clearance.
1162
haematologica | 2016; 101(10)
A perioperative population pharmacokinetic model
The random parameters inter-individual variability (IIV) and inter-occasion variability (IOV) of the PK parameters were estimated using an exponential function according to: đ??śđ??żđ?&#x2018;&#x2013; = đ?&#x153;&#x192;đ??śđ??ż Ă&#x2014;đ?&#x2018;&#x2019;(đ?&#x153;&#x201A;đ?&#x2018;&#x2013;+đ?&#x153;&#x2026;đ?&#x2018;&#x2013;) where Ρi and ki represent the IIV and IOV, respectively, and are assumed to be symmetrically distributed with a mean of 0 and an estimated variance of Ď&#x2030;2 and p2. IIV and IOV were included in the model if shrinkage was less than 20%.25 The structural model also accounted for under prediction of plasma concentrations of a Bdomain deleted product (RefactoÂŽ) due to known discrepancies and influence of one-stage laboratory assays on plasma concentrations,26,27 as described:
Cpred,bdp = Cpred x (1 â&#x20AC;&#x201C; θbdp) Where Cpred,bdp and Cpred are the predicted concentrations of the B-domain deleted product (bdp) and other products, respectively, and θ is the fractional decrease in concentration. Residual variability in FVIII concentration was described using a combined error model.
Covariate search After obtaining the structural model individual empirical Bayesian estimates were obtained for all PK parameters. Correlations between these parameters and patientsâ&#x20AC;&#x2122; and surgical
Figure 2. Visualization of NONMEM analysis and outcomes. Allometric scaling based on body weight was applied with an allometric exponent of 0.75 for the clearance parameters and 1 for the volume terms; age in years; IIV: inter-individual variability.
Table 4. Parameter estimates for the final model and bootstrap analysis.
Parameter Structural model 1 - Clearance (CL; mL/h/68 kg) 2 - Volume of central compartment (V1; mL/68 kg) 3 - Inter-compartmental clearance (Q; mL/h/68 kg) 4 - Volume of peripheral compartment (V2; mL/68 kg) B-domain deleted product Covariate parameters 5 - CL â&#x20AC;&#x201C; Age (change with increasing age) 6 - CL â&#x20AC;&#x201C; Blood group O (% difference) 7 - CL â&#x20AC;&#x201C; Major surgical procedure (% difference) 8 - V1 â&#x20AC;&#x201C; Age (change with increasing age) Inter-individual variability Clearance (% CV) Volume of central compartment (% CV) Residual variability Additive residual error (SD; IUml-1) Center 1,2,3 Center 4,5 Proportional residual error (% CV) Center 1,2,3 Center 4,5
Structural model Mean (%RSE)
Final model Mean (%RSE)
190 3030 170 1930 0.32
150 2810 160 1900 0.34
(8) (4) (20) (11) (13)
160 2810 170 1890 0.33
(5) (3) (15) (8) (10)
-0.17 26 -7 -0.09
(22) (7) (6) (28)
-0.16 27 -7 -0.09
(13) (22) (34) (18)
37 27
(14) (14)
36 26
(10) (11)
0.15 0.05
(12) (28)
0.14 0.05
(9) (20)
0.18 0.23
(15) (9)
0.18 0.23
(9) (7)
45 29
(5) (3) (17) (12) (11)
(13) (13)
Bootstrap analysis final model Mean (%RSE)
RSE: relative standard error; CL: clearance V1: volume of central compartment; Q: inter-compartmental clearance; V2: volume of peripheral compartment; CV: coefficient of variation; SD: standard deviation.
haematologica | 2016; 101(10)
1163
H. Hazendonk et al.
characteristics, and the use of concomitant medication were explored graphically. All covariates were tested in a univariate analysis. The most clinically relevant and statistically significant covariate was retained in the model: a stepwise forward approach was used to determine clinical and statistically significant covariates with P<0.05. Backward elimination was performed to confirm that all included covariates in the final model were statistically significant with P<0.01. As the occurrence of a bleeding complication could not be related to actual FVIII plasma concentrations,9 occurrence of a bleeding complication was not included in the final model. Moreover, only a limited difference in clearance was observed between patients with and without a bleeding complication (7%). Also, time dependent changes in clearance were tested during the perioperative period.
Final model and model evaluation The stability and performance of the final model was checked using an internal validation procedure via the bootstrap resampling technique in which 1000 bootstrap datasets were generated by random sampling with replacement.28 Visual predictive check plots obtained after Monte Carlo simulations of the study population were used to evaluate if the final model adequately described observed data.29
Results Patients and treatment in the perioperative setting Our cohort consisted of 119 hemophilia A patients
Table 5. Model equations describing the perioperative population PK model. CL CL (mL/h) = 150 x ((body weight / 68)0.75) x ((age / 40)-0.17) x (1.26blood group) x (0.93severity of surgical procedure) V1 V1 (mL) = 2810 x (body weight / 68) x ((age / 40)-0.09) Q Q (ml/h) = 160 x ((body weight / 68)0.75) V2 V2 (mL) = 1900 x (body weight / 68) Body weight (kilograms); Age (years); blood group equals one in case of blood group O and zero in case of blood group non O; severity of surgical procedure equals one in case of a major surgical procedure and zero in case of a minor surgical procedure.
A
Blood group O
Blood group non O
B
major
major
minor
Severity of surgical procedure 1164
minor
Figure 3. Graphical visualization of variability of the clearance and covariates. Visualization of variability of the clearance. (A) As a function of blood group O versus blood group non O. (B) As a function of severity of surgical procedure.
haematologica | 2016; 101(10)
A perioperative population pharmacokinetic model
undergoing a total of 198 surgical procedures, as described previously.9 Patients were treated for up to two weeks after surgery according to the National Hemophilia Consensus (Table 1).3 Treatment consisted of a pre-operative bolus infusion of approximately 50 IU kg-1 followed by a treatment scheme with either bolus infusions or continuous infusion therapy based on a clearance rate of 3-4 mL kg-1 hour-1. General characteristics of these included patients are shown in Table 2. Seventy-five patients underwent only one surgical procedure. Half of all patients had blood group O (51%). In 3% of all surgical procedures a severe bleeding complication occurred, defined as necessity of a red blood cell transfusion (RBCT) and/or necessity of a second surgical intervention, which could not be related to FVIII plasma concentrations. In total, 1389 FVIII measurements were obtained, equally distributed on consecutive days in the perioperative setting (Figure 1). Approximately 7 samples per patient were taken in the perioperative period. In summary, 45% of FVIII plasma concentrations were below the target range in the first 24 hours and 75% were above the target range after six days of hospitalization (Table 1).
Pharmacokinetic modeling Structural model development: time profiles of FVIII plasma concentrations were best described by a two-compartment model with allometric scaling for body weight (Figure 2). By allometric scaling, all estimated PK parameters were normalized for a body weight of 68 kg. Model building steps that resulted in significant decrease of the OFV, and consequently a better fit of the model, are shown in Table 3. In the structural model, typical values for CL and V1 were 190 mL/hour/68 kg and 3030 mL/68 kg (Table 4). It was possible to estimate IIV for CL and V1, whereas estimates for IIV of Q and V2 were imprecise and accompanied by a large shrinkage of more than 40%.25 Although this may suggest that there was no inter-patient variability in Q and V2, this is due to the fact that the available data were not sufficiently informative. The IIV for CL and V1 were respectively 45% and 29%, underlining the importance of tailoring therapy to the individual. Estimation of IOV on CL and V1 resulted in high shrinkage values for both parameters (34% and 46%, respectively); consequently IOV was not included in the model. Inclusion of individual endogenous baseline FVIII plasma concentrations and inclusion of a structural underprediction of plasma concentrations using a B-domain deleted product improved the model. A proportional underprediction of 0.34 (34%) in FVIII plasma concentration was estimated for this product. The residual error was described using a combined error model. Covariate search: in the univariate analysis, significant covariates of clearance were age (P<0.001), blood group (P<0.01), severity of surgical procedure (P<0.01), lean body mass (P<0.01), use of tranexamic acid and heparin (P<0.05), historically measured VWF antigen and activity levels (P<0.05). Treatment center and type of product were not significant covariates. After the step forward analysis, only age, blood group, and severity of surgical procedure were significantly associated with clearance. After the inclusion of age in the model, VWF antigen and activity levels were no longer statistically significant. Age was also associated with V1 (Table 3). Different models were used to test possible time dependent changes in clearance during the perioperative period; however, no haematologica | 2016; 101(10)
Figure 4. Clearance of FVIII in major and minor surgical procedures after stratification for age. Post hoc estimates of FVIII clearance, normalized for total body weight, and stratified for age (<4 years or >4 years) were categorized according to severity of surgical procedure. *A Spearmanâ&#x20AC;&#x2122;s correlation test was performed to test for clearance differences between major and minor surgical procedures. The median age of children included in the study was used as cut-off value for analysis. This was supported by results of Figure 5A.
differences were observed. Differences in residual error were detected for the different centers. In the final model, IIV of CL decreased from 45% towards 37% after inclusion of these covariates. IIV of V1 decreased from 29% to 27%. The PK parameter estimates of the final model are presented in Table 4. Typical PK parameter estimates were described with the equations presented in Table 5. According to the equation, clearance was 214, 169, 150 and 142 mL/h/68 kg for a typical patient (with blood group non-O undergoing a minor surgical procedure) with an age of 5, 20, 40 and 55 years, respectively. In case of a major surgical procedure, a small decrease in CL was observed of 7% (Table 4). Interestingly, individual post hoc clearances were higher in patients with a major surgical procedure (Figure 3B). This was, however, explained by collinearity between covariates; older patients underwent more major surgical procedures (Figure 4). Clearance increased by 26% in patients with blood group O. CL and elimination half-life are depicted as functions of age and body weight in Figure 5. The adequacy of the derived final model is shown in Figure 6. Population and individually predicted concentrations for all patients were plotted against the measured concentrations in Figure 6. A good agreement was observed between FVIII concentrations predicted by the model and those assessed by laboratory measurements. Overall, standardized weighted residuals revealed a random distribution around zero, within a range of -2 to +2, indicative of an unbiased estimation (Figure 6C). Model evaluation: a good agreement was found between parameter estimates of the final model and parameter estimates of the bootstrap analysis (Table 4). A visual predic1165
H. Hazendonk et al.
A
B
C
D
Figure 5. Clearance and elimination half-life as functions of age and body weight. (A) Clearance of FVIII, normalized for total body weight, as a function of age. (B) Clearance of FVIII as a function of body weight. (C) The elimination half-life of FVIII as a function of age. (D) The elimination half-life of FVIII as a function of body weight. Eta shrinkage was 10% and 20%, respectively, for the estimates of inter-individual variability of clearance and volume of the central compartment.
tive check was conducted by 1000 simulations based on the final model (Figure 1). It confirmed adequateness of the model, as 7% of the measured concentrations were calculated above the 95th percentile of the simulated concentrations and 9% of the measured concentrations were found to be below the 5th percentile of the simulated concentrations.
Discussion In this study, a population PK model was constructed describing the perioperative PK of several FVIII concentrates in current use. The majority of these factor VIII concentrates were FVIII recombinant products (77% of surgical procedures), of which 14% were a B-domain deleted FVIII concentrate, as well as plasma-derived FVIII concentrates (23% of surgical procedures). In the population PK model, a difference in results due to the B-domain deleted FVIII concentrate (Refacto AFÂŽ) was accounted for. No other differences were observed between products. As this difference is incorporated into the population PK model, this perioperative FVIII population PK model can be used for all described FVIII concentrates. The developed model will facilitate Bayesian adaptive dosing, allowing individualization of FVIII dosing during the entire perioperative period. So far, only a few studies have reported application of PK-guided dosing during the perioperative period. Unfortunately, in all these studies, only the FVIII loading dose was based on an individual PK-profile obtained several days before surgery.30-35 Iterative perioperative FVIII dosing-adjustments after first loading dose 1166
could not be performed as there was no population PK model. The perioperative population PK model presented here will now make Bayesian adaptive dosing in this setting possible. Moreover, it will consider all important patientsâ&#x20AC;&#x2122; characteristics associated with clearance in the surgical setting. The model presented here consists of a two-compartment model with allometric scaling of the PK parameters according to body weight. Both increasing age and increased severity of surgical procedure were overall significantly associated with a lower FVIII clearance, although individual clearance rates showed that patients with a major surgical procedure did demonstrate higher clearance rates. This contradiction may be due to the fact that included covariates in the PK model were confounders, e.g. older patients with a decreased CL of FVIII concentrate underwent major surgical procedures more often than younger patients. Also, increased consumption of concentrates due to blood loss and activation of coagulation are other possible modifying factors. In addition, blood group O was associated with higher FVIII clearance, which will be discussed in the following sections. Although it should be underlined that this population PK model represents an important development, it is important to realize that it does not account for pharmacodynamic outcome measures, as the occurrence of a bleeding complication could not be related to actual FVIII plasma concentrations due to scarcity of FVIII plasma concentrations during an acute bleeding event. As in most resource rich countries, current perioperative replacement therapy in hemophilia A in the Netherlands consists of a FVIII loading dose followed by either continhaematologica | 2016; 101(10)
A perioperative population pharmacokinetic model
uous FVIII infusion or treatment with FVIII bolus infusions while targeting predefined peak and trough FVIII plasma concentrations, as stated in the National Hemophilia Consensus.3 The retrospective study performed to collect data for this PK model has been described earlier.9 Results show the challenges of current perioperative dosing of FVIII replacement therapy in daily clinical practice when targeting prescribed FVIII plasma concentrations, as significant underdosing and overdosing were demonstrated. Moreover, it underlines the necessity of alternative more individualized dosing strategies in the perioperative setting; this is possible when PK-guided dosing based on a population PK model is applied. PK-guided dosing based on population PK models has mainly been studied in the long-term prophylactic setting. However, in order to apply Bayesian adaptive dosing, it is necessary to utilize a population PK model appropriate for the individual patient and the specific setting concerned. In analyses preceding the construction of this perioperative population PK model, it was confirmed that the mean estimated PK parameters for prophylactic dosing, as reported by Bjรถrkman et al.,12 did not reliably predict observed perioperative FVIII plasma concentrations. Using the prophylactic model, calculations showed an underprediction of perioperative FVIII concentrations of less than 1.00 IUml-1 as well as an overprediction of FVIII concentrations of more than 1.00 IUml-1. In other words, actual FVIII plasma concentrations were respectively higher and lower than those predicted by a prophylactic population PK model (data not shown). Therefore, it was concluded that prophylactic population PK models can not be applied in the perioperative setting. Use of the prophylactic model in this setting would generate a bias of predicted perioperative FVIII plasma concentrations. In the prophylactic setting, a similarly constructed population PK model has already been applied.12 CL, V1 and Q were actually in accordance when a comparison was made between perioperative and prophylactic PK population model (CL: 150 vs. 222 mL/h/68 kg; V1: 2810 vs. 3520 mL/68 kg; and Q: 160 vs. 256 mL/h/68 kg, respectively). However, in the present perioperative model, a value of 1880 mL/68 kg was found for V2 in contrast to a value of 240 mL/68 kg found in the prophylactic situation, suggesting a rapid redistribution of FVIII concentrate following intravenous administration.12 Due to increased V2, calculated distribution half-life and elimination half-life are significantly larger (as half-life is a derivative of the distribution volume) in the perioperative setting in comparison with the prophylactic state (4 hours and 25 hours vs. 0.6 hours and 12 hours, respectively). These calculated halflifes are in accordance with previously described half-life observed immediately after surgery and half-life observed at steady state of 10 surgical patients described with a surgical model (9.6 and 17.8 hours, respectively) in comparison to 10 surgical patients described with an estimated half-life of 10.1 hours described with a non-surgical model.15 Unfortunately, the rapid redistribution was not quantifiable, due to minimal data of laboratory assessment after infusion. Previously, it has been suggested that V2 may reflect the FVIII distribution into extravascular spaces or within an intravascular compartment, more specifically as a reflection of adhesion to the vessel wall, or that it may reflect the process of a rapid initial elimination.36,37 We hypothesized that an extra intravascular component resulting in a large V2 may be the result of haematologica | 2016; 101(10)
A
B
C
Figure 6. Observed and model-predicted FVIII plasma concentrations. NONMEM model diagnostic plots, observed and model predicted FVIII plasma concentrations plotted against each other. (A) Population predicted FVIII plasma concentrations. (B) Individually predicted FVIII plasma concentrations. (C) Conditionally weighted residuals versus time.
the high affinity and stoichiometry of FVIII to VWF,38 combined with the significant increase of VWF after surgery due to inflicted endothelial damage and its role in the acute phase reaction.39 In addition, Deitcher et al. have shown that volume of distribution increases after desmopressin administration, which, of course, results in an overall increase in VWF levels.40 Moreover, we believe that VWF may play a crucial role in the perioperative setting with regard to FVIII PK parameters, as previous studies have demonstrated a clear association between VWF plasma concentrations and FVIII halflife.41,42 This is not surprising, as VWF protects FVIII against proteolytic degradation by expression of ABH antigens on N-linked glycans and the uptake of the copper-binding protein ceruloplasmin.43,44 In addition, it has been shown that, in healthy individuals undergoing orthopedic surgery, VWF 1167
H. Hazendonk et al.
decreases significantly intraoperatively and rises immediately after surgery.39 Therefore, we suspected a time-dependent FVIII clearance in the presented PK model, with an increased clearance during the surgical procedure itself and a decrease in clearance directly after surgery. However, no time-dependent clearance could be established. Unfortunately, it was not possible to investigate the role of VWF plasma concentrations in our analyses in more detail, as VWF measurements are currently not routine practice in the perioperative setting and only historically measured VWF plasma concentrations were available in half of the study population. However, a 26% higher clearance rate was observed in blood group O patients in the perioperative setting, underlining the potential importance of measurement of VWF plasma concentrations in the perioperative setting if PK-guided dosing is implemented. This is supported by earlier reports that blood group O patients have around 25% lower VWF levels in comparison to patients with blood group non-O.43 Strikingly, this effect of blood group on clearance was not significant in the prophylactic population PK model as shown by Björkman et al.12 However, we are not informed if VWF levels were available for those analyses. In contrast, higher VWF levels may also help explain the unexpected overall lower clearance found in patients undergoing major surgical procedures. The ongoing prospective randomized controlled “OPTI-CLOT” trial (RCT) (described in more detail elsewhere)45 will provide, among other things, an insight into the pathophysiology of VWF in hemophilia patients during the perioperative setting, and the relationship between VWF levels and estimates of FVIII PK parameters. These data will further validate the perioperative PK population model presented here, refining its applicability, and further defining the influence of possible modifying factors of PK parameters. Moreover, extending this population PK model, in combination with extended half-life (EHL) products in the near future could be of great value. However, first of all, studies are needed to document in detail associations between clearance of current FVIII products and EHL products within individuals. Clinically, in the perioperative setting, adaptive Bayesian dosing can be used to optimize and individualize dosing in order to obtain desired target FVIII plasma concentrations with increased certainty. Bayesian analysis combines individual PK data with information from an available population PK model. Such a population PK
References 1. Fijnvandraat K, Cnossen MH, Leebeek FW, Peters M. Diagnosis and management of haemophilia. BMJ. 2012;344:e2707. 2. Collins PW, Blanchette VS, Fischer K, et al. Break-through bleeding in relation to predicted factor VIII levels in patients receiving prophylactic treatment for severe hemophilia A. J Thromb Haemost. 2009;7(3): 413-420. 3. Leebeek FWG, Mauser-Bunschoten EP, Editors. [Richtlijn Diagnostiek en behandeling van hemofilie en aanverwante hemostasestoornissen]. Van Zuiden Communications BV. 2009;1-197.
1168
model is constructed from PK data of many individuals, and not only embodies defined patients’ characteristics known to influence clearance and other PK parameters, but also as yet unidentified patients’ characteristics which cannot be quantified. Individual patient information that is entered into the model must include dose and time point of factor concentrate administration, as well as the FVIII plasma concentrations achieved. Incorporation of the patient’s body weight, blood group, age and severity of surgical procedure will improve estimation of the individual clearance of factor concentrate. In clinical practice, individual clearance and other PK parameter estimates can be made by a clinical pharmacologist experienced in this methodology and iteratively updated, leading to calcu-lated dose adjustments. We are currently planning to develop a PK tool to implement this perioperative population PK model in daily clinical practice. The first dose of FVIII concentrate, still in steady state, will be based on individual PK parameters deducted from an individual PK profile constructed according to the prophylactic population PK model. As we were not able to demonstrate timedependent changes in PK parameters during the perioperative setting, the perioperative population PK model described here can be applied to the complete perioperative period with varying target FVIII plasma concentrations, as described by National Guidelines. In conclusion, we have constructed a perioperative population PK model facilitating iterative dose-adjustments by Bayesian analysis. We believe this model will prove its value as it will lead to optimization of current dosing strategies by reducing underdosing and overdosing, and, therefore, both a decrease of bleeding risk and an expected overall reduction of factor concentrate consumption with a subsequent reduction in costs. Acknowledgments This study is part of the “OPTI-CLOT” research program (Patient tailOred PharmacokineTIc-guided dosing of CLOTting factor concentrate in bleeding disorders)”, an (inter)national multicenter study aiming to implement PK-guided dosing of clotting factor replacement therapy by initiating studies to prove the implications of PK-guided dosing, to construct perioperative and prophylactic PK population models and to evaluate the cost-effectiveness of a PK-guided approach. A complete list of the members of the OPTI-CLOT research program appears in the Online Supplementary Appendix.
4. Johnson KA, Zhou ZY. Costs of care in hemophilia and possible implications of health care reform. Hematology Am Soc Hematol Educ Program. 2011;2011:413418. 5. Schramm W, Berger K. Economics of prophylactic treatment. Haemophilia. 2003;9 Suppl 1:111-115; dicussion 116. 6. Feldman BM, Aledort L, Bullinger M, et al. The economics of haemophilia prophylaxis: governmental and insurer perspectives. Proceedings of the Second International Prophylaxis Study Group (IPSG) symposium. Haemophilia. 2007;13(6):745-749. 7. Nederlandse Zorgautoriteit. [Onderzoek naar de toegankelijkheid en betaalbaarheid van geneesmiddelen in de medisch special-
istische zorg]. Available from: https://www.nza.nl/publicaties/1048188/ Onderzoeksrapport__Toegankelijkheid_en _betaalbaarheid_van_geneesmiddelen_in_ de_medisch_specialistis, 29-06-2015:1-110. 8. Traore AN, Chan AK, Webert KE, et al. First analysis of 10-year trends in national factor concentrates usage in haemophilia: data from CHARMS, the Canadian Hemophilia Assessment and Resource Management System. Haemophilia. 2014; 20(4):e251-259. 9. Hazendonk HC, Lock J, Mathot RA, et al. Perioperative treatment of hemophilia A patients: blood group O patients are at risk of bleeding complications. J Thromb Haemost. 2016;14(3):468-478.
haematologica | 2016; 101(10)
A perioperative population pharmacokinetic model
10. Hazendonk HCAM, Lock J, Fijnvandraat K, et al. A retrospective observational multicenter study on peri-operative Factor IX consumption in Hemophilia B ("OPTICLOT" studies). J Thromb Haemost. 2013;11:Suppl 2. 11. Carlsson M, Berntorp E, Bjรถrkman S, Lethagen S, Ljung R. Improved cost-effectiveness by pharmacokinetic dosing of factor VIII in prophylactic treatment of haemophilia A. Haemophilia. 1997;3(2):96101. 12. Bjorkman S, Folkesson A, Jonsson S. Pharmacokinetics and dose requirements of factor VIII over the age range 3-74 years: a population analysis based on 50 patients with long-term prophylactic treatment for haemophilia A. Eur J Clin Pharmacol. 2009;65(10):989-998. 13. Bjorkman S, Oh M, Spotts G, et al. Population pharmacokinetics of recombinant factor VIII: the relationships of pharmacokinetics to age and body weight. Blood. 2012;119(2):612-618. 14. Collins PW, Fischer K, Morfini M, Blanchette VS, Bjorkman S; International Prophylaxis Study Group Pharmacokinetics Expert Working G. Implications of coagulation factor VIII and IX pharmacokinetics in the prophylactic treatment of haemophilia. Haemophilia. 2011;17(1):2-10. 15. Longo G, Messori A, Morfini M, et al. Evaluation of factor VIII pharmacokinetics in hemophilia-A subjects undergoing surgery and description of a nomogram for dosing calculations. Am J Hematol. 1989;30(3):140-149. 16. Boer P. Estimated lean body mass as an index for normalization of body fluid volumes in humans. Am J Physiol. 1984;247(4 Pt2):F632-636. 17. Peters AM, Snelling HL, Glass DM, Bird NJ. Estimation of lean body mass in children. Br J Anaesth. 2011;106(5):719-723. 18. Garrow JS, Webster J. Quetelet's index (W/H2) as a measure of fatness. Int J Obes. 1985;9(2):147-153. 19. Koshy M, Weiner SJ, Miller ST, et al. Surgery and anesthesia in sickle cell disease. Cooperative Study of Sickle Cell Diseases. Blood. 1995;86(10):3676-3684. 20. U.S. Department of Health and Human Services FaDA. Guidance for Industry. Population Pharmacokinetics. 21. Boeckmann AJ, Sheiner LB, Beal SL. NONMEM Users Guide. NONMEM Project Group. 2011;University of California at San Francisco (ICON Development Solutions Ellicott City, Maryland ):1-165.
haematologica | 2016; 101(10)
22. Jonsson EN, Karlsson MO. Xpose--an S-PLUS based population pharmacokinetic/pharmacodynamic model building aid for NONMEM. Comput Methods Programs Biomed. 1999;58(1):51-64. 23. Keizer RJ, van Benten M, Beijnen JH, Schellens JH, Huitema AD. Pirana and PCluster: a modeling environment and cluster infrastructure for NONMEM. Comput Methods Programs Biomed. 2011;101(1): 72-79. 24. Anderson BJ, Holford NH. Mechanismbased concepts of size and maturity in pharmacokinetics. Annu Rev Pharmacol Toxicol. 2008;48:303-332. 25. Savic RM, Karlsson MO. Importance of shrinkage in empirical bayes estimates for diagnostics: problems and solutions. AAPS J. 2009;11(3):558-569. 26. Hubbard AR, Weller LJ, Bevan SA. A survey of one-stage and chromogenic potencies in therapeutic factor VIII concentrates. Br J Haematol. 2002;117(1):247-248. 27. Hubbard AR, Sands D, Sandberg E, Seitz R, Barrowcliffe TW. A multi-centre collaborative study on the potency estimation of ReFacto. Thromb Haemost. 2003;90(6):1088-1093. 28. Ette EI. Stability and performance of a population pharmacokinetic model. J Clin Pharmacol. 1997;37(6):486-495. 29. Bergstrand M, Hooker AC, Wallin JE, Karlsson MO. Prediction-corrected visual predictive checks for diagnosing nonlinear mixed-effects models. AAPS J. 2011;13(2):143-151. 30. Batorova A, Martinowitz U. Continuous infusion of coagulation factors. Haemophilia. 2002;8(3):170-177. 31. Bidlingmaier C, Deml MM, Kurnik K. Continuous infusion of factor concentrates in children with haemophilia A in comparison with bolus injections. Haemophilia. 2006;12(3):212-217. 32. Dingli D, Gastineau DA, Gilchrist GS, Nichols WL, Wilke JL. Continuous factor VIII infusion therapy in patients with haemophilia A undergoing surgical procedures with plasma-derived or recombinant factor VIII concentrates. Haemophilia. 2002;8(5):629-634. 33. Mulcahy R, Walsh M, Scully MF. Retrospective audit of a continuous infusion protocol for haemophilia A at a single haemophilia treatment centre. Haemophilia. 2005;11(3):208-215. 34. Srivastava A. Choice of factor concentrates for haemophilia: a developing world perspective. Haemophilia. 2001;7(1):117-122.
35. Stieltjes N, Altisent C, Auerswald G, et al. Continuous infusion of B-domain deleted recombinant factor VIII (ReFacto) in patients with haemophilia A undergoing surgery: clinical experience. Haemophilia. 2004;10(5):452-458. 36. Bjorkman S, Carlsson M, Berntorp E, Stenberg P. Pharmacokinetics of factor VIII in humans. Obtaining clinically relevant data from comparative studies. Clin Pharmacokinet. 1992;22(5):385-395. 37. Noe DA. A mathematical model of coagulation factor VIII kinetics. Haemostasis. 1996;26(6):289-303. 38. Vlot AJ, Koppelman SJ, van den Berg MH, Bouma BN, Sixma JJ. The affinity and stoichiometry of binding of human factor VIII to von Willebrand factor. Blood. 1995;85(11):3150-3157. 39. Kahlon A, Grabell J, Tuttle A, et al. Quantification of perioperative changes in von Willebrand factor and factor VIII during elective orthopaedic surgery in normal individuals. Haemophilia. 2013;19(5):758764. 40. Deitcher SR, Tuller J, Johnson JA. Intranasal DDAVP induced increases in plasma von Willebrand factor alter the pharmacokinetics of high-purity factor VIII concentrates in severe haemophilia A patients. Haemophilia. 1999;5(2):88-95. 41. Vlot AJ, Mauser-Bunschoten EP, Zarkova AG, et al. The half-life of infused factor VIII is shorter in hemophiliac patients with blood group O than in those with blood group A. Thromb Haemost. 2000;83(1):65-69. 42. Fijnvandraat K, Peters M, ten Cate JW. Inter-individual variation in half-life of infused recombinant factor VIII is related to pre-infusion von Willebrand factor antigen levels. Br J Haematol. 1995;91(2):474476. 43. Klarmann D, Eggert C, Geisen C, et al. Association of ABO(H) and I blood group system development with von Willebrand factor and Factor VIII plasma levels in children and adolescents. Transfusion. 2010;50(7):1571-1580. 44. Lenting PJ, van Mourik JA, Mertens K. The life cycle of coagulation factor VIII in view of its structure and function. Blood. 1998;92(11):3983-3996. 45. Hazendonk HC, van Moort I, Fijnvandraat K, et al. The "OPTI-CLOT" trial. A randomised controlled trial on periOperative PharmacokineTIc-guided dosing of CLOTting factor concentrate in haemophilia A. Thromb Haemost. 2015;114(3):639644.
1169
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Platelet Biology & Its Disorders
Ferrata Storti Foundation
Haematologica 2016 Volume 101(10):1170-1179
Whole exome sequencing identifies genetic variants in inherited thrombocytopenia with secondary qualitative function defects
Ben Johnson,1 Gillian C. Lowe,1 Jane Futterer,1 Marie Lordkipanidzé,1 David MacDonald,1 Michael A. Simpson,2 Isabel Sanchez-Guiú,3 Sian Drake,1 Danai Bem,1 Vincenzo Leo,4 Sarah J. Fletcher,1 Ban Dawood,1 José Rivera,3 David Allsup,5 Tina Biss,6 Paula HB Bolton-Maggs,7 Peter Collins,8 Nicola Curry,9 Charlotte Grimley,10 Beki James,11 Mike Makris,4 Jayashree Motwani,12 Sue Pavord,13 Katherine Talks,6 Jecko Thachil,7 Jonathan Wilde,14 Mike Williams,12 Paul Harrison,15 Paul Gissen,16 Stuart Mundell,17 Andrew Mumford,18 Martina E. Daly,4 Steve P. Watson,1 and Neil V. Morgan1 on behalf of the UK GAPP Study Group
Institute for Cardiovascular Sciences, College of Medical and Dental Sciences, University of Birmingham, UK; 2Division of Genetics and Molecular Medicine, King's College, London, UK; 3Centro Regional de Hemodonación, Universidad de Murcia, IMIBArrixaca, Murcia, Spain; 4Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield Medical School, University of Sheffield, UK; 5Hull Haemophilia Treatment Centre, Hull and East Yorkshire Hospitals NHS Trust, Castle Hill Hospital, Hull, UK; 6Department of Haematology, Royal Victoria Infirmary, Newcastle Upon Tyne, UK; 7Department of Haematology, Manchester Royal Infirmary, Manchester, UK; 8Arthur Bloom Haemophilia Centre, School of Medicine, Cardiff University, UK; 9 Oxford Haemophilia & Thrombosis Centre, Churchill Hospital, Oxford, UK; 10Nottingham Haemophilia Centre, Nottingham University Hospital, UK; 11Regional Centre for Paediatric Haematology, Leeds Children’s Hospital, UK; 12Department of Haematology, Birmingham Children's Hospital, UK; 13Department of Haematology, Oxford University Hospitals NHS Foundation Trust, UK; 14Adult Haemophilia Centre, Queen Elizabeth Hospital, Birmingham, UK; 15School of Immunity and Infection, College of Medical and Dental Sciences, University of Birmingham, UK; 16Medical Research Council, Laboratory for Molecular Cell Biology, University College London, UK; 17School of Physiology, Pharmacology and Neuroscience, University of Bristol, UK; and 18School of Cellular and Molecular Medicine, University of Bristol, UK 1
Correspondence: n.v.morgan@bham.ac.uk
ABSTRACT Received: March 17, 2016. Accepted: June 10, 2016. Pre-published: June 16, 2016. doi:10.3324/haematol.2016.146316
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1170
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1170
I
nherited thrombocytopenias are a heterogeneous group of disorders characterized by abnormally low platelet counts which can be associated with abnormal bleeding. Next-generation sequencing has previously been employed in these disorders for the confirmation of suspected genetic abnormalities, and more recently in the discovery of novel disease-causing genes. However its full potential has not yet been exploited. Over the past 6 years we have sequenced the exomes from 55 patients, including 37 index cases and 18 additional family members, all of whom were recruited to the UK Genotyping and Phenotyping of Platelets study. All patients had inherited or sustained thrombocytopenia of unknown etiology with platelet counts varying from 11x109/L to 186x109/L. Of the 51 patients phenotypically tested, 37 (73%), had an additional secondary qualitative platelet defect. Using whole exome sequencing analysis we have identified “pathogenic” or “likely pathogenic” variants in 46% (17/37) of our index patients with thrombocytopenia. In addition, we report variants of uncertain significance in 12 index cases, including novel candidate genetic variants in previously unreported genes in four index cases. These results demonstrate that whole exome sequencing is an efficient method for elucidating potential pathogenic genetic variants in inherited thrombocytopenia. Whole exome sequencing also has the added benefit of discovering potentially pathogenic genetic variants for further study in novel genes not previously implicated in inherited thrombocytopenia. haematologica | 2016; 101(10)
Whole exome sequencing in inherited thrombocytopenia
Introduction
Platelet counts, morphology and white blood cell counts
Inherited thrombocytopenias (IT) are a heterogeneous group of disorders characterized by platelet counts of less than 150x109/L in whole blood. Platelet counts are considered normal when maintained at levels between 150x109/L and 450x109/L. This is achieved by homeostatic processes controlling platelet production (thrombopoiesis), platelet senescence and platelet consumption/destruction. Pathogenic mutations can result in a disruption of these balanced processes causing IT. However, the clinical manifestations are often dependent on both a decreased platelet count and a qualitative or acquired platelet defect and can vary dramatically from severe and potentially life-threatening bleeding to no symptoms. This variation is noted among individuals shown to have the same underlying genetic causes of disease, suggesting that bleeding risk and phenotype are complex traits.1 The average incidence of IT is estimated to be approximately 270 cases per 1x106 live births.2 To date there are 27 individual IT disorders with known causative mutations registered within the Online Mendelian Inheritance in Man (OMIM) catalog, although 33 disease-causing genes have been described.3 Genetic studies have played a major role in the diagnosis and progressive understanding of IT. The genes implicated in the disease encode proteins that vary widely in function and include transcription factors (ETV6, FLI1, GATA1, GFI1B and RUNX1) and proteins involved in cytoskeleton rearrangement and organization (ACTN1, FLNA, GP1BA, GP1BB, GP9, TUBB1 and WAS). However, some protein functions currently remain unknown (SLFN14 and GNE).4-9. Although our knowledge of the causes of IT continues to grow, presently a genetic diagnosis is only reported in approximately 50% of individuals.10-12 So far, genetic investigation into IT has focused on candidate gene sequencing and individual cases of whole exome sequencing (WES) when a causative gene is not obvious.9 With 50% of patients currently undiagnosed, a change in the way we approach genetic diagnosis is necessary. Here we present the first, large-scale, WES-only approach to patients with suspected IT. We demonstrate its application in determining possible genetic origins of IT including identification of variants in novel candidate causative genes. We combine this with an approach implemented by the Genotyping and Phenotyping of Platelets (GAPP) study, which combines WES analysis with extensive platelet phenotyping to create a complete method of diagnosis and gene discovery in this subset of patients.
Results from patients’ samples were compared to ranges for healthy volunteers for the specific method of morphology used. Platelet counts for light transmission aggregometry and flow cytometry analysis as well as mean platelet volume in platelet-rich plasma were originally measured using the Beckman Coulter counter (n=44). Subsequently, platelet counts, morphology and white blood cell counts in whole blood were determined using the Sysmex XN-1000 (n=11). The PLT-F channel was used to measure platelet counts in whole blood and the immature platelet fraction. Mean platelet volume was determined from the impedance PLT-I channel. White blood cell counts were obtained using the Sysmex XN-DIFF channel. All samples were tested against a normal range which was established by measuring the counts for 40 healthy individuals using the Sysmex XN-1000.
Methods
Platelet preparation and platelet function testing Platelet function was assessed by light transmission aggregometry, including lumiaggregometry, for samples with platelet counts in platelet-rich plasma of >1x108/mL (n=13). An in-house flowcytometry assay was developed to assess platelet function in patients with platelet counts in platelet-rich plasma <1x108/mL (n=22). Platelets from individuals with borderline platelet counts in platelet-rich plasma, between 1.0 and 1.5x108/mL, were assessed using both assays (n=16). Aggregometry was performed as previously described.13,14 For flow cytometry, resting surface levels of CD42b, CD41 and GPVI were assessed. The platelet-rich plasma was then stimulated with ADP (3 and 30 μM), CRP (0.3 and 3 μg/mL) and PAR-1 peptide (10 and 100 μM). Membrane expression of P-selectin (FITC-conjugated mouse anti-human CD62P antibody, BD Pharmingen), a marker of platelet alpha granule release, as well as fluorescent fibrinogen binding (a marker of integrin activation) was assessed by flow cytometry on an Accuri C6 flow cytometer. Incubation took place at 37ºC for 2 min and was terminated by adding a 5-fold excess of ice-cold phosphate-buffered saline.
Whole exome sequencing WES and bioinformatics analysis were performed as described previously8,15,16 (Figure 1). The pathogenicity of variants was determined and called using the consensus guidelines as set out by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG guidelines).17 Segregation was determined by Sanger sequencing of candidate variants in both affected and unaffected family members, when available, and the classification was adapted appropriately for the specific study and small sample size.
Sanger sequencing To verify candidate mutations and examine their segregation among family members Sanger sequencing was performed using standard methods on an ABI 3730 automated sequencer, as described previously.8
Study approval The UK-GAPP study was approved by the National Research Ethics Service Committee of West Midlands–Edgbaston (REC reference: 06/MRE07/36) and participants gave written informed consent in accordance with the Declaration of Helsinki. This study was registered at www.isrctn.org as #ISRCTN 77951167. The GAPP study is included in the National Institute of Health Research Non-Malignant Haematology study portfolio (ID9858). haematologica | 2016; 101(10)
Results Recruitment of patients To date, 55 patients with a suspected IT or sustained reduced platelet counts have been enrolled from 25 UK Haemophilia Care Centres and investigated as part of the GAPP study. Before enrollment in the study, all patients underwent clinical and genetic work-up to exclude known 1171
B. Johnson et al.
platelet disorders (including Bernard-Soulier syndrome and MYH9-related disorders, analyzed initially by blood film), idiopathic thrombocytopenic purpura and other non-platelet disorders including von Willebrand disease and inherited coagulation factor deficiencies. The patientsâ&#x20AC;&#x2122; bleeding phenotypes are displayed in Table 1. WES was performed on genomic DNA from all patients, including 37 index cases, all of whom met the studyâ&#x20AC;&#x2122;s entry criteria. All patients, excluding F35.I and F35.II, were of white British or mixed British ethnicity. All results following platelet function testing and WES were reported back to the referring hematology consultants to aid in genetic counselling and disease management.
Platelet counts, morphology and function testing Patients were recruited with a platelet count in whole blood, at the time of enrollment, of less than 150x109/L. Patients with platelet counts in the range of 150x109/L to 200x109/L remained enrolled in the study if they showed a similar phenotype to related affected family members and a platelet count below 150x109/L had been observed prior to enrollment (patients F4.II, F11.III, F13.I and F30.II). Platelet counts, mean platelet volume and immature platelet fraction are displayed in Table 1. Of the 55 recruited patients, 12 were deemed to have a macrothrombocy-
topenia and three a microthrombocytopenia (Table 1). White cell counts were within the normal range (3.78x109/L - 10.11x109/L, n=40) in all patients analyzed (n=11). Platelet function studies revealed the presence of a secondary qualitative defect in addition to the low platelet count in 37/51 (73%) of the 55 patients whose DNA underwent WES and who were also available for platelet function testing (Table 1). Of the 37 patients with a secondary qualitative defect, 89% (33/37) displayed defects in both alpha and dense granule secretion. Five of these patients with an observed granule secretion defect were also suspected to have an additional Gi defect because of reduced responses to all concentrations of ADP. The remaining four patients without an observable granule secretion defect showed abnormalities in alternative pathways (integrin activation, cyclooxygenase pathway and GPVI surface levels) in addition to low platelet counts (Table 1).
Whole exome sequencing WES was performed on genomic DNA from all 55 patients, including 37 index cases, following platelet function testing. An average fold-coverage of 111 was observed across all DNA samples analyzed by WES with
Figure 1. Bioinformatics pipeline analysis of whole exome sequencing data. Initial WES analysis focused on comparison with a panel of 358 genes (Online Supplementary Table S1), after which screening of exome variants focused on novel variants. Variants were classified using the ACMG consensus guidelines.
1172
haematologica | 2016; 101(10)
Whole exome sequencing in inherited thrombocytopenia
an average of 91% of target sequences having >20x coverage. Areas of poor coverage were analyzed manually when occurring in previously IT-associated genes. WES revealed between 24,000 and 25,000 variants (single nucleotide variants, small scale insertions/deletions, and splice site variations) in the DNA from each patient, with an average of 197 novel variants per exome. On average, per individual, 2401 variants with a mean allele frequency of <0.01 were observed, excluding synonymous variants. By evaluating the specificity of the pipeline in calling small variations, it was found that the sensitivity was over 99% and the false discovery rate was approximately 3%. The percentage of the genes with ≤20x coverage for a panel of 358 platelet-related genes is included within Online Supplementary Table S1. Copy number variations were detected using ExomeDepth.18 The analysis revealed an average of 137 copy number variations per exome (n=32). No copy number variations were deemed potential candidates either because of a high allele frequency or a lack of expression or functional role of the gene within the megakaryocyte/platelet lineage.
Variants in known thrombocytopenia-causing genes WES and downstream analysis identified variants within 33 known IT-causing genes in 25 index cases (68%). All variants exceeded 30x sequence coverage at the point of variation and have been confirmed by Sanger sequencing. Variants were selected from positive hits to genes within the panel of 358 IT-associated genes (Online Supplementary Table S1). On average, 37 variants per individual (range, 11-52) were noted in genes from the panel of 358 IT-associated genes, of which on average four (range, 0-7) variants were significant per exome analyzed. In total 28 variants were noted in 14 genes previously known to cause IT (Table 2). Twenty-one index cases possessed a single variant in a gene previously known to cause IT. Four index cases possessed two variants in genes previously known to cause IT. One variant, RUNX1; c.270+1G>T, was noted in two index cases (F13.I and F14.I). Candidate variations were present within ACTN1, the 5’-UTR of ANKRD26, CYCS, FLI1, GFI1B, ITGB3, GP1BA (heterozygous), MYH9, NBEAL2, RUNX1, SLFN14, STIM1, TPM4 and TUBB1. All but six variants were novel and not present within the variant databases previously mentioned. Three variants, ANKRD26; c.126T>G in F2.I, MYH9; c.3493C>T (rs80338829) in F9.I and RUNX1; c.530G>A in F19.I and F19.II have been previously associated with IT.1,19,20 The remaining three variants that have been previously observed occurred at frequencies of <0.005 (0.05%) in available databases. One of the databases scrutinized was that of the ExAC consortium (http://exac.broadinstitute.org) which may include data from individuals with low platelet counts who were either undiagnosed or recruited through an unrelated study (Table 2). Seven variants have previously been published as part of two separate publications from the UKGAPP study group.8,15 Classification of the 28 variants occurring within the known IT-related genes, following the interpretation guidelines set out by Richards et al.,17 revealed four variants to be “pathogenic”, 13 to be “likely pathogenic” and 11 to be of “uncertain significance”. Variants classified as “pathogenic” were either already known to be a genetic cause of IT; ANKRD26; c.-126T>G in F2.I and MYH9; haematologica | 2016; 101(10)
p.Arg1165Cys in F9.I, or were predicted to be loss-of-function variants in genes for which a loss of function is known to cause disease; FLI1; p.Asn331Thr fs*4, in F4.I and F4.II and RUNX1; pTrp79* in P12.I. On average, less than one novel variant was expected to be observed in the known IT-causing genes in which variants were observed. The number of variants occurring also exceeds the expected number when extending the analysis to cover variants with a mean allele frequency of <0.01. Of the 37 index patients, four presented with two candidate variations in known disease-linked genes, which in one case were present in the same gene. These were as follows: F6.I (GFI1B; c.676+1G>A and STIM1; p.Ala610Thr), F10.I (NBEAL2; (p.Leu459Arg fs*13 and p.Asn2298Ser), F12.I (RUNX1; p.Trp79* and ITGB3 p.Arg117Trp) and F23.I (TPM4; p.Ala183Val and TUBB1; p.Phe242Leu). Of the 25 index cases with variants in known diseasecausing genes, nine were observed to have variants within the RUNT1-related transcription factor gene; RUNX1. One variant, RUNX1; p.Arg177Gln, observed in F19.I and F19.II has been previously reported as a causative germline mutation of a familial platelet disorder in two individuals from the same pedigree.20 The variations consisted of five missense variants, two splice-site variants and one nonsense variant. One splice-site variation, c.270+1G>T, was present within three affected individuals from two separate families (F13 and F14). All variants, with the exception of a missense substitution (p.D6N), lie within the genetic region encoding the RUNT homology domain (RHD) which mediates DNA binding and heterodimerization with CBFβ (Figure 2).21 Platelets from the majority of these patients (10/13) demonstrated a reduction in ATP secretion and, in keeping with previous reports, several of these patients displayed additional clinical features. Variations in RUNX1 are associated with a propensity to myelodysplastic syndrome and acute myeloid leukemia. To date, hematologic malignancies have not been reported in any patients; however, the brother of F16.1 did have a history of acute myeloid leukemia but was unavailable for testing.
Potentially damaging variants in novel candidate genes After scrutinizing individuals for variants within the panel of 358 platelet-associated genes (Online Supplementary Table S1), individuals without a variant in a previously IT-associated gene were analyzed for variants in novel genes. WES analysis revealed potentially damaging candidate variants in three families with currently unknown genetic etiology (Table 3). All candidate variants are novel (excluding a previously annotated variant in MKL1; p.Val575Met, which occurs at a frequency of 0.007718 within the ExAC consortium), segregate with the disease status and have been confirmed by Sanger sequencing.
Variants within ANKRD18A, GNE and FRMPD1 in two related individuals from consanguineous relationships WES analysis of two related patients (F35.I and F35.II) of South Asian ethnicity was approached differently to that of other patients in this study. Both patients displayed a similarly severe clinical phenotype with a significant reduction in circulating platelets (15x109/L). Platelet function testing revealed a reduction in P-selectin (CD62P) expression upon stimulation and variable fluorescent fibrinogen binding which was consistent across both affected 1173
B. Johnson et al.
individuals. The patients were cousins born from consanguineous relationships within a single consanguineous kindred so the analysis was focused on identification of a shared homozygous variant due to the recessive segregation of disease. Three variants occurring within ANKRD18A; p.Glu799del, GNE; p.Gly447Arg and FRMPD1; p.Ala509Val were present in both affected individuals and within a tightly linked region of homozygosity on chromosome 9p. The variations within ANKRD18A and GNE were novel according to the previously mentioned databases whereas the variant in FRMPD1 has been observed at a frequency of 0.0003708 including 39 times within the South Asian population (rs571037699). There is no ClinVar entry for this variant and all three variants are classified as variants of “uncertain significance”.
One missense variant in the recently proposed inherited thrombocytopenia-linked gene, MKL1 One individual was shown to harbor a rare (frequency <0.01) missense variant within the Megakaryoblastic Leukaemia (translocation) 1 gene; MKL1. The variant was the only variant occurring within a gene of hemostatic relevance within 109 significant novel variants. The variant;
MKL1; c.1723G>A, p.Val575Met present in patient F37.I has been noted previously at a frequency of 0.0007718 (allele count of 6/7774 in the ExAC consortium). The patient has a mild reduction in platelet count (130x109/L) and no secondary qualitative defects in platelet function were observed. The variant is classified as of “unknown significance”.
Novel missense candidate variants in PADI2 and TTF2 Three affected individuals and four unaffected related individuals of a large kindred were recruited to the study. Mild thrombocytopenia was observed within the family with platelet counts ranging from 80x109/L to 186x109/L in the three affected individuals. All three affected individuals presented with a normal platelet size (7.9-8.6 fL) and a mild reduction in secretion was observed in F30.I and F30.III but not in F30.II. All affected individuals shared a similar bleeding phenotype, suffering from spontaneous epistaxis, excessive bruising and prolonged bleeding from minor wounds. WES analysis revealed 14 novel or rare (frequency <0.01) variants shared between the three affected individuals. Sanger sequencing of all 14 variants in four unaffected related individuals narrowed down candi-
Table 1. Platelet and bleeding phenotypes of 55 patients recruited to the UK-GAPP study.
Family Patient 1
I
Platelet count MPV (fL) (x109/L) 73
IPF (%)
9.6
Secondary defect
Bleeding phenotype
Yes (Fibrinogen)
Cutaneous bruising/bleeding, menorrhagia
2
I
50
8.6
No
3
I II III
80 50 98
10.3 12 10.5
Yes (Cyclooxygenase) Yes (Cyclooxygenase) N/A
Cutaneous bruising, menorrhagia Cutaneous bruising, epistaxis, purpura Cutaneous bruising, epistaxis, purpura
4
I II
142 157
11.8+ 11.4+
Yes (Secretion and Gi) Yes (Secretion)
Oral cavity bleeding, epistaxis, menorrhagia Cutaneous bruising, oral cavity bleeding
5
I II
92 100
8.8 8.6
Yes (Secretion) Yes (Secretion)
Cutaneous bruising, epistaxis, bleeding into joints Cutaneous bruising, life-threatening bleeding following surgery
6
I II
110 100
8.9
Yes (Secretion) Yes (Secretion)
Cutaneous bruising, excessive bleeding following surgery Cutaneous bruising, epistaxis
7
I
50
10.4
Yes (Secretion)
Cutaneous bruising, menorrhagia
8
I II
70 70
10.7+ 10.1
No No
Cutaneous bruising, menorrhagia, post-partum hemorrhage Cutaneous bruising, epistaxis, hematuria, menorrhagia, post-partum hemorrhage
9
I
35
11.4+
No
Epistaxis, cutaneous bruising.
3.2
Cutaneous bruising
10
I
55
N/A
11
I II III
62 N/A 146
Yes (Secretion) Yes (Secretion) Yes (Secretion)
Cutaneous bruising Cutaneous bruising, epistaxis Cutaneous bruising, epistaxis, menorrhagia
Cutaneous bruising, epistaxis
12
I
100
8
Yes (Secretion)
Excessive cutaneous bleeding
13
I II
163 45
9.1 11.9+
No Yes (GPVI)
14
I
139
8
Yes (Secretion)
Epistaxis, hematoma
15
I
90
7.6-
Yes (Secretion)
Cutaneous bruising, petechiae
16
I II
130 70
7.17.5-
Yes (Secretion and Gi) Yes (Secretion and Gi)
17
I
N/A
N/A
N/A
18
I
110
8.1
Yes (Secretion)
Cutaneous bruising, epistaxis, hematoma Cutaneous bruising, epistaxis
Cutaneous bruising, epistaxis, oral cavity bleeding Cutaneous bruising, epistaxis, oral cavity bleeding Excessive bruising/bleeding Cutaneous bruising/bleeding, petechiae, hematoma continued on the next page
1174
haematologica | 2016; 101(10)
Whole exome sequencing in inherited thrombocytopenia continued from the previous page
19
I II
100 100
9 9.2
No No
Cutaneous bruising/bleeding Cutaneous bruising/bleeding
20
I
89
13+
17.5
Yes (Secretion and Gi)
Cutaneous bruising
21
I II I II
63 83 74 62
11.9 11.9 11.2 12.7+
19.1 24.3
Yes (Secretion) Yes (Secretion) Yes (Secretion and Gi) Yes (Secretion and Gi)
III
109
11
Yes (Secretion)
Cutaneous bruising, epistaxis, hematoma Cutaneous bruising, epistaxis, hematoma Cutaneous bruising/bleeding, hematuria Cutaneous bruising/bleeding, menorrhagia, post-partum hemorrhage, hematoma Cutaneous bruising, hematoma, menorrhagia
23
I
119
11.1
Yes (Secretion)
Cutaneous bruising/bleeding
24
I II
104 133
9.6 8.6
No No
25
I
11
13.4+
Yes (Secretion and Gi)
26
I
43
14+
No
27
I
100
10.3
Yes (Secretion)
Cutaneous bruising/bleeding, epistaxis, oral cavity bleeding
28
I
25
8.5
Yes (Secretion)
Cutaneous bruising
29
I
15
9.4
Yes (Secretion)
Hematomas
30
I II III
137 186 80
7.9 8.6 8.2
Yes (Secretion) No Yes (Secretion)
Cutaneous bruising, epistaxis, menorrhagia Cutaneous bruising, menorrhagia, hematuria Cutaneous bruising
31
I
20
9.7
32
I
15
9.5
20.2
No
Cutaneous bruising
33
I
66
9.9
1.8
Yes (Secretion)
Cutaneous bleeding
34
I
93
14.4+
20.5
Yes (Secretion)
Cutaneous bleeding, epistaxis
35
I II
15 14
10.4 15+
87 83
Yes (Secretion and other) Yes (Secretion and other)
36
I
104
13.3+
17
No
Menorrhagia
37
I
130
9.7
No
Cutaneous bruising, epistaxis, menorrhagia
22
20.8
N/A
Menorrhagia, post-partum hemorrhage Epistaxis Cutaneous bruising Cutaneous bruising, menorrhagia, oral cavity bleeding
Cutaneous bruising, epistaxis, oral cavity bleeding
Cutaneous bruising, epistaxis, hematomas Cutaneous bleeding
Average platelet count = 85x10 /L (normal range 147-327x10 /L, n=40). Average mean platelet volume (MPV) = 10 fL (normal range 7.8-12.69 fL, n=40). Immature platelet fraction (IPF) was available for 11 patients and varied between 1.8-87% (normal range 1.3-10.8%, n=40). Patients with an observed macro and micro thrombocytopenia are denoted by a + and -, respectively, following their most recent analyzed MPV. Secondary qualitative defects are abbreviated to the following; (Gi) - reduction in response upon ADP stimulation indicating a possible defect in the Gi pathway, (GPVI) – reduction in surface GPVI quantity. Each individual bleeding diathesis is summarized under bleeding phenotype. 9
9
dates to only two missense variants; PADI2 (p.Lys499Arg) and TTF2 (p.His1089Asp). Both variants segregate with disease, not being present in the unaffected individuals. Both variants have been observed previously at a low frequency (<0.01) within the EXaC database (Table 3) and are currently classified as being of “uncertain significance”.
Discussion Here we present the first, large-scale application of WES analysis to patients with inherited bleeding diatheses presenting with thrombocytopenia of unknown etiology. Platelet counts and phenotypic presentations varied considerably among the patients studied, which is consistent with the variability observed in the spectrum of IT. However, the majority of patients (73%) were noted to have a secondary qualitative defect in platelet function which may explain the disproportionate bleeding when compared to the patients’ platelet counts. A lack of consistency was noted in families 13 and 30, which apparently included affected individuals both with and without defects in platelet function. Clinical complications are shared among the affected family members so this most haematologica | 2016; 101(10)
likely represents limitations in the sensitivity of platelet function testing or intra-familial variability. Overall, when considering pathogenicity WES analysis positively predicted pathogenicity in 46% of index cases (17/37) (results classified as “pathogenic” or “likely pathogenic” in a gene consistent with the patients’ phenotype and zygosity consistent with expected inheritance). Twenty-two percent of the index cases (8/37) had uncertain/possible pathogenicity (results classified as being of “uncertain significance” in known IT-causing genes). The remaining 32% of index cases (12/37) had a negative prediction of pathogenicity (no convincing variants identified in known IT-causing genes). WES is not without its limitations and, as with any genetic analysis, all variants must be functionally confirmed as deleterious to the coded protein. However, our positive variant discovery rate is comparable to or exceeds the rates in previous large-scale WES clinical multicenter studies of Mendelian disorders.22,23 Focusing our genetic analysis on patients with unknown etiology of disease with minor prior genetic testing has produced a spectrum of variants different from that from previous, large-scale, targeted genetic studies of IT. Patients were recruited to the study with clinically diagnosed bleeding disorders of unknown etiology. One 1175
B. Johnson et al.
caveat about this approach is the possible exclusion of individuals with known Bernard-Soulier syndrome or MYH9-related disorders as these two forms of IT are routinely tested for in many hematology centers in the UK. However, three index cases with variants in either GP1BA or MYH9 were noted in our analysis; these patients had atypical presentations of Bernard-Soulier syndrome or MYH9-related disorder and were, therefore, potentially falsely-negatively reported cases. The individuals with variants within GP1BA and MYH9 showed a slight increase in mean platelet volume; however, this was not at the magnitude of giant platelets normally attributed to this group of disorders and only patient F9.I showed any secondary syndromic symptoms with the individual suffering from congenital cataracts. One attribute of excluding patients with known variants in GP1BA, GP1BB, MYH9 and potentially GP9 was the discovery of a relatively large percentage of individuals analyzed (24% of index cases) with variants in RUNX1 as a primary likely cause of disease. With the exception of one predicted loss-of-function variant, the variants present within RUNX1 are currently classified as either “likely
pathogenic” or of “uncertain significance” and need functional confirmation to be considered the cause of disease. However, the presence of these variants in a large number of individuals with an often shared secondary functional defect in secretion does suggest that the prevalence of RUNX1 variants may be higher than previously thought. This raises the issue of whether they should be considered as clinically significant as Bernard-Soulier syndrome and MYH9-related disorders and be searched for in a primary genetic screening at the initial diagnosis of IT. An advantage of using WES is the possibility of finding candidate variations in novel genes in subjects who do not possess variants in known IT-causing genes. The determination of whether these candidate variants are in fact pathogenic relies on functional confirmation of the deleterious effect of the variant. However, WES analysis, especially with combined segregation analysis by Sanger sequencing in extensive kindreds, can provide indications as to which variants may be of scientific and clinical relevance. This strategy has recently been utilized in the discovery of novel candidate variations in SLFN14 initially as part of the GAPP study.15,24
Table 2. Results of whole exome sequencing analysis of 55 patients with inherited thrombocytopenia showing variants in known thrombocytopenia-causing genes. 68% of individuals have a predicted genetic etiology in a previously IT-associated gene. When a variant has been previously observed it is annotated in the prevalence column with the database in which it is included. The ACMG consensus guideline results are also displayed in the final classification column.17
Family Patient
Gene(s)
Genomic variation
Protein effect
1 2 3
ACTN1 ANKRD26 CYCS CYCS CYCS FLI1 FLI1 FLI1 FLI1 GFI1B STIM1 GFI1B STIM1 GP1BA GP1BA GP1BA MYH9 NBEAL2 NBEAL2 RUNX1 RUNX1 RUNX1 RUNX1 ITGB3 RUNX1 RUNX1 RUNX1 RUNX1 RUNX1 RUNX1
c.2647G>C c.-126T>G c.155C>T c.155C>T c.155C>T c.992_995del c.992_995del c.1028A>G c.1028A>G c.814+1G>A c.1828G>A c.814+1G>A c.1828G>A c.1761A>C c.413G>T c.413G>T c.3493C>T c.1376delT c.6893A>G c.16G>A c.16G>A c.16G>A c.236G>A c.349C>T c.270+1G>T c.270+1G>T c.270+1G>T c.322G>A c.427+1G>T c.427+1G>T
p.Gly883Arg
4 5 6
I I I II III I II I II I II
7 8 9 10 11
12 13 14 15 16
I I II I I I II III I I II I I I II
p.Ala52Val p.Ala52Val p.Ala52Val p.Asn331Thr fs*4 p.Asn331Thr fs*4 p.Tyr343Cys p.Tyr343Cys p.Ala610Thr p.Ala610Thr p.Gln587His p.Gly138Val p.Gly138Val p.Arg1165Cys p.Leu459Arg fs*13 p.Asn2298Ser p.Asp6Asn p.Asp6Asn p.Asp6Asn p.Trp79* p.Arg117Trp
p.Gly108Ser
Variation type
Prevalence
Classification
Missense Novel 5'-UTR variation Known Missense Novel Missense Novel Missense Novel Frameshift deletion Novel Frameshift deletion Novel Missense Novel Missense Novel Splicing Novel Missense 0.00019 (1k) Splicing Novel Missense 0.00019 (1k) Missense 0.00043 (EXaC) (rs570515282) Missense Novel Missense Novel Missense Known (rs80338829) Frameshift deletion Novel Missense Novel Missense Novel Missense Novel Missense Novel Nonsense Novel Missense Novel Splicing Novel Splicing Novel Splicing Novel Missense Novel Splicing Novel Splicing Novel
Likely pathogenic Pathogenic Likely pathogenic
Pathogenic Likely pathogenic Likely pathogenic Uncertain significance Likely pathogenic Uncertain significance Uncertain significance Likely pathogenic Pathogenic Likely pathogenic Likely pathogenic Likely pathogenic
Pathogenic Uncertain significance Likely pathogenic
Uncertain significance Likely pathogenic continued on the next page
1176
haematologica | 2016; 101(10)
Whole exome sequencing in inherited thrombocytopenia continued from the previous page
17 18 19 20 21 22
23 24 25 26 27 28 29 30
31 32 33 34 35 36 37
I I I II I I II I II III I I II I I I I I I II III I I I I I II I I
RUNX1 RUNX1 RUNX1 RUNX1 SLFN14 SLFN14 SLFN14 SLFN14 SLFN14 SLFN14 TPM4 TUBB1 TUBB1 TUBB1 TUBB1 Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown Unknown
c.505A>G c.512A>T c.530G>A c.530G>A c.652A>G c.657A>T c.657A>T c.659T>A c.659T>A c.659T>A c.548C>T c.726C>G c.721C>T c.721C>T c.1080_1081insG
p.Thr169Ala p.Asp171Val p.Arg177Gln p.Arg177Gln p.Lys218Glu p.Lys219Asn p.Lys219Asn p.Val220Asp p.Val220Asp p.Val220Asp p.Ala183Val p.Phe242Leu p.Arg241Trp p.Arg241Trp p.Leu361Ala fs*19
Missense Novel Missense Novel Missense Known Missense Known Missense Novel Missense Novel Missense Novel Missense Novel Missense Novel Missense Novel Missense Novel Missense Novel Missense 0.0001071 (ExAC)(rs368923302) Missense 0.0001071 (ExAC)(rs368923302) Frameshift insertion Novel
Family 35 is an interesting case of two affected related individuals born from consanguineous relationships. The molecular function of ANKRD18A is currently unknown, while FRMPD1 regulates the subcellular localization of activator of G-protein signaling 3 (AGS3).25 Both genes are expressed weakly in hematopoietic cells. However, GNE, coding for an enzyme in the sialic acid biosynthetic pathway, is expressed in all cells of the hematopoietic lineage. There are currently 88 registered mutations in GNE in the Human Genome Mutation Database (www.hgmd.cf.ac.uk). Mutations are known to be the genetic cause of sialuria (OMIM269921) and hereditary inclusion body myopathy (OMIM600737).26,27 Recently, two separate groups have reported patients with compound heterozygous variations in GNE, causing GNErelated myopathy with congenital thrombocytopenia.28,29 The platelet counts of the four reported affected individuals were below 45x109/L; platelet volume measurements were not recorded. None of the patients displayed signs of myopathy until mid-adolescence/early adulthood; F35.I and F35.II are currently aged 10 and 6, respectively. Without functional characterization of the effects of each variation, we cannot definitively conclude the genetic etiology of these two individuals’ severe thrombocytopenia. However, WES analysis has allowed us to focus our efforts on three potentially pathogenic variants in novel genes. haematologica | 2016; 101(10)
Uncertain significance Uncertain significance Likely pathogenic Uncertain significance Uncertain significance Likely pathogenic
Uncertain significance Uncertain significance Uncertain significance Likely pathogenic
MKL1 was initially included in our panel of 358 genes for post-WES analysis due to its role in megakaryocyte maturation elucidated via its binding partner, serum response factor (SRF).30-32 Recently, the first case of a homozygous mutation in MKL1 in a patient with severe immunodeficiency and no hematologic malignancies was reported.33 One interesting phenotypic presentation of the affected individual was an intermittent mild thrombocytopenia with low platelet counts in whole blood of between 50x109/L and 150x109/L. Here we present a novel variant within MKL1, at a highly conserved genetic site. The missense variant observed in F37.I represents the only variant to occur in a gene with previous hematologic implications. One further variant in MKL1 was observed in addition to a “likely pathogenic” frameshift causing insertion within TUBB1 in patient F25.I. Due to the predicted loss of function of the frameshift causing the TUBB1 variant it is unlikely that the variant with MKL1 is additive to the phenotype of patient F25.I. However, the variant of uncertain significance in patient F37.I is an interesting candidate to take forward for functional studies. WES and segregation determination using Sanger sequencing revealed candidate variants in PADI2 and TTF2 that segregate with disease in F30.I, F30.II and F30.III. The phenotypic presentations vary between the patients but clinical presentations are consistent, which 1177
B. Johnson et al. Table 3. Potentially damaging variants in novel candidate genes.
FamilyGene 30
Variant
Protein effect
PADI2
c.1496A>G
p.Lys499Arg
0.000008681
1.647
1
Diseas-e causing
Tolerated
TTF2
c.3265C>G
p.His1089Asp
1.65E-05
5.131
1
Diseasecausing
Damaging Deleterious Damaging
Novel
0.772
0.965
Polymorphism
NA
Deleterious
NA
PM2, Uncertain PP (segregation), PM6 significance PM2, Uncertain PP (segregation), PM6 sgnificance
35 ANKRD18A c.2395_2397delhom p.Glu799delhom
37
Prevalence PhyloP PhastCons Mutation taster
SIFT
Provean PolyPhen-2 Neutral
Benign
GNE
c.1339G>Ahom
p.Gly447Arghom
Novel
5.343
1
Diseasecausing
Damaging
Neutral
Damaging
FRMPD1
c.1526C>Thom
p.Ala509Valhom
0.0003708
-1.459
0
Polymorphism
Tolerated
Neutral
Benign
MKL1
c.1723G>A
p.Val575Met
0.0007718
3.358
1
Diseasecausing
Damaging
Neutral
Damaging
ACMG
Classification
PM (segregation)
Uncertain significance
PM (segregation) , PP3
Uncertain significance
PP (segregation), PM6
Uncertain significance Uncertain significance
When a variant has been previously observed it is annotated in the prevalence column with the database in which it is included. PhyloP scores vary between -14 and +6 and measure conservation at each individual base, sites predicted to be conserved are assigned a positive score, fast evolving sites are assigned a negative score. Mutationtaster uses a Bayes classifier to predict the effect of a mutation. SIFT damaging prediction score= <0.05. Provean deleterious score = <-2.5. PolyPhen-2 predictions are appraised qualitatively as benign or damaging. The ACMG consensus guidelines, including supporting evidence, are also shown.
Figure 2. Spatial amino acid locations of all thrombocytopenia-causing variants present within RUNT transcription factor 1 (RUNX1) (RefSeq NP_001001890). Previously disease-causing variants found the HGMD (www.hgmd.cf.ac.uk) and ClinVar (www.ncbi.nlm.nih.gov/clinvar/) databases are denoted above. The eight variants found within RUNX1 in the GAPP cohort of 54 patients who have undergone whole exome sequencing are denoted below and the effect on the protein or predicted splice-site is shown.
may reflect limitations in the sensitivity of platelet function testing. Neither gene has previously been implicated in hematologic abnormalities: mutations in PADI2 have been associated with schizophrenia, breast cancer and rheumatoid arthritis, while mutations in TTF2 have been associated with thyroid dysgenesis.34-37 WES analysis has therefore provided us with the first steps for determining the impact of these two variants of uncertain significance and whether they have the propensity to be disease causing. In summary, we show that WES can be applied to identify the underlying genetic cause in known IT-causing genes for patients with thrombocytopenia and unclear disease etiology. We show similar positive detection rates when compared to prior targeted studies and, with the addition of complementary functional studies, show an improved detection rate when compared to WES analysis of other developmental disorders. We also suggest the applicability of WES in providing preliminary insight into novel genes and their potential mechanism of action through candidate variations of unknown significance.
References 1. Noris P, Perrotta S, Seri M, et al. Mutations in ANKRD26 are responsible for a frequent form of inherited thrombocytopenia: analy-
1178
This approach provides a foundation to enhance our current knowledge on megakaryopoiesis, platelet function and platelet senescence/death through subsequent functional studies. Acknowledgments We thank the families for providing samples and our clinical and laboratory colleagues for their help. This work was supported by the British Heart Foundation (RG/PG/13/36/30275; RG/09/007), an MRC Doctoral Training Partnership grant (BJ), a Wellcome Trust Combined Training Programme Fellowship (093994) (GCL), the Healing Foundation (PH) and the Platelet Charity. We thank the NIHR Haematology Specialty Group for their help in recruiting to the study, and all our clinical investigators and collaborators. The authors also acknowledge support from the Department of Health via the National Institute for Health Research (NIHR) Comprehensive Biomedical Research Centre Award to Guy's & St Thomas' NHS Foundation Trust in partnership with King's College London and King's College Hospital NHS Foundation Trust. We thank the Queen Elizabeth Hospital Charity for funding the Sysmex XN-1000.
sis of 78 patients from 21 families. Blood. 2011;117(24):6673-6680. 2. Balduini CL. Diagnosis and management of inherited thrombocytopenias. European Human Genetics Conference 2014; 2014; Milan, Italy.
3. Johnson B, Fletcher SF, Morgan NV. Inherited thrombocytopenia: novel insights into megakaryocyte maturation, proplatelet formation and platelet lifespan. Platelets. 2016:1-7. Epub ahead of print. 4. Canales ML, Mauer AM. Sex-linked heredi-
haematologica | 2016; 101(10)
Whole exome sequencing in inherited thrombocytopenia
5.
6.
7.
8.
9.
10. 11. 12. 13.
14.
15.
tary thrombocytopenia as a variant of Wiskott-Aldrich syndrome. N Engl J Med. 1967;277(17):899-901. Kunishima S, Kobayashi R, Itoh TJ, Hamaguchi M, Saito H. Mutation of the beta1-tubulin gene associated with congenital macrothrombocytopenia affecting microtubule assembly. Blood. 2009;113(2):458-461. Kunishima S, Okuno Y, Yoshida K, et al. ACTN1 mutations cause congenital macrothrombocytopenia. Am J Hum Genet. 2013;92(3):431-438. Nichols KE, Crispino JD, Poncz M, et al. Familial dyserythropoietic anaemia and thrombocytopenia due to an inherited mutation in GATA1. Nat Genet. 2000;24(3): 266-270. Stockley J, Morgan NV, Bem D, et al. Enrichment of FLI1 and RUNX1 mutations in families with excessive bleeding and platelet dense granule secretion defects. Blood. 2013;122(25):4090-4093. Zhang MY, Churpek JE, Keel SB, et al. Germline ETV6 mutations in familial thrombocytopenia and hematologic malignancy. Nat Genet. 2015;47(2):180-185. Balduini CL, Pecci A, Noris P. Inherited thrombocytopenias: the evolving spectrum. Hamostaseologie. 2012;32(4):259-270. Balduini CL, Savoia A. Genetics of familial forms of thrombocytopenia. Hum Genet. 2012;131(12):1821-1832. Savoia A. Molecular basis of inherited thrombocytopenias. Clin Genet. 2016;89(2): 154-162. Dawood BB, Wilde J, Watson SP. Reference curves for aggregation and ATP secretion to aid diagnose of platelet-based bleeding disorders: effect of inhibition of ADP and thromboxane A(2) pathways. Platelets. 2007;18(5):329-345. Dawood BB, Lowe GC, Lordkipanidze M, et al. Evaluation of participants with suspected heritable platelet function disorders including recommendation and validation of a streamlined agonist panel. Blood. 2012;120(25):5041-5049. Fletcher SJ, Johnson B, Lowe GC, et al. SLFN14 mutations underlie thrombocytopenia with excessive bleeding and platelet secretion defects. J Clin Invest. 2015;125(9): 3600-3605.
haematologica | 2016; 101(10)
16. Leo VC, Morgan NV, Bem D, et al. Use of next-generation sequencing and candidate gene analysis to identify underlying defects in patients with inherited platelet function disorders. J Thromb Haemost. 2015;13(4): 643-650. 17. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-424. 18. Plagnol V, Curtis J, Epstein M, et al. A robust model for read count data in exome sequencing experiments and implications for copy number variant calling. Bioinformatics. 2012;28(21):2747-2754. 19. Seri M, Cusano R, Gangarossa S, et al. Mutations in MYH9 result in the MayHegglin anomaly, and Fechtner and Sebastian syndromes. The MayHegglin/Fechtner Syndrome Consortium. Nature Genetics. 2000;26(1):103-105. 20. Preudhomme C, Renneville A, Bourdon V, et al. High frequency of RUNX1 biallelic alteration in acute myeloid leukemia secondary to familial platelet disorder. Blood. 2009;113(22):5583-5587. 21. Kamachi Y, Ogawa E, Asano M, et al. Purification of a mouse nuclear factor that binds to both the A and B cores of the polyomavirus enhancer. J Virol. 1990;64(10): 4808-4819. 22. Chong JX, Buckingham KJ, Jhangiani SN, et al. The genetic basis of Mendelian phenotypes: discoveries, challenges, and opportunities. Am J Hum Genet. 2015; 97(2):199215. 23. Yang Y, Muzny DM, Reid JG, et al. Clinical whole-exome sequencing for the diagnosis of Mendelian disorders. N Engl J Med. 2013;369(16):1502-1511. 24. Marconi C, Di Buduo CA, Barozzi S, et al. SLFN14-related thrombocytopenia: identification within a large series of patients with inherited thrombocytopenia. Thromb Haemost. 2016;115(5):1076-1079. 25. An N, Blumer JB, Bernard ML, Lanier SM. The PDZ and band 4.1 containing protein Frmpd1 regulates the subcellular location of activator of G-protein signaling 3 and its
26.
27.
28.
29.
30. 31.
32.
33.
34.
35.
36.
37.
interaction with G-proteins. J Biol Chem. 2008;283(36):24718-24728. Seppala R, Lehto VP, Gahl WA. Mutations in the human UDP-N-acetylglucosamine 2epimerase gene define the disease sialuria and the allosteric site of the enzyme. Am J Hum Genet. 1999;64(6):1563-1569. Eisenberg I, Avidan N, Potikha T, et al. The UDP-N-acetylglucosamine 2-epimerase/Nacetylmannosamine kinase gene is mutated in recessive hereditary inclusion body myopathy. Nat Genet. 2001;29(1):83-87. Izumi R, Niihori T, Suzuki N, et al. GNE myopathy associated with congenital thrombocytopenia: a report of two siblings. Neuromuscul Disord. 2014;24(12):10681072. Zhen C, Guo F, Fang X, Liu Y, Wang X. A family with distal myopathy with rimmed vacuoles associated with thrombocytopenia. Neurol Sci. 2014;35(9):1479-1481. Cheng EC, Luo Q, Bruscia EM, et al. Role for MKL1 in megakaryocytic maturation. Blood. 2009;113(12):2826-2834. Smith EC, Thon JN, Devine MT, et al. MKL1 and MKL2 play redundant and crucial roles in megakaryocyte maturation and platelet formation. Blood. 2012;120(11):2317-2329. Halene S, Gao Y, Hahn K, et al. Serum response factor is an essential transcription factor in megakaryocytic maturation. Blood. 2010;116(11):1942-1950. Record J, Malinova D, Zenner HL, et al. Immunodeficiency and severe susceptibility to bacterial infection associated with a lossof-function homozygous mutation of MKL1. Blood. 2015;126(13):1527-1535. Watanabe Y, Nunokawa A, Kaneko N, et al. A two-stage case-control association study of PADI2 with schizophrenia. J Hum Genet. 2009;54(7):430-432. McElwee JL, Mohanan S, Griffith OL, et al. Identification of PADI2 as a potential breast cancer biomarker and therapeutic target. BMC Cancer. 2012;12:500. Chang X, Xia Y, Pan J, Meng Q, Zhao Y, Yan X. PADI2 is significantly associated with rheumatoid arthritis. PLoS One. 2013;8 (12):e81259. Castanet M, Polak M. Spectrum of human Foxe1/TTF2 mutations. Horm Res Paediatr. 2010;73(6):423-429.
1179
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Bone Marrow Failure
Ferrata Storti Foundation
Marked overlap of four genetic syndromes with dyskeratosis congenita confounds clinical diagnosis Amanda J. Walne,1 Laura Collopy, 1 Shirleny Cardoso,1 Alicia Ellison, 1 Vincent Plagnol,2 Canan Albayrak,3 Davut Albayrak,3 Sara Sebnem Kilic,4 Turkan Patıroglu,5 Haluk Akar,5 Keith Godfrey,6 Tina Carter,7 Makia Marafie,8 Ajay Vora,9 Mikael Sundin,10,11 Thomas Vulliamy,1 Hemanth Tummala,1 and Inderjeet Dokal1
Centre for Genomics and Child Health, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, Barts NHS Trust, London, UK; 2University College London Genetics Institute, UK; 3Department of Pediatric Hematology, Ondokuz Mayis University, Samsun, Turkey; 4Department of Pediatric Immunology Uludag University, Bursa, Turkey; 5Department of Pediatric Immunology Erciyes University Medical Facility, Kayseri, Turkey; 6Department of Pediatric Dermatology and NIHR Southampton Biomedical Research Center, University Hospital, Southampton and University of Southampton, UK; 7Department of Oncology and Haematology, Princess Margaret Hospital, Perth, WA, Australia; 8Clinical Cancer and Community Genetics, Kuwait Medical Genetics Center, Al-Sabah Medical area, Kuwait; 9 Department of Haematology, Sheffield Children’s NHS foundation Trust, Sheffield, UK; 10 Section of Pediatric Hematology/Immunology/SCT, Astrid Lindgren Children’s Hospital, Karolinska University Hospital, Stockholm, Sweden; and 11Division of Pediatrics, CLINTEC, Karolinska Institutet, Stockholm, Sweden 1
Haematologica 2016 Volume 101(10):1180-1189
ABSTRACT
Correspondence: a.walne@qmul.ac.uk
Received: April 20, 2016. Accepted: June 21, 2016. Pre-published: September 9, 2016. doi:10.3324/haematol.2016.147769
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1180
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is required for any other use.
1180
D
yskeratosis congenita is a highly pleotropic genetic disorder. This heterogeneity can lead to difficulties in making an accurate diagnosis and delays in appropriate management. The aim of this study was to determine the underlying genetic basis in patients presenting with features of dyskeratosis congenita and who were negative for mutations in the classical dyskeratosis congenita genes. By whole exome and targeted sequencing, we identified biallelic variants in genes that are not associated with dyskeratosis congenita in 17 individuals from 12 families. Specifically, these were homozygous variants in USB1 (8 families), homozygous missense variants in GRHL2 (2 families) and identical compound heterozygous variants in LIG4 (2 families). All patients had multiple somatic features of dyskeratosis congenita but not the characteristic short telomeres. Our case series shows that biallelic variants in USB1, LIG4 and GRHL2, the genes mutated in poikiloderma with neutropenia, LIG4/Dubowitz syndrome and the recently recognized ectodermal dysplasia/short stature syndrome, respectively, cause features that overlap with dyskeratosis congenita. Strikingly, these genes also overlap in their biological function with the known dyskeratosis congenita genes that are implicated in telomere maintenance and DNA repair pathways. Collectively, these observations demonstrate the marked overlap of dyskeratosis congenita with four other genetic syndromes, confounding accurate diagnosis and subsequent management. This has important implications for establishing a genetic diagnosis when a new patient presents in the clinic. Patients with clinical features of dyskeratosis congenita need to have genetic analysis of USB1, LIG4 and GRHL2 in addition to the classical dyskeratosis congenita genes and telomere length measurements. haematologica | 2016; 101(10)
DC or not DC? A clinical dilemma
Introduction Dyskeratosis congenita (DC) is a highly heterogeneous genetic and clinical syndrome. The classical presentation of DC is characterized by the mucocutaneous triad of abnormal skin pigmentation (hyper/hypopigmentation with atrophy and telangiectasia, termed poikiloderma), oral leukoplakia and nail dystrophy. Individuals with DC frequently develop bone marrow (BM) failure, are at a high risk of developing cancer and can develop disease features in virtually every system in the body. Pathologically, DC is characterized by selective exhaustion of highly proliferative cells1 that have critically short telomeres and exhibit an abnormal DNA damage response.2 The Dyskeratosis Congenita Registry (DCR, London, UK) is a collection of patients that have a clinical diagnosis of DC or an overlapping phenotype as defined by Dokal et al.3 The diagnostic inclusion criteria are: (1) All three mucocutaneous features (abnormal skin pigmentation, nail dystrophy and oral leukoplakia). (2) One mucocutaneous feature plus BM failure and two other somatic features of DC. (3) Aplastic anaemia (AA), myelodysplastic syndrome (MDS) or idiopathic pulmonary fibrosis (IPF) associated with a pathogenic telomerase variant. (4) Hoyeraal-Hreidarsson syndrome (HHS, growth retardation, developmental delay, microcephaly, cerebellar hypoplasia, BM failure and immunodeficiency). (5) Two or more features seen in DC plus very short telomeres (<1st centile). Based on these criteria the phenotype of DC is highly variable. Furthermore, not all physical malformations are present at the point of diagnosis, with many additional features developing with age or, in the case of severe disease, death may occur before more typical clinical presentations become apparent. To date, pathogenic mutations have been identified in 11 genes that cause DC according to the criteria above: dyskeratosis congenita 1, dyskerin (DKC1), telomerase RNA component (TERC), telomerase reverse transcriptase (TERT), NOP10 ribonucleoprotein (NOP10), NHP2 ribonucleoprotein (NHP2), TERF1 (TRF1)-interacting nuclear factor 2 (TINF2), WD repeat containing antisense to TP53 (WRAP53, also known as TCAB1), CST telomere maintenance complex component 1 (CTC1), regulator of telomere elongation helicase 1 (RTEL1), adrenocortical dysplasia homolog (mouse) (ACD), and poly(A)-specific ribonuclease (PARN). DKC1, TERC, TERT, NOP10 and NHP2 all encode components of the telomerase complex which is involved in telomere elongation. TINF2 and ACD encode components of the shelterin complex which is involved in telomere protection. CTC1 encodes a member of the CST complex which facilitates recruitment and docking of telomerase on to the telomere. WRAP53 is involved in telomerase trafficking and RTEL1 is involved in telomere replication, and DNA replication and repair.3-7 The final gene, PARN, is not exclusively involved in telomere biology but is mutated in a small number of cases of DC, and is an exoribonuclease involved in the control of mRNA stability and the maturation of TERC snoRNA.8-10 As the majority of DC genes are involved in telomere biology and DC patients usually have short telomeres, this has led to DC being classed as a telomeropathy.11 However these mutations do not explain all the cases that are present in the DCR (Data from the DCR, London, UK, April 2016). The advent of next-generation sequencing (NGS), haematologica | 2016; 101(10)
particularly whole-exome sequencing (WES), has become an invaluable tool for determining the causal variant(s) underlying a specific disease. By sequencing all the coding regions and comparing variants in many individuals with an overlapping phenotype, it is possible to elucidate potential disease genes in a way that would not have been possible five years ago.12 This can give the clinical diagnosis a genetic basis. The main problem in obtaining a definitive clinical diagnosis is the degree of phenotypic overlap that exists between many different diseases. The diagnosis, in many cases, depends on the interpretation of the initial clinician and this is often dependent on their specialty. Although subsequent opinions may be sought, the preliminary diagnosis tends to remain with the patient until proven otherwise. An accurate genetic diagnosis can aid clinical diagnosis and subsequent patient management. This has the potential to suggest alternative treatment avenues that may not have been considered, and also aid in the selection of healthy sibling donors for hematopoietic stem cell transplantation, should this become necessary. Here we report on 12 families with an initial diagnosis of DC who were found to have biallelic variants in one of three genes linked to a disease distinct from DC. The genes identified were USB1 (U6 snRNA biogenesis 1) associated with poikiloderma with neutropenia (PN), GRHL2 (grainyhead-like transcription factor 2) which is associated with ectodermal dysplasia/short stature syndrome (ECTDS), and LIG4 (Ligase IV, DNA, ATP-dependent) which is predominantly associated with LIG4 syndrome and occasionally Dubowitz syndrome. The key presenting features of PN are poikiloderma with noncycling neutropenia, recurrent infections, short stature and nail abnormalities.13 The typical clinical features associated with ECTDS are short stature, nail dystrophy, abnormal oral pigmentation, and keratoderma and hyperkeratosis of the hands and feet.14 LIG4 syndrome is characterized by immune deficiency and developmental and growth delay.15 Patients can also display unusual facial features, microcephaly, pancytopenia and various skin abnormalities. Dubowitz syndrome is characterized by growth failure/short stature, characteristic facial features, microcephaly, mild mental retardation and eczema.16 All of these disorders have a high degree of phenotypic overlap with DC, thus leading to difficulties in making an accurate clinical diagnosis. The identification of biallelic USB1, GRHL2 and LIG4 mutations in 12 different families, initially diagnosed to have DC, therefore demonstrate the marked phenotypic overlap of â&#x20AC;&#x2DC;classical DCâ&#x20AC;&#x2122; with four other genetic syndromes.
Methods Patient samples Exome capture or targeted gene screening was performed on a series of genetically uncharacterized index cases in the Dyskeratosis Congenita Registry (held at Barts and The London Hospital, London, UK). Presenting features included some/all of the classic mucocutaneous abnormalities (abnormal skin pigmentation, nail dystrophy, leukoplakia), with or without bone marrow failure. Peripheral blood samples were obtained with written consent under the approval of our local research ethics committee (London â&#x20AC;&#x201C; City and East). Genomic DNA was extracted from these peripheral blood samples for use in all the subsequent analyses (Puregene, Qiagen). 1181
A.J. Walne et al. Table 1. Overlap between DC, PN and patients with homozygous variants in USB1.
Syndrome/ Family DC (#305000, PN Family 1 (OMIM) #127500, (#604173) #224230) Consanguinity Country of Origin Family history Age at report (years) Mutation
Various*
USB1
Yes Turkey Yes 24,14 p.H179M fsTer86 Yes Yes
Nail dystrophy >60% cases** Yes Abnormal skin >60% cases Yes pigmentation BM failure >60% cases Yes Yesa Leukoplakia >40% cases Yes Developmental delay >20% cases Yes Microcephaly >20% cases Growth restriction† >20% cases Yes Yes Hair loss >20% cases Pulmonary disease >10% cases Yes Yes Cancer ~10% cases~10% cases Abnormal dentition <10% cases Yes Yes Gonadal abnormalities<10% cases Yes Deafness/ear <10% cases abnormalities Eye abnormalities <10% cases Abnormal facies <10% cases Yes Skeletal abnormalities <10% cases Immune deficiency <10% cases Yes Short telomeres <1st centile Unknown Normal
Family 237
Family 338
Turkey Yes 12,7,5 p.H179M fsTer86 Yes Yes
Turkey
Yesb
Family 4
Family 5
Family 6 Family 7
Family 8
‡ Turkey
Yes Australia Afghanistan
Yes Afghanistan
7 p.H179M fsTer86 Yes Yes
Yes Turkey Yes 9, 7 p.H179M fsTer86 Yes Yes
Yesc
Yesd
Yes
Yes Yes Yes Yes Yes
Yes Yes
Yes
6 p.H179M fsTer86 Yes Yes
10 4 p.Q225Ter p.H208R
16 p.H208R
Yes Yes
Yes Yes
Yes Yes
Yese Yes
Yesf
Yesg
Yesh Yes
Yes
Yes Yes
Yes MDS Yes
Yes Yes
Yes
Normal
Yes
Yes
Yes
Yes Yes Yes
Normal
Normal
Yes Normal
Normal
Yes Yes Normal
Yes Normal
All clinical features reported within a family are detailed. DC-Dyskeratosis congenita; PN-poikiloderma with neutropenia; *mutations identified in CTC1, DKC1, NHP2, NOP10, PARN, RTEL1, TERC, TERT,TINF2,TPP1 and WRAP53; **as calculated from the index cases of the 1st 400 families included in the DCR, London, UK. †Term used to cover short stature, intra uterine growth restriction and low birth weight; ‡parents reported as being from the same village but no statement of consanguinity. The blood counts for the index case at presentation are as follows: a hemoglobin 106g/l, leucocytes 2.3x109/L, neutrophils 0.3x109/L, platelets 246x109/L and bone marrow showed decreased cellularity; bhemoglobin 125g/L, leucocytes 3.2x109/L, neutrophils 0.1x109/L platelets 311x109/L and bone marrow showed decreased cellularity with approximately 10% blasts; c-hemoglobin 113g/l, leucocytes 3.2x109/L, neutrophils 0.8x109/L platelets 308x109/L; d hemoglobin 115g/L, leucocytes 3.4x109/L, neutrophils 0.8x109/L platelets 138x109/l; ehemoglobin 121g/l, leucocytes 3.68x109/L, neutrophils 0.7x109/L, platelets 269x109/L; fhemoglobin 70g/L, leucocytes 2.1x109/L, neutrophils 1.1x109/L, platelets 5x109/L and bone marrow showed increased cellularity, decreased megakaryocytes, dysmyelopoiesis with nuclear changes and abnormal granularity; ghemoglobin 115g/L, leucocytes 4.5x109/L, neutrophils 0.9x109/L, platelets 204x109/L; hhemoglobin 114g/L, leucocytes 1.7x109/L, neutrophils 0.8x109/L, platelets 112x109/l. MDS- myelodysplastic syndrome.
Telomere length measurement
Targeted gene screening
Telomere lengths were measured using the monochrome multiplex quantitative PCR (MMqPCR) method (modified from Cawthon),17 on a LightCycler 480 real-time thermocycler (Roche). Briefly, in each well, amplification of telomeric DNA (T) and a single copy gene (S) were quantified against standard curves obtained from the dilution of a reference DNA sample. The T/S ratio, obtained in triplicate for each sample, is proportional to the telomere length. This ratio was normalized to the T/S ratio of a second reference sample that was run on every plate to give a relative T/S ratio.
We designed an NGS assay covering the coding regions and some 5’UTR’s from 31 genes associated with genetic bone marrow failure syndromes (Online Supplementary Table S1). We used the Illumina TruSeq custom amplicon kit for library preparation and capture according to the manufacturer’s instructions. The resultant targeted fragments were indexed by dual barcodes and then sequenced on the Illumina MiSeq platform. Read alignment was performed automatically using the online MSR: TruSeq amplicon software. Annotation of the variants was performed using ANNOVAR.
Whole exome sequencing (WES)
Variant filtering
50ng genomic DNA was subjected to library preparation and exome capture using the Nextera Rapid Capture Exome kit (Illumina). Sequencing was performed using the Illumina HiSeq 2000 system and 100bp paired-end reads were generated and the data processed through the Illumina pipeline. Variants were called as described previously.8 All relevant variants were confirmed by Sanger sequencing.
Exome data from our patients were jointly called with 2500 WES internal control samples (UCL-EX consortium) with unrelated conditions to minimize artificial batch effects. Where there was a family history, we assumed an autosomal recessive mode of inheritance as all parents were reportedly asymptomatic. Filtering was performed using the following criteria: all variants had to pass the Illumina filter, have a read depth of >10X, be present in the
1182
haematologica | 2016; 101(10)
DC or not DC? A clinical dilemma
Table 2. Clinical overlap between the different genetic syndromes and the families with biallelic variants identified in this study.
Syndrome/Family (OMIM reference) Consanguinity Country of origin Family history Age at report (years) Mutated gene (variant identified)
DC (#305000, #127500, #224230)
Various*
Nail dystrophy >60% cases** Abnormal skin pigmentation >60% cases BM failure >60% cases Leukoplakia >40% cases Developmental delay >20% cases Microcephaly >20% cases † Growth restriction >20% cases Abnormal dentition <10% cases Oesophogeal stricture <10% cases Gonadal abnormalities <10% cases Ataxia <10% cases Immune deficiency <10% cases Deafness/ear abnormalities <10% cases Kidney abnormalities <10% cases Abnormal facies Short telomeres
<10% cases <1st centile
ECTDS (#616029)
Family 9 Family 1025 LIG4 syndrome Dubowitz (#606593) syndrome (#223370)
Yes Kuwait Yes 10, 5 GRHL2 (p.Y398H, GRHL2 p.I482K) (p.I482K)
Yes Yes Yes
Yes Yes Yes
Yes Turkey 27 GRHL2 (p.P405T)
Yes Yes Noa Yes
Yes Yes
Yes Yes Yesb Yes
Yes Yes Yes
Family 11
Family 12
Guadeloupe
UK
5 5 LIG4 LIG4 LIG4 LIG4 (various) (p.R814Ter, (p.R814Ter, (p.R814Ter, p.S205LfsTer29) p.K424RfsTer20) p.K424RfsTer20)
Yes Yes
Yes
Yes Yes Yes
Yes Yes Yes Yes
Yes
Yes
Yes
1 café au lait spot Yesc
Yesd
Yes Yes
Yes Dwarfism
Yes
Yes
Yes Normal
Normal
Yes
Unknown
Recurrent infections
Renal agenesis
Normal
Normal
Yes Unknown
Yes <1st centilee
All clinical features reported within a family are detailed. ECTDS: Ectodermal dysplasia/short stature syndrome. *Mutations identified in CTC1, DKC1, NHP2, NOP10, PARN, RTEL1, TERC, TERT, TINF2, TPP1 and WRAP53;**as calculated from the index cases of the 1st 400 families in the DCR, London, UK; †term used to cover short stature, intra uterine growth restriction and low birth weight. The blood counts for the index case at presentation are as follows: ahemoglobin 141g/l, leucocytes 7.2x109/L, neutrophils 4.4x109/L, platelets 307x109/L (normal blood counts); bhemoglobin 89g/L, leucocytes 4.9x109/L, platelets 216x109/L and bone marrow showed reduced cellularity with megaloblastic changes; chemoglobin 96g/L neutrophils 0.5x109/L, platelets 20x109/L and bone marrow showed reduced cellularity and reduced megakaryocytes; dpancytopenia with B lymphopenia and bone marrow showed reduced cellularity. e- reported in 3 patients only.
general population at a frequency of <0.0002 as reported on The Exome Aggregation Consortium database, (ExAC), 1000 genome project, and from the UCL-EX consortium. Variants that were predicted to be tolerated and benign by the Sorting Intolerant from Tolerant algorithm (SIFT) and Polymorphism Phenotyping v2 (PolyPhen-2) were then removed from further analysis. All relevant variants were confirmed by Sanger sequencing.
Results Telomere length analysis It has been well established that in the majority of patients with “classical” DC who have mutations in components of the telomere maintenance pathway have short telomeres, usually below the 1st centile.18,19 In fact, this measure is often used as a diagnostic screening tool to perform a differential diagnosis of DC from other bone marrow failure syndromes such as Fanconi anemia.20 Telomere lengths were therefore measured in our patients by MMqPCR, and this was used as an indicator of whether we could be looking for a mutation in a known telomere biology related gene or not. As none of the patients reported in this study had short telomeres when compared with haematologica | 2016; 101(10)
Figure 1. Telomere lengths in patients with GRHL2, LIG4 and USB1 mutations are not short. Relative T/S ratios are within the normal range. Open circles: controls (n=130); green diamonds: patients with biallelic USB1 variants (n=12), blue triangles: patients with homozygous GRHL2 variants (n=2); inverted red triangles – patients with compound heterozygous LIG4 variants (n=2). For comparison short telomeres are seen in patients with mutations in the known telomere associated genes DKC1 (n=50, open squares), TERC (n=66, open triangles), TERT (n=36, open inverted triangles) and TINF2 (n=33, open diamonds). The black line represents the median for each data set.
1183
A.J. Walne et al. A
B
C
D
E
F
G
H
Figure 2. Identification and segregation of disease-causing variants in USB1. (A) The c.531delA variant is identified in 5 families with a DC phenotype. Representative sequencing traces are shown for the wild-type (+/+), heterozygous (+/-) and homozygous (-/-) forms of the mutation. (B) Family tree and sequence traces of c.673C>T variant seen in Family 6. A representative heterozygous trace and the homozygous trace are shown. (C) Family tree and sequence traces of the novel c.623A>G homozygous variant (G/G) as seen in Family 7 and 8. NA: sample not available. D) Protein crystal structure of human USB1 (PDB id: 4W7H) depicted in ribbon form indicating both the transit and terminal lobes indicated in green α helices. The position of identified residues mutated in USB1 deficiency patients is indicated as a stick model in blue. (E and F) In silico modeling of truncation mutations reveals the loss of several α helices and β sheets in both the transit and terminal lobes of the USB1 secondary structure. (G and H) The invariant H-x-S motif in the active site is indicated as a stick model. The missense change histidine to arginine at position 208 (red) introduces an extra hydrogen bond (dotted black line indicated by arrows), with Ser210 (blue), which in turn might cause aberrant oligoadenylation of the RNA substrate.
controls (Figure 1), this suggested that we were looking for mutations outside the spectrum of those usually associated with DC, but they could still be associated with other bone marrow failure diseases or else have a previously unidentified association to another pathology.
Identification of causal genetic variants Analysis of our data has identified significant biallelic variants in three genes as detailed below. Five families underwent WES (Families 4, 9-12), and the remaining underwent targeted sequencing either using a 31 bone marrow failure disease gene panel (Online Supplementary Table S1) or direct sequencing.
USB1 (U6 snRNA biogenesis 1) Biallelic USB1 mutations have been previously described in patients with DC and the overlapping disease poikiloderma with neutropenia (PN).21,22 Since our previous publication,21 we have identified homozygous USB1 variants in an additional 8 families with features of DC. In 7 of the 8 families, the variant was identified by targeted sequencing and the remaining family (2 cases) underwent WES. The only variant that was homozygous and shared by both affected siblings was in USB1 so this was deemed to be causal. Five out of the 8 index cases had the previously documented recurring homozygous variant c.531delA, p.His179MetfsTer8613 (Figure 2A) which is not 1184
described on ExAC. The 6th index case was found to have the homozygous variant c.673C>T, p.Gln225Ter (Figure 2B), which has been reported previously as part of a compound heterozygote.23 The remaining 2 families (Family 7 and Family 8) have the same homozygous missense mutation c.623A>G, p.His208Arg (Figure 2C). This is the first report of a homozygous missense being associated with disease in USB1, as to date all 19 disease-causing variants reported in USB1 are predicted to be loss of function (splicing, nonsense or frame shift).13 Segregation analysis confirmed an autosomal recessive mode of inheritance, and homozygous variants were confirmed in an additional 4 affected siblings within these families (Figure 2A). In silico analysis of mutations on USB1 crystal structure (PDB id: 4W7H, Figure 2D) revealed that both p.His179MetfsTer86 and p.Gln225Ter truncates several α-helices and β-sheets in USB1 (Figure 2E,F). Furthermore the homozygous missense change p.His208Arg identified in families 7 and 8 affects the highly conserved histidine residue His208, (Online Supplementary Figure S1 and Figure 2G,H). This change is predicted to be probably damaging by PolyPhen-2; (HumVar score 1.0) and disease causing by Mutation Taster. This His208 residue is implicated in the catalytic mechanism of USB1, where it facilitates the displacement of uridine nucleoside on RNA substrates to stabilize U6 small nuclear RNA, which in turn plays a critical role in RNA splicing.24 haematologica | 2016; 101(10)
DC or not DC? A clinical dilemma
Figure 3. Similar clinical presentation is observed between DC and PN highlighting the marked overlap between these two syndromes. Panels A, D, G are from a male patient with p.Ala353Val DKC1 mutation, panels B, E and H are from a male patient with the homozygous USB1 p.His127MetfsTer86 mutation and panels C, F and I are from a child with the homozygous USB1 p.His208Arg mutation . Panels A-C show similar patterns of reticular pigmentation on the legs. Panels D-F show nail dystrophy on the toes. Panels G and H show reticular pigmentation on the trunk and panel I shows pigmentation on the arm along with nail dystrophy on the fingers. It is notable that the mucocutaneous features of the patient with USB1 p.His179MetfsTer86 mutation appear more like those of the patient with DKC1 p.Ala353Val mutation than the one with the USB1 p.His208Arg mutation.
Table 1 summarizes the clinical presentation of these families, and the similarity between the clinical presentation of DC and PN is further highlighted in Figure 3. The degree of overlap between the reticular pigmentation observed on the legs (Figure 3A-C), and nail dystrophy of the toes (Figure 3D-F) is marked. Similar pigmentation is seen on the trunk (Figure 3G,H) and nail dystrophy and blistering is shown on the hands (Figure 3I). In addition to highlighting the similarity between the cutaneous presentation of DC and PN, panels B, C, E, F, H and I also emphasize the difference in presentation that can be caused by mutations within the same gene. The blood results detailed in Table 1 show that although there is neutropenia in all the patients, the overall blood pathology is more global rather than being restricted to just one lineage. This investigation has expanded the repertoire of USB1 variants identified to date, by adding a recurrent novel homozygous missense variant.
GRHL2 (grainyhead-like transcription factor 2) The index case from both families 9 and 1025 underwent WES, and as both families were consanguineous, we haematologica | 2016; 101(10)
assumed an autosomal recessive mode of inheritance. Using the filtering approach described, it was noted that both families had homozygous non synonymous variants in the gene GRHL2; c.1445T>A p.Ile482Lys in Family 9 and c.1213C>A p.Pro405Thr in Family 10 (other homozygous calls that fulfilled the filtering criteria are detailed in the Online Supplementary Table S2). Segregation analysis confirmed that both variants were inherited in an autosomal recessive manner (Figure 4A). Both of these single base substitutions affect highly conserved nucleotides (Online Supplementary Figure S2). As neither variant is reported on ExAC and both variants are predicted by PolyPhen-2 to be probably damaging (HumVar scores p.P405T - 0.995 and p.I482K - 0.917) and disease causing by Mutation Taster, GRHL2 is the disease causing gene in these two families. GRHL2 is a member of a highly conserved family of transcription factors that are essential for epithelial development. At the molecular level, the GRHL transcription factors regulate the expression of proteins involved in cell proliferation, differentiation, adhesion and polarity. These factors adopt a DNA-binding immunoglobulin fold homologous to the DNA-binding 1185
A.J. Walne et al. A
B
Figure 4. Identification of disease-causing variants in two different genes in patients with clinical features of DC. (A) Segregation and sequence traces for the homozygous GRHL2 variants identified in Families 9 and 10. Representative heterozygous and homozygous traces are shown for both variants. Schematic representation of the protein shows the identified domains and the location of the variants. The red arrow highlights the variant identified in this study. The black arrow highlights a variant identified in a previous study. The green arrow highlights the variant observed in both this and a previous study. NA-sample not available. (B) Segregation and sequence traces for the compound heterozygous LIG4 variants identified in Families 11 and 12. A representative wild-type and heterozygous trace is shown for each. The nucleotides deleted are indicated by a line above the corresponding bases on the wild-type (+/+) trace. Schematic representation of the protein showing the precise location of variants in different domains. Black arrows highlight variants identified in previous studies. Green arrows highlight variants observed in both this and previous studies.
domain of key tumor suppressor p53.26 Recently, Petrof and colleagues described homozygous mutations in GRHL2 in two families with ectodermal dysplasia/short stature syndrome (ECTDS).14 One of these was p.Ile482Lys, which is the same variant as identified in Family 9. All the variants described by us and Petrof et al., affect the DNA binding domain of the protein (Figure 4A). Phenotypically there are many shared features between our two families (Families 9 and 10) and the ECTDS families reported in the literature (Table 2). Another recent study27 examined the role of GRHL2 in the developing kidney, and clarified the involvement of GRHL2 in a network that controls the development of lumen expansion and barrier formation in renal epithelia, which may explain the previously reported renal agenesis seen in Family 10.25 This suggests that homozygous variants in the gene GRHL2, previously linked to ECTDS, are disease-causing in these 2 families.
LIG4 (Ligase IV, DNA, ATP-dependent) Analysis of exome data from the index cases of Families 11 and 12 failed to reveal any homozygous variants that passed the filtering strategy. However, a biallelic analysis showed that they both shared the same compound heterozygous variants in LIG4, namely c.2440C>T, p.Arg814Ter and c.1270_1274 delAAAAG, p.Lys424ArgfsTer20 (Figure 4B, other biallelic calls are listed in the Online Supplementary Table S2). Although both 1186
variants observed in Families 11 and 12 are reported on ExAC at a very low frequency (0.000082 and 0.00014, respectively) in the heterozygous state, the likelihood of them occurring together (as is the case here) by chance is 2.4x10-8. LIG4 is involved in DNA non-homologous end joining and V(D)J recombination.28 Biallelic mutations in LIG4 are rare, with only 28 cases being reported to date in the literature,29 and are usually associated with LIG4 syndrome, and more recently in Dubowitz syndrome (DS).16 Both LIG4 syndrome and DS belong to a group of disorders that are associated with impaired DNA damage response mechanisms.28 LIG4 syndrome is a hereditary disorder associated with impaired DNA double-strand break repair mechanisms. It is characterized by growth restriction, developmental delay, microcephaly, facial dysmorphism, pancytopenia, variable immune deficiency and an increased predisposition to leukemia. DS is a rare multiple congenital syndrome characterized by cognitive delay, growth failure, microcephaly, distinctive facial dysmorphism, immune defects, pancytopenia, hematological malignancy and neuroblastoma. The degree of clinical overlap between LIG4 syndrome, DS and our 2 patients is marked (Table 2). Furthermore, the fact that both of the variants in LIG4 described here have been previously documented to be pathogenic and have been shown to occur in the same combination,30 suggests that these are causal genetic defects in these two unrelated individuals. haematologica | 2016; 101(10)
DC or not DC? A clinical dilemma
Figure 5. Biological overlap of classical DC proteins (TERT, DKC1, NOP10, NHP2, PARN and RTEL1) and those (GRHL2, USB1 and LIG4) mutated in patients in this study. Proteins with similar functions are grouped together. The schematic shows that GRHL2 can be grouped broadly with the molecules which have a role in telomerase expression, USB1 with small nuclear ribonucleoprotein (snRNP) processing and LIG4 with genomic stability. Previous studies have shown that biallelic mutations in GRHL2, USB1 and LIG4 are associated with ectodermal dysplasia, poikiloderma with neutropenia (or USB1 deficiency syndrome) and Lig4/Dubowitz syndrome, respectively. Mutations in DKC1, NHP2, NOP10, PARN and RTEL1 have been linked to dyskeratosis congenita.
Discussion Historically the diagnosis of disease is defined by the presenting clinical features, but it is becoming more apparent that this can give conflicting diagnoses as shown in Tables 1 and 2. Recent advances in whole exome and targeted sequencing have enabled a genetic diagnosis to be obtained more readily than ever before. This now enables the clinician to consider the possibility of diseases with overlapping phenotype based on a genetic result, rather than solely relying on the clinical presentation. In this report we have elucidated the underlying genetic causal variant(s) in 12 families (comprising 17 affected individuals) who were given a clinical diagnosis of dyskeratosis congenita. Initially we undertook whole exome sequencing on uncharacterized patients who presented with sufficient clinical features to be diagnosed as DC. As targeted sequencing methodologies improved, we then chose to use this approach as a screening tool to identify patients who had mutations in the known DC and bone marrow failure genes (Online Supplementary Table S1). The data reported herein represent a subset of all the cases analyzed, where we believe the disease-causing gene has been identified and where the identified gene has been previously associated with another genetic syndrome. By using a combination of these two strategies we identified disease-causing variants in the genes USB1 (8 families, 12 cases), GRHL2 (2 families, 3 cases) and LIG4 (2 families, 2 cases). Clinically, all patients show a high degree of overlap with the phenotypic profile for DC as described by Dokal et al.3 and the other diseases with which the disease-causing variants have been associated. This presents a problem in defining the disease: is it best to use phenotype or genotype? Significantly short telomeres are often used as a biomarker for DC. A telomere length below the 1st centile is considered diagnostic for DC, but this measurement is not routinely reported in patients with other bone marrow failure diseases. Figure 1 shows how telomeres in patients with variants in the known telomere genes (DKC1, TERC, TERT and TINF2) are significantly short when compared haematologica | 2016; 101(10)
with a control population. Telomere lengths in all affected cases reported herein were not significantly short when compared with controls, regardless of the underlying genetic variant (Figure 1). This measure should perhaps be used as a screening tool where a diagnosis of classical DC is suspected, in order to decide the best course of action to determine a genetic diagnosis. Reduced telomere length has also been associated with an increased risk of cancer, and DC is regarded as a cancer prone syndrome. Patients with USB1 mutations tend to have an earlier presentation of myelodysplastic syndrome (MDS) than the general population, as most reports describe the disease in children. In this study and in our previous study21 we report 6 patients with MDS and 2 with acute myeloid leukemia, suggesting there is an increased cancer risk despite the patients not having short telomeres. Based on the small number of cases in our series, there does not appear to be an increased cancer risk for patients with either GRHL2 or LIG4 mutations. It is notable, however, that in the literature LIG4 syndrome is described as a disorder with a predisposition to leukemia. Although an increased cancer risk is not observed in our families, this may be due to the small numbers of patients described herein. Due to the extensive overlap in clinical features between DC and PN as described in Table 1, telomere length seems to be one way of separating these two diagnostic definitions. Given that these cases do not have short telomeres but do possess DC features (Table 1), present us with a perplexing situation in accurately defining these patients. We propose it is timely to combine features of these two entities as a new syndrome, â&#x20AC;&#x153;USB1 deficiency syndromeâ&#x20AC;?, characterized by bone marrow failure, abnormal skin pigmentation (poikiloderma), nail dystrophy, growth restriction, cancer predisposition and normal length telomeres. It is also notable that the genes mutated in these overlapping syndromes have similarities with the biological functions of the known DC genes (Figure 5). USB1 has a role in snRNA processing, as is the case for dyskerin, NHP2 and NOP10.31 It also functions as an exoribonuclease reminiscent to the function of PARN that has been recently found to be mutated in both DC and pulmonary 1187
A.J. Walne et al. fibrosis.32 Both LIG4 and RTEL1 are involved in the maintenance of genomic stability, as they are both involved in pathways that repair double-strand breaks. LIG4 is an ATP-dependent DNA ligase that joins double-strand breaks during non-homologous end joining, and RTEL1 (a protein that is mutated in a subset of DC) is an ATPdependent DNA helicase, which has an important role in telomere-length regulation. Both function via interactions with another shelterin component TRF2 (Telomere repeat binding-factor 2).33,34 Phenotypically, many of the features seen in LIG4/Dubowitz syndrome are also seen in DC patients with biallelic RTEL1 mutations (Table 2).7 This supports the idea that mutations in LIG4 can give rise to a phenotype that is reminiscent of DC as caused by a gene that has a known association with telomere biology. GRHL2 affects TERT expression by acting on its promoter35,36 thereby providing an important biological link to classical DC. While these proteins do have different functions or interactions, we speculate that it is this similarity, as suggested in Figure 5, that gives rise to the overlapping phenotypes. It is also notable that many of the DC genes have functions other than telomere maintenance. It is therefore possible that those overlapping shared functions may also be contributing to the similar clinical features observed herein. In the future it will be interesting to determine whether therapeutic agents that work in one disease have a beneficial effect in any of the overlapping syndromes. The findings of this study raise a potential issue as to what constitutes DC. Taking into account the clinical features, germline genetic mutation (s) and telomere length we can recognize three categories of patients. “Category 1”: patients with sufficient clinical features (such as abnormal skin pigmentation, nail dystrophy, leukoplakia and bone marrow failure) to be labelled as DC, who have a germline mutation (s) in a telomere biology gene (such as TERT, TERC, DKC1, TINF2, RTEL1 and so forth) and who have short telomeres (described as being less than 1st centile). “Category 2”: patients with some clinical features of DC (but not sufficient to be classified as DC) who are found to have a germline mutation (s) in one of the telom-
References 1. Pereboeva L, Westin E, Patel T, et al. DNA damage responses and oxidative stress in dyskeratosis congenita. PLoS One. 2013;8(10):e76473. 2. Kirwan M, Beswick R, Walne AJ, et al. Dyskeratosis congenita and the DNA damage response. Br J Haematol. 2011;153(5):634-643. 3. Dokal I, Vulliamy T, Mason P, Bessler M. Clinical utility gene card for: Dyskeratosis congenita - update 2015. Eur J Hum Genet. 2015;23(4). 4. Gramatges MM, Bertuch AA. Short telomeres: from dyskeratosis congenita to sporadic aplastic anemia and malignancy. Transl Res. 2013;162(6):353-363. 5. Savage SA. Human telomeres and telomere biology disorders. Prog Mol Biol Transl Sci. 2014;125:41-66. 6. Kocak H, Ballew BJ, Bisht K, et al. Hoyeraal-Hreidarsson syndrome caused by a germline mutation in the TEL patch of the
1188
7.
8.
9.
10.
11.
ere biology genes and who have short telomeres. “Category 3”: patients with sufficient clinical features of DC but who harbor a germline mutation (s) in a nontelomere biology gene (such as USB1, LIG4 and GRHL2) and have normal length telomeres. Category 1 patients can be considered to represent “classical or pure DC”. Category 1 and Category 2 patients can both be considered to represent disorders of telomeres – “telomeropathies”. Category 3 could be considered as “DC-like” or “DC-overlap”. In summary, in 9 out of the 12 families highlighted in this study, the biallelic variants identified in USB1, GRHL2 and LIG4 have been shown to be pathogenic in previous studies.13,14,20,23,24 In the remaining families we have provided strong genetic and in silico evidence that the identified mutations are pathogenic. This study has demonstrated that the pleotropic clinical phenotype of DC markedly overlaps with the recognized disease entities LIG4 syndrome, Dubowitz syndrome and PN as well as the recently recognized ECTDS. In doing so it has substantiated the last category as a significant disease entity. The marked overlap of features of DC with PN, ECTDS, LIG4 and Dubowitz syndromes has important implications for establishing genetic diagnosis when a new patient presents in the clinic, specifically in patients with a clinical diagnosis of DC the genetic analysis needs to include GRHL2, LIG4 and USB1 in addition to the classical DC genes. Acknowledgments The authors would like to thank all the clinicians and patients who have helped us over the years, particularly Drs Babik, Balci, Ghoulam and Leblanc. We would also like to thank the staff at Barts and The London Genome Centre for Sanger sequencing analysis. Funding Financial support is provided by The Medical Research Council-MR/K000292/1, Children with Cancer- 2013/144 and Blood Wise-14032 (AJW, LC, SC, AE, TV, HT and ID). KMG is supported by the National Institute for Health Research through the NIHR Southampton Biomedical Research Centre.
telomere protein TPP1. Genes Dev. 2014;28(19):2090-2102. Walne AJ, Vulliamy T, Kirwan M, Plagnol V, Dokal I. Constitutional mutations in RTEL1 cause severe dyskeratosis congenita. Am J Hum Genet. 2013;92(3):448-453. Tummala H, Walne A, Collopy L, et al. Poly(A)-specific ribonuclease deficiency impacts telomere biology and causes dyskeratosis congenita. J Clin Invest. 2015;125(5):2151-2160. Moon DH, Segal M, Boyraz B, et al. Poly(A)-specific ribonuclease (PARN) mediates 3'-end maturation of the telomerase RNA component. Nat Genet. 2015;47(12):1482-1488. Dhanraj S, Gunja SM, Deveau AP, et al. Bone marrow failure and developmental delay caused by mutations in poly(A)-specific ribonuclease (PARN). J Med Genet. 2015;52(11):738-748. Townsley DM, Dumitriu B, Young NS. Bone marrow failure and the telomeropathies. Blood. 2014;124(18):2775-2783.
12. Ng SB, Buckingham KJ, Lee C, et al. Exome sequencing identifies the cause of a mendelian disorder. Nat Genet. 2010;42 (1):30-35. 13. Koparir A, Gezdirici A, Koparir E, et al. Poikiloderma with neutropenia: genotypeethnic origin correlation, expanding phenotype and literature review. Am J Med Genet A. 2014;164A(10):2535-2540. 14. Petrof G, Nanda A, Howden J, et al. Mutations in GRHL2 result in an autosomalrecessive ectodermal Dysplasia syndrome. Am J Hum Genet. 2014;95(3):308-314. 15. Chistiakov DA. Ligase IV syndrome. Adv Exp Med Biol. 2010;685:175-185. 16. Stewart DR, Pemov A, Johnston JJ, et al. Dubowitz syndrome is a complex comprised of multiple, genetically distinct and phenotypically overlapping disorders. PLoS One. 2014;9(6):e98686. 17. Cawthon RM. Telomere length measurement by a novel monochrome multiplex quantitative PCR method. Nuc Acids Res. 2009;37(3):e21.
haematologica | 2016; 101(10)
DC or not DC? A clinical dilemma
18. Alter BP, Baerlocher GM, Savage SA, et al. Very short telomere length by flow fluorescence in situ hybridization identifies patients with dyskeratosis congenita. Blood. 2007;110(5):1439-1447. 19. Vulliamy TJ, Kirwan MJ, Beswick R, et al. Differences in disease severity but similar telomere lengths in genetic subgroups of patients with telomerase and shelterin mutations. PLoS One. 2011;6(9):e24383. 20. Alter BP, Giri N, Savage SA, Rosenberg PS. Telomere length in inherited bone marrow failure syndromes. Haematologica. 2015; 100(1):49-54. 21. Walne AJ, Vulliamy T, Beswick R, Kirwan M, Dokal I. Mutations in C16orf57 and normal-length telomeres unify a subset of patients with dyskeratosis congenita, poikiloderma with neutropenia and RothmundThomson syndrome. Hum Mol Genet. 2010;19(22):4453-4461. 22. Arnold AW, Itin PH, Pigors M, et al. Poikiloderma with neutropenia: a novel C16orf57 mutation and clinical diagnostic criteria. Br J Dermatol. 2010;163(4):866-869. 23. Clericuzio C, Harutyunyan K, Jin W, et al. Identification of a novel C16orf57 mutation in Athabaskan patients with Poikiloderma with Neutropenia. Am J Med Genet A. 2011;155A(2):337-342. 24. Hilcenko C, Simpson PJ, Finch AJ, et al. Aberrant 3' oligoadenylation of spliceosomal U6 small nuclear RNA in poikiloderma with neutropenia. Blood. 2013;121(6):10281038.
haematologica | 2016; 101(10)
25. Balci S, Engiz O, Erekul A, Gozdasoglu S, Vulliamy T. An atypical form of dyskeratosis congenita with renal agenesis and no mutation in DKC1, TERC and TERT genes. J Eur Acad Dermatol Venereol. 2009; 23(5):607-608. 26. Mlacki M, Kikulska A, Krzywinska E, Pawlak M, Wilanowski T. Recent discoveries concerning the involvement of transcription factors from the Grainyhead-like family in cancer. Exp Biol Med (Maywood). 2015;240(11):1396-1401. 27. Aue A, Hinze C, Walentin K, et al. A Grainyhead-Like 2/Ovo-Like 2 Pathway Regulates Renal Epithelial Barrier Function and Lumen Expansion. J Am Soc Nephrol. 2015;26(11):2704-2715. 28. Chistiakov DA, Voronova NV, Chistiakov AP. Ligase IV syndrome. Eur J Med Genet. 2009;52(6):373-378. 29. Tamura S, Higuchi K, Tamaki M, et al. Novel compound heterozygous DNA ligase IV mutations in an adolescent with a slowly-progressing radiosensitive-severe combined immunodeficiency. Clin Immunol. 2015;160(2):255-260. 30. Murray JE, Bicknell LS, Yigit G, et al. Extreme growth failure is a common presentation of ligase IV deficiency. Hum Mutat. 2014;35(1):76-85. 31. Meier UT. The many facets of H/ACA ribonucleoproteins. Chromosoma. 2005; 114(1):1-14. 32. Berndt H, Harnisch C, Rammelt C, et al. Maturation of mammalian H/ACA box
33.
34.
35.
36.
37.
38.
snoRNAs: PAPD5-dependent adenylation and PARN-dependent trimming. RNA. 2012;18(5):958-972. Chapman JR, Taylor MR, Boulton SJ. Playing the end game: DNA double-strand break repair pathway choice. Mol Cell. 2012;47(4):497-510. Arnoult N, Karlseder J. Complex interactions between the DNA-damage response and mammalian telomeres. Nat Struct Mol Biol. 2015;22(11):859-866. Chen W, Dong Q, Shin KH, et al. Grainyhead-like 2 enhances the human telomerase reverse transcriptase gene expression by inhibiting DNA methylation at the 5'-CpG island in normal human keratinocytes. J Biol Chem. 2010; 285(52):40852-40863. Kang X, Chen W, Kim RH, Kang MK, Park NH. Regulation of the hTERT promoter activity by MSH2, the hnRNPs K and D, and GRHL2 in human oral squamous cell carcinoma cells. Oncogene. 2009;28(4):565574. Patiroglu T, Akar HH. Clericuzio-type Poikiloderma with Neutropenia Syndrome in a Turkish Family: a Three Report of Siblings with Mutation in the C16orf57 gene. Iran J Allergy Asthma Immunol. 2015;14(3):331-337. Kilic SS, Cekic S. Juvenile Idiopathic Inflammatory Myopathy in a Patient With Dyskeratosis Congenita Due to C16orf57 Mutation. J Pediatr Hematol Oncol. 2015; 38(2):e75-77.
1189
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Myeloproliferative Disorders
Ferrata Storti Foundation
Stat5 is critical for the development and maintenance of myeloproliferative neoplasm initiated by Nf1 deficiency
Zohar Sachs,1,2* Raha A. Been,2,3* Krista J. DeCoursin,2 Hanh T. Nguyen,1 Nurul A. Mohd Hassan,2 Klara E. Noble-Orcutt,1 Craig E. Eckfeldt,1 Emily J. Pomeroy,1 Ernesto Diaz-Flores,4,5 Jennifer L. Geurts,2 Miechaleen D. Diers,2,6 Diane E. Hasz,2 Kelly J. Morgan,6 Margaret L. MacMillan,6,7 Kevin M. Shannon,4,5 David A. Largaespada,2,6,7 and Stephen M. Wiesner2,8
Division of Hematology, Oncology, and Transplantation, Department of Medicine, University of Minnesota, Minneapolis, MN; 2Masonic Cancer Center, University of Minnesota, Minneapolis, MN; 3College of Veterinary Medicine and Department of Comparative and Molecular Biosciences, University of Minnesota, St. Paul, MN; 4 Department of Pediatrics, University of California, San Francisco, CA; 5Helen Diller Family Comprehensive Cancer Center, University of California, San Francisco, CA; 6 Department of Pediatrics, University of Minnesota, Minneapolis, MN; 7Blood and Marrow Transplantation Program, University of Minnesota, Minneapolis, MN; 8Center for Allied Health Programs, University of Minnesota, Minneapolis, MN, USA 1
Haematologica 2016 Volume 101(10):1190-1199
*ZS and RAB contributed equally to this work.
ABSTRACT
J
Correspondence: sachs038@umn.edu
Received: September 7, 2015. Accepted: June 15, 2016. Pre-published: June 14, 2016. doi:10.3324/haematol.2015.136002
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1190
Š2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1190
uvenile myelomonocytic leukemia is a rare myeloproliferative neoplasm characterized by hyperactive RAS signaling. Neurofibromin1 (encoded by the NF1 gene) is a negative regulator of RAS activation. Patients with neurofibromatosis type 1 harbor loss-of-function mutations in NF1 and have a 200- to 500-fold increased risk of juvenile myelomonocytic leukemia. Leukemia cells from patients with juvenile myelomonocytic leukemia display hypersensitivity to certain cytokines, such as granulocyte-macrophage colony-stimulating factor. The granulocyte-macrophage colony-stimulating factor receptor utilizes pre-associated JAK2 to initiate signals after ligand binding. JAK2 subsequently activates STAT5, among other downstream effectors. Although STAT5 is gaining recognition as an important mediator of growth factor signaling in myeloid leukemias, the contribution of STAT5 to the development of hyperactive RAS-initiated myeloproliferative disease has not been well described. In this study, we investigated the consequence of STAT5 attenuation via genetic and pharmacological approaches in Nf1deficient murine models of juvenile myelomonocytic leukemia. We found that homozygous Stat5 deficiency extended the lifespan of Nf1deficient mice and eliminated the development of myeloproliferative neoplasm associated with Nf1 gene loss. Likewise, we found that JAK inhibition with ruxolitinib attenuated myeloproliferative neoplasm in Nf1-deficient mice. Finally, we found that primary cells from a patient with KRAS-mutant juvenile myelomonocytic leukemia displayed reduced colony formation in response to JAK2 inhibition. Our findings establish a central role for STAT5 activation in the pathogenesis of juvenile myelomonocytic leukemia and suggest that targeting this pathway may be of clinical utility in these patients.
Introduction Juvenile myelomonocytic leukemia (JMML) is a rare myeloproliferative neoplasm (MPN) with no effective chemotherapy or targeted therapy options. Hematopoietic stem cell transplantation, with its considerable morbidity and morality burden, remains the only modality that can improve survival in patients haematologica | 2016; 101(10)
Stat5 is critical for Nf1-initiated MPN
with this condition.1,2 Nearly all patients (80-90%) harbor somatic or germline mutations that lead to hyperactive RAS signaling.2-4 Recent deep sequencing efforts have discovered that some patients harbor two, co-occurring RAS-pathway activating mutations and that these compound mutations are associated with more aggressive disease,3,4 underscoring the importance of hyperactive RAS in JMML. Neurofibromin, encoded by NF1, negatively regulates RAS activity.5 Patients with inherited mutations of NF1 have a 200- to 500-fold increased risk of developing JMML.1 Mice harboring activated Ras genes or Nf1 deficiency develop MPN that resembles human JMML.6-14 Likewise, mice that harbor compound activating mutations that activate the RAS pathway also display a more aggressive JMML phenotype.15 Notably, transplantation of Nf1-null fetal liver cells or somatic deletion of Nf1 in the hematopoietic compartment results in progressive myeloid expansion.9,10,16,17 Furthermore, induced pluripotent stem cells, generated from two patients with JMML, differentiated into myeloid cells with high proliferative capacity and enhanced basal ERK (a well-known mediator of RAS activation) and STAT5 activation.18 Malignant cells from JMML patients and JMML mouse models display hypersensitivity to certain cytokines, in particular granulocyte-macrophage colony-stimulating factor (GMCSF).5,9,14,19 The absence of GM-CSF receptor signaling prevents the development of MPN in recipient mice receiving hematopoietic stem cells doubly deficient for Nf1 and the GM-CSF receptor common β chain.16 Similarly, in an NrasG12D/+ model of MPN, β common chain deficiency did not prevent initiation of disease, but reduced splenomegaly and spontaneous colony formation and prolonged survival.20 GM-CSF receptor signaling promotes proliferation and differentiation by activating a variety of signal transduction pathways including Janus kinase 2 - signal transducer and activator of transcription 5 (Jak2-Stat5) and Ras.21,22 Mek inhibitors to modulate RAS activation have had variable therapeutic efficacy in JMML models. Myeloid cells, derived from the induced pluripotent stem cells described above, displayed reduced GM-CSF independence in response to Mek inhibition. In an activated Kras model of MPN, Mek inhibition abrogated the disease.23 In mouse models of Nf1-deficient or Kras-mutant MPN, Mek inhibition enhanced erythropoiesis and reduced spleen size, but failed to eradicate Nf1-deficient or Krasmutant cells.23,24 These studies support a central role of aberrant Raf/MEK/ERK signaling in the abnormal growth of JMML cells. The importance of STAT5a/b activation in JAK2mutant MPN has been well described. STAT5 is an important contributor to hematopoiesis and cancer.25-28 Hyperphosphorylation of STAT5 in response to minimal concentrations of GM-CSF is a hallmark of JMML.29 JAK2 mutations are common in other MPN, including 95% of cases of polycythemia vera and 50-60% of cases of primary myelofibrosis and essential thrombocythemia.30 Treatment with the JAK2 inhibitor ruxolitinib improves the clinical parameters and symptoms associated with these disorders31-34 and leads to a reduction of STAT5 activation in the cells of treated patients.35 JAK2 inhibition reduces the viability of primary cells from patients with chronic myelomonocytic leukemia displaying hypersensitivity to GM-CSF signaling.36 Likewise, Stat5 deficiency abrogates disease in mouse models of JAK2V617F MPN.37,38 haematologica | 2016; 101(10)
These findings highlight the critical role of STAT5 signaling in JAK2-mutant and other MPN featuring hyperactive GM-CSF signaling. The possible contribution of the JAK2/STAT5 pathway to MPN with hyperactive RAS signaling, such as JMML, has not been well described. JMML cells derived from NF1-deficient patients display differential STAT5 activation,29 implicating this pathway in diseases with hyperactive RAS signaling. In a mouse model of NrasG12D CMML, a subset of cells developed hyperactive Erk and Stat5 activation in response to GM-CSF signaling.39 Mek inhibition prolonged the life of 40% of CMML mice harboring NrasG12D/G12D, while combined Mek inhibition with Jak2 inhibition abolished the disease in these mice.40 These findings implicate STAT5 as a potential contributor to the pathogenesis of MPN with activated RAS. Since the therapeutic options in JMML are severely limited, identifying effective drug targets in this devastating disease of infancy is an important clinical priority. To elucidate the contribution of the Jak2-Stat5a/b signaling pathway to MPN derived from loss of Nf1, we attenuated Stat5 signaling in Nf1-deficient mice using either a genetic Stat5a/b hypomorphic knockout41,42 (which harbors a loss of both Stat5a and Stat5b genes) or pharmacological Jak2 inhibition with ruxolitinib.
Methods Mice Animals were treated in accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Minnesota A complex breeding scheme was established to generate animals of the appropriate genotypes (Figure 1A). The Stat5a and Stat5b alleles used in this study produce low amounts of an N-terminally deleted, partially functional form of their respective proteins.42,43 Henceforth, Stat5 refers to both Stat5a and Stat5b loci on mouse chromosome 11, with the status of both alleles indicated simultaneously as either + for both wild-type alleles or DN for the hypomorphic double knockout. The murine Stat5 loci map approximately 15 cM away from the Nf1 locus on chromosome 11. Therefore, two separate recombinant chromosomes were generated, one chromosome with the Stat544,45 combined with the Nf1Fcr (null) allele46 and the other with Stat5 combined with the Nf1flox allele.47 Breeding was complicated because Stat5DN/DN females are infertile and Stat5DN/DN offspring often fail to thrive. The low ratio of useful animals per litter necessitated transplantation of donor bone marrow into histocompatible recipient animals. Mx1-Cre transgenic animals (C57BL/6) were crossed with Nf1flox mice (C57BL/6) to generate Nf1flox/+/Mx1-Cre animals. Separately, Stat5DN mice on a C57BL/6 x 129/Sv background were crossed with Nf1Fcr mice (C57BL/6) to generate Nf1Fcr/+/Stat5DN//+ animals and with Nf1flox/+/Mx1-Cre animals to generate Nf1flox/+/Stat5DN//+/Mx1-Cre animals. These animals were crossed to provide donor animals of the following genotypes: Nf1flox/Fcr/Stat5+/+/Mx1-Cre, Nf1flox/+/Stat5DN//+/Mx1-Cre, Nf1flox/Fcr/Stat5DN//+/Mx1-Cre, Nf1flox/Fcr/Stat5DN/DN/Mx1-Cre and Nf1flox/+/Stat5DN/DN/Mx1-Cre animals (Figure 1A). Stat5DN/DN/Nf1 heterozygous mice, whether with the Nf1Fcr or Nf1flox allele, had particularly poor health and frequently died by 6 to 8 weeks of age. Transplants involving these genotypes were, therefore, done with single donors, rather than donor cells pooled from multiple mice. Multiple transplants were performed to achieve adequate numbers of experimental transplant recipients. 1191
Z. Sachs et al.
For these and all of the other genotypes, all donor animals also carried the Mx1-Cre transgene. Donor animals were F1 offspring from a cross of two strain backgrounds, C57BL/6 and 129/Sv, both of which express Ly5.2 (Ptpcra) on the surface of hematopoietic cells. Recipient animals were generated as F1 offspring from a cross between 129/Sv (Ly5.2, Ptpcra) and C57BL/6J (Ly5.1, Ptpcrb) animals. The recipient offspring were therefore congenic at the Ly5.1 locus, providing a mechanism by which to distinguish recipient Ly5.1+Ly5.2+ cells from donor Ly5.1- Ly5.2+ cells by immune staining for surface expression of Ly5.1 and Ly5.2.
Colony forming assays Methylcellulose cultures were performed as previously described.5 Briefly, peripheral blood mononuclear cells were isolated using LymphoprepTM (Stemcell Technologies, Vancouver, BC, Canada) according to the manufacturer’s instructions and plated at 5x103 cells/mL with MethoCult (Stemcell Technologies, Vancouver, BC, Canada), 100 U/mL penicillin G, 10 mg/mL streptomycin, and inhibitor. The inhibitors used were ruxolitinib at 0.4 μM, PD325901 at 13 μM (both from Selleckchem, Houston, TX, USA), or dimethylsulfoxide vehicle. Cultures were incubated at 37°C in 5% CO2 and scored 7 days later. Experiments in each condition were performed in triplicate.
A
Results Generation of Stat5, Nf1 double-knockout mice Mice were bred to generate five experimental groups: wild-type at the Stat5 loci and homozygous deficient at the Nf1 locus (Stat5+/+/Nf1flox/Fcr), heterozygous at the Stat5 loci and either heterozygous or homozygous deficient at the Nf1 locus (Stat5DN/+/Nf1+/Fcr or Stat5DN/+/Nf1flox/Fcr), and homozygous deficient at the Stat5 loci and either heterozygous or homozygous deficient at the Nf1 locus (Stat5DN/DN/Nf1+/Fcr or Stat5DN/DN/Nf1flox/Fcr) (Figure 1A). All donor animals also carried the Mx1-Cre transgene. Bone marrow from mice from each of these groups was transplanted into histocompatible recipients. One week after the transplant, recipient animals were injected with polyinosinicpolycytidylic acid (pIpC) to induce expression of the interferon responsive Mx1-Cre transgene; this led to deletion of the Nf1flox allele (Nf1D). Recipient animals heterozygous at the Ptpcr locus, Ptpcra/b, expressed both Ly5.1 and Ly5.2, and donor animals homozygous at the Ptpcr locus, Ptpcra/a, expressed only Ly5.2. This difference allowed us to identify cell origin by cell surface immune-staining. Recipient animals for all genotypes tested showed 70-90% engraftment 4 weeks after transplantation by flow cytometric
B
C
Figure 1. Stat5/Nf1-deficient bone marrow engrafts recipient animals. (A) Diagram depicting the breeding scheme to generate the five genetic backgrounds used in these studies. Stat5DN/+/Nf1Fcr/+ were generated and crossed with Stat5DN/+/Nf1Fcr/+/Mx1-Cre animals to generate the required genotypes. (B) Bone marrow was harvested from mice in each group and transplanted into syngeneic recipients. Four weeks after transplant and 2 weeks after induction of Cre recombinase, peripheral blood of recipient animals was immune-stained to measure the level of engraftment by Ly5.2+/Ly5.1- donor cells. Recipient mice showed greater than 70% engraftment by donor cells. Typical results are shown. (C) Eight weeks after transplantation, DNA was extracted from peripheral blood nucleated cells of recipient animals. Polymerase chain reaction analysis was performed on genomic DNA from each animal to determine the degree of deletion of the floxed Nf1 allele. A band indicating deletion was detected in all animals from which adequate DNA was obtained. Typical results are shown for three animals. W: water; D: recombined flox allele.
1192
haematologica | 2016; 101(10)
Stat5 is critical for Nf1-initiated MPN
analysis of circulating white blood cells (Figure 1B). Eight weeks after transplantation, peripheral blood was analyzed for deletion of the floxed Nf1 allele (Figure 1C). In all animals from which adequate DNA was obtained, Cremediated recombination was detected in peripheral blood mononuclear cell DNA (data not shown).
Stat5 deficiency attenuates Nf1-deficient myeloproliferative neoplasm To determine the potential contribution of Stat5 to the development of Nf1-deficient MPN, cells derived from Nf1flox/flox mice induced with pIpC to cause biallelic Nf1 deletion (Nf1D/D) were used. Baseline levels of STAT5 phospho-
A
B
C
D
Figure 2. Stat5 insufficiency alleviates MPN in Nf1-deficient mice. Bone marrow was harvested from donor mice, transplanted into syngeneic recipients, and allowed to engraft (as described in Figure 1). (A) Bone marrow was harvested from recipient animals [Nf1∆/∆ (n=4) and Nf1+/+ (n=3)], serum- and cytokine-starved, then stimulated with GM-CSF (10 ng/mL) for 10, 30 and 60 min. Levels of phosphorylated STAT5 (pSTAT5), ERK1/2 (pErk1/2), and STAT3 (pSTAT3) were measured using phospho-specific, intracellular flow cytometry of c-Kit+/lineage- cells. A representative histogram is shown. Geometric mean of fluorescence is normalized for each experiment by dividing the geometric mean of the fluorescence of each sample by the average of the geometric mean of all the samples in each experiment. Induction of phosphorylation is reported as the fraction of basal levels. Induction of phosphorylation is calculated by subtracting the basal geometric mean of fluorescence from the geometric mean of fluorescence from each GM-CSF-stimulated sample; this difference is normalized for each sample by dividing by the basal geometric mean of fluorescence. (B) Peripheral blood from transplant recipients was collected every 6 weeks for the duration of experiments and with increased frequency in diseased animals. Total white blood cells counts (WBC) and peripheral blood smears (data not shown) were used to monitor the development of myeloproliferative disease in recipient mice. P<0.01, one way ANOVA followed by the Bartlett test for equal variance and the Tukey multiple comparison were performed for the comparison between Stat5+/+/Nf1∆N/Fcr bone marrow and Stat5∆N/+/Nf1∆N/+, Stat5∆N/∆N/Nf1∆N/Fcr, and Stat5∆N/∆N/Nf1∆N/+ bone marrow. (C) Moribund animals were sacrificed and spleen weights assessed. ANOVA followed by the Tukey test for significant differences were performed. (D) Kaplan-Meier survival plot comparing overall survival of recipients with the indicated genetic background. n ≥ 7 recipients per group; Log-rank followed by chi square testing was performed. In all figure panels, error bars represent standard errors of the mean.
haematologica | 2016; 101(10)
1193
Z. Sachs et al. A
B
C
Figure 3. STAT5 insufficiency reverses immature myeloid expansion in Nf1-deficient mice. Surface immunophenotyping was performed on bone marrow and spleen mononuclear cells. (A) Representative immunophenotyping flow cytometry plot of Nf1-deficient bone marrow. (B) The proportion of cells from each immunophenotypic compartment in the bone marrow is indicated for each genotype. The inset shows the ratio of Mac1+Gr1High to Mac1+Gr1Low cells for each genotype. Two-way ANOVA followed by the Bonferroni post-test for significance was used for the main panel, and one way ANOVA followed by the Tukey multiple comparison test was used for the inset Mac1+Gr1High/Low ratio. (C) The proportion of Mac1+Gr1Low cells in the spleen is indicated for each genotype. ANOVA followed by the Tukey test was used to determine significant differences. In all figure panels, n = 10 for Stat5+/+/ Nf1∆N/Fcr, n = 4 for Stat5∆N/+/ Nf1∆N/Fcr, and n = 6 for Stat5∆N/∆N/ Nf1∆N/Fcr; error bars represent standard errors of the mean.
rylation were comparable between Nf1D/D and Nf1+/+ controls. GM-CSF-stimulation of cKit+ lineage-/Low populations (enriched for stem cells and progenitor cells) from these mice led to increased levels of STAT5 phosphorylation, a measure of STAT5 activation, in both populations. Notably, Nf1D/D cells achieved maximal levels of STAT5 phosphorylation within 10 min of stimulation, while Nf1+/+ populations took 60 min to achieve comparable levels of STAT5 activation (Figure 2A). A more rapid response to cytokines in Nf1D/D cells tightly correlates with the hypersensitivity to GM-CSF observed in the hematopoietic compartment of these Nf1-mutant mice10 and recapitulates the cytokine hyperresponsiveness described in other studies of MPN.5,9,14,48-50 In contrast, these cells displayed a trend toward elevated levels of phosphorylated ERK (pERK) under basal conditions, without significant differences in induction of pERK. Phosphorylated STAT3 levels were similar between the two genotypes; GM-CSF stimulation did not significantly increase STAT3 phosphorylation in either group. Transplantation of Stat5+/+/Nf1D/Fcr bone marrow in this study resulted in MPN similar to that found in previous studies as assessed by elevated white blood cell counts and spleen weights9,51 (Figure 2B,C and Online Supplementary Tables S1-S5). Hemoglobin concentration did not vary significantly by genotype and remained within the normal range (Online Supplementary Figure S1A). The platelet count remained in the low-normal range for all genotypes for the first year after transplantation (Online Supplementary Figure S1B), likely reflect1194
ing low-level radiation-induced bone marrow toxicity. After 1 year after transplantation, Stat5DN/DN/ Nf1D/+ animals developed platelet counts that were significantly higher than those of the animals with other genotypes, but well within the normal range. Since these animals had an intact Nf1 allele, this phenotype likely reflects the effect of isolated Stat5 deficiency on the platelet count. Stat5+/+/Nf1D/Fcr animals succumbed to MPN at a median of 55 weeks after transplantation (Figure 2D and Online Supplementary Figure S1C). In contrast, animals with a single, intact Nf1 allele (Nf1D/+) did not develop MPN, as assessed by white blood cell counts and spleen weights (Figure 2B,C). Animals that received bone marrow harboring a single, intact Stat5 allele and homozygous Nf1 deficiency (Stat5D/+/Nf1D/Fcr) also developed MPN. These Stat5D/+/Nf1D/Fcr recipients had a comparable median survival to that of Stat5+/+/Nf1D/Fcr recipients (58 versus 55 weeks) and comparable spleen sizes (average spleen 1.03 g versus 0.75 g, P=0.25) but did display a delay in the development of MPN as measured by peripheral white blood cell counts (Figure 2B). In contrast, recipients of Nf1D/Fcr bone marrow lacking both copies of wild-type Stat5 (Stat5D/D/ Nf1D/Fcr) did not develop MPN (as determined by white blood cell counts and spleen weights) and had a prolonged median survival of 79 weeks. The experiment was terminated 80 weeks after the transplants. Three Stat5D/D/ Nf1D/Fcr animals died prior to this time point with no obvious cause of death but their premature death may be attributed to the complications of radiation exposure. The survival of animals of all genohaematologica | 2016; 101(10)
Stat5 is critical for Nf1-initiated MPN
A
B
C
D
Figure 4. The Jak/Stat inhibitor, ruxolitinib significantly reduces disease burden in mice with MPN. Mice were treated daily with ruxolitinib for 6 weeks and sacrificed at the completion of treatment or when moribund. (A) Spleens were harvested and weighed at the time of sacrifice. Average spleen weights are indicated for each genotype. (B) For each mouse with MPN, reduction of disease burden was assessed by dividing the white blood cell count (WBC) at the time of sacrifice to maximum WBC measured during the animal’s lifetime. The average of these ratios was compared between mice with MPN treated with vehicle and ruxolitinib. (C) Kaplan-Meier survival curves are shown for vehicle- or ruxolitinibtreated Nf1∆N/∆N animals with MPN. (D) The average percentage of Mac1+Gr1Low cells in the bone marrows of MPN mice is shown for mice treated with vehicle and ruxolitinib. The number of mice (n) per group is indicated. In all figure panels, error bars represent standard errors of the mean.
types along with a comprehensive table of their clinical status is shown in the Online Supplementary Material (Online Supplementary Figure S1C and Online Supplementary Table S1). Chimerism was measured via Ly5.1/5.2 mismatching for all animals at the termination of the experiment and necropsy. All surviving Stat5DN/DN animals were sufficiently reconstituted with donor hematopoietic cells (average % donor cells ± SE in bone marrow = 83.3 ± 13.4 and 73.1 ± 16.2 for Stat5DN/DNNf1D/Fcr and Stat5DN/DNNf1D/+mice, respectively), yet did not develop disease. Only one animal showed chimerism with less than 60% donor cells (49.7%). Thus, the failure to observe MPN in Stat5DN/DN recipients could not be attributed to engraftment failure. These data show that the absence of any wild-type STAT5 abrogates the development of NF1-deficient MPN and haematologica | 2016; 101(10)
demonstrate that STAT5 activity is a critical contributor to NF1-deficient MPN.
Stat5 deficiency reverses myeloid precursor accumulation characteristic of myeloproliferative neoplasm
MPN in Stat5DN/+/Nf1D/Fcr animals was typical of MPN associated with Stat5+/+/Nf1-/- animals and did not differ significantly from that reported by other investigators10 in terms of absolute leukocytosis, splenomegaly, and morphology (Figure 2 and Online Supplementary Figure S1). Furthermore, as has been described in other models of Nf1-deficient MPN,10 homozygous Nf1-deficient animals displayed an expansion of immature myeloid precursors (Mac1+Gr1Low double-positive cells) in the bone marrow (Figure 3A,B), spleen (Figure 3A,C), and peripheral blood 1195
Z. Sachs et al.
(data not shown). Homozygous Stat5 deficiency reversed this expansion (Figure 3). Interestingly, mice with Stat5 heterozygous, Nf1-deficient MPN displayed an expansion of the Mac1+Gr1Low (immature precursors) compartment that was intermediate between the expansion of this compartment in Stat5DN/DN and Stat5+/+ animals. This intermediate phenotype suggests that haploinsufficiency at the Stat5 locus may abrogate MPN as well.
A
The Jak2 inhibitor, ruxolitinib, diminishes myeloproliferative neoplasm Next, we investigated whether pharmacological inhibition of the Jak/Stat pathway abrogates MPN initiated by Nf1 deficiency. We used ruxolitinib because Jak inhibition with ruxolitinib has been shown to attenuate Stat5 activation.52 We analyzed the bone marrow and spleen compartments of Mx1-Cre, Nf1flox/flox animals.10 Treatment with pIpC homozygously ablates the Nf1 locus (Nf1D/D) in the hematopoietic compartment of these mice leading to JMML-like MPN.10 These mice were treated with pIpC at 2 months of age and aged for an additional 6 months to allow MPN to develop before treatment with ruxolitinib. The mice were treated twice daily with ruxolitinib for 6 weeks. Complete blood counts were obtained weekly throughout the treatment period. At the completion of treatment, mice were sacrificed and bone marrow and spleens were harvested. Nf1-deficient animals developed MPN characterized by splenomegaly, leukocytosis, and anemia (Figure 4A, Online Supplementary Figure S2A,B, Online Supplementary Tables S6-S9). Platelet counts remained within the normal range but were higher in Nf1-deficient, vehicle-treated animals than in animals retaining a wild-type copy of the Nf1 allele (Online Supplementary Figure S2C). Ruxolitinib treatment attenuated MPN in mice, as evidenced by reduced spleen size (Figure 4A). White blood cell counts varied significantly among the mice with MPN (Online Supplementary Figure S2A). Ruxolitinib therapy reduced the white blood cell count by 50% in these MPN animals (Figure 4B). Ruxolitinib treatment was also associated with worsening anemia in animals with MPN (as has been described in clinical trials with this agent) but did not reduce the hemoglobin concentration of animals without MPN (Online Supplementary Figure S2B). Additionally, ruxolitinib was associated with a reduction in platelet count that was more pronounced in animals with MPN (Online Supplementary Figure S2C). All of the ruxolitinib-treated Nf1D/D mice survived until completion of the experiment. In contrast, 5/13 (38%) of vehicle-treated Nf1D/D mice succumbed to MPN during the treatment course (Figure 4C). Although this experiment was not designed to detect the effect of ruxolitinib on survival, the difference that we observed is statistically significant (P<0.05). Since the mice were sacrificed at completion of therapy, the difference in survival between the two treatment groups likely under-estimates the effect of ruxolitinib on survival. In accordance with our findings in STAT5D/D mice, Jak/Stat inhibition with ruxolitinib tended to reduce the percentage of Mac1+Gr1Low cells in the bone marrow of mice with Nf1D/D MPN, although this trend was not statistically significant (Figure 4D). These data demonstrate that a targeted inhibitor of Jak/Stat signaling is efficacious in attenuating the clinical features of Nf1-deficient MPN in mice. 1196
B
C
Figure 5. Jak/Stat inhibition inhibits ERK phosphorylation and colony formation in RAS-activated MPN. (A) Intracellular, phospho-specific flow cytometry was performed on bone marrow cells from vehicle-treated Nf1∆/+ control mice (n=12), vehicle-treated Nf1∆/∆ MPN mice (n=3), and ruxolitinib-treated Nf1∆/∆ MPN mice (n=7). Average mean fluorescence intensity representing levels of phospho-Erk are shown for each cohort. (B) Primary bone marrow mononuclear cells from a patient with KRAS-mutant (KRASG13D) JMML were incubated with vehicle or ruxolitinib (4 μm) for 30 min before fixation and permeabilization. Levels of phospho-Erk were assessed by mass cytometry. (C) Peripheral blood mononuclear cells from a patient with KRAS-mutant JMML were plated in methylcellulose containing MEK inhibitor, JAK inhibitor, both inhibitors, or vehicle. Each condition was plated in three replicates. Colony formation was scored after 7 days. In all figure panels, error bars represent standard errors of the mean. 95% confidence intervals are indicated.
haematologica | 2016; 101(10)
Stat5 is critical for Nf1-initiated MPN
Stat5 activity is implicated in maintaining signaling through the Ras/Nf1 pathway.
Bone marrow cells of Nf1-deficient (Mx1-Cre, Nf1D/D) mice with MPN displayed hyperactive Erk signaling in comparison to bone marrow from Mx1-Cre, Nf1D/+ controls with no MPN (Figure 5A). Ruxolitinib treatment, which inhibits the Jak/Stat pathway, reduced levels of phosphorylated Erk in Nf1-deficient bone marrow (Figure 5A). This finding, along with our data from Stat5DN/DN/Nf1D/Fcr mice (Figures 2 and 3), suggests that sustained Stat5 signaling may be required to maintain hyperactive Mek/Erk signaling conferred by Nf1 deficiency. Indeed, mononuclear cells of mice treated with ruxolitinib showed a trend to reduced pSTAT5 induction in response to in vitro GM-CSF stimulation, although this trend was not statistically significant (Online Supplementary Figure S2D). To investigate whether the STAT5 and RAS pathways are similarly inter-connected in human MPN, we studied a patient with JMML harboring a KRAS mutation (KRASG13D). As in our murine model, ruxolitinib treatment led to reduced levels of pErk in bone marrow mononuclear cells from this patient (Figure 5B). Treatment with a MEK inhibitor (PD325901) or a JAK inhibitor (ruxolitinib) led to a decrease in colony formation in methylcellulose by peripheral blood mononuclear cells from this patient (Figure 5C). Notably, JAK inhibition had a more profound effect on colony formation than had MEK inhibition. Simultaneous treatment with both inhibitors gave results similar to those with inhibition of JAK alone. These data suggest that active JAK/STAT is required for the proliferative phenotype of MPN with hyperactive RAS.
Discussion In this study, we used genetic and pharmacological approaches to demonstrate the importance of Stat5 in the pathogenesis of MPN initiated by Nf1 inactivation. We showed that MPN in Nf1-deficient, Stat5 hypomorphic mice is significantly diminished, leading to prolonged survival, improvement in blood count indices, and reduced spleen size in comparison to Nf1-deficient mice with intact Stat5 genes. Similarly, Nf1-deficient mice treated with ruxolitinib, an inhibitor of Jak/Stat signaling, had attenuated MPN with reduced white blood cell counts and smaller spleens. Both approaches tended to reverse the Mac1+GrLow immature myeloid cell accumulation seen in Nf1-deficient MPN bone marrow. We showed that ruxolitinib treatment diminished Erk signaling in these mice and in the bone marrow of a KRAS-mutated JMML patient. Ruxolitinib treatment also inhibited colony formation of primary cells from this JMML patient. Our ruxolitinib data implicate the Jak/Stat pathway in the pathogenesis of MPN but do not rule out effects of other STAT in the phenotype we observed. However, the ruxolitinib data, together with the data from our genetic Stat5-deficient model, suggest that STAT5 can modulate RAS-activated MAPK pathway activity. The Stat5 alleles utilized in these experiments express an N-terminally deleted form of Stat5 that retains partial Stat5 function.42,43 Nevertheless, we showed that attenuation of Stat5 with retention of residual Stat5 function was sufficient to alleviate MPN in our genetic model. In a mouse model of MPN mediated by Mpl mutation, which leads to tonic activation of the Jak2/Stat5 pathway, conditional genetic ablation of Jak2 (via floxed alleles) was sufficient to haematologica | 2016; 101(10)
induce complete remission of MPN, while ruxolitinib treatment of these Mpl mutant mice could only attenuate the disease.53 This study indicates that ruxolitinib does not completely inhibit Jak2 signaling, a finding that is consistent with clinical trials that show that ruxolitinib improves the clinical parameters of MPN but does not cure the disease.3234,54 Likewise, incomplete inactivation of Stat5 with ruxolitinib alleviated many features of MPN in our model. Our data indicate a reduction of STAT5 activity, as is clinically attainable with ruxolitinib, may be sufficient to alleviate disease. Activation of JAK-STAT and RAS signaling are both common features of myeloid leukemias.30,55,56 Previous work has demonstrated that MEK inhibition can attenuate myeloid neoplasia but is insufficient to cure this disease in either mouse models or human patients.23,56-58 KrasG12D myeloid cells remain hypersensitive to cytokines (according to colonyforming assays) despite MEK inhibition.23 Jak inhibition abrogates GM-CSF-dependent ERK phosphorylation in KrasG12D myeloid cells.59 These results indicate that other pathways contribute to disease in leukemias with hyperactive RAS signaling. Our work suggests activation of the STAT5 pathway may provide these critical signals in leukemia. Two recent reports from independent groups show that activated NRAS directs self-renewal in hematopoietic60 and leukemia stem cells.61 Gene expression analyses by both of these groups revealed that oncogenic NRAS led to activation of Stat5-mediated gene transcription and confirmed a relationship between RAS and STAT5 activity. Early T-precursor acute lymphoblastic leukemia (ETP ALL) is a treatment-resistant, fatal leukemia with a mutational and gene expression signature comparable to that of poor-risk acute myeloid leukemia.62-65 Like acute myeloid leukemia, most ETP ALL harbor mutations that activate RAS and RAS pathway components.63,65,66 JAK1 and JAK3 are also commonly mutated in ETP ALL.65,66 Recently, activated STAT5 (phospho-STAT5) levels were found to be elevated in all ETP-ALL cases tested. This elevation was not related to JAK mutational status but to surface levels of interleukin-7 receptor (a receptor known to activate RAS, JAK/STAT, and PI3K signaling67,68). Intriguingly, ruxolitinib treatment of ETP-ALL xenografts led to profound reduction in disease, independently of JAK mutational status.67 Analogous to our work, this study also demonstrated the efficacy of STAT5 inhibition in leukemia with hyperactive RAS signaling. Despite the considerable strides in the development of targeted therapies for treating myeloid neoplasms, there are no chemotherapy or targeted treatment options that have been shown to improve outcomes in JMML. As >90% of JMML patients exhibit activated RAS signaling, our data suggest that combination therapy with RAS-pathway inhibitors and ruxolitinib may be an effective, rational therapeutic strategy in this disease. MEK inhibition in mice with Nf1-deficient or Ras-activated MPN led to improvements in disease parameters but failed to eradicate leukemia cells.24,40 Likewise, MEK inhibition in early phase clinical trials in acute myelogenous leukemia has yielded largely disappointing results.57 In contrast, combined MEK and JAK/STAT inhibition in an NRASG12D/G12D model of MPN significantly improved survival of these mice.40 Our JMML patient harbors a KRAS mutation, yet ruxolitinib was more effective than MEK inhibition at controlling colony formation of the patientâ&#x20AC;&#x2122;s cells. These data provide a rationale for 1197
Z. Sachs et al.
clinical trials combining ruxolitinib with RAS-pathway inhibitors to control activated RAS myeloid neoplasia. Acknowledgments The authors would like to thank and acknowledge the Comparative Pathology, Flow Cytometry and Biostatistics and Informatics Shared Resources of the Masonic Cancer Center and the Mass Cytometry Shared Resource (which is supported by the Office of the Vice President for Research) at the University of Minnesota and Michael Franklin, a scientific writing editor supported by the Division of Hematology, Oncology, and Transplantation, Department of Medicine, University of Minnesota, for his editorial assistance.
References 1. Dvorak CC, Loh ML. Juvenile myelomonocytic leukemia: molecular pathogenesis informs current approaches to therapy and hematopoietic cell transplantation. Front Pediatr. 2014;2(25):2296-2360. 2. Locatelli F, Niemeyer CM. How I treat juvenile myelomonocytic leukemia (JMML). Blood. 2015;125(7):1083-1090. 3. Stieglitz E, Taylor-Weiner AN, Chang TY, et al. The genomic landscape of juvenile myelomonocytic leukemia. Nat Genet. 2015;47(11):1326-1333. 4. Caye A, Strullu M, Guidez F, et al. Juvenile myelomonocytic leukemia displays mutations in components of the RAS pathway and the PRC2 network. Nat Genet. 2015;47(11):1334-1340. 5. Bollag G, Clapp DW, Shih S, et al. Loss of NF1 results in activation of the Ras signaling pathway and leads to aberrant growth in haematopoietic cells. Nat Genet. 1996;12(2):144-148. 6. Amir el AD, Davis KL, Tadmor MD, et al. viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia. Nat Biotechnol. 2013;31(6):545-552. 7. Braun BS, Tuveson DA, Kong N, et al. Somatic activation of oncogenic Kras in hematopoietic cells initiates a rapidly fatal myeloproliferative disorder. Proc Natl Acad Sci USA. 2004;101(2):597-602. 8. Chan IT, Gilliland DG. Oncogenic K-ras in mouse models of myeloproliferative disease and acute myeloid leukemia. Cell Cycle. 2004;3(5):536-537. 9. Largaespada DA, Brannan CI, Jenkins NA, Copeland NG. Nf1 deficiency causes Rasmediated granulocyte/macrophage colony stimulating factor hypersensitivity and chronic myeloid leukaemia. Nat Genet. 1996;12(2):137-143. 10. Le DT, Kong N, Zhu Y, et al. Somatic inactivation of Nf1 in hematopoietic cells results in a progressive myeloproliferative disorder. Blood. 2004;103(11):4243-4250. 11. Li Q, Haigis KM, McDaniel A, et al. Hematopoiesis and leukemogenesis in mice expressing oncogenic NrasG12D from the endogenous locus. Blood. 2011;117(6):20222032. 12. Wang J, Liu Y, Li Z, et al. Endogenous oncogenic Nras mutation initiates hematopoietic malignancies in a dose- and cell type-dependent manner. Blood. 2011;118(2):368-379.
1198
Funding This work was supported in part by gifts made to the Minnesota Medical Foundation for the Matthew Kyle Nordos Pecha Memorial Fund and Dylan's Wish Memorial Fund. This work was partially funded by a Leukemia and Lymphoma Society of America Specialized Center of Research Grant (LLS 7019-04) and National Cancer Institute (U01 CA84221) grant to DAL. This was supported in part by an NIH Heart, Lung, and Blood Institute training grant (T32HL007062) (ZS), an NIH/NCATS training grant (ULI RR033183 & KL2 RR0333182) (ZS), and funds from the Division of Hematology, Oncology, and Transplantation, Department of Medicine, University of Minnesota (ZS).
13. Wang J, Liu Y, Tan LX, et al. Distinct requirements of hematopoietic stem cell activity and Nras G12D signaling in different cell types during leukemogenesis. Cell Cycle. 2011;10(17):2836-2839. 14. Zhang YY, Vik TA, Ryder JW, et al. Nf1 regulates hematopoietic progenitor cell growth and ras signaling in response to multiple cytokines. J Exp Med. 1998;187(11):18931902. 15. Liu YL, Yan Y, Webster C, et al. Timing of the loss of Pten protein determines disease severity in a mouse model of myeloid malignancy. Blood. 2016;127(15):1912-1922. 16. Kim A, Morgan K, Hasz DE, et al. Beta common receptor inactivation attenuates myeloproliferative disease in Nf1 mutant mice. Blood. 2007;109(4):1687-1691. 17. Mahgoub N, Taylor BR, Gratiot M, et al. In vitro and in vivo effects of a farnesyltransferase inhibitor on Nf1-deficient hematopoietic cells. Blood. 1999;94(7): 2469-2476. 18. Gandre-Babbe S, Paluru P, Aribeana C, et al. Patient-derived induced pluripotent stem cells recapitulate hematopoietic abnormalities of juvenile myelomonocytic leukemia. Blood. 2013;121(24):4925-4929. 19. Emanuel PD, Bates LJ, Castleberry RP, Gualtieri RJ, Zuckerman KS. Selective hypersensitivity to granulocyte-macrophage colony-stimulating factor by juvenile chronic myeloid leukemia hematopoietic progenitors. Blood. 1991;77(5):925-929. 20. Zhang J, Ranheim EA, Du J, et al. Deficiency of beta common receptor moderately attenuates the progression of myeloproliferative neoplasm in NrasG12D/+ mice. J Biol Chem. 2015;290 (31):19093-19103. 21. Guthridge MA, Stomski FC, Thomas D, et al. Mechanism of activation of the GM-CSF, IL-3, and IL-5 family of receptors. Stem Cells. 1998;16(5):301-313. 22. Okuda K, Foster R, Griffin JD. Signaling domains of the beta c chain of the GMCSF/IL-3/IL-5 receptor. Ann N Y Acad Sci. 1999;872:305-312; discussion 312-313. 23. Lyubynska N, Gorman MF, Lauchle JO, et al. A MEK inhibitor abrogates myeloproliferative disease in Kras mutant mice. Sci Transl Med. 2011;3(76):76ra27. 24. Chang T, Krisman K, Theobald EH, et al. Sustained MEK inhibition abrogates myeloproliferative disease in Nf1 mutant mice. J Clin Invest. 2013;123(1):335-339. 25. Grimley PM, Fang H, Rui H, et al. Prolonged STAT1 activation related to the growth arrest of malignant lymphoma cells
26.
27.
28.
29.
30. 31.
32.
33.
34.
35.
36.
37.
38.
by interferon-alpha. Blood. 1998;91(8): 3017-3027. Hennighausen L, Robinson GW. Interpretation of cytokine signaling through the transcription factors STAT5A and STAT5B. Genes Dev. 2008;22(6):711-721. Paukku K, Silvennoinen O. STATs as critical mediators of signal transduction and transcription: lessons learned from STAT5. Cytokine Growth Factor Rev. 2004;15(6): 435-455. Sternberg DW, Gilliland DG. The role of signal transducer and activator of transcription factors in leukemogenesis. J Clin Oncol. 2004;22(2):361-371. Kotecha N, Flores NJ, Irish JM, et al. Singlecell profiling identifies aberrant STAT5 activation in myeloid malignancies with specific clinical and biologic correlates. Cancer Cell. 2008;14(4):335-343. Muxi PJ, Oliver AC. Jak-2 positive myeloproliferative neoplasms. Curr Treat Options Oncol. 2014;15(2):147-156. Harrison C, Kiladjian JJ, Al-Ali HK, et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med. 2012;366(9):787-798. Harrison C, Vannucchi AM. Ruxolitinib: a potent and selective Janus kinase 1 and 2 inhibitor in patients with myelofibrosis. An update for clinicians. Ther Adv Hematol. 2012;3(6):341-354. Vannucchi AM, Kiladjian JJ, Griesshammer M, et al. Ruxolitinib versus standard therapy for the treatment of polycythemia vera. N Engl J Med. 2015;372(5):426-435. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med. 2012;366(9):799-807. Quintas-Cardama A, Verstovsek S. Molecular pathways: Jak/STAT pathway: mutations, inhibitors, and resistance. Clin Cancer Res. 2013;19(8):1933-1940. Padron E, Painter JS, Kunigal S, et al. GMCSF-dependent pSTAT5 sensitivity is a feature with therapeutic potential in chronic myelomonocytic leukemia. Blood. 2013;121 (25):5068-5077. Walz C, Ahmed W, Lazarides K, et al. Essential role for Stat5a/b in myeloproliferative neoplasms induced by BCR-ABL1 and JAK2(V617F) in mice. Blood. 2012;119(15): 3550-3560. Yan D, Hutchison RE, Mohi G. Critical requirement for Stat5 in a mouse model of polycythemia vera. Blood. 2012;119(15): 3539-3549.
haematologica | 2016; 101(10)
Stat5 is critical for Nf1-initiated MPN
39. Wang J, Liu Y, Li Z, et al. Endogenous oncogenic Nras mutation promotes aberrant GM-CSF signaling in granulocytic/monocytic precursors in a murine model of chronic myelomonocytic leukemia. Blood. 2010;116 (26):5991-6002. 40. Kong G, Wunderlich M, Yang D, et al. Combined MEK and JAK inhibition abrogates murine myeloproliferative neoplasm. J Clin Invest. 2014;124(6):2762-2773. 41. Cain JA, Xiang Z, O'Neal J, et al. Myeloproliferative disease induced by TELPDGFRB displays dynamic range sensitivity to Stat5 gene dosage. Blood. 2007;109 (9):3906-3914. 42. Li G, Wang Z, Zhang Y, et al. STAT5 requires the N-domain to maintain hematopoietic stem cell repopulating function and appropriate lymphoid-myeloid lineage output. Exp Hematol. 2007;35(11):1684-1694. 43. Cui Y, Riedlinger G, Miyoshi K, et al. Inactivation of Stat5 in mouse mammary epithelium during pregnancy reveals distinct functions in cell proliferation, survival, and differentiation. Mol Cell Biol. 2004;24(18): 8037-8047. 44. Socolovsky M, Nam H, Fleming MD, Haase VH, Brugnara C, Lodish HF. Ineffective erythropoiesis in Stat5a(-/-)5b(-/-) mice due to decreased survival of early erythroblasts. Blood. 2001;98(12):3261-3273. 45. Teglund S, McKay C, Schuetz E, et al. Stat5a and Stat5b proteins have essential and nonessential, or redundant, roles in cytokine responses. Cell. 1998;93(5):841-850. 46. Brannan CI, Perkins AS, Vogel KS, et al. Targeted disruption of the neurofibromatosis type-1 gene leads to developmental abnormalities in heart and various neural crest-derived tissues. Genes Dev. 1994;8(9): 1019-1029. 47. Zhu Y, Romero MI, Ghosh P, et al. Ablation of NF1 function in neurons induces abnormal development of cerebral cortex and reactive gliosis in the brain. Genes Dev. 2001;15(7):859-876. 48. Irish JM, Hovland R, Krutzik PO, et al. Single cell profiling of potentiated phospho-protein
haematologica | 2016; 101(10)
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
networks in cancer cells. Cell. 2004;118(2): 217-228. Krutzik PO, Irish JM, Nolan GP, Perez OD. Analysis of protein phosphorylation and cellular signaling events by flow cytometry: techniques and clinical applications. Clin Immunol. 2004;110(3):206-221. Van Meter ME, Diaz-Flores E, Archard JA, et al. K-RasG12D expression induces hyperproliferation and aberrant signaling in primary hematopoietic stem/progenitor cells. Blood. 2007;109(9):3945-3952. Kogan SC, Ward JM, Anver MR, et al. Bethesda proposals for classification of nonlymphoid hematopoietic neoplasms in mice. Blood. 2002;100(1):238-245. Quintas-Cardama A, Vaddi K, Liu P, et al. Preclinical characterization of the selective JAK1/2 inhibitor INCB018424: therapeutic implications for the treatment of myeloproliferative neoplasms. Blood. 2010;115(15): 3109-3117. Bhagwat N, Koppikar P, Keller M, et al. Improved targeting of JAK2 leads to increased therapeutic efficacy in myeloproliferative neoplasms. Blood. 2014;123(13): 2075-2083. Verstovsek S, Passamonti F, Rambaldi A, et al. A phase 2 study of ruxolitinib, an oral JAK1 and JAK2 inhibitor, in patients with advanced polycythemia vera who are refractory or intolerant to hydroxyurea. Cancer. 2014;120(4):513-520. Spiekermann K, Biethahn S, Wilde S, Hiddemann W, Alves F. Constitutive activation of STAT transcription factors in acute myelogenous leukemia. Eur J Haematol. 2001;67(2):63-71. Ward AF, Braun BS, Shannon KM. Targeting oncogenic Ras signaling in hematologic malignancies. Blood. 2012;120(17):33973406. Jain N, Curran E, Iyengar NM, et al. Phase II study of the oral MEK inhibitor selumetinib in advanced acute myelogenous leukemia: a University of Chicago phase II consortium trial. Clin Cancer Res. 2014;20(2):490-498. Lauchle JO, Kim D, Le DT, et al. Response
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
and resistance to MEK inhibition in leukaemias initiated by hyperactive Ras. Nature. 2009;461(7262):411-414. Diaz-Flores E, Goldschmidt H, Depeille P, et al. PLC-gamma and PI3K link cytokines to ERK activation in hematopoietic cells with normal and oncogenic Kras. Sci Signal. 2013;6(304):ra105. Li Q, Bohin N, Wen T, et al. Oncogenic Nras has bimodal effects on stem cells that sustainably increase competitiveness. Nature. 2013;504(7478):143-147. Sachs Z, LaRue RS, Nguyen HT, et al. NRASG12V oncogene facilitates self-renewal in a murine model of acute myelogenous leukemia. Blood. 2014;124(22):3274-3283. Coustan-Smith E, Mullighan CG, Onciu M, et al. Early T-cell precursor leukaemia: a subtype of very high-risk acute lymphoblastic leukaemia. Lancet Oncol. 2009;10(2):147156. Neumann M, Coskun E, Fransecky L, et al. FLT3 mutations in early T-cell precursor ALL characterize a stem cell like leukemia and imply the clinical use of tyrosine kinase inhibitors. PloS One. 2013;8(1):e53190. Neumann M, Heesch S, Gokbuget N, et al. Clinical and molecular characterization of early T-cell precursor leukemia: a high-risk subgroup in adult T-ALL with a high frequency of FLT3 mutations. Blood Cancer J. 2012;2(1):e55. Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature. 2012;481 (7380):157-163. Neumann M, Heesch S, Schlee C, et al. Whole-exome sequencing in adult ETP-ALL reveals a high rate of DNMT3A mutations. Blood. 2013;121(23):4749-4752. Maude SL, Dolai S, Delgado-Martin C, et al. Efficacy of JAK/STAT pathway inhibition in murine xenograft models of early T-cell precursor (ETP) acute lymphoblastic leukemia. Blood. 2015;125(11):1759-1767. Reth M, Nielsen P. Signaling circuits in early B-cell development. Adv Immunol. 2014; 122:129-175.
1199
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Chronic Myeloid Leukemia
Ferrata Storti Foundation
Haematologica 2016 Volume 101(10):1200-1207
Nilotinib 300 mg twice daily: an academic single-arm study of newly diagnosed chronic phase chronic myeloid leukemia patients
Fausto Castagnetti,1 Massimo Breccia, 2 Gabriele Gugliotta, 1 Bruno Martino,3 Mariella D’Adda,4 Fabio Stagno,5 Angelo Michele Carella,6 Paolo Avanzini,7 Mario Tiribelli,8 Elena Trabacchi,9 Giuseppe Visani,10 Marco Gobbi,11 Marzia Salvucci,12 Luciano Levato,13 Gianni Binotto,14 Silvana Franca Capalbo,15 Maria Teresa Bochicchio,1 Simona Soverini,1 Michele Cavo,1 Giovanni Martinelli,1 Giuliana Alimena,2 Fabrizio Pane,16 Giuseppe Saglio,17 Gianantonio Rosti,1 and Michele Baccarani18 on behalf of the GIMEMA CML Working Party
Institute of Hematology “L. and A. Seràgnoli”, Department of Experimental, Diagnostic and Specialty Medicine, “S. Orsola-Malpighi” University Hospital, University of Bologna; 2 Department of Cellular Biotechnologies and Hematology, "Sapienza" University of Rome; 3Hematology Unit, Azienda Ospedaliera “Bianchi-Melacrino-Morelli”, Reggio Calabria; 4Hematology Unit, Azienda Ospedaliera “Spedali Civili”, Brescia; 5Chair of Hematology, University of Catania; 6U.O. Ematologia I, IRCCS AOU S. Martino-IST, Genova; 7Hematology Unit, Arcispedale Santa Maria Nuova, IRCCS, Reggio Emilia; 8 Division of Hematology and BMT, Department of Experimental and Clinical Medical Sciences, Azienda Ospedaliero-Universitaria di Udine; 9Hematology and Bone Marrow Transplantation Unit, Department of Hematology and Oncology, “G. da Saliceto” Hospital, Piacenza; 10Hematology and Stem Cell Transplantation Unit, Azienda Ospedaliera Ospedali Riuniti Marche Nord (AORMN), Pesaro; 11Clinical Hematology Unit, IRCCS AOU S. Martino-IST, Genova; 12Hematology Unit, "Santa Maria delle Croci" Hospital, Ravenna; 13Hematology Unit, "Pugliese-Ciaccio” Hospital, Catanzaro; 14 Hematology and Clinical Immunology Unit, University of Padova, Padova; 15Hematology Unit, Azienda Ospedaliero-Universitaria Ospedali Riuniti, Foggia; 16Chair of Hematology, Department of Biochemistry and Medical Biotechnologies, “Federico II” University, Napoli; 17Chair of Hematology, Department of Clinical and Biological Sciences, “S. Luigi Gonzaga” University Hospital, University of Torino, Orbassano (TO); 18Department of Hematology and Oncology “L. and A. Seràgnoli", University of Bologna, Italy 1
Correspondence: fausto.castagnetti@unibo.it
ABSTRACT Received: February 23, 2016. Accepted: July 28, 2016. Pre-published: July 28, 2016. doi:10.3324/haematol.2016.144949
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1200
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1200
T
he introduction and the extended clinical use of nilotinib in the first-line treatment of chronic myeloid leukemia have been based on company-sponsored trials. Independent confirmations are extremely important. We report an investigator-sponsored study of nilotinib 300 mg twice daily in 130 chronic myeloid leukemia patients in early chronic phase. A deep molecular response was achieved in 46% (MR4.0) and 17% (MR4.5) of patients at 2 years; 58% of the enrolled patients achieved a MR4.0 at least once, with a sustained MR4.0 in 52% of them. With a median observation of 29 months (range 24-37 months), 77% of patients were still on treatment with nilotinib. The reasons for permanent discontinuation were: 3% progression, 5% failure or suboptimal response, 8% adverse events, 1% treatment-free remission, and 5% other reasons. Thirteen thrombotic arterial events were reported in 12 patients. A prospective evaluation of metabolic effects showed an increase of fasting glucose without significant variations of glycated hemoglobin, an increase of total cholesterol (both low density lipoprotein and high density lipoprotein fractions) and a decrease of triglycerides. This study confirms a high and rapid efficacy of nilotinib 300 mg twice daily and provides detailed information on the type and incidence of non-hematologic and metabolic adverse events (clinicaltrials.gov identifier:01535391). haematologica | 2016; 101(10)
Efficacy and safety of frontline nilotinib in CML
Introduction
Table 1. Patient characteristics at diagnosis.
Patients, N Nilotinib is a second generation BCR-ABL1 tyrosine kinase inhibitor (TKI).1 It has been approved for the first-line treatment of newly diagnosed, chronic phase (CP) Philadelphia chromosome-positive (Ph+), BCR-ABL1-positive (BCR-ABL1+) chronic myeloid leukemia (CML), following the report of a company-sponsored phase 3 prospective randomized trial (ENESTnd) comparing nilotinib to imatinib 400 mg once daily (OD).2 Several updates of the study, over 6 years, have confirmed the initial findings that nilotinib was superior to imatinib for any degree of molecular response, and for the rapidity of the response.3,4 The progression-free survival (PFS) was reported to be marginally improved and no difference in overall survival (OS) was detectable.3,4 In the ENESTnd trial, two different nilotinib doses were tested, namely 300 mg twice daily (TD) and 400 mg TD.2-4 The 300 mg TD dose was selected for approval because it was reported to be as effective as, but less toxic than, the 400 mg TD dose. A relevant issue was the cardiovascular toxicity, with focus on arterial occlusive events (ischemic heart disease (IHD), peripheral arterial occlusive disease (PAOD) and ischemic cerebrovascular events, ICVEs), that at 5 years were reported in 2.1% of patients in the imatinib arm, in 6.8% of patients in the nilotinib 300 mg TD arm, and in 12.6% of patients in the nilotinib 400 mg TD arm.3 This concern was raised and reinforced by other independent studies, for the most part retrospective and mainly in second-line treatment, reporting a significant incidence of cardio-vascular adverse events (CVAEs) during nilotinib treatment.5-9 With a minimum observation of 24 months, the molecular response rates in a second single-arm company-sponsored study of nilotinib 300 mg TD, the ENEST1st trial, were even higher compared to the ENESTnd results, with consistent safety data.10 There are no independent, investigator-sponsored studies of the drug in first-line treatment, with the exception of the two small pilot studies that were performed prior to the approval of nilotinib in firstline treatment, at the dose registered for second-line treatment (400 mg TD).11-14 When the 300 mg TD dose became the standard in first-line therapy, the GIMEMA CML Working Party designed a prospective phase 3b single-arm trial of nilotinib 300 mg TD, to independently assess the efficacy and safety. Since all patients have now been followed for a minimum of two years, we herein report the results of the main analysis of response and adverse events (AEs).
Methods A phase 3b single-arm study of nilotinib, 300 mg TD was conducted in adult patients with newly diagnosed CP BCR-ABL1+ CML (clinicaltrials.gov identifier:01535391). A dose increase to 400 mg TD was scheduled for suboptimal response or failure (2009 European LeukemiaNet, ELN, criteria),15 excluding disease progression, in the absence of toxicity or relevant BCR-ABL1 mutations. Pre-treatment with imatinib for up to 30 days was permitted. The primary endpoint was the rate of MR4.0 at 24 months. The secondary endpoints are detailed in the Online Supplementary Table S1. The study was reviewed and approved by the Internal Review Board of all the participating institutions. The cut-off date for this analysis was June 30, 2015. The detection of a Ph chromosome and/or a BCR-ABL1 fusion haematologica | 2016; 101(10)
Age, years; median (range) Age ≥ 65 years, N (%) Sex, male, N (%) ECOG ≥ 1, N (%) Hb level, g/dL; median (range) Platelet count, 103/ L; median (range) WBC count, 103/ L; median (range) Blast cells in PB, %; median (range) Eosinophils in PB, %; median (range) Basophils in PB, %; median (range) Spleen, cm; median (range) Palpable spleen, N (%) Sokal score, N (%): Low Intermediate High Euro score, N (%): Low Intermediate High EUTOS score, N (%): Low High CCA/Ph+, present, N (%) Variant translocations, present, N (%) Pre-treatment, N (%): Imatinib Hydroxyurea Anagrelide
130 50 (18 - 85) 25 (19) 84 (65) 21 (16) 12.0 (7.6 - 16.3) 396 (101 - 4093) 41.5 (1.8 – 363.2) 1.0 (0 – 14.0) 2.0 (0 – 10.0) 2.2 (0 – 13.0) 1 (0 - 26) 67 (52) 56 (43) 47 (36) 27 (21) 61 (47) 62 (48) 7 (5) 120 (92) 10 (8) 4 (3) 8 (6) 7 (5) 69 (53) 1 (1)
ECOG: performance status according to the Eastern Co-operative Oncology Group grading; Hb: hemoglobin; WBC: white blood cells; PB: peripheral blood; EUTOS: European Treatment and Outcome Study; CCA/Ph+ : clonal chromosome abnormalities in Ph-positive cells.
gene associated with consistent morphologic features were required to confirm the CML diagnosis and the chronic (CP), accelerated (AP) or blast disease phase (BP) were defined according to current ELN criteria.16 Risk scores were calculated according to Sokal,17 Euro,18 and EUTOS19 formulations. The molecular response (MR) was assessed by peripheral blood RT-PCR, according to the International Scale (IS).20 Definitions: early molecular response (EMR), BCR-ABL1 transcript ≤ 10% at 3 months; major molecular response (MMR or MR3.0), BCR-ABL1 transcript ≤ 0.1%; MR4.0 and MR4.5, BCR-ABL1 transcript ≤ 0.01%, and ≤ 0.0032%, respectively, in samples with > 10,000, and > 32,000 ABL1 copies, respectively;21,22 sustained MR4.0 or MR4.5, stable response for > 1 year with > 3 evaluable tests. Molecular tests were performed every 3 months. The cytogenetic response was assessed by chromosome banding analysis at 3, 6 and 12 months; if there were < 20 available metaphases, a fluorescence in situ hybridization (FISH) analysis on peripheral blood cells was accepted (complete cytogenetic response, CCyR, ≤1% of BCR-ABL1 positive nuclei, > 200 nuclei analyzed).16 Mutational analysis of BCRABL1 kinase domain point mutations (Sanger Sequencing) was performed in case of progression, failure or suboptimal response.15,23 OS, PFS, and event-free survival (EFS, or survival without treatment discontinuation) were calculated from treatment start until death (OS), until death or progression to AP or BP 1201
F. Castagnetti et al.
(PFS), or until death, progression to AP or BP, failure on nilotinib or nilotinib treatment discontinuation for any cause, except treatment-free remission (EFS), respectively. Probabilities of OS, PFS and EFS were calculated using the Kaplan-Meier method.24 Patients who discontinued nilotinib and patients who underwent allogeneic stem cell transplantation were not censored for OS and PFS. The rates of molecular and cytogenetic response "at" milestones were calculated by dividing the number of responders at that timepoint by the number of all enrolled patients. The time to response was calculated from treatment start until the first achievement of the response. The cumulative probability of response was calculated taking into consideration the presence of competing risks (failure, progression or death).25,26 The AEs were graded according to the NCI Common Terminology Criteria for Adverse Events version 4.0 (CTCAE v4.0). Lipid modifications were graded according to adapted American Association of Clinical Endocrinologists (AACE) criteria,27 and glucose abnormalities according to adapted American Diabetes Association (ADA) criteria.28
intermediate Sokal score, developed a lymphoid blast crisis with T315I mutation and was treated with conventional chemotherapy followed by allogeneic stem cell transplantation (alloSCT). A second patient, a 46 year old with an intermediate Sokal score, progressed to AP with V280A mutation and was treated with dasatinib. The third of these patients, a 32 year old with a low Sokal score, progressed to AP with clonal evolution and myelodysplastic features, without BCR-ABL1 mutations, and was submitted to alloSCT. The fourth and final patient, a 43 year old with a low Sokal score, progressed to a myeloid blast phase without BCR-ABL1 mutations and is on second-line treatment with dasatinib (Online Supplementary Table S2). Eleven patients (8%) discontinued nilotinib for toxicity, one patient (1%) decided to discontinue the treatment because of stable deep molecular response, achieving a treatment-free remission (duration of nilotinib treatment and duration of stable MR4.0 before discontinuation, 35 and 23 months, respectively), and 5 patients (4%) discontinued for other reasons, including withdrawal of informed consent and pregnancy. Two patients (2%) were lost to follow-up. Overall, 3
Results A
Baseline characteristics One hundred and thirty patients were enrolled between December 2011 and November 2012 at 32 GIMEMA Clinical Centers. The baseline characteristics of the patients are shown in Table 1. The median age was 50 years. Sixtyfive percent were males. High-risk patients were 21% by Sokal, 5% by EURO, and 8% by EUTOS scores. Clonal chromosome abnormalities in Ph+ cells were reported in 3% of patients.
Patient disposition The median follow-up was 29 months (range 24-37 months). The patient disposition at the last contact is shown in Table 2. Overall, 77% of patients were still on nilotinib (80% at 2 years), mainly at the initial 300 mg TD dose. In 6 patients the nilotinib dose was escalated to 400 mg TD due to the absence of an optimal response: 4 patients are still on treatment, 2 of them with an optimal response and 2 without significant improvements, while 2 patients definitively discontinued nilotinib as a result of treatment failure. Twenty-eight patients permanently discontinued nilotinib, of whom 11 (8%) were due to failure. Four of these patients progressed to AP or BP after 1, 4, 6, and 19 months, respectively; all these patients were alive at last contact. One of these patients, a 41 year old with an
B
Table 2. Patient disposition at last contact.
Patients, N Still on nilotinib, N (%) 600 mg < 600 mg 800 mg Discontinued nilotinib, N (%) Progression to advanced phase Failure or suboptimal response Adverse events Treatment-free remission Other causesa Lost to follow-up, N (%)
130 100 (77%) 89 (68%) 7 (5%) 4 (3%) 28 (22%) 4 (3%) 7 (5%) 11 (8%) 1 (1%) 5 (4%) 2 (2%)
details of other causes: pregnancy, 2 patients, withdrawal of informed consent, 1 patient, low compliance, 1 patient, alcohol abuse, 1 patient.
a
1202
Figure 1. Cumulative incidence of deep molecular response. (A) MR4.0: the 2-year estimated cumulative incidence of MR4.0 was 51% (95% CI, 42-60%), B) MR4.5: the 2-year estimated cumulative incidence of MR4.5 was 24% (95% CI, 1631%). The cumulative probability of achieving a response was calculated under consideration of competing risks, defined as failure followed by treatment change, progression or death.
haematologica | 2016; 101(10)
Efficacy and safety of frontline nilotinib in CML
patients underwent alloSCT, 2 after transformation to AP (previously described) and 1 after resistance to all TKIs, including ponatinib, without progression and without detectable BCR-ABL1 mutations.
Responses and outcome According to the intention-to-treat (ITT) principle, 70% of patients were in CCyR at 6 months, and 77% at 12 months. The 3-month EMR was achieved in 80% of patients (ITT). The rates of MMR, MR4.0, and MR4.5 at 3, 6, 12, 18, and 24 months are shown in Table 3. At 24 months, according to the ITT principle, 65% of patients were in MMR, 46% were in MR4.0, and 17% were in MR4.5. Overall, 75/130 patients (58%) achieved a MR4.0 at least once and 39/75 (52%) achieved a sustained MR4.0; 40/130 patients achieved a MR4.5 and 11/40 (28%) a sustained MR4.5. The cumulative probability of achieving MR4.0 and MR4.5 is shown in Figure 1; after two years of treatment the probabilities were 51% (95% CI, 42-60%) and 24% (95% CI, 1631%), respectively. The duration of observation is still too short for a detailed analysis of the stability of deep molecular response. Only 1 patient died, due to the worsening of a pre-existing comorbidity (chronic obstructive pulmonary disease: a worsening of general clinical conditions, followed by a switch to imatinib with a stable response, was observed). As shown in Figure 2, the 30-month OS was 99% (95% CI 91-100%), the 30-month PFS was 96% (95% CI 89-98%) and the 30-month EFS (survival without treatment discontinuation, except for treatment-free remission) was 76% (95% CI 67-83%).
patients. Twenty-five AEs listed under a comprehensive definition of cardiovascular AEs (CVAEs) were reported in 21 patients (16%): arterial thrombosis, 13 events (Table 4); venous thrombosis, 4 events (1 deep venous thrombosis, 3 superficial thrombophlebitis); QTc prolongation, 3 events (only 1 case > 500 msec); arrhythmias, 3 events (2 atrial fibrillations, 1 atrial-ventricular blockade); congestive heart failure, 1 event; acute pericarditis, 1 event. They were reported as a grade 3 event in nine cases, and as a grade 4 event in one case. The treatment was temporarily discontinued for 9 events in 8 patients (36% of CVAEs, 6% of all patients), and it was permanently discontinued for 8 events (32% of CVAEs, 6% of all patients). A medical treatment without hospitalization was administered for 13 events; hospitalization was required for 8 events. Major surgery was performed in 2 patients: substitution of the femoral head and amputation of the right lower limb, respectively. One case of myocardial infarction was treated with coronary stents. No patient died of toxicity. The characteristics of patients with (N = 12) or without (N = 118) arterial thrombotic events are shown in Table 5. The patients with high or very high cardiovascular risk according to the European Guidelines on cardiovascular disease prevention in clinical practice,29 retrospectively assessed, had a significantly higher probability of arterial thrombotic events (Online Supplementary Figure S1). Among the other nonhematologic and non-cardiovascular AEs, only fatigue (17%
Safety A grade 3-4 thrombocytopenia and neutropenia were recorded in 8% and 9% of patients. Hematologic AEs caused early permanent treatment discontinuation in 2 Table 3. Molecular response at milestones.
BCR-ABL1IS
≤ 0.1%
≤ 0.01%
≤ 0.0032%
3 months 6 months 12 months 18 months 24 months
23% 53% 57% 63% 65%
2% 12% 28% 31% 46%
0 2% 7% 11% 17%
‘BCR-ABL1IS’: BCR-ABL/ABL% ratio, according to the International Scale. The response rates were calculated by dividing the number of patients with that response at that time by the total number of enrolled patients (N=130).
Figure 2. Outcome of all the 130 enrolled patients. The estimated 30-month overall survival was 99% (95% CI, 91-100%), the 30-month estimated progression-free survival was 96% (95% CI, 89-98%), the estimated 30-month survival without treatment discontinuation was 76% (95% CI, 67-83%).
Table 4. List of arterial thrombotic adverse events by year, all grades, irrespective of causality.
Type of event
Lower limbsa Coronaryb Carotidc Optic Avascular necrosis of the femoral head Total
1st year*
2nd year*
3rd year*
Total
Permanent treatment discontinuation
Temporary treatment discontinuation
1 2 1 0 1 5
1 2 0 1 1 5
1 2 0 0 0 3
3 6 1 1 2 13+
1 4 0 0 1 6
1 1 1 0 0 3
*Duration of nilotinib treatment: ≤12 months, 130 patients; 13-24 months; 115 patients; > 24 months, 67 patients; a Amputation, 1 case; b Myocardial infarction, 1 case, stable or unstable angina, 4 cases, troponin increase after atrial fibrillation, 1 case; c Transient ischemic attack, 1 case; + The 13 events were observed in 12 patients (one patient had two events): the number of patients remaining on nilotinib was 6 (three patients assuming 600 mg daily, three patients assuming 300 mg daily).
haematologica | 2016; 101(10)
1203
F. Castagnetti et al. Table 5. Characteristics of patients with and without arterial thrombotic adverse events.
Variables At baseline Age, years, median (range) Sex, male, N (%) Body mass index, N (%) < 25 25-29.9 30-34.9 ≥ 35 ECOG ≥ 1, N (%) Smokers, N (%) Arterial hypertension, N (%) Diabetes, N (%) HbA1c, %, median (range) Total cholesterol, mg/dL, median (range) LDL, mg/dL, median (range) HDL, mg/dL, median (range) High or very high EURO CV score, N (%) On treatmenta HbA1c, %, median (range) Total cholesterol, mg/dL, median (range) LDL, mg/dL, median (range) HDL, mg/dL, median (range)
Patients WITH events N = 12
Patients WITHOUT events N = 118
P
62 (49-85) 7 (58)
47 (18-81) 77 (65)
0.002 0.753
6 (50) 3 (25) 2 (17) 1 (8) 4 (33) 4 (33) 4 (33) 3 (25) 6.2 (4.7-7.0) 170 (126-315) 111 (57-211) 36 (15-75) 9 (75)
69 (58) 35 (30) 10 (8) 4 (3) 17 (14) 22 (19) 24 (20) 9 (8) 5.9 (3.6-8.7) 168 (81-260) 103 (36-186) 36 (11-79) 28 (24)
0.162
0.104 0.257 0.287 0.082 0.809 0.425 0.692 0.435 < 0.001
6.4 (5.2-8.2)b 206 (140-272) 128 (55-183) 49 (29-90)
5.6 (3.8-8.7)c 210 (107-328) 133 (42-215) 55 (10-103)
0.023 0.754 0.598 0.414
ECOG: performance status according to the Eastern Cooperative Oncology Group grading; HbA1c: glycated hemoglobin; EURO CV score: cardiovascular risk according to the European Guidelines on cardiovascular disease prevention in clinical practice, version 2012 (31). LDL: low density lipoprotein; HDL: high density lipoprotein. aData at 6 months (the results do not change considering other milestones, e.g. 3 months or 12 months); b66% of patients with HbA1c > 6.5% at 6 months were diabetic at baseline; c50% of patients with HbA1c > 6.5% at 6 months were diabetic at baseline.
grade 1-2 and 1% grade 3), bone and muscle and joint pain (22% grade 1-2 and 1% grade 3), and skin rash (29% grade 1-2 and 1% grade 3) were reported in more than 10% of patients. Grade 3-4 laboratory abnormalities were as follows: grade 3 transaminase increase 2% (permanent treatment discontinuation, 1 case), grade 3 bilirubin increase 5%, grade 3 and grade 4 lipase increase 12% and 2%, respectively, and grade 3 amylase increase 1% (no pancreatitis). According to adapted ADA criteria, and considering the maximum grade reached by each patient while on study, 47%, 11%, 5% and 6% of patients experienced a grade 1 (101-125 mg/dl), grade 2 (126-150 mg/dl), grade 3 (151-200 mg/dl) and grade 4 (>200 mg/dl) hyperglycemia, respectively; 29% of patients had an increase of fasting glucose at 1 year, compared to baseline levels (P<0.001). According to adapted ADA criteria, 47% of patients had a grade 1 (5.76.4%), 10% a grade 2 (6.5-6.9%), 3% a grade 3 (7-7.9%) and 5% a grade 4 (≥8%) glycated hemoglobin (HbA1c), respectively, (maximum grade reached on our study); no significant increase of glycated hemoglobin has been observed from baseline. According to the AACE criteria, 40 patients (31%) experienced a borderline (200-239 mg/dL) and 54 (42%) a high-risk (≥240 mg/dL) hypercholesterolemia (maximum grade reached on our study). A significant increase of cholesterol, both in LDL and HDL fractions (P<0.001), and a significant reduction of triglycerides (P<0.001) were detected at 1 year from baseline. The serum glucose, HbA1c and lipid concentrations at each milestone are shown in Figure 3 and Online Supplementary Table S3.
Discussion The introduction and the extended clinical use of second 1204
generation TKIs is becoming a very important issue in firstline therapy for CML.2-4,10,30-38 They offer a treatment choice that must be weighed for short- and long-term efficacy and toxicity, and for cost-efficacy.16,39,40 The information on nilotinib is limited to data coming from two company-sponsored studies, of which one was designed to compare nilotinib and imatinib, and the other to confirm the rate of deep molecular response on nilotinib.2-4,10 There are no data from independent, investigator-sponsored, studies. This GIMEMA trial was made possible thanks to the support of Novartis Oncology Italy, who provided the drug free of charge and a small unrestricted support. The study was designed, conducted and analyzed by GIMEMA, so that this is the first study providing company-independent data on the treatment of newly diagnosed CP CML patients with nilotinib 300 mg TD. The study has some strengths and some limitations. The strengths are the independence and the involvement of several clinical centers all over Italy. The major limitation is the short time of observation, with a minimum of 2 years, a median of 2.5 years, and a maximum of 3 years. The ENESTnd and the ENEST1st trials were reported with a minimum observation period of 6 and 2 years, respectively.4,10 Two major issues are important in the treatment of CML. One issue is the rapidity31,41-44 and the depth45 of the molecular response. We found that EMR was achieved less frequently than in both ENEST trials, but that MMR (MR3.0), MR4.0 and MR4.5 rates were in the same range of ENEST1st, and even higher than in ENESTnd (Online Supplementary Table S4).2-4,10 The significance of these differences cannot be assessed because the three studies are different, with different age distribution and proportion of high-risk (Sokal) patients (28% in ENESTnd, 18% in ENEST1st, 21% in this study), with different enrolment criteria and different guidelines for dose reduction or treatment discontinuation. haematologica | 2016; 101(10)
Efficacy and safety of frontline nilotinib in CML
It should be noted that the molecular response rates have been calculated in two ways and reported as response rates "at" and "by" different time points, where the rate "at" the time point gives the actual proportion of patients who are in that response at the given time point, and the rate "by" the same time point gives the cumulative proportion of patients achieving the response at least once before the given time point. The estimated cumulative incidence of response (response "by") helps to make a comparison between two treatments, as in the ENESTnd study,2-4 but overestimates the actual proportion (response "at") of the patients who are in that response at that time, which is important for clinical decisions. The response rates "at" each time point are frequently lower than the values "by" the same time points. The second major issue is the so-called cardiovascular toxicity, that includes different events, with different physiopathologic mechanisms and different clinical relevance: myocardial infarction, atrial fibrillation, superficial throm-
A
B
C
D
E
F
bophlebitis, arterial thrombosis, congestive heart failure, stroke, and many others. The incidence, the severity and the consequences of these complications are difficult to assess and to compare, because they may depend on different variables, including not only the patients characteristics, the baseline cardiovascular risk, the prior treatments and the drug dose,46-48 but also the definition of the events, and, importantly, the accuracy of the event reports, that depends on the retrospective or prospective nature of the data collection. When the GIMEMA study was designed, the cardiovascular toxicity was not yet pointed out, apart from QTc prolongation, and the cardiovascular risk at baseline was not routinely assessed. However, the cardiovascular toxicity was revealed as soon as patient enrolment began, so that the identification and the reporting of CVAEs, in facts, became prospective. However, monitoring, prophylaxis, and the treatment of CVAEs were left to local investigators, because it was not possible to provide guidelines. Several reports indicated a possible metabolic non-target effect of
Figure 3. Metabolic effects of nilotinib by time. (A) Distribution of fasting glucose at milestones; (B) Distribution of glycated hemoglobin at milestones; (C) Distribution of triglycerides at milestones; (D) Distribution of total cholesterol at milestones; (E) Distribution of low density lipoprotein at milestones; (F) Distribution of high density lipoprotein at milestones.
haematologica | 2016; 101(10)
1205
F. Castagnetti et al.
nilotinib, potentially related to CVAEs; importantly, in our study, fasting glucose, glycated hemoglobin and serum lipids were prospectively assessed. Moreover, to evaluate the clinical impact of metabolic effects, we decided to classify the abnormalities according to specific criteria, as recommended by ADA and AACE guidelines.27,28 With these caveats, it is possible to describe, but not to compare, the 2-year incidence of major CVAEs in the present trial and in the ENEST1st one:10 ischemic heart events were 4.6% and 3.4% respectively; arterial thrombotic events were 4.6% (including one case of retinal artery occlusion and two cases of avascular necrosis of the femoral head) and 1.9%, respectively; arrhythmias were 2.3% and 0.7%, respectively; congestive heart failure was reported in 0.8% and 0.3%, respectively. The incidence of QTc prolongation (2.3% in the GIMEMA study) and of venous thrombosis (3.1%) was not reported in the ENEST1st study. With a median follow-up of 29 months, the number of reported arterial thrombotic events was higher during the first and the second year, if compared to the third year of treatment, but considering the number of patients on nilotinib treatment, the yearly incidence was comparable (Table 4). A linear increase of the cumulative incidence of CVAEs over time was also described in the 5-year update of the ENESTnd trial; the yearly incidence of CVAE continues unabated after 5 years and may even increase.3 In our study the cumulative probability of developing a CVAE was 7 % (95% CI, 4-13%) at 1 year and 13 % (95% CI, 8-20%) at 2 years (data not shown). We confirmed a significant increase of fasting glucose level during nilotinib treatment, but we were not able to demonstrate any significant increase of glycated hemoglobin. Similar results were reported in the ENESTnd trial, in a sub-analysis including only patients with normal glycemic status at baseline:49 the majority of patients with hyperglycemia did not meet the glycated hemoglobin criteria for diabetes. The patients with arterial thrombotic events had a higher incidence of cardiovascular risk factors at baseline (in particular, a higher incidence of diabetes was observed), and were elderly (Table 5, Online Supplementary Figure S1). The exact pathogenesis of fasting glucose alterations during nilotinib is still controversial and probably related to insulin resistance: both in the ENESTnd study49 and in the ENIGMA 2 study50 an increase in insulin production and a decrease of fasting C-peptide were observed, with increased levels of HOMA-IR and HOMA-β values. In conclusion, this independent study highlights the therapeutic efficacy of nilotinib, confirming the rates, the velocity, and the depth of molecular response; moreover, it confirms that the risk of cardiovascular toxicity, including several different events, is higher in patients with high cardiovascular risk, requiring specific measures of prophylaxis and monitoring.
References 1. Weisberg E, Manley PW, Breitenstein W, et al. Characterization of AMN107, a selective inhibitor of native and mutant Bcr-Abl. Cancer Cell. 2005;7(2):129-141. 2. Saglio G, Kim DW, Issaragrisil S, et al. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J
1206
Acknowledgments The authors would like to thank the valuable assistance of Miriam Fogli and Michela Apolinari. The following members of the ‘GIMEMA Working Party on CML’ actively participated in this study, enrolling patients and collecting clinical data: Rupoli S, Scortechini AR (Hematology Department, University of Ancona, Azienda Ospedaliero Universitaria Ospedali Riuniti di Ancona, Ancona); Cantore N, Palmieri F (Hematology Division, Ospedale Civile ‘San Giuseppe Moscati’, Avellino); Luatti S, Testoni N (Institute of Hematology ‘Seràgnoli’, Department of Experimental, Diagnostic and Specialty Medicine, University of Bologna, Bologna); Rossi G, Giuliani G (Hematology Unit, Azienda Ospedaliera ‘Spedali Civili’, Brescia); Di Raimondo F, Vigneri P (Hematology Unit, ‘Ferrarotto’ Hospital, Catania); Molica S, Lentini M (Hematology Unit, ‘Pugliese’ Hospital, Catanzaro); Cuneo A, Cavazzini F (Chair of Hematology, Dipartimento di Scienze Mediche, ‘Arcispedale S Anna’ University Hospital, Ferrara); Spinosa G, Palumbo G (Hematology Unit, Azienda Ospedaliero-Universitaria Ospedali Riuniti, Foggia); Ibatici A, Beltrami G (Hematology I Unit, IRCCS AOU San Martino-IST, Genova); Pierri I, Bergamaschi M (Clinical Hematology Unit, IRCCS AOU San Martino-IST, Genova); Ciceri F, Lunghi F (Hematology Unit, “San Raffaele” University Hospital, Milano); Bassan R, Maino E (Hematology Unit, Ospedale dell’Angelo, Mestre, VE); Luppi M, Marasca R (Chair of Hematology, University of Modena and Reggio Emilia, Modena); Semenzato G, Cason G (Department of Internal Medicine, University of Padova, Padova); Fabbiano F, Turri D (Hematology Unit, ‘V Cervello’ Hospital, Palermo); Isidori A, Barulli S (Hematology Unit, ‘San Salvatore’ Hospital, Pesaro); Vallisa D, Arcari A (Hematology Division, ‘Guglielmo da Saliceto’ Hospital, Piacenza); Zaccaria A, Zuffa E (Hematology Unit, ‘Santa Maria delle Croci’ Hospital, Ravenna); Ronco F, Ielo D (Hematology Unit, Ospedali Riuniti, Reggio Calabria); Merli F, Capodanno I (Hematology Unit, Arcispedale Santa Maria Nuova, Reggio Emilia); Tosi P, Merli A (Hematology Unit, Ospedale Infermi Azienda Unità Sanitaria, Rimini); Musto P, Pietrantuono G (Hematology Unit, IRCCS Centro di Riferimento Oncologico della Basilicata, Rionero in Vulture, PZ); Latagliata R (Chair of Hematology, ‘La Sapienza’ University, Roma); De Fabritiis P, Trawiska M (Hematology Unit, ‘S. Eugenio’ Hospital, Roma); Amadori S, Cantonetti M (Department of Hematology, ‘Tor Vergata’ University, Roma); Ronci B, Cedrone M (Hematology Unit, Ente Ospedaliero San Giovanni Addolorata, Roma); Falcone AP, Sgherza N (Hematology Unit, Casa Sollievo della Sofferenza, San Giovanni Rotondo, FG); Bocchia M, Defina M (Chair of Hematology, University of Siena, Siena); Vitolo U, Pregno P (Hematology Unit, Azienda Ospedaliero Universitaria Città della Salute e della Scienza, University of Torino, Torino); Gherlinzoni F, Calistri E (Hematology Unit, ‘Ca' Foncello’ Hospital, Treviso); Fanin R, Medeot M (Chair of Hematology, University of Udine, Udine); Pizzolo G, Bonifacio M (Chair of Hematology, University of Verona, Verona, Italy).
Med. 2010;362(24):2251-2259. 3. Hochhaus A, Saglio G, Hughes TP, et al. Long-term benefits and risks of frontline nilotinib vs imatinib for chronic myeloid leukemia in chronic phase: 5-year update of the randomized ENESTnd trial. Leukemia. 2016;30(5):1044-1054. 4. Larson R, Kim D, Issaragrisil S, et al. Efficacy and Safety of Nilotinib (NIL) vs Imatinib (IM) in Patients (pts) With Newly Diagnosed
Chronic Myeloid Leukemia in Chronic Phase (CML-CP): Long-Term Follow-Up of ENESTnd [abstract]. Blood. 2014;124(21). Abstract 4541. 5. Aichberger KJ1, Herndlhofer S, Schernthaner GH, et al. Progressive peripheral arterial occlusive disease and other vascular events during nilotinib therapy in CML. Am J Hematol. 2011;86(7):533-539. 6. Le Coutre P, Rea D, Abruzzese E, et al.
haematologica | 2016; 101(10)
Efficacy and safety of frontline nilotinib in CML
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Severe peripheral arterial disease during nilotinib therapy. J Natl Cancer Inst. 2011;103(17):1347-1348. Giles FJ, Mauro MJ, Hong F, et al. Rates of peripheral arterial occlusive disease in patients with chronic myeloid leukemia in the chronic phase treated with imatinib, nilotinib, or non-tyrosine kinase therapy: a retrospective cohort analysis. Leukemia. 2013;27(6):1310-1315. Kim TD, Rea D, Schwarz M, et al. Peripheral artery occlusive disease in chronic phase chronic myeloid leukemia patients treated with nilotinib or imatinib. Leukemia. 2013;27(6):1316-1321. Valent P, Hadzijusufovic E, Schernthaner GH, Wolf D, Rea D, le Coutre P. Vascular safety issues in CML patients treated with BCR/ABL1 kinase inhibitors. Blood. 2015;125(6):901-906. Hochhaus A, Rosti G, Cross NC, et al. Frontline nilotinib in patients with chronic myeloid leukemia in chronic phase: results from the European ENEST1st study. Leukemia. 2016;30(1):57-64. Kantarjian HM, Giles F, Gattermann N, et al. Nilotinib (formerly AMN107), a highly selective BCR-ABL tyrosine kinase inhibitor, is effective in patients with Philadelphia chromosome-positive chronic myelogenous leukemia in chronic phase following imatinib resistance and intolerance. Blood. 2007;110(10):3540-3546. Rosti G, Palandri F, Castagnetti F, et al. Nilotinib for the frontline treatment of Ph(+) chronic myeloid leukemia. Blood. 2009;114 (24):4933-4938. Gugliotta G, Castagnetti F, Breccia M, et al. Long-term outcome of a phase 2 trial with nilotinib 400 mg twice daily in first-line treatment of chronic myeloid leukemia. Haematologica. 2015;100(9):1146-1150. Cortes JE, Jones D, O'Brien S, et al. Nilotinib as front-line treatment for patients with chronic myeloid leukemia in early chronic phase. J Clin Oncol. 2010;28(3):392-397. Baccarani M, Cortes J, Pane F, et al. Chronic myeloid leukemia: an update of concepts and management recommendations of European LeukemiaNet. J Clin Oncol. 2009;27(35):6041-6051. Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia: 2013. Blood. 2013;122(6):872-884. Sokal JE, Cox EB, Baccarani M, et al. Prognostic discrimination in 'good-risk' chronic granulocytic leukemia. Blood. 1984;63(4):789-799. Hasford J, Pfirrmann M, Hehlmann R, et al. A new prognostic score for survival of patients with chronic myeloid leukemia treated with interferon alfa. J Natl Cancer Inst. 1998;90(11):850-858. Hasford J, Baccarani M, Hoffmann V, et al. Predicting complete cytogenetic response and subsequent progression-free survival in 2060 patients with CML on imatinib treatment: the EUTOS score. Blood. 2011;118(3):686-692. Hughes T, Deininger M, Hochhaus A, et al. Monitoring CML patients responding to treatment with tyrosine kinase inhibitors: review and recommendations for harmonizing current methodology for detecting BCRABL transcripts and kinase domain mutations and for expressing results. Blood. 2006;108(1):28-37. Cross NC, White HE, Muller MC, Saglio G, Hochhaus A. Standardized definitions of molecular response in chronic myeloid leukemia. Leukemia. 2012;26(10):2172-2175.
haematologica | 2016; 101(10)
22. Cross NC, White HE, Colomer D, et al. Laboratory recommendations for scoring deep molecular responses following treatment for chronic myeloid leukemia. Leukemia. 2015;29(5):999-1003. 23. Soverini S, Hochhaus A, Nicolini FE, et al. BCR-ABL kinase domain mutation analysis in chronic myeloid leukemia patients treated with tyrosine kinase inhibitors: recommendations from an expert panel on behalf of European LeukemiaNet. Blood. 2011;118(5): 1208-1215. 24. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481. 25. Pfirrmann M, Hochhaus A, Lauseker M, Saussele S, Hehlmann R, Hasford J. Recommendations to meet statistical challenges arising from endpoints beyond overall survival in clinical trials on chronic myeloid leukemia. Leukemia. 2011;25(9):1433-1438. 26. Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the presence of competiting risks: new representations of old estimators. Stat Med. 1999;18(6):695-706. 27. Mechanick JI, Camacho PM, Garber AJ, et al. American association of clinical endocrinologists and american college of endocrinology protocol for standardized production of clinical practice guidelines, algorithms and checklists-2014 update and the AACE G4G program. Endocrine Practice. 2014;20(7):692702. 28. American Diabetes Association. Classification and diagnosis of diabetes. Diabetes Care. 2015;38 Suppl:S8-S16. 29. Perk J, De Backer G, Gohlke H, et al. European Guidelines on cardiovascular disease prevention in clinical practice (version 2012). The Fifth Joint Task Force of the European Society of Cardiology and Other Societies on Cardiovascular Disease Prevention in Clinical Practice (constituted by representatives of nine societies and by invited experts). Eur Heart J. 2012;33(13):1635-1701. 30. Kantarjian H, Shah NP, Hochhaus A, et al. Dasatinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia. N Engl J Med. 2010;362(24):22602270. 31. Jabbour E, Kantarjian HM, Saglio G, et al. Early response with dasatinib or imatinib in chronic myeloid leukemia: 3-year follow-up from a randomized phase 3 trial (DASISION). Blood. 2014;123(4):494-500. 32. Cortes J, Saglio G, Kantarjian HM, et al. Final 5-Year Study Results of DASISION: The Dasatinib Versus Imatinib Study in Treatment-NaĂŻve Chronic Myeloid Leukemia Patients Trial. J Clin Oncol. 2016 May 23, DOI 10.1200/JCO.2015.64.8899 [Epub ahead of print]. 33. Cortes JE, Kim DW, Kantarjian HM, et al. Bosutinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukemia: results from the BELA trial. J Clin Oncol. 2012;30(28):3486-3492. 34. BrĂźmmendorf TH, Cortes JE, de Souza CA, et al. Bosutinib versus imatinib in newly diagnosed chronic-phase chronic myeloid leukaemia: results from the 24-month follow-up of the BELA trial. Br J Haematol. 2015;168(1):69-81. 35. Lipton JH, Chuah C, Guerci-Bresler A, et al. Ponatinib versus imatinib for newly diagnosed chronic myeloid leukaemia: an international, randomised, open-label, phase 3 trial. Lancet Oncol. 2016 Apr 12, DOI 10.1016/S1470-2045(16)00080-2 [Epub ahead of print].
36. Jain P, Kantarjian H, Jabbour E, et al. Ponatinib as first-line treatment for patients with chronic myeloid leukaemia in chronic phase: a phase 2 study. Lancet Haematol. 2015;2(9):e376-e383. 37. Hoffmann VS, Baccarani M, Hasford J, et al. The EUTOS population-based registry: incidence and clinical characteristics of 2904 CML patients in 20 European Countries. Leukemia. 2015;29(6):1336-1343. 38. Castagnetti F, Gugliotta G, Breccia M, et al. Long-term outcome of chronic myeloid leukemia patients treated frontline with imatinib. Leukemia. 2015;29(9):1823-1831. 39. Experts in Chronic Myeloid Leukemia. The price of drugs for chronic myeloid leukemia (CML) is a reflection of the unsustainable prices of cancer drugs: from the perspective of a large group of CML experts. Blood. 2013;121(22):4439-4442. 40. O'Brien S, Radich JP, Abboud CN, et al. Chronic myelogenous leukemia, version 1.2015. J Natl Compr Canc Netw. 2014;12(11):1590-1610. 41. Marin D, Ibrahim AR, Lucas C, et al. Assessment of BCR-ABL1 Transcript Levels at 3 Months Is the Only Requirement for Predicting Outcome for Patients With Chronic Myeloid Leukemia Treated With Tyrosine Kinase Inhibitors. J Clin Oncol. 2012;30(3):232-238. 42. Hanfstein B, Muller MC, Hehlmann R, et al. Early molecular and cytogenetic response is predictive for long-term progression-free and overall survival in chronic myeloid leukemia (CML). Leukemia. 2012;26(9): 2096-2102. 43. Jain P, Kantarjian H, Nazha A, et al. Early responses predict better outcomes in patients with newly diagnosed chronic myeloid leukemia: results with four tyrosine kinase inhibitor modalities. Blood. 2013;121 (24):4867-4874. 44. Hughes TP, Saglio G, Kantarjian HM, et al. Early molecular response predicts outcomes in patients with chronic myeloid leukemia in chronic phase treated with frontline nilotinib or imatinib. Blood. 2014;123(9):1353-1360. 45. Hehlmann R, Muller MC, Lauseker M, et al. Deep molecular response is reached by the majority of patients treated with imatinib, predicts survival, and is achieved more quickly by optimized high-dose imatinib: results from the randomized CML-study IV. J Clin Oncol. 2014;32(5):415-423. 46. Moslehi JJ, Deininger M. Tyrosine Kinase Inhibitor-Associated Cardiovascular Toxicity in Chronic Myeloid Leukemia. J Clin Oncol. 2015;33(35):4210-4218. 47. Valent P, Hadzijusufovic E, Schernthaner GH, Wolf D, Rea D, le Coutre P. Vascular safety issues in CML patients treated with BCR/ABL1 kinase inhibitors. Blood. 2015;125(6):901-906. 48. Rea D, Mirault T, Cluzeau T, et al. Early onset hypercholesterolemia induced by the 2nd-generation tyrosine kinase inhibitor nilotinib in patients with chronic phasechronic myeloid leukemia. Haematologica 2014;99(7):1197-1203. 49. Rea D, Gautier JF, Breccia M, et al. Incidence of hyperglycemia by 3 years in patients with newly diagnosed chronic myeloid leukemia in chronic phase treated with nilotinib or imatinib in ENESTnd. Blood. 2012;120(21). Abstract 1686. 50. Zdenek R, Belohlavkova P, Cetkovsky P, et al. Comparison of Glucose and Lipid Metabolism Abnormality during Nilotinib, Imatinib and Dasatinib Therapy - Results of Enigma 2 Study. Blood. 2014;124(21). Abstract 1813.
1207
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Acute Myeloid Leukemia
Ferrata Storti Foundation
Ciprofloxacin versus colistin prophylaxis during neutropenia in acute myeloid leukemia: two parallel patient cohorts treated in a single center Michele Pohlen,1* Julia Marx,1* Alexander Mellmann,2 Karsten Becker,3 Rolf M. Mesters,1 Jan-Henrik Mikesch,1 Christoph Schliemann,1 Georg Lenz,1,4,5 Carsten Müller-Tidow,1,6 Thomas Büchner,1 Utz Krug,1,7 Matthias Stelljes,1 Helge Karch,2,5 Georg Peters,3,5 Hans U. Gerth,8** Dennis Görlich,9** and Wolfgang E. Berdel1,5**
Department of Medicine A, Hematology and Oncology, University Hospital of Muenster; Institute of Hygiene, University Hospital Muenster; 3Institute of Medical Microbiology, University Hospital Muenster; 4Translational Oncology, University Hospital Muenster; 5 Cluster of Excellence EXC 1003, Cells in Motion; 6Department of Medicine IV, Hematology and Oncology, University Hospital Halle; 7Department of Medicine 3, Klinikum Leverkusen; 8Department of Medicine D, Nephrology and Rheumatology, University Hospital Muenster; 9Institute of Biostatistics and Clinical Research, University Muenster, Germany 1 2
Haematologica 2016 Volume 101(10):1208-1215
*MP and JM contributed equally to this work. **HUG, DG and WEB contributed equally to this work.
ABSTRACT
Correspondence: michele.pohlen@ukmuenster.de
Received: April 22, 2016. Accepted: July 19, 2016. Pre-published: July 28, 2016. doi:10.3324/haematol.2016.147934
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1208
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1208
P
atients undergoing intensive chemotherapy for acute myeloid leukemia are at high risk for bacterial infections during therapyrelated neutropenia. However, the use of specific antibiotic regimens for prophylaxis in afebrile neutropenic acute myeloid leukemia patients is controversial. We report a retrospective evaluation of 172 acute myeloid leukemia patients who received 322 courses of myelosuppressive chemotherapy and had an expected duration of neutropenia of more than seven days. The patients were allocated to antibiotic prophylaxis groups and treated with colistin or ciprofloxacin through 2 different hematologic services at our hospital, as available. The infection rate was reduced from 88.6% to 74.2% through antibiotic prophylaxis (vs. without prophylaxis; P=0.04). A comparison of both antibiotic drugs revealed a trend towards fewer infections associated with ciprofloxacin prophylaxis (69.2% vs. 79.5% in the colistin group; P=0.07), as determined by univariate analysis. This result was confirmed through multivariate analysis (OR: 0.475, 95%CI: 0.236-0.958; P=0.041). The prophylactic agents did not differ with regard to the microbiological findings (P=0.6, not significant). Of note, the use of ciprofloxacin was significantly associated with an increased rate of infections with pathogens that are resistant to the antibiotic used for prophylaxis (79.5% vs. 9.5% in the colistin group; P<0.0001). The risk factors for higher infection rates were the presence of a central venous catheter (P<0.0001), mucositis grade III/IV (P=0.0039), and induction/relapse courses (vs. consolidation; P<0.0001). In conclusion, ciprofloxacin prophylaxis appears to be of particular benefit during induction and relapse chemotherapy for acute myeloid leukemia. To prevent and control drug resistance, it may be safely replaced by colistin during consolidation cycles of acute myeloid leukemia therapy. haematologica | 2016; 101(10)
Antibiotic prophylaxis in neutropenic AML patients
Introduction Bacterial infections are the most common cause of treatment-related mortality in patients with neutropenia after chemotherapy, particularly when the expected duration of neutropenia is seven or more days.1-3 Patients with acute myeloid leukemia (AML) are at a particularly high risk. In addition to disease- and therapy-induced myelosuppression, disease-related conditions, such as alteration of the host defenses secondary to infiltration of the bone marrow and therapy-induced side effects (such as mucositis or diarrhea after breakdown of the mucosal barrier), further contribute to the high risk of infections.4 Fluoroquinolone has partially replaced the previous use of non-absorbable antibiotics, such as colistin/polymyxin B and oral vancomycin, for the prophylaxis of neutropenia-related infections. Initially, this was based on better tolerance for fluoroquinolone rather than on a proven decrease in the infection rate.5-11 Although subsequent double-blind, placebo-controlled, randomized trials have shown a decrease in the infection rates,12,13 evidence of significant benefits of fluoroquinolones in preventing infection-related mortality is still limited to meta-analyses.14 Nevertheless, fluoroquinolones have been included in some, but not all, guidelines for the treatment of neutropenic AML patients.1,3,15 The prevention and control of drug-resistant and multidrug-resistant pathogens are becoming increasingly challenging. Widespread antibiotic prophylaxis may promote the development of drug resistance. In addition, fluoroquinolones are increasingly being linked to serious sideeffects, thus leading the Food and Drug Administration (FDA) to raise concerns regarding their use and to introduce the term Fluoroquinolone-associated Disability.16 Thus, the use of antibiotic prophylaxis in afebrile neu-
tropenic AML patients remains controversial.17 Here, we report a single-institution analysis comparing the effects and benefits of two common antibiotic prophylaxis regimens, colistin and ciprofloxacin, in a cohort of AML patients with a high risk of chemotherapy-induced neutropenic infections.
Methods Patients A total of 172 consecutive patients with AML who received inpatient, intensive chemotherapy in the Department of Medicine A of the University Hospital of Muenster, Germany, and were at risk of developing chemotherapy-induced neutropenia lasting more than seven days, were included in this retrospective analysis. All patients provided written informed consent prior to the initiation of the anti-leukemic therapy. Approval for this analysis was obtained from the Ethics Board of the Westfalian WilhelmsUniversity Muenster, Germany, and the Physicians Chamber of Westphalia-Lippe, Germany (approval number 2015-695-f-S). On the basis of service availability, all patients were allocated to one of two different physician services within the same department; one treatment team used oral colistin, and the other administered ciprofloxacin. Individual patientsâ&#x20AC;&#x2122; characteristics (e.g. disease status, treatment regimen) had no influence on his/her assignment to each treatment group. All other therapies and supportive care were provided to both groups, according to identical institutional guidelines. Intensive induction or consolidation chemotherapy and an expected duration of neutropenia of more than seven days were prerequisites for eligibility in this study. Neutropenia was defined as a neutrophil count less than <0.5x109 cells/L, and leukocytopenia was defined as a leukocyte count less than 1.0x109 cells/L when differential leukocyte counts were not available.
Figure 1. Flow chart displaying the treatment courses administered and the patient numbers (n,N).AML: acute myeloid leukemia.
haematologica | 2016; 101(10)
1209
M. Pohlen et al.
Prophylactic regimen
divided into four doses. Patients in the ciprofloxacin group were given 500 mg twice daily. Both drugs were taken orally. In addition, both groups received prophylaxis for Pneumocystis jirovecii pneumonia with 960 mg of trimethoprim-sulfamethoxazole twice daily for two days per week.
Patients received antibiotic prophylaxis simultaneously with the start of chemotherapy. Upon hematologic recovery (<0.5x109 neutrophils/L or >1.0x109 leukocytes/L), prophylaxis was stopped. Patients in the colistin group were given 8 million IU daily,
Table 1A. Patients’ baseline characteristics.
Characteristics
All patients (n=156)
Age at first treatment Mean ± SD 58.1±13.9 Median (IQR) 60 (49-69) Range 18-85 Sex, n (%) Male 85 (54.5) Female 71 (45.5) FAB classification, n (%) M0 13 (8.3) M1 14 (9.0) M2 42 (26.9) M3 2 (1.3) M4 41 (26.3) M5 28(18.0) M6 6 (3.9) M7 1 (0.6) Undetermined 9 (5.8) Therapy stage, n (%) Induction 133 (85.3) Consolidation 14 (9.0) Relapse 9 (5.8) Charlson comorbidity index at first treatment, n (%) 2 87 (55.8) 3 35 (22.4) 4 18 (11.5) ≥5 16 (10.3) Number of courses analyzed (%) 1 84 (53.9) 2 27 (17.3) 3 20 (12.8) 4 18 (11.5) 5 7 (4.5) Prophylactic regimen, n (%) Without antibiotic prophylaxis 35 (22.4) With antibiotic prophylaxis 138 (88.5) Ciprofloxacin 57 (36.5) Colistin 72 (46.2) Switch between prophylaxes 9 (5.8)
Without prophylaxis (n=35)
Ciprofloxacin (n=57)
Prophylaxis Colistin (n=72)
54.5±15.6 60 (47.2-65) 18-76
57.1±14.0 56 (48-67) 18-84
59.2±13.3 62 (54.5-69) 19-85
19 (54.3) 16 (45.7)
33 (57.9) 24 (42.1)
39 (54.2) 33 (45.8)
2 (5.7) 3 (8.6) 14 (40.0) 8 (22.9) 0 (0.0) 6 (17.1) 1 (2.9) 1 (2.9) 0 (0.0)
6 (10.5) 3 (5.3) 15 (26.3) 2 (3.5) 15 (26.3) 9 (15.8) 3 (5.3) 0 (0.0) 4 (7.0)
5 (6.9) 7 (9.7) 18 (25.0) 0 (0.0) 20 (27.8) 15 (20.8) 2 (2.8) 0 (0.0) 5 (6.9)
18 (51.4) 12 (34.3) 5 (14.3)
28 (49.1) 23 (40.4) 6 (10.5)
41 (56.9) 26 (36.1) 5 (6.9)
21 (60.0) 10 (28.6) 2 (5.7) 2 (5.7)
34 (59.6) 13 (22.8) 5 (8.8) 5 (8.8)
41 (56.9) 12 (16.7) 11 (15.3) 8 (11.1)
16 (45.7) 5 (14.3) 4 (11.4) 6 (17.1) 4 (11.4)
26 (45.6) 10 (17.5) 6 (10.5) 10 (17.5) 5 (8.8)
41 (56.9) 13 (18.1) 12 (16.7) 5 (6.9) 1 (1.4)
-
-
-
P 0.299
0.672 0.683
0.609
0.576
0.076
-
FAB: French-American-British classification; SD: standard deviation; IQR: Interquartile range; n: number.
Table 1B. Status at the time of treatment group allocation: the treatment courses according to therapy stage.
Treatment course (%) 1st* 2nd 3rd 4th 5th 6th 10th 11th
Induction (n=148 courses)
Consolidation (n=129 courses)
Relapse (n=19 courses)
126 (85.1) 21 (14.2) 1 (0.7) 0 0 0 0 0
0 26 (20.2) 53 (41.1) 31 (24.0) 19 (14.7) 0 0 0
0 0 3 (15.8) 4 (21.1) 8 (42.1) 2 10.5) 1 (5.3) 1 (5.3)
*Includes courses of double-induction; n: number.
1210
haematologica | 2016; 101(10)
Antibiotic prophylaxis in neutropenic AML patients
End points used in the analysis The primary end point was a clinically documented infection requiring empirical antibacterial therapy, which was defined as the presence of at least two of the following criteria: a) fever during neutropenia (oral temperature above 38.3°C in a single measurement or ≥ 38.0°C in measurements taken over at least one hour); b) clinical signs of infections (e.g. hypotension, tachypnea, or tachycardia); and c) laboratory (e.g. an increase in C-reactive protein or procalcitonin levels) or microbiological findings. Because antibiotic prophylaxis may interfere with the culture results, microbiological findings were not mandatory because this would have resulted in under-reporting of infections.13 The secondary end points were microbiological findings (positive culture results), an infection-related need for intensive care medicine, and mortality as a result of any type of infection.
Statistical analysis Distributions of patient baseline characteristics in both prophylactic groups were compared using χ2 tests for categorical variables and Mann-Whitney U tests for the continuous variables. The differences between groups were analyzed through statistical methods capable of modeling repeated measurements. Here, generalized estimation equations (GEEs) were applied. All statistical analyses were performed with IBM SPSS Statistics for Windows, v.22.0 (IBM Corp., Armonk, NY, USA) and SAS software (v.9.4, for Windows, SAS Institute Inc., Cary, NC, USA). A detailed description of Materials and Methods is included in the Online Supplementary Appendix.
with at least one chemotherapy course per stay (Figure 1). Courses with antibiotic treatments prior to the start of chemotherapy (n=26) were excluded, and the data for 296 treatment courses (156 patients) were used for the subsequent analyses. The patients’ baseline characteristics at the first treatment course are presented in Table 1A. During a total of 44 courses (14.9%) antibiotic prophylaxis was not administered in 35 patients, mostly at the request of the patient. However, this group of patients was also analyzed and separately compared with patients who received prophylaxis. The remaining 138 patients received antibiotic prophylaxis over 252 treatment courses: 72 patients received colistin (in 122 courses), and 57 patients received ciprofloxacin (in 130 courses). Nine patients switched treatment group (5.8% crossover) mainly due to capacity reasons of one team and to ensure continuation of chemotherapy. Patients received a median of 3 treatment courses. A complete standard therapy usually included 3-5 courses per patient (1-2 induction courses and 2-3 consolidation courses). Deviations from standard therapy were mostly due to courses outside of the study period, the need for allogeneic stem cell transplantation, death, or exclusion of courses with antibiotic pre-treatment. Induction therapy accounted for 50.0% of all treatments, followed by consolidation (43.6%) and relapse (6.4%) treatments. This distribution was similar between the colistin and ciprofloxacin groups (Table 1B).
Infection rates Results Patients’ characteristics A total of 172 patients received 322 treatment courses
In the absence of antibiotic prophylaxis, clinically documented infections occurred significantly earlier (P=0.0001) (Online Supplementary Figure S1) and more often (88.6% vs. 74.2%) with prophylaxis (Table 2). Infections during col-
Table 2. Infection-related data.
Prophylaxis (n=252 courses) Ciprofloxacin (n=130) Onset of infection, median in days Infection, n (%) Induction, n (%)a Consolidation, n (%)a Relapse, n (%)a Infection with detection of pathogen, n (%) Gram-positive Gram-negative Fungal Viral Resistant to prophylaxis Infection with multidrug-resistant pathogen, n (%) Vancomycin-resistant Enterococcus Extended-spectrum-betalaktamase Pseudomonas aeruginosa Mucositis (grade III/IV), n (%) Central venous catheter, n (%) Need for intensive care, n (%) Infection-related Length of hospital stay, median, days Death during hospital stay, n (%)b Infection-related
15.5 90 (69.2) 55 (88.7) 29 (48.3) 6 (75.0) 39 (30.0) 31 7 1 4 31 4 (3.1) 2 1 1 34 (26.2) 71 (54.6) 5 (3.8) 5 29.0 3 (5.1) 3
Without prophylaxis (n=44 courses) Colistin (n=122) 13.0 97 (79.5) 63 (96.9) 28 (54.9) 6 (100) 42 (34.4) 27 18 3 2 4 6 (4.9) 5 1 0 21 (17.2) 74 (60.7) 5 (4.1) 4 32.0 6 (8.5) 6
10.0 39 (88.6) 19 (90.5) 15 (83.3) 5 (100) 22 (50.0) 15 9 2 1 2 (4.5) 1 0 1 9 (20.5) 27 (61.4) 4 (9.1) 4 29.0 3 (20.0) 3
Percentage of infections in each treatment group according to therapy stage (induction, consolidation, or relapse); bpercentage of patients (not treatment courses); n: number.
a
haematologica | 2016; 101(10)
1211
M. Pohlen et al. Table 3. Infection rates in the entire patient cohort (prophylaxis and no prophylaxis): univariate GEE analysis of 296 treatment courses (156 patients).
Prophylaxis No Yes Sex Male Female Central venous catheter No Yes Mucositis (grade III/IV) No Yes Therapy stage Induction Consolidation Relapse
N
Estimated risk of infection in % (95% CI)
P
44 252
89.7 (78.2-95.5) 77.6 (72.0-82.5)
0.0403
170 152
84.7 (77.2-88.8) 73.6 (64.6-81.0)
0.0420
128 194
57.5 (46.8-67.7) 91.9 (86.7-95.2)
<0.0001
254 67
76.3 (70.1-81.6) 93.7 (82.6-97.9)
0.0111
167 133 22
93.4 (87.7-96.5) 56.4 (46.1-66.2) 89.1 (65.0-97.3)
<0.0001
Charlson index (per point) Age (per year)
OR (95% CI)
P
0.8682 (0.6861-1.0985) 0.9927 (0.9716-1.0143)
0.2391 0.5018
GEE: generalized estimation equation; CI: confidence interval; OR: odds ratio; N: number.
istin prophylaxis were observed in 79.5% of courses compared with 69.2% of courses in the case of ciprofloxacin prophylaxis (P=0.0727). Patients who received ciprofloxacin prophylaxis developed infections on day 15 (median), which was later than the time at which patients developed infections in the colistin group (median on day 13; P=0.0266) (Online Supplementary Figure S2A) The median length of hospital stay was substantially influenced (P=0.0283, GEE model) by the choice of prophylaxis (32 vs. 29 days in the colistin and ciprofloxacin groups, respectively). However, ciprofloxacin had no effect on the probability of discharge without infection (ciprofloxacin vs. colistin; P=0.0747) (Online Supplementary Figure S2B). The highest infection rates were observed during the induction and relapse reinduction courses (96.9% and 100% in the colistin group compared with 88.7% and 75.0% in the ciprofloxacin group). In contrast, infections occurred less frequently during consolidation courses (54.9% with colistin and 48.3% with ciprofloxacin; P<0.0001). Approximately one-quarter (26.2%) of the patients who received ciprofloxacin suffered from mucositis, compared with 17.2% in the colistin group (P=0.1683). Central venous catheters were used in slightly over half of the courses in both groups (54.6% in the ciprofloxacin vs. 60.7% in the colistin group).
Microbiological findings In cases of infection, the detection rate of isolated microorganisms was similar in both prophylactic groups (30.0% in the ciprofloxacin group vs. 34.4% in the colistin group; P=0.6436) (Table 2). In the colistin group, most of the micro-organisms were Gram-positive bacteria (64.3%), followed by Gram-negative bacteria (42.9%). The use of ciprofloxacin caused a clear shift towards Gram-positive pathogens (79.5% vs. 17.9% Gram-negative). The frequency of multidrug-resistant bacteria was not significantly different between the groups (4.9% in the colistin group vs. 3.1% in the ciprofloxacin group; P=0.4727). Although vancomycin-resistant enterococci appeared more often in the colistin group (5 vs. 2 courses; 1212
P=0.6667), multidrug-resistant Pseudomonas aeruginosa was not observed in this group (compared with 1 isolate in the ciprofloxacin group). Notably, the rate of pathogens with resistance to the assigned prophylactic drug was significantly higher in the ciprofloxacin group (79.5%, 31 of 39 vs. 9.5%, 4 of 42 in the colistin group; P<0.0001). Concerning the microbiologic milieu on both wards, results of routine monitoring display a similar spectrum of germs, especially concerning resistant pathogens. In detail, ciprofloxacin-resistance was predominantly found in samples with E. coli and Pseudomonas aeruginosa (approx. 40%), colistin-resistant bacteria have not become evident in significant quantity.
Outcome The need for intensive care was reduced by the application of prophylaxis (4.0% of patients who received prophylaxis vs. 9.1% who did not receive prophylaxis; P=0.2747), but there was no difference between the two prophylactic agents (3.8% of patients who received ciprofloxacin vs. 4.1% who received colistin; P=0.9245). Although there was a trend, mortality was not significantly influenced by the application of prophylaxis (7.0% mortality among patients who received prophylaxis vs. 20.0% among patients who did not receive prophylaxis; P=0.4219) or the type of prophylaxis (5.1% in the ciprofloxacin group vs. 8.6% in the colistin group; P=0.2857). All cases of death were described as infection-related (Table 3).
Risk factors for infections After the use of either antibiotic, the infection rate decreased from 88.6% to 74.2% (P=0.0403) (Table 2). Furthermore, the presence of a central venous catheter, mucositis, and induction/relapse therapy were associated with increased infection rates in the univariate analysis of the entire cohort (all P<0.05) (Table 3). The same parameters were significantly associated with infections in the univariate analysis of patients who received antibiotic prophylaxis (Table 4). Among these risk factors, only the incidence of mucositis increased the infection rates [odds ratio haematologica | 2016; 101(10)
Antibiotic prophylaxis in neutropenic AML patients
Table 4. Infection rates among patients treated with antibiotic prophylaxis: univariate GEE analysis of 252 treatment courses (138 patients).
Prophylaxis Colistin Ciprofloxacin Sex Male Female Central venous catheter No Yes Mucositis (grade III/IV) No Yes Therapy stage Induction Consolidation Relapse
N
Estimated risk of infection in % (95% CI)
P
122 130
81.4 (73.7-87.2) 71.4 (62.2-79.0)
0.0727
133 119
81.3 (73.9-86.9) 70.6 (61.0-78.6)
0.0531
107 145
52.9 (42.0-63.0) 91.5 (85.6-95.1)
<0.0001
196 55
72.3 (65.6-78.1) 93.4 (82.2-97.8)
0.0039
127 111 14
93.1 (86.8-96.5) 51.8 (41.5-62.0) 85.3 (57.1-96.2)
<0.0001
OR (95% CI)
P
0.8894 (0.6995-1.1308) 0.9957 (0.9733-1.0185)
0.3386 0.7077
Charlson index (per point) Age (per year) GEE: generalized estimation equation; CI: confidence interval; OR: odds ratio; N: number.
Table 5. Infection rates among patients treated with antibiotic prophylaxis: multivariate GEE model of 252 treatment courses (138 patients).
Type of prophylaxis Ciprofloxacin Colistin Sex Male Female Central venous catheter Yes No Mucositis (grade III/IV) Yes No Age (per year) Charlson score (per point) Therapy stage Induction or relapse Consolidation
Estimated risk of infection in % (95% CI)
OR (95% CI)
83.1 (70.3-91.1) 91.2 (83.8-95.4)
0.475 (0.236-0.958)
90.1 (82.7-95.2) 83.9 (71.6-91.5)
1.867 (0.902-3.864)
91.3 (81.6-96.2) 82.8 (65.3-92.5)
2.183 (0.573-3.086)
94.7 (84.5-98.3) 74.1 (65.8-80.9)
6.229 (1.773-21.883)
P 0.0405 0.0969 0.3157 0.0045
1.002(0.975-1.029) 0.818 (0.609-1.100) 74.2 (54.9-87.2) 94.6 (86.9-97.9)
0.9099 0.2047 0.0757
0.163 (0.042-0.637)
The reference categories are in bold. GEE: generalized estimation equation; CI: confidence interval; OR: odds ratio.
(OR) 6.229, 95% confidence interval (CI) 1.773-21.883; P=0.0045] in multivariate analysis, whereas prophylaxis with ciprofloxacin significantly decreased (0.4475, 95%CI: 0.236-0.958; P=0.0405) the infection rate in the multivariate analysis (Table 5). Furthermore, a subgroup analysis of different disease stages (induction, relapse, and consolidation) was performed. Here, prophylaxis with ciprofloxacin was independently associated with decreased infection rates only during the induction or relapse courses (OR 0.097, 95%CI: 0.017-0.556; P=0.0038) and not during the consolidation courses (OR 0.650, 95%CI: 0.285-1.481; P=0.2941) (Table 6). In contrast, mucositis was a significant predictor of infection in the consolidation courses (OR 4.398, 95%CI: 1.593-12.141; P=0.0089) but marginally missed significance in the induction/relapse courses (OR 5.357, 95%CI: 0.759-37.843; P=0.0511). haematologica | 2016; 101(10)
Discussion We report the results of a retrospective, single-institution analysis comparing the effects and benefits of two common antibiotics, colistin and ciprofloxacin, that were administered prophylactically in a cohort of AML patients at high risk of infection due to chemotherapy-induced neutropenia. Although the comparison was not based on a prospective randomization, the allocation of the patients to the 2 different prophylactic drugs was random, and all other therapy and supportive care was provided to both groups according to identical institutional guidelines. First, our data confirm that antibiotic prophylaxis is advantageous in preventing febrile neutropenia compared with no prophylaxis. Limiting allocation to courses without prophylaxis was neither planned nor random. Despite this and the small size of this group, it is remarkable that 1213
M. Pohlen et al. Table 6. Influence of the therapy regimen/disease stage on the infection rates among patients treated with antibiotic prophylaxis (multivariate GEE model).
Type of prophylaxis Ciprofloxacin Colistin Sex Male Female Central venous catheter Yes No Mucositis (grade III/IV) Yes No Age (per year) Charlson score (per point)
Induction or relapse (N=125 patients, n=141 courses) Estimated risk OR (95% CI) P of infection in % (95% CI)
Consolidation (N=61 patients, n=110 courses) Estimated risk OR (95% CI) of infection in % (95% CI)
93.4 (71.9-98.7) 99.3 (93.7-99.9)
0.097 (0.017-0.556)
0.0038
65.6 (42.5-83.1) 74.6 (55.2-87.4)
0.650 (0.285-1.481)
0.2941
97.1 (84.4-99.5) 98.4 (89.2-99.8)
1.839 (0.541-6.250)
0.3354
62.6 (39.8-80.9) 77.0 (57.4-89.2)
1.998 (0.834-4.785)
0.1274
95.6 (89.1-99.0) 98.7 (77.7-99.9)
0.380 (0.021-6.962)
0.4279
76.9 (46.2-92.8) 62.7 (47.7-75.6)
1.975 (0.519-7.509)
0.3559
99.1 (90.8-99.9) 95.2 (78.1-99.1)
5.357 (0.759-37.843)
0.0511
83.2 (61.2-94.0) 53.0 (35.6-69.7)
4.398 (1.593-12.141)
0.0089
1.036(0.-989-1.086) 0.710 (0.510-0.988)
0.0991 0.2349
0.994 (0.959-1.031) 0.855 (0.582-1.257)
0.7602 0.4226
P
The reference categories are in bold. GEE: generalized estimation equation; CI: confidence interval; OR: odds ratio; N, n: number.
this finding was not dependent on the agent that was chosen and is in accordance with previous observations.12,13 Second, the application of ciprofloxacin decreases the infection rates in the induction and relapse courses more than colistin. This result is in accordance with some previous studies comparing fluoroquinolones with nonabsorbable agents, although none of these studies investigated ciprofloxacin versus colistin in a high-risk cohort of AML patients.7,11 Concerning microbiological findings, the type of prophylaxis did not influence the infection rates, but the application of ciprofloxacin induced a shift from Gramnegative to Gram-positive organisms in the microbiological findings, as previously described.18,19 The effects of fluoroquinolones on the incidence of Methicillin-resistant Staphylococcus aureus (MRSA) have been described previously.20,21 Furthermore, patients without antibiotic prophylaxis and those who received colistin presented with a similar spectrum of micro-organisms, which may be explained by the lack of systemic activity of colistin and its narrower spectrum compared with ciprofloxacin. The type of prophylaxis did not significantly influence major clinical events, such as the requirement for intensive care or infection-related mortality. Thus, our results are in accordance with other studies showing a slight but insignificant trend towards lower mortality rates in patients who received antibiotic prophylaxis (vs. no prophylaxis) and in those who received ciprofloxacin (vs. colistin).12,13 However, based on the observed mortality rates in this study, a 5-fold increase in the number of patients would have been necessary to reveal statistically significant differences. Because of the limited AML incidence, no study to date has overcome this obstacle, and statistically significant differences in mortality rates have been observed only in larger meta-analyses.14,22,23 However, we found a significant difference in the number of pathogens that were resistant to the applied prophylaxis. Prophylaxis-resistant pathogens were identified more frequently in patients treated with ciprofloxacin (79.5% vs. 9.5% with colistin). This result may raise concerns regarding the general usage of broad-spectrum 1214
antibiotics such as fluoroquinolones for prophylaxis. Because these drugs are part of the standard therapy for many other infections, a prior application can considerably impair their efficacy in this context. Thus, it is more likely that an even broader empiric regimen may be chosen if an infection occurs.17 In our study, no relevant differences in the rates of multidrug-resistant pathogens were observed, and only vancomycin-resistant enterococci were observed at non-significant levels in the colistin group. However, the choice of prophylaxis must also be taken into account with regard to the development of multidrug-resistant organisms, which are an increasing challenge in the health care system.24-28 Consequently, a reduction of the likelihood of infections in neutropenic cancer patients must be weighed against the additional effects of drug resistance on the morbidity and mortality of hospital-acquired infections.24,29-32 Notably, patients who received ciprofloxacin prophylaxis were discharged earlier than patients who received colistin prophylaxis. Because the time to infection was also prolonged for patients who received ciprofloxacin prophylaxis, these data suggest a clinically meaningful benefit of ciprofloxacin prophylaxis. In addition to antimicrobial prophylaxis, two other factors influenced the infection rates: the use of a central venous catheter and the incidence of mucositis. Both factors are potential ports of entry for bacterial invasion into the bloodstream and have been described previously.4 However, only mucositis emerged as an independent predictor of a higher infection rate, particularly in patients receiving consolidation therapies. The influence of prophylaxis clearly differed between the induction/relapse and consolidation courses. Although the choice of prophylactic agent was an independent parameter for lower infection rates in the induction/relapse courses, the infection rates in the consolidation courses were predominantly influenced by mucositis. This result may be explained by the observation that patients are at a higher risk at the time of primary diagnosis or relapse because they previously suffered from functional neutropenia for an unknown period of time due haematologica | 2016; 101(10)
Antibiotic prophylaxis in neutropenic AML patients
to overt leukemia. Furthermore, the differences in the chemotherapy protocols used for induction and consolation may also play a role. Some limitations of this study deserve discussion. First, although other factors disturb the mucosal barrier, this analysis was limited to mucositis. Second, the duration of neutropenia was not included in the analysis, thus, a potential association with the differences in the induction or consolidation courses could not be revealed. Third, this analysis can provide only suggestions for antibiotic prophylaxis in a cohort of selected high-risk patients but cannot provide general recommendations. Antimicrobial resistance is influenced by an indi-
References 1. Neumann S, Krause SW, Maschmeyer G, et al. Primary prophylaxis of bacterial infections and Pneumocystis jirovecii pneumonia in patients with hematological malignancies and solid tumors: guidelines of the Infectious Diseases Working Party (AGIHO) of the German Society of Hematology and Oncology (DGHO). Ann Hematol. 2013;92(4):433-442. 2. Link H, Bohme A, Cornely OA, et al. Antimicrobial therapy of unexplained fever in neutropenic patients--guidelines of the Infectious Diseases Working Party (AGIHO) of the German Society of Hematology and Oncology (DGHO), Study Group Interventional Therapy of Unexplained Fever, Arbeitsgemeinschaft Supportivmassnahmen in der Onkologie (ASO) of the Deutsche Krebsgesellschaft (DKG-German Cancer Society). Ann Hematol. 2003;82(Suppl 2):S105-17. 3. Freifeld AG, Bow EJ, Sepkowitz KA, et al. Clinical practice guideline for the use of antimicrobial agents in neutropenic patients with cancer: 2010 update by the infectious diseases society of america. Clin Infect Dis. 2011;52(4):e56-93. 4. Bow EJ, Meddings JB. Intestinal mucosal dysfunction and infection during remission-induction therapy for acute myeloid leukaemia. Leukemia. 2006;20(12):20872092. 5. Gluckman E, Roudet C, Hirsch I, et al. Prophylaxis of bacterial infections after bone marrow transplantation. A randomized prospective study comparing oral broad-spectrum nonabsorbable antibiotics (vancomycin-tobramycin-colistin) to absorbable antibiotics (ofloxacin-amoxicillin). Chemotherapy. 1991;37(Suppl 1):3338. 6. Winston DJ, Ho WG, Nakao SL, Gale RP, Champlin RE. Norfloxacin versus vancomycin/polymyxin for prevention of infections in granulocytopenic patients. Am J Med. 1986;80(5):884-890. 7. Winston DJ, Ho WG, Bruckner DA, Gale RP, Champlin RE. Ofloxacin versus vancomycin/polymyxin for prevention of infections in granulocytopenic patients. Am J Med. 1990;88(1):36-42. 8. Archimbaud E, Guyotat D, Maupas J, et al. Pefloxacin and vancomycin vs. gentamicin, colistin sulphate and vancomycin for prevention of infections in granulocytopenic patients: a randomised double-blind study. Eur J Cancer. 1991;27(2):174-178. 9. Arning M, Wolf HH, Aul C, Heyll A, Scharf RE, Scheider W. Infection prophylaxis in neutropenic patients with acute leukaemia-a randomized, comparative study with ofloxacin, ciprofloxacin and co-trimoxazole/colistin. J Antimicrob Chemother.
haematologica | 2016; 101(10)
vidual patientâ&#x20AC;&#x2122;s characteristics as well as hospital/environmental conditions. The data from this analysis enable a hypothesis to be made, and further prospective trials are warranted. In summary, ciprofloxacin prophylaxis appears to be of particular benefit during induction and relapse chemotherapy for AML, but it may be safely replaced by colistin during consolidation cycles of AML therapy. The selection of prophylactic agents should take into account variables such as therapy stage (induction/relapse vs. consolidation), the risk of developing mucositis, and the local distribution of resistant pathogens.
1990;26(Suppl D):137-142. 10. Jansen J, Cromer M, Akard L, Black JR, Wheat LJ, Allen SD. Infection prevention in severely myelosuppressed patients: a comparison between ciprofloxacin and a regimen of selective antibiotic modulation of the intestinal flora. Am J Med. 1994; 96(4):335-341. 11. Moriuchi Y, Kamihira S, Yamamura M, et al. Comparison of ciprofloxacin with polymyxin B for infection prophylaxis in neutropenic patients with acute non-lymphocytic leukemia. Rinsho Ketsueki. 1990;31(10):1664-1669. 12. Bucaneve G, Micozzi A, Menichetti F, et al. Levofloxacin to prevent bacterial infection in patients with cancer and neutropenia. N Engl J Med. 2005;353(10):977-987. 13. Cullen M, Steven N, Billingham L, et al. Antibacterial prophylaxis after chemotherapy for solid tumors and lymphomas. N Engl J Med. 2005;353(10):988-998. 14. Gafter-Gvili A, Fraser A, Paul M, et al. Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy. Cochrane Database Syst Rev. 2012;1:CD004386. 15. Flowers CR, Seidenfeld J, Bow EJ, et al. Antimicrobial prophylaxis and outpatient management of fever and neutropenia in adults treated for malignancy: American Society of Clinical Oncology clinical practice guideline. J Clin Oncol. 2013;31(6):794810. 16. Food and Drug Administration Center for Drug Evaluation and Research (FDA). FDA Briefing Document: The Benefits and Risks of Systemic Fluoroquinolone Antibacterial Drugs for the Treatment of Acute Bacterial Sinusitis (ABS), Acute Bacterial Exacerbation of Chronic Bronchitis in Patients Who Have Chronic Obstructive Pulmonary Disease (ABECB-COPD), and Uncomplicated Urinary Tract Infections (uUTI). November 5, 2015. 17. Slavin MA, Lingaratnam S, Mileshkin L, et al. Use of antibacterial prophylaxis for patients with neutropenia. Australian Consensus Guidelines 2011 Steering Committee. Intern Med J. 2011;41(1b):102109. 18. Gudiol C, Bodro M, Simonetti A, et al. Changing aetiology, clinical features, antimicrobial resistance, and outcomes of bloodstream infection in neutropenic cancer patients. Clin Microbiol Infect. 2013;19(5):474-479. 19. Hammond SP, Baden LR. Antibiotic prophylaxis for patients with acute leukemia. Leuk Lymphoma. 2008;49(2):183-193. 20. Evans ME, Titlow WB. Selection of fluoroquinolone-resistant methicillin-resistant Staphylococcus aureus with ciprofloxacin and trovafloxacin. Antimicrob Agents Chemother. 1998;42(3):727.
21. Venezia RA, Domaracki BE, Evans AM, Preston KE, Graffunder EM. Selection of high-level oxacillin resistance in heteroresistant Staphylococcus aureus by fluoroquinolone exposure. J Antimicrob Chemother. 2001;48(3):375-381. 22. Leibovici L, Paul M, Cullen M, et al. Antibiotic prophylaxis in neutropenic patients: new evidence, practical decisions. Cancer. 2006;107(8):1743-1751. 23. Imran H, Tleyjeh IM, Arndt CA, et al. Fluoroquinolone prophylaxis in patients with neutropenia: a meta-analysis of randomized placebo-controlled trials. Eur J Clin Microbiol Infect Dis. 2008;27(1):53-63. 24. Kern WV, Klose K, Jellen-Ritter AS, et al. Fluoroquinolone resistance of Escherichia coli at a cancer center: epidemiologic evolution and effects of discontinuing prophylactic fluoroquinolone use in neutropenic patients with leukemia. Eur J Clin Microbiol Infect Dis. 2005;24(2):111-118. 25. Carratala J, Fernandez-Sevilla A, Tubau F, Dominguez MA, Gudiol F. Emergence of fluoroquinolone-resistant Escherichia coli in fecal flora of cancer patients receiving norfloxacin prophylaxis. Antimicrob Agents Chemother. 1996;40(2):503-505. 26. Baum HV, Franz U, Geiss HK. Prevalence of ciprofloxacin-resistant Escherichia coli in hematologic-oncologic patients. Infection. 2000;28(5):278-281. 27. Razonable RR, Litzow MR, Khaliq Y, Piper KE, Rouse MS, Patel R. Bacteremia due to viridans group Streptococci with diminished susceptibility to Levofloxacin among neutropenic patients receiving levofloxacin prophylaxis. Clin Infect Dis. 2002;34(11): 1469-1474. 28. Somolinos N, Arranz R, Del Rey MC, Jimenez ML. Superinfections by Escherichia coli resistant to fluoroquinolones in immunocompromised patients. J Antimicrob Chemother. 1992;30 (5):730-731. 29. Rangaraj G, Granwehr BP, Jiang Y, Hachem R, Raad I. Perils of quinolone exposure in cancer patients: breakthrough bacteremia with multidrug-resistant organisms. Cancer. 2010;116(4):967-973. 30. Gafter-Gvili A, Paul M, Fraser A, Leibovici L. Effect of quinolone prophylaxis in afebrile neutropenic patients on microbial resistance: systematic review and metaanalysis. J Antimicrob Chemother. 2007;59(1):5-22. 31. Chong Y, Yakushiji H, Ito Y, Kamimura T. Clinical impact of fluoroquinolone prophylaxis in neutropenic patients with hematological malignancies. Int J Infect Dis. 2011;15(4):e277-81. 32. Bow EJ. Fluoroquinolones, antimicrobial resistance and neutropenic cancer patients. Curr Opin Infect Dis. 2011;24(6):545-553.
1215
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Acute Myeloid Leukemia
Ferrata Storti Foundation
Haematologica 2016 Volume 101(10):1216-1227
Acute myeloid leukemia cells polarize macrophages towards a leukemia supporting state in a Growth factor independence 1 dependent manner Yahya S. Al-Matary,1* Lacramioara Botezatu,1* Bertram Opalka,1 Judith M. Hönes,1 Robert F. Lams,1 Aniththa Thivakaran,1 Judith Schütte,1 Renata Köster,1 Klaus Lennartz,2 Thomas Schroeder,3 Rainer Haas,3 Ulrich Dührsen,1 and Cyrus Khandanpour1
Department of Hematology, University Hospital of Essen, West German Cancer Center (WTZ); 2Institute of cell biology (Tumor Research), University Hospital Essen, University of Duisburg-Essen; 3Department of Hematology, Oncology and Clinical Immunology, Heinrich Heine University Düsseldorf, University Hospital, Germany 1
*Y.S.A-M. and L.B. contributed equally to this work.
ABSTRACT
T
Correspondence: cyrus.khandanpour@uk-essen.de
Received: January 20, 2016. Accepted: July 07, 2016. Pre-published: July 07, 2016. doi:10.3324/haematol.2016.143180
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1216
he growth of malignant cells is not only driven by cell-intrinsic factors, but also by the surrounding stroma. Monocytes/Macrophages play an important role in the onset and progression of solid cancers. However, little is known about their role in the development of acute myeloid leukemia, a malignant disease characterized by an aberrant development of the myeloid compartment of the hematopoietic system. It is also unclear which factors are responsible for changing the status of macrophage polarization, thus supporting the growth of malignant cells instead of inhibiting it. We report herein that acute myeloid leukemia leads to the invasion of acute myeloid leukemia-associated macrophages into the bone marrow and spleen of leukemic patients and mice. In different leukemic mouse models, these macrophages support the in vitro expansion of acute myeloid leukemia cell lines better than macrophages from non-leukemic mice. The grade of macrophage infiltration correlates in vivo with the survival of the mice. We found that the transcriptional repressor Growth factor independence 1 is crucial in the process of macrophage polarization, since its absence impedes macrophage polarization towards a leukemia supporting state and favors an anti-tumor state both in vitro and in vivo. These results not only suggest that acute myeloid leukemia-associated macrophages play an important role in the progression of acute myeloid leukemia, but also implicate Growth factor independence 1 as a pivotal factor in macrophage polarization. These data may provide new insights and opportunities for novel therapies for acute myeloid leukemia.
Introduction ©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1216
The growth of various solid tumors, lymphomas and leukemias is not only the result of cell-specific changes at the genetic and epigenetic level, but is also affected by the surrounding microenvironment, the stroma and the cells therein.1-4 The stroma is composed of many different cell types, among them fibroblasts, mesenchymal stem cells, vascular cells and a variety of immune cells including T and B lymphocytes, natural killer cells (NK-cells), neutrophils and macrophages.4 Tumor cells induce the stroma and immune cells to express and partially secrete various factors and cytokines that promote the growth of the tumor cells, instead of activating the immune system to battle the malignant cells.5,6 This process of “polarization” is the result of a complex bidirectional interaction between the tumor and the stroma cells. haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
Hence, the polarized macrophages in tumors are called tumor-associated macrophages (TAMs).5 The plasticity of macrophages is mostly tissue-specific and regulated by local and systemic signals.7 In response to different signals derived from the surrounding tissue, bacteria or activated lymphocytes, macrophages can differentiate into various polarization states with distinct functional phenotypes.8 Although considered a simplification,9 the M1/M2 is a straightforward classification for functionally distinct types of macrophages. M1 macrophages, known as classically activated macrophages, are stimulated by bacterial lipopolysaccharide (LPS), interferon-γ (IFN-γ), tumor necrosis factor (TNF)-α or granulocyte-macrophage colony-stimulating factor (GM-CSF), and are characterized by the production of numerous antimicrobial agents and inflammatory mediators, such as interleukin 6 (IL-6), reactive oxygen species (ROS) and nitric oxide (NO).10 The M1 macrophages are involved in the host defense against different pathogens and play a role in anti-tumor immunity. In contrast, M2 macrophages or alternatively activated macrophages have anti-inflammatory activity and are stimulated by interleukin 4 (IL-4) or interleukin 13 (IL-13). They secrete arginase, metalloproteinases, transforming growth factor-β (TGFβ), interleukin 10 (IL-10) and other cytokines that cause immune suppression, angiogenesis and tissue repair.11 M2 macrophages have been further subdivided into M2a, M2b, M2c and M2d macrophages, according to the polarizing cytokines.12 In contrast to M1 macrophages, which suppress tumor growth, M2 macrophages play an important role in the development and progression of different tumors,13,14 and are therefore also known as TAMs. Despite a good understanding of the role of macrophages in solid tumors, little is known about the interaction between stroma cells and leukemic cells. Leukemic stem cells (LSCs) can modify the bone marrow (BM) niche in such a way that it supports the growth of LSCs instead of hematopoietic stem cells (HSCs).15 This might enhance the LSCs quiescence, leading to chemotherapy resistance.1,16-19 A recent study reported that the inhibition of SIRPα signalling in macrophages impairs engraftment of human LSCs in immunocompromised NSG mice.20 Clinically, the accumulation of TAMs in the lymph nodes of patients with classic Hodgkin lymphoma was associated with a poor prognosis.21 The most common form of adult leukemia is acute myeloid leukemia (AML),22 which is characterized by an accumulation of myeloid blast cells in the BM. As AML patients have a poor prognosis,22 novel therapy approaches are urgently needed. Furthermore, the function of AML-associated macrophages (AAMs) and their role in AML progression remains to be further investigated. Transcription factors, key elements of gene regulation, show a distinct expression pattern and organ specificity. One such transcription factor is Growth factor independence 1 (Gfi1), a transcriptional repressor that plays an important role in HSCs maintenance and quiescence, and is crucial for normal lymphoid and myeloid hematopoiesis.23-25 Gfi1-deficient mice are characterized by severe neutropenia and an overproduction of TNF-α and other inflammatory mediators of macrophages when exposed to bacterial endotoxin or LPS.26 Using different mouse models of human AML we report herein that AAMs support the expansion of AML cells both in vivo and in vitro. Furthermore, we show that Gfi1 has an important role in the process of macrophage polarization. haematologica | 2016; 101(10)
Methods Human BM samples Human BM samples were obtained following the informed consent of all subjects. All experiments with human samples were carried out in accordance with the approved protocol of the University of Duisburg-Essen ethics committee. The diagnosis of AML was confirmed based on cytological and flow cytometry examination.22,27
Mouse strains NUP98-HOXD13 transgenic mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). The Gfi1-KO mice have been previously described.28 Wild-type (WT) mice (C57BL/6J) were provided by the animal facility of the University Hospital Essen. All animals were housed in single ventilated cages and specific pathogen-free conditions at the animal facility of University Hospital Essen. All animal experiments were carried out in accordance with the protocol of the government ethics committee for animal use, which on 21.07.2011 approved all studies on animals under document number G1196/11.
AML cell lines C1498GFP, a murine AML cell line,29 was a kind gift from Dr. Justin Kline from the University of Chicago, USA. The cells were maintained in DMEM (Gibco, Life Technologies, Darmstadt, Germany), supplemented with 10% fetal bovine serum (FBS) (PANTM BIOTECH, Aidenbach, Germany) and 1% penicillin/streptomycin (Gibco).
Statistics A student's t-test was applied to calculate the differences between various groups. For the survival analysis, a Kaplan-Meier test was performed. Differences were considered to be significant when the P-value was <0.05. The Graph Pad (version 6) software was used for applying all significance tests.
Results AAMs proliferate and accumulate in the BM of AML patients The expression of CD163 has been reported to be restricted to monocytes/macrophage lineages.30 Recently, CD163+ M2 TAMs have been reported to be involved in tumor progression in several hematological malignancies such as multiple myeloma31 or classical Hodgkin lymphoma (CHL).32 A common cell surface marker identified in TAMs is CD206.33 To explore the ability of AML cells to educate macrophages and affect their polarization, we examined the rate of infiltration of CD163+CD206+ M2-like macrophages in the BM of AML patients and healthy volunteers (Online Supplementary Table S1). The frequency of CD163+CD206+ M2-like macrophages in the BM of AML patients was significantly elevated compared to healthy volunteers (Online Supplementary Figures S1A-S1C).
Leukemic cells polarize non-leukemic monocytes/macrophages that proliferate and accumulate in BM and spleen of recipient mice To investigate the molecular mechanisms and the role of monocytes/macrophages in the development of AML, we used different established murine models of human AML. AML1-ETO9a, the product of the t(8;21)(q22;q22) translocation, and MLL-AF9, the product of the t(9;11)(p22;q23) translocation, are commonly involved in AML pathogenici1217
Y.S. Al-Matary et al. ty in humans, and are also used to model AML in mice.34,35 While AML1-ETO9a-induced AML is associated with a rather good prognosis, MLL-AF9-driven AML has a rather To study the role of bad prognosis.34,35 monocytes/macrophages in AML, we transduced lineage
negative (Lin–) BM cells from WT mice with retroviruses encoding MLL-AF9 or AML1-ETO9a cDNA fused to an IRES-GFP gene cassette, and transplanted these cells into lethally irradiated mice together with 1.5x105 competitive BM cells. Leukemic BM cells were then re-transplanted into
A
B
C
D
E
F
Figure 1. AML-associated monocytes/macrophages (AAM) proliferate and accumulate in the BM and spleen of AML mice. (A) Lin– BM cells from WT mice were transduced either with MLL-AF9 or AML1-ETO9a retroviruses and 1*105 MLL-AF9 or 5-7*105 AML1-ETO9a GFP+ cells were transplanted into lethally irradiated (10Gy) primary recipient mice together with 5*105 competitive BM cells. Leukemic BM cells (1*105 GFP+ cells) were then re-transplanted into secondary sublethally irradiated (3Gy) mice. Macrophage surface markers from leukemic mice were subsequently analyzed by flow cytometry. (B) Representative gating strategy for GFP–CD11bhiGr1int monocytes/macrophages in BM cells derived from mice transplanted with non-transduced (left panel) or AML1-ETO9a-transduced cells (right panel). (C) The frequency of non-leukemic GFP–CD11bhiGr1int monocytes/macrophages in the BM (left panel) and spleen (right panel) of leukemic mice transplanted with MLL-AF9 (n=5) or AML1-ETO9a transduced cells (n=5) compared to mice transplanted with non-transduced cells (n=4), (***P<0.0008 for BM, **P<0.001 for spleen). (D) Representative gating strategy for GFP–CD11b+Ly6G– monocytes/macrophages in BM cells derived from mice transplanted with non-transduced or AML1ETO9a-transduced cells. (E) The frequency of non-leukemic GFP–D11b+Ly6G– macrophages in the BM (left) or spleen (right) of transplanted leukemic mice (n=5 for MLL-AF9 and n=5 for AML1-ETO9a), compared to mice transplanted with non-transduced cells (n=4), (***P<0.0001 for BM, *P=0.04 and ***P=0.0002 for spleen). (F) 2-3x105 BMDMs from mice transplanted with non-transduced or AML1-ETO9a or MLL-AF9-transduced cells were co-cultured with 5*104 C1498GFP cells for 6 days (left panel). Fold change of C1498GFP live cell numbers is given (right panel). Results from triplicates of 3 independent experiments for mice transplanted with MLL-AF9 (n=9) and AML1-ETO9a (n=9) transduced cells and 4 independent experiments for mice transplanted with non-transduced cells (n=12) are shown, *P=0.03 for AML-ETO9a and ***P<0.001 for MLL-AF9 transgenic cells). BM: bone marrow; AAMs: acute myeloid leukemia associated macrophages; WT: wild-type; BMDM: bone marrow-derived macrophage; AML: acute myeloid leukemia.
1218
haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
secondary, sublethally irradiated recipient mice (Figure 1A). The expression of GFP alongside the expression of either of the two different oncofusion proteins by the transduced pre-leukemic cells enabled the differentiation between leukemic and non-leukemic cells. To minimize any potential bias as a result of the irradiation, we used control mice that were sublethally irradiated but received only WT BM cells from healthy mice. In the BM and spleen of leukemic secondary recipient mice we first determined the fraction of GFP– AAMs defined as GFP–CD11bhiGr1int.28 The frequency of GFP– AAMs in the BM and spleen of leukemic mice was significantly higher than in sublethally irradiated mice transplanted with competitive normal BM cells only (Figure 1B,C). Also, when we defined AAMs as GFP–CD11b+Ly6G– cells36 (Figure 1D), we found similar results (Figure 1E). To confirm our findings and in order to rule out any effects of irradiation, we used the NUP98-HOXD13 transgenic mouse model that mimics the t(2;11)(q31;p15) translocation, which is associated with human myeloid malignan-
cies. These mice show features of human myelodysplastic syndrome (MDS), and some mice develop AML.37 Similarly, the percentage of AAMs in the BM and spleen of leukemic NUP98-HOXD13 transgenic mice was higher than in WT non-leukemic mice (Online Supplementary Figures S2A and S2B). We confirmed that, phenotypically, in both the GFP–CD11bhiGr1int and GFP–CD11b+Ly6G– monocyte population the expression of F4/80, the typical marker for BM macrophages, was more than 90% and 70%, respectively (Online Supplementary Figure S2C and S2D). We then tested whether these AAMs would support the growth of murine AML cells in vitro. We co-cultured BMderived macrophages (BMDMs) with the murine AML cell line C1498GFP for 6 days, counted the non-adherent C1498GFP cells and determined the number of GFPexpressing leukemic cells by flow cytometry. BMDMs from transplanted leukemic mice supported the proliferation of the C1498GFP cells better than BMDMs from nonleukemic mice (Figure 1F).
A
B
C
D
E
F *
haematologica | 2016; 101(10)
Figure 2.Characterization of AAMs by flow cytometry. (A) Schematic illustration of the experimental design. BM cells from MLL-AF9 or AML1-ETO9a leukemic primary recipient mice or C1498GFP murine AML cells were transplanted into sublethally irradiated (3Gy) secondary recipient mice. When moribund, the mice were sacrificed and different macrophage classes were analyzed by flow cytometry. (B) Representative FACS plots from the BM of mice transplanted either with nontransduced or with AML1-ETO9a transduced cells showing the gating strategy used for classifying different types of macrophages according to the expression of Ly6C and MHCII markers. Cells with a GFP–CD11b+Ly6G–MHCII-Ly6C– phenotype were considered AAM1. C) The frequency of AAM1 in the BM (left panel) and spleen (right panel) of leukemic mice transplanted with AML1-ETO9a, (n=4), MLL-AF9 (n=4) or C1498GFP (n=3) compared to mice transplanted with non-transduced cells (n=5), (***P<0.0001, **P=0.001). (D) Representative FACS plots showing macrophage classes in the BM of the Gfi1-WTxNUP98-HOXD13 mouse model. (E) The frequency of AAM1 cells in the BM of leukemic NUP98-HOXD13 mice (n=6) compared to WT mice (n=3), (*P=0.04). F) Survival of the leukemic NUP98HOXD13 mice is inversely correlated with the percentage of AAM1 in the BM (R square=0.92). BM: bone marrow, AML: acute myeloid leukemia, AAMs: acute myeloid leukemia associated macrophages, WT: wild-type; Gfi1: growth factor independent 1.
1219
Y.S. Al-Matary et al.
Characterization of AAMs Macrophages are characterized by specific gene expression patterns, cytokine secretion and cell surface molecules.7
By using a similar gating strategy for studying TAMs in lung cancer, as reported earlier,36 we quantified the different mononuclear phagocyte subsets in the BM and spleen of
A
B
n=6 C
E D
n=8
n=6
F
Figure 3. Characterization of AAMs by RT-PCR and ELISA. (A) Schematic illustration of the experimental design. 1x105-4x105 C1498GFP were transplanted into sublethally irradiated (3Gy) secondary recipient mice. When the mice developed AML, GFP–CD11b+Ly6G– BM macrophages were sorted for further experiments. (B) Cytospins were prepared from sorted AAMs (GFP–CD11b+Ly6G–) and stained according to the May-Grunwald Giemsa protocol. Bar represents 20μm. (C) Fold change of Arg1, IL-6 and Nos2 mRNA levels in sorted AAMs from non-leukemic mice transplanted with WT BM cells (n=6) and leukemic mice transplanted with C1498GFP cells (n=6), normalized to GAPDH. Results of duplicates from three independent experiments are shown (**P=0.006 for Arg1, **P=0.005 for IL-6 and ***P<0.0001 for Nos2). (D) 5x105 AAMs sorted from leukemic mice transplanted with C1498GFP cells (n=8) or 5x105 CD11b+Ly6G– non-leukemic macrophages sorted from mice transplanted with WT BM cells (n=8) were cultured in DMEM/glutamax supplemented with 10% FBS and 1% Pen/Strep. After 24 hours medium was collected, filtered and the levels of IL-10 were measured using an ELISA commercial kit. Results of duplicates from four independent experiments are shown (*P=0.01). (E) Fold change of Gfi1 mRNA level in sorted AAMs from leukemic mice transplanted with C1498GFP cells (n=6) and non-leukemic mice transplanted with WT BM cells (n=6), normalized to GAPDH. Results of duplicates from three independent experiments are shown (**P=0.001). (F) 5x104 C1498GFP+ cells were co-cultured with 1.5x104 sorted GFP–CD11b+Ly6G– cells (left panel). The numbers of C1498GFP+ cells in the presence (n=9) or absence (n=9) of sorted AAMs are shown (right panel). Results of triplicates from three independent experiments are given (***P=0.0009). BM: bone marrow, AML: acute myeloid leukemia, AAMs: acute myeloid leukemia associated macrophages, WT: wild-type; Gfi1: growth factor independent 1; BMDM: bone marrow derived macrophage; Arg1: arginase 1; Nos2: nitric oxide synthase 2; IL-6: interleukin 6; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; IL-10: interleukin 10.
1220
haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
sublethally irradiated mice transplanted either with C1498GFP cell line or with MLL-AF9 or AML1-ETO9a leukemic BM cells from primary recipient mice (Figure 2A). Depending on the expression levels of Ly6C and MHCII surface markers, the GFP–CD11b+Ly6G– monocytes/
macrophages from non-leukemic and leukemic mice were divided into six populations (Figure 2B).36,38 In all leukemic mouse models, we found that not only the frequency (Figure 2C) but also the absolute numbers (Online Supplementary Figure S3A and S3B) of AAM1 cells, which are
A
B
C
E
D
F
G
I
H
Figure 4. Gfi1 enhances M2 polarization by IL-4 and suppresses the M1 polarization of macrophages by LPS in vitro. (A) Schematic representation of the in vitro polarization experiment. Gfi1-WT or Gfi1-KO BMDMs were stimulated with LPS (100 ng/ml) or IL-4 (20ng/ml) for 48 hours. The medium was collected for ELISA and M1 and M2 macrophages were characterized by flow cytometric and gene expression analysis. (B) Representative FACS plots of Ly6C+CD206– M1(LPS) macrophages from Gfi1-WT and Gfi1-KO BMDMs (left panel). The frequency of polarized Ly6C+CD206– M1(LPS) macrophages from Gfi1-WT (n=6) and Gfi1-KO (n=6) BMDMs (right panel), (***P<0.0001). Results of duplicates from three independent experiments are shown. (C) Fold change of IL-6 and Nos2 mRNA levels in Gfi1-WT (n=4) and Gfi1-KO (n=4) M1(LPS) macrophages, normalized to GAPDH. Results of duplicates from two independent experiments are shown (*P=0.03 for IL-6, **P=0.002 for Nos2). (D) The levels of IL-1B (left panel) and IL-6 (right panel) in the supernatants of Gfi1-WT (n=8) and Gfi1-KO (n=8) M1(LPS) macrophages. Results of duplicates from four independent experiments are shown (*P=0.05 for IL-1B, ***P<0.0001). (E) Fold change of Arg1 mRNA level in Gfi1-WT (n=4) and Gfi1-KO (n=4) M2(IL-4) macrophages, normalized to GAPDH. Results of duplicates from two independent experiments are shown (*P=0.04). (F) The levels of IL-10 in supernatants of Gfi1WT (n=8) and Gfi1-KO (n=8) M2(IL-4) macrophages. Results of duplicates from four independent experiments are shown (**P=0.004). G) Schematic representation of the experimental design for simultaneous in vitro polarization of M1 and M2 macrophages. Gfi1-WT or Gfi1-KO BMDMs were stimulated with both LPS (100 ng/ml) and IL-4 (20ng/ml) for 48 hours and M1 and M2 macrophages were characterized by flow cytometry. (I) Representative FACS plots showing different macrophage classes derived from Gfi1-WT or Gfi1-KO mice polarized by both LPS and IL-4. H) BMDMs from Gfi1-WT (n=4) and Gfi1-KO (n=4) mice were polarized for 48 hours with LPS and IL-4. The frequency of Ly6C–CD206+ M2 macrophages (left panel), (***P<0.0001), Ly6C+ CD206+ macrophages (middle, **P=0.002) and Ly6C+CD206–M1 macrophages (right, P<0.0001) are shown. Results of duplicates from two independent experiments are shown. LPS: lipopolysaccharide, BMDMs: bone marrow derived macrophages; Arg1: arginase1; Nos2: nitric oxide synthase 2; IL-6: interleukin 6; IL-4: interleukin 4; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; WT: wild-type; Gfi1: growth factor independent 1; IL-10: interleukin 10: IL-1B: interleukin 1β.
haematologica | 2016; 101(10)
1221
Y.S. Al-Matary et al. equivalent to the TAM1 phenotype (Ly6C-MHCII-), as well as the frequency of Ly6CintMHCIIlow immature leukemic macrophages38 (Online Supplementary Figure S4A and S4B) were significantly increased in the BM and spleen, whereas the frequency of Ly6C+MHCII– monocytes and the other macrophage subsets were decreased or not significantly changed (Online Supplementary Figure S4A and S4B). We confirmed our findings in the NUP98-HOXD13 mouse model, where the frequency of AAM1 in the BM and spleen of leukemic transgenic mice was higher than in the WT non-leukemic mice (Figure 2D,E). Notably, the survival of the leukemic NUP98-HOXD13 mice was inversely correlated with the percentage of AAM1 in the BM (Figure 2F). Evaluation of Wright-Giemsa stained cytospin preparations of sorted GFP–CD11b+Ly6G– cells derived from C1498GFP transplanted leukemic mice, confirmed that these cells were indeed macrophages (Figure 3A,B). Furthermore, they expressed significantly higher levels of Arg1 mRNA (Figure 3C, left panel), which is characteristic for M2 macrophages with tumor-promoting functions.39 In contrast, the expression of IL-6 and Nos2 mRNA, character-
istic for M1 macrophages,10 were decreased compared to macrophages sorted from non-leukemic mice (Figure 3C, middle and right panel). To further investigate the status of macrophage polarization, GFP–CD11b+Ly6G– sorted cells were cultured in DMEM-Glutamax medium supplemented with 10% FBS, and after 24 hours the level of IL-10 secreted in the culture medium was measured. The production of IL10, which is characteristic of the M2 activation profile, was significantly increased in AAMs from leukemic mice compared to macrophages from non-leukemic mice (Figure 3D). There were no significant differences with regard to the secretion of IL-6 and IL-1β that are characteristic of M1 macrophages10 (data not shown). Since Gfi1 is a transcription factor with an important role in macrophage development,25,40 we next examined its expression in AAMs. Gfi1 expression was about two-fold upregulated in AAMs compared to non-leukemic macrophages (Figure 3E), indicating that higher levels of Gfi1 might be necessary for macrophage polarization. To investigate whether these AAMs can support the growth of leukemic cells in vitro, we co-cultured sorted GFP–CD11b+Ly6G– AAMs from leukemic mice with the
A
B
C
D
E
Figure 5. Gfi1 is involved in the polarization of M2 macrophages by C1498GFP AML cell line in vitro. (A) Schematic representation of the procedure for co-culturing of BMDMs from Gfi1-WT or Gfi1-KO mice with C1498GFP murine AML cell line followed, after 3 days, by flow cytometric and gene expression analysis. (B) Representative FACS plots showing the frequency of Ly6C–CD206+ M2 macrophages derived from Gfi1-WT mice co-cultured in the presence or absence of C1498GFP cells (left panel) and the corresponding quantification of MFI for CD206 surface marker expression (right panel), (*P=0.02). Results of duplicates from three independent experiments are shown. (C) Fold change of Arg1 and Gfi1 mRNA expression in Gfi1-WT BMDMs cultured in the presence (n=6) or absence (n=6) of C1498GFP cells, normalized to GAPDH. RT-PCR results of duplicates from three independent experiments are shown (*P=0.02 for Arg1 and ***P<0.0001 for Gfi1). (D) Fold change in Arg1 mRNA expression in Gfi1-WT (n=4) and Gfi1-KO (n=4) BMDMs co-cultured with C1498GFP cells, normalized to GAPDH. RT-PCR results of duplicates from two independent experiments are shown (**P=0.004). (E) The level of IL-6 in supernatants of macrophages from Gfi1-WT (n=4) or Gfi1-KO (n=4) co-cultured with C1498GFP cells for 3 days. Results of duplicates from two independent experiments are shown (*P=0.02 and **P=0.003). AML: acute myeloid leukemia; BMDMs: bone marrow derived macrophages; MFI: mean fluorescence intensity; Arg1: arginase1; IL-6: interleukin 6; IL-4: interleukin 4; GAPDH: glyceraldehyde 3-phosphate dehydrogenase; WT: wild-type; Gfi1: growth factor independent 1; RT-PCR: real-time PCR.
1222
haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
murine C1498GFP AML cell line for 48 hours. The growth/proliferation of C1498GFP cells was significantly increased in the presence of AAMs (Figure 3F). Together, these results indicate that the frequency and absolute numbers of AAM1 are increased in the BM of leukemic mice. Furthermore, these AAMs exhibit features of M2 macrophages.
The role of Gfi1 in macrophage polarization in vitro To assess whether Gfi1 can affect macrophage polarization in response to M1 or M2 stimuli, Gfi1-KO and Gfi1-WT BMDMs were cultured in the presence of either LPS or INF-γ, which are both M1 stimulators, or IL-4, an M2 stimulator8,11 (Figure 4A, Online Supplementary Figure S5A). In the absence of Gfi1, LPS or INF-γ activation resulted in a M1 response as demonstrated by a 2-4-fold increase in the frequency of Ly6C+CD206– M1 macrophages (Figure 4B, Online Supplementary Figure S5B and S5C). Furthermore,
Gfi1-KO M1(LPS) macrophages expressed significantly increased IL-6 and Nos2 mRNA levels and secreted more IL6 and IL-1B (Figure 4C,D). Also, in Gfi1-KO M1(INF-γ), there was an almost 3-fold increase in Nos2 mRNA levels, (Online Supplementary Figure S5D) and 2-fold increase in IL1B secretion (Online Supplementary Figure S5E). Although, phenotypically, there was no difference between the frequencies of M2-polarized macrophages derived from Gfi1WT and Gfi1-KO mice (data not shown), IL-4 stimulation resulted in an M2 response in Gfi1-WT but not in the Gfi1KO macrophages, as demonstrated by a significant increase in Arg1 mRNA expression in Gfi1-WT macrophages (Figure 4E) and IL-10 secretion (Figure 4F). In vivo, polarization of M1 and M2 macrophages can take place simultaneously depending on the signals and cytokines secreted from the tumor microenvironment. In an attempt to mimic the in vivo conditions, Gfi1-WT and Gfi1-KO BMDMs were challenged in vitro with both LPS and IL-4, and M1 and M2 sur-
A
C
B
P=0.01
D
E
Figure 6. The role of Gfi1 in polarization of AAMs in transplanted leukemic mice. (A) Schematic illustration of the experimental design. Sublethally irradiated (3Gy) Gfi1-WT or Gfi1-KO mice were transplanted with 1x105 Gfi1-WT MLL-AF9 GFP+ leukemic BM cells derived from primary recipient mice. The mice were monitored and sacrificed and analyzed when moribund. BMDMs from Gfi1-WT or Gfi1-KO mice were co-cultured with C1498GFP AML cells and after 6 days, C1498GFP counts were evaluated. (B) Kaplan-Meier survival curve of Gfi1-KO (n=3) and Gfi1-WT (n=6) transplanted with Gfi1-WT MLL-AF9 leukemic cells (P=0.01). (C) Total white blood cell count (WBC) in peripheral blood (left) (*P=0.02) and the number of GFP+ leukemic blast cells in the BM (right) (*P=0.04) of Gfi1-WT (n=4) and Gfi1-KO (n=3) leukemic mice. (D) The frequency of GFP–CD11bhiGr-1int non-malignant macrophages in the BM (left panel) and spleen (right panel) of Gfi1-WT (n=6) and Gfi1-KO (n=3) transplanted with MLL-AF9 transduced cells compared to mice transplanted with non-transduced cells (n=4) (*P<0.01, **P=0.001). (E) Fold change of live C1498GFP cell number after 6 days of co-culturing with BMDMs from Gfi1-WT, Gfi1-KO MLL-AF9 transplanted leukemic mice or from mice transplanted with non-transduced cells. Results of triplicates from 3 and 4 independent experiments for Gfi1-WT leukemic (n=9) and non-leukemic mice (n=12) and from 1 experiment for Gfi1-KO leukemic mice (n=3) are shown (**P=0.008 and ***P=0.0004). BM: bone marrow; BMDM: bone marrow-derived macrophage; AML: acute myeloid leukemia; WT: wild-type; Gfi1: growth factor independent 1; GFP: green fluorescent protein: AAM: acute myeloid leukemia associated macrophage.
haematologica | 2016; 101(10)
1223
Y.S. Al-Matary et al.
face marker expressions were examined by flow cytometry (Figure 4G). In the presence of both stimuli, more than 60% of Gfi1-WT BMDMs were polarized into Ly6C–CD206+ M2like macrophages without any differentiation into Ly6C+CD206– M1 macrophages (Figure 4H,I), whereas Gfi1-
KO BMDMs showed less efficient CD206+Ly6C– M2 polarization and enhanced differentiation into Ly6C+CD206+ and Ly6C+CD206– M1 macrophages (Figure 4H,I). Together, these findings suggest that Gfi1 directs macrophage polarization towards a M2-like macrophage state.
A
C
B
P=0.02
E D
F
Figure 7. The role of Gfi1 in polarization of AAMs in vivo. (A) Schematic representation of the experimental design. Gfi1-WT and Gfi1-KO mice were crossed to NUP98HOXD13 MDS/AML mouse model. Double transgenic mice were monitored for AML onset and survival. Leukemic mice were analyzed to determine the frequency of different macrophage types. (B) Kaplan-Meier survival curve of Gfi1-KO (n=17) and Gfi1-WT (n=39) NUP98-HOXD13 AML mice (P=0.02). (C) Total white blood cells count (WBC) in peripheral blood (left) (*P=0.02) and the percentage of blasts in the BM (right) (*P=0.04) of Gfi1-WT (n=6) and Gfi1-KO (n=5) NUP98-HOXD13 leukemic mice. (D) Representative FACS plots showing the frequency of AAM1 in a Gfi1-WTxNUP98-HOXD13 and a Gfi1-KOxNUP98-HOXD13 leukemic mouse. (E) The frequency of Ly6C–MHCII– AAM1 in the BM (right) and spleen (left) of Gfi1-WTxNUP98-HOXD13 (n=6) and Gfi1-KOxNUP98-HOXD13 (n=5) leukemic mice (*P=0.02 for BM and *P=0.05 for spleen). (F) The frequency of Ly6Chi monocytes in the BM (right) and spleen (left) of Gfi1-WTxNUP98-HOXD13 (n=6) and Gfi1KOxNUP98-HOXD13 (n=5) leukemic mice (**P=0.005 for BM and *P=0.01 for spleen). AML: acute myeloid leukemia; AAM: AML-associated macrophages; BM: bone marrow; Gfi1: growth factor independent 1; WT: wild-type.
1224
haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
To investigate the effect of AML cells on the macrophage phenotypes in vitro, we co-cultured Gfi1-WT and Gfi1-KO BMDMs with C1498GFP cells for 3 days (Figure 5A). Coculture of Gfi1-WT BMDMs with C1498GFP cells significantly upregulated CD206 expression on macrophages (Figure 5B) and resulted in an increased expression level of Arg1 mRNA (Figure 5C, left panel). Interestingly, Gfi1 was found to be highly upregulated in Gfi1-WT BMDMs co-cultured with C1498GFP cells (Figure 5C, right panel). Although, phenotypically, there was no difference in M1 or M2 macrophages polarization between Gfi1-WT and Gfi1KO cultured in the presence of C1498GFP cells, Gfi1-KO BMDMs showed a M1 response, as demonstrated by lower levels of Arg1 mRNA (Figure 5D) and a significant increase in IL-6 secretion compared to Gfi1-WT BMDMs (Figure 5E), confirming that the loss of Gfi1 shifts the macrophage phenotype towards an M1-like activation profile.
The role of Gfi1 in polarization of AAMs in vivo To test the relevance of these findings and to investigate the effect of Gfi1 ablation on the growth of leukemic cells in vivo, we transplanted Gfi1-WT MLL-AF9-expressing BM cells into sublethally irradiated secondary Gfi1-WT and Gfi1-KO mice (Figure 6A). Gfi1-KO mice that received MLL-AF9-expressing cells survived longer (Figure 6B) and had a significantly lower white blood cell (WBC) count in peripheral blood (PB) (Figure 6C, left panel), reduced numbers of GFP+ leukemic cells in the BM (Figure 6C, right panel) and decreased frequency of non-malignant macrophages (GFP–CD11bhiGr1int) in the BM and spleen (Figure 6D), compared to Gfi1-WT mice transplanted with MLL-AF9-expressing cells. To further study the role of Gfi1 in macrophage function, we co-cultured BMDMs from Gfi1-WT and Gfi1-KO leukemic mice with C1498GFP cells and found that Gfi1-KO BMDMs did not support the growth of C1498GFP cells in vitro to the same extent as Gfi1WT BMDMs (Figure 6E). We validated these results in the NUP98-HOXD13 transgenic mouse model. We crossed these mice with Gfi1-WT or Gfi1-KO mice and analyzed their survival and the frequency of different macrophage classes in the BM and spleen of NUP98-HOXD13-expressing mice that developed AML (Figure 7A). In agreement with the results presented above, the Gfi1-KOxNUP98-HOXD13-expressing leukemic mice survived longer (Figure 7B), and were characterized by lower numbers of WBCs in PB and decreased frequency of blast cells in the BM (Figure 7C), compared to Gfi1WTxNUP98-HOXD13-expressing leukemic mice. Furthermore, Gfi1-KOxNUP98-HOXD13 leukemic mice had a significantly decreased frequency of AAM1 in the BM and spleen compared to Gfi1-WTxNUP98-HOXD13 leukemic mice (Figure 7D,E). Other macrophage populations such as immature macrophages, AAM2s and AMLassociated dendritic cells (ADCs) were also decreased in Gfi1-KOxNUP98-HOXD13-expressing leukemic mice (Online Supplementary Figure S6). The frequency of Ly6C+MHCII– monocytes from which the different macrophage populations are derived was increased in the BM and spleen of Gfi1-KOxNUP98-HOXD13 compared to Gfi1-WTxNUP98-HOXD13 leukemic mice (Figure 7F), suggesting that monocytes from Gfi1-KOxNUP98-HOXD13 mice differentiate less efficiently into more mature macrophages than monocytes from Gfi1-WTxNUP98HOXD13 mice. Taken together, all of these results suggest that AAMs haematologica | 2016; 101(10)
play an important role in the progression of AML, and Gfi1 is crucial in the process of macrophage polarization, since its absence impedes macrophage polarization towards a leukemia-supporting state and favors an anti-tumor state.
Discussion We investigated the interaction between AAMs and murine AML cells in vivo and in vitro. We observed an increased accumulation of monocytes/macrophages in the BM of AML patients and in the BM and spleen of several AML mouse models, indicating that the leukemic cells might induce BM monocyte/macrophage proliferation and/or infiltration. In addition, we found the same pattern of monocytes/macrophages infiltration in a NUP98HOXD13 transgenic MDS/AML mouse model. This suggests that the presence of AML and the leukemic environment leads to an infiltration of monocytes/macrophages and promotes their differentiation into AAMs. In the case of the very aggressive type of the MLL-AF9 induced AML, the absolute number of AAMs in the BM of the leukemic mice is lower than in the BM of healthy mice (data not shown). Our hypothesis is that the MLL-AF9 leukemic cells overgrow all other cells, including the AAMs. However, in all cases, the relative percentage of AAMs in the BM of leukemic mice was always increased compared to the situation found in the BM of healthy mice, and the functional changes of AAMs, with regard to supporting the growth of leukemic cells, were similar from one type of AML to the next. The supporting role of TAMs in the growth of tumor cells has been studied in a number of different types of solid cancers.41 Initially, the concept of M1 and M2 macrophages have been helpful in exploring the new field of TAMs,13,41-43 but it has been recently redefined. For example, what we describe herein as M2 macrophages44 has recently been proposed to be IL-4 macrophages, and the M1 macrophages as LPS or IFN-γ macrophages.41 Also, distinct expression profiles and secretion patterns have been used to better characterize different macrophage classes.9,45 Although TAMs are mostly M2-like macrophages, some studies showed that TAMs have a gene expression profile similar to both, M1- or M2-like macrophages.36 We have demonstrated that, phenotypically, AAMs derived from the BM and spleen of leukemic mice were M2-like macrophages (Ly6C–MHCII–)33,36 that express higher levels of Arg1 and lower levels of IL-6 and Nos2 mRNA, and secrete more IL-10 than non-leukemic macrophages. The decrease in the frequency of Ly6C+MHCII– monocytes in the BM and spleen of leukemic mice, and the increased numbers of Ly6C–MHCI– AAMs compared to non-leukemic mice, suggest that AAMs might be derived from Ly6C+MHCII– monocytes. On the other hand, the accumulation of Ly6CintMHCII– immature macrophages, which are the intermediate stage between Ly6C+MHCII– monocytes and Ly6C–MHCII– AAMs36,38 in the BM and spleen of leukemic mice, indicates that the differentiation process of Ly6C+MHCII– monocytes towards an AAM phenotype is active during leukemia development. In our first set of experiments, mice were subjected to sublethal irradiation to enable the engraftment of leukemic cells. It is known that irradiation can alter the stroma microenvironment to support the malignant transformation46 or to alter the macrophage subtypes.13,47 However, to 1225
Y.S. Al-Matary et al.
ensure comparability, we always correlated our findings to sublethally irradiated mice transplanted with wild-type, non-malignant BM cells. In terms of the functional characterization of AAMs in vitro, we cannot exclude that the differentiation of AAMs via M-CSF might alter their function, but as we obtained similar results in a murine model of AML in which AAMs were sorted and co-cultured with AML cells without prior M-CSF co-culture, we believe that the cytokine-induced differentiation is not per se artificial. As Gfi1 is required for the differentiation and maturation of HSCs into myeloid and lymphoid cells,25,40 we hypothesized that Gfi1 might play an important role in the polarization of macrophages in leukemic mice. It is known that within the myeloid lineage/compartment, Gfi1 favors the differentiation towards granulocytes and impedes monocyte development.24-26,28 However, it has been shown that there is a discrepancy between reduced Gfi1 mRNA levels and elevated Gfi1 protein levels in monocytes.48 Thus, despite lower Gfi1 expression at the mRNA level, Gfi1 is present at the protein level, and is required for the proper differentiation of monocytes towards macrophages and other monocyte-derived cell types.48 In our experiments, Gfi1 was 2-fold upregulated at mRNA levels in AAMs derived from the BM of transplanted leukemic mice and in macrophages co-cultured with AML cells, indicating that Gfi1 indeed plays a role in macrophage differentiation. Leukemic Gfi1-KO mice survived longer, and had a lower percentage of leukemic cells in PB and BM and decreased numbers of AAMs than Gfi1-WT leukemic mice. These results indicate that various Gfi1-deficient stroma elements, including AAMs, were not well polarized to support the growth of AML cells in vivo. This might be explained by the fact that the loss of Gfi1 shifts the cells toward a M1-like activation profile, which counteracts the growth of malignant cells rather than supporting it. It could be argued that Gfi1-deficient macrophages are too different from their WT counterparts. A number of publications have examined Gfi1-WT and Gfi1-KO macrophages and found that Gfi1KO macrophages might differ on a quantitative level with regard to certain pathways, but overall they can be regarded as macrophages.28,48-50 Our finding that Gfi1-KO AAMs express more IL-6, Nos2 and other inflammatory mediators at mRNA level in vitro and in vivo when exposed to LPS, is in line with reports demonstrating a hyper-reactive response in Gfi1-deficient macrophages after exposure to LPS.28,49 Gfi1 exerts this function by its inhibitory effect on the Toll-like receptor 4 (TLR4) pathway through antagonizing the nuclear transcription factor k-light-chain-enhancer of activated B cells (NF-kB ).49 In contrast to the inhibitory effect of Gfi1 on M1 macrophage polarization, our results indicate that Gfi1 enhances the polarization of AAMs (M2-like macrophages) in vivo and in vitro. The upregulation of Gfi1 in
References 1. Colmone A, Amorim M, Pontier AL, Wang S, Jablonski E, Sipkins DA. Leukemic cells create bone marrow niches that disrupt the behavior of normal hematopoietic progenitor cells. Science. 2008;322(5909):18611865. 2. Turley SJ, Cremasco V, Astarita JL.
1226
response to M2 stimuli underlines this. We observed that transgenic Gfi1-KOxNUP98-HOXD13 leukemic mice had a lower frequency of AAMs and a higher percentage of Ly6C+ monocytes than Gfi1-WTxNUP98-HOXD13 leukemic mice. We hypothesize that in the absence of Gfi1, the differentiation of immature macrophages into AAMs is disturbed. In vitro, Gfi1-KO macrophages co-cultured with C1498GFP cells expressed higher levels of IL-6 and lower levels of Arg1 mRNA than Gfi1-WT macrophages. Gfi1 might regulate M1 and M2 polarization through its suppressive function on genes that are associated with M1 polarization. The increased Gfi1 expression in AAMs in vivo might impede M1 macrophage polarization and function, resulting in a shift of polarization towards a M2 phenotype. Additionally, Gfi1 is required by AAMs or M2 macrophages to secrete enzymes and cytokines, such as Arg1 and IL-10, which play important roles in the suppression of the immune system. There are, however, many open questions on how Gfi1 polarizes AAMs and which pathways might be involved.25,49 On a functional level, we characterized the interaction between macrophages and AML cells by using established procedures applied for the analysis of the interaction between macrophages and solid cancers.6,36 AML cells induce the expansion and/or migration of tissue-resident macrophages. They function as AAMs since they support the growth of AML cells both in vivo and in vitro. Furthermore, the polarization of AAMs depends on the presence of Gfi1, which is a potential new regulator of AAMs and macrophage polarization. We show one possibility of how the polarization of AAMs might be regulated, and targeting Gfi1 could be a novel approach to AML therapy by inhibiting the function of AAMs, expanding the possibility of stroma targeting approaches.51 Despite recent advances in the field of immunotherapy of solid cancers, a better understanding on how macrophages contribute to the growth of AML might open new AML therapy approaches. Acknowledgments The authors would like to thank Justin Kline, Chicago, USA, for providing the AML cell line C1498GFP, Saskia Grunwald for excellent technical assistance and the team of the animal facility of University Hospital Essen for genotyping, technical and administrative assistance during the whole mouse project. The authors would also like to thank Joachim Göthert, André Görgens and Namir Shaabani for sharing resources, mice or expertise. Funding CK was supported by the IFORES fellowship of the University Clinic of Essen, a Max-Eder fellowship from the German Cancer fund (Deutsche Krebshilfe) as well as of the Dr. Werner JackstädtStiftung. YSA-M was supported by an IDB (Islamic Development Bank) PhD scholarship.
Immunological hallmarks of stromal cells in the tumour microenvironment. Nat Rev Immunol. 2015;15(11):669-682. 3. McAllister SS, Weinberg RA. The tumourinduced systemic environment as a critical regulator of cancer progression and metastasis. Nat Cell Biol. 2014;16(8):717-727. 4. Whiteside TL. The tumor microenvironment and its role in promoting tumor
growth. Oncogene. 2008;27(45):5904-5912. 5. Franklin RA, Liao W, Sarkar A, et al. The cellular and molecular origin of tumor-associated macrophages. Science. 2014;344 (6186):921-925. 6. Chen SY, Yang X, Feng WL, et al. Organspecific microenvironment modifies diverse functional and phenotypic characteristics of leukemia-associated macrophages in mouse
haematologica | 2016; 101(10)
Gfi1 and polarization of AML-associated macrophages
7. 8.
9.
10.
11.
12.
13.
14.
15. 16.
17.
18.
19.
20.
21.
T cell acute lymphoblastic leukemia. J Immunol. 2015;194(6):2919-2929. Davies LC, Jenkins SJ, Allen JE, Taylor PR. Tissue-resident macrophages. Nat Immunol. 2013;14(10):986-995. Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. 2004;25(12):677-686. Xue J, Schmidt SV, Sander J, et al. Transcriptome-based network analysis reveals a spectrum model of human macrophage activation. Immunity. 2014;40(2):274-288. Kigerl KA, Gensel JC, Ankeny DP, Alexander JK, Donnelly DJ, Popovich PG. Identification of two distinct macrophage subsets with divergent effects causing either neurotoxicity or regeneration in the injured mouse spinal cord. J Neurosci. 2009;29(43):13435-13444. Mantovani A, Sozzani S, Locati M, Allavena P, Sica A. Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol. 2002;23(11):549-555. Duluc D, Delneste Y, Tan F, et al. Tumorassociated leukemia inhibitory factor and IL-6 skew monocyte differentiation into tumor-associated macrophage-like cells. Blood. 2007;110(13):4319-4330. Chittezhath M, Dhillon MK, Lim JY, et al. Molecular profiling reveals a tumor-promoting phenotype of monocytes and macrophages in human cancer progression. Immunity. 2014;41(5):815-829. Franklin RA, Li MO. The ontogeny of tumor-associated macrophages: a new understanding of cancer-elicited inflammation. Oncoimmunology. 2014;3(9):e955346. Civini S, Jin P, Ren J, et al. Leukemia cells induce changes in human bone marrow stromal cells. J Transl Med. 2013;11:298. Lane SW, Scadden DT, Gilliland DG. The leukemic stem cell niche: current concepts and therapeutic opportunities. Blood. 2009;114(6):1150-1157. Geyh S, Rodriguez-Paredes M, Jager P, et al. Functional inhibition of mesenchymal stromal cells in acute myeloid leukemia. Leukemia. 2016;30(3):683-691. Kim JA, Shim JS, Lee GY, et al. Microenvironmental remodeling as a parameter and prognostic factor of heterogeneous leukemogenesis in acute myelogenous leukemia. Cancer Res. 2015;75(11): 2222-2231. Chandran P, Le Y, Li Y, et al. Mesenchymal stromal cells from patients with acute myeloid leukemia have altered capacity to expand differentiated hematopoietic progenitors. Leuk Res. 2015;39(4):486-493. Theocharides AP, Jin L, Cheng PY, et al. Disruption of SIRPalpha signaling in macrophages eliminates human acute myeloid leukemia stem cells in xenografts. J Exp Med. 2012;209(10):1883-1899. Steidl C, Lee T, Shah SP, et al. Tumor-associated macrophages and survival in classic Hodgkin's lymphoma. N Engl J Med.
haematologica | 2016; 101(10)
2010;362(10):875-885. 22. Estey E, Dohner H. Acute myeloid leukaemia. Lancet. 2006;368(9550):18941907. 23. Moroy T. The zinc finger transcription factor Growth factor independence 1 (Gfi1). Int J Biochem Cell Biol. 2005;37(3):541-546. 24. Moroy T, Vassen L, Wilkes B, Khandanpour C. From cytopenia to leukemia: the role of Gfi1 and Gfi1b in blood formation. Blood. 2015;126(24):2561-2569. 25. Phelan JD, Shroyer NF, Cook T, Gebelein B, Grimes HL. Gfi1-cells and circuits: unraveling transcriptional networks of development and disease. Curr Opin Hematol. 2010;17(4):300-307. 26. Person RE, Li FQ, Duan Z, et al. Mutations in proto-oncogene GFI1 cause human neutropenia and target ELA2. Nat Genet. 2003;34(3):308-312. 27. Valk PJ, Verhaak RG, Beijen MA, et al. Prognostically useful gene-expression profiles in acute myeloid leukemia. N Engl J Med. 2004;350(16):1617-1628. 28. Karsunky H, Zeng H, Schmidt T, et al. Inflammatory reactions and severe neutropenia in mice lacking the transcriptional repressor Gfi1. Nat Genet. 2002;30(3):295300. 29. Zhang L, Gajewski TF, Kline J. PD-1/PD-L1 interactions inhibit antitumor immune responses in a murine acute myeloid leukemia model. Blood. 2009;114(8):15451552. 30. Nguyen TT, Schwartz EJ, West RB, Warnke RA, Arber DA, Natkunam Y. Expression of CD163 (hemoglobin scavenger receptor) in normal tissues, lymphomas, carcinomas, and sarcomas is largely restricted to the monocyte/macrophage lineage. Am J Surg Pathol. 2005;29(5):617-624. 31. Beider K, Bitner H, Leiba M, et al. Multiple myeloma cells recruit tumor-supportive macrophages through the CXCR4/CXCL12 axis and promote their polarization toward the M2 phenotype. Oncotarget. 2014;5(22):11283-11296. 32. Harris JA, Jain S, Ren Q, Zarineh A, Liu C, Ibrahim S. CD163 versus CD68 in tumor associated macrophages of classical Hodgkin lymphoma. Diagn Pathol. 2012;7:12. 33. Quatromoni JG, Eruslanov E. Tumor-associated macrophages: function, phenotype, and link to prognosis in human lung cancer. Am J Transl Res. 2012;4(4):376-389. 34. Yan M, Kanbe E, Peterson LF, et al. A previously unidentified alternatively spliced isoform of t(8;21) transcript promotes leukemogenesis. Nat Med. 2006;12(8):945949. 35. Krivtsov AV, Twomey D, Feng Z, et al. Transformation from committed progenitor to leukaemia stem cell initiated by MLLAF9. Nature. 2006;442(7104):818-822. 36. Laoui D, Van Overmeire E, Di Conza G, et al. Tumor hypoxia does not drive differentiation of tumor-associated macrophages but rather fine-tunes the M2-like macrophage population. Cancer Res. 2014;74(1):24-30. 37. Lin YW, Slape C, Zhang Z, Aplan PD.
38.
39.
40.
41.
42.
43.
44.
45. 46.
47.
48.
49.
50.
51.
NUP98-HOXD13 transgenic mice develop a highly penetrant, severe myelodysplastic syndrome that progresses to acute leukemia. Blood. 2005;106(1):287-295. Movahedi K, Laoui D, Gysemans C, et al. Different tumor microenvironments contain functionally distinct subsets of macrophages derived from Ly6C(high) monocytes. Cancer Res. 2010;70(14):57285739. Umemura N, Saio M, Suwa T, et al. Tumorinfiltrating myeloid-derived suppressor cells are pleiotropic-inflamed monocytes/macrophages that bear M1and M2-type characteristics. J Leukoc Biol. 2008;83(5):1136-1144. van der Meer LT, Jansen JH, van der Reijden BA. Gfi1 and Gfi1b: key regulators of hematopoiesis. Leukemia. 2010;24(11): 1834-1843. Biswas SK, Mantovani A. Macrophage plasticity and interaction with lymphocyte subsets: cancer as a paradigm. Nat Immunol. 2010;11(10):889-896. Colegio OR, Chu NQ, Szabo AL, et al. Functional polarization of tumour-associated macrophages by tumour-derived lactic acid. Nature. 2014;513(7519):559-563. Galdiero MR, Garlanda C, Jaillon S, Marone G, Mantovani A. Tumor associated macrophages and neutrophils in tumor progression. J Cell Physiol. 2013;228(7):14041412. Murray PJ, Allen JE, Biswas SK, et al. Macrophage activation and polarization: nomenclature and experimental guidelines. Immunity. 2014;41(1):14-20. Qian BZ, Pollard JW. Macrophage diversity enhances tumor progression and metastasis. Cell. 2010;141(1):39-51. Barcellos-Hoff MH, Park C, Wright EG. Radiation and the microenvironment tumorigenesis and therapy. Nature reviews Cancer. 2005;5(11):867-875. Klug F, Prakash H, Huber PE, et al. Lowdose irradiation programs macrophage differentiation to an iNOS(+)/M1 phenotype that orchestrates effective T cell immunotherapy. Cancer Cell. 2013;24(5): 589-602. Marteijn JA, van der Meer LT, Van Emst L, de Witte T, Jansen JH, van der Reijden BA. Diminished proteasomal degradation results in accumulation of Gfi1 protein in monocytes. Blood. 2007;109(1):100-108. Sharif-Askari E, Vassen L, Kosan C, et al. Zinc finger protein Gfi1 controls the endotoxin-mediated Toll-like receptor inflammatory response by antagonizing NFkappaB p65. Mol Cell Biol. 2010;30 (16):3929-3942. Spooner CJ, Cheng JX, Pujadas E, Laslo P, Singh H. A recurrent network involving the transcription factors PU.1 and Gfi1 orchestrates innate and adaptive immune cell fates. Immunity. 2009;31(4):576-586. Ben-Batalla I, Schultze A, Wroblewski M, et al. Axl, a prognostic and therapeutic target in acute myeloid leukemia mediates paracrine crosstalk of leukemia cells with bone marrow stroma. Blood. 2013;122 (14):2443-2452.
1227
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Acute Myeloid Leukemia
Ferrata Storti Foundation
Co-operative leukemogenesis in acute myeloid leukemia and acute promyelocytic leukemia reveals C/EBPα as a common target of TRIB1 and PML/RARA Karen Keeshan,1 Pauline Vieugué,2*†† Shahzya Chaudhury,1* Loveena Rishi,1‡ Coline Gaillard,2¥ Lu Liang,1‡‡ Elaine Garcia,2† Takuro Nakamura,3 Nader Omidvar,4 and Scott C. Kogan2
Paul O’Gorman Leukaemia Research Centre, Institute of Cancer Sciences, University of Glasgow, UK; 2Department of Laboratory Medicine, University of California San Francisco, CA, USA; 3Division of Carcinogenesis, The Cancer Institute, Japanese Foundation for Cancer Research, Tokyo, Japan; 4Department of Haematology, School of Medicine, Cardiff University, UK
1
Haematologica 2016 Volume 101(10):1228-1236
†† ‡
Current Address: Cancer Research Center of Lyon (CRCL), Lyon, France.
Current Address: Cancer Research UK Beatson Institute, Glasgow, UK.
¥
Current Address: Genentech Inc, South San Francisco, USA.
‡‡ †
Current Address: National Engineering Research Center of Cell Products, Tianjin, PR China.
Current Address: Biological and Biomedical Sciences Program, Harvard Medical School, USA.
*
SC and PV contributed equally to this work.
ABSTRACT
Correspondence: scott.kogan@ucsf.edu/ karen.keeshan@glasgow.ac.uk
Received: October 27, 2015. Accepted: June 24, 2016. Pre-published: July 6, 2016. doi:10.3324/haematol.2015.138503
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1228
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1228
T
he PML/RARA fusion protein occurs as a result of the t(15;17) translocation in the acute promyelocytic leukemia subtype of human acute myeloid leukemia. Gain of chromosome 8 is the most common chromosomal gain in human acute myeloid leukemia, including acute promyelocytic leukemia. We previously demonstrated that gain of chromosome 8-containing MYC is of central importance in trisomy 8, but the role of the nearby TRIB1 gene has not been experimentally addressed in this context. We have now tested the hypothesis that both MYC and TRIB1 have functional roles underlying leukemogenesis of trisomy 8 by using retroviral vectors to express MYC and TRIB1 in wild-type bone marrow and in marrow that expressed a PML/RARA transgene. Interestingly, although MYC and TRIB1 readily co-operated in leukemogenesis for wild-type bone marrow, TRIB1 provided no selective advantage to cells expressing PML/RARA. We hypothesized that this lack of co-operation between PML/RARA and TRIB1 reflected a common pathway for their effect: both proteins targeting the myeloid transcription factor C/EBPα. In support of this idea, TRIB1 expression abrogated the all-trans retinoic acid response of acute promyelocytic leukemia cells in vitro and in vivo. Our data delineate the common and redundant inhibitory effects of TRIB1 and PML/RARA on C/EBPα providing a potential explanation for the lack of selection of TRIB1 in human acute promyelocytic leukemia, and highlighting the key role of C/EBPs in acute promyelocytic leukemia pathogenesis and therapeutic response. In addition, the co-operativity we observed between MYC and TRIB1 in the absence of PML/RARA show that, outside of acute promyelocytic leukemia, gain of both genes may drive selection for trisomy 8.
haematologica | 2016; 101(10)
MYC, TRIB1 and PML/RARA co-operation in APL and AML
Introduction Acute myeloid leukemia (AML) is the most common leukemia subtype in adults representing 80% of cases for a worldwide incidence of 3.8/100,000 cases per year.1 The relapse risk for AML remains unacceptably high and relapse is the most common cause of death. Multiple courses of chemotherapy including combinations of anthracycline and cytarabine remain the mainstay of treatment but a ceiling of benefit has been reached and toxicity is significant. One particular exception is patients with acute promyelocytic leukemia (APL), a distinctive subtype representing 5%-15% of AML cases.2 APL patients typically express the PML/RARA fusion protein as a result of a t(15;17)(q22;q12) translocation, which renders cells exquisitely sensitive to all-trans retinoic acid (ATRA) leading to remission in the majority of cases.3 APL is characterized by clonal proliferation of myeloid blasts that have lost their differentiation capacity, and ATRA is thought to bypass this maturation block by relieving the repressive properties of PML/RARA on myeloid differentiation genes.4 Alteration of transcription factors often seen in AMLs such as RUNX1, GATA or CEBPA are commonly associated with mutations in tyrosine kinase receptors or other important signaling molecules, such as RAS or FLT3. Recent studies and the development of whole-genome sequencing technologies have revealed a complex process of leukemic transformation.5,6 These studies highlight the importance of transcriptional regulators controlling gene expression and differentiation, mutations affecting cell signaling and, most recently, mutations affecting epigenetic modifiers as co-operating factors in leukemic disease. It therefore appears that the path to cellular transformation can be a complex and multistep-process. In APL, secondary karyotypic lesions are often seen, with trisomy 8 being the most common. Trisomy 8 is also the most common unbalanced gain in AML in general.7,8 Interestingly, the segment that is often gained carries the well-known MYC proto-oncogene.9 MYC encodes a transcription factor controlling expression of downstream targets such as cyclins, thereby promoting proliferation, but is also able to limit cellular differentiation, including via deregulation of the master regulator of myeloid differentiation C/EBPα.10 Importantly, MYC and C/EBPα expression require tight regulation to maintain myeloid and stem cell homeostasis.11 Previous analyses in our laboratory have shown that cells characterized by a gain of MYC through trisomy 8 display approximately 45% higher MYC RNA levels. Using a PML/RARA transgenic model, we also showed that MYC overexpression both accelerated the development of leukemia and impaired myeloid cell maturation, and that gain of MYC underlines the recurrent trisomy of this gene commonly seen in APL.9 Interestingly, TRIB1 is located contiguously to MYC on chromosome 8, and is thus expected to be duplicated in the chromosomal gain containing the MYC fragment. Not surprisingly, overexpression of TRIB1 has been found in several AML patients12 and TRIB proteins have been implicated in AML pathogenesis.13 Initially the Tribbles gene was identified in drosophila (dTribbles) and mammalian TRIB genes are comprised of three human homologs: TRIB1, TRIB2 and TRIB3.14 Supporting a role for TRIB1 in disease initiation, murine recipients of hematopoietic stem cells transduced with TRIB1 or TRIB2, but not TRIB3 haematologica | 2016; 101(10)
developed AML and mediated COP1 ubiquitin ligasedependent C/EBPα degradation.15-17 This TRIB-mediated degradation of C/EBPα was critical for TRIB-induced AML, and possibly necessary for the maturation block seen in AMLs. In APL, the PML/RARA fusion protein retains the DNA binding domain of the endogenous RARA, therefore acting as a chimeric transcription factor.18 By dimerizing and associating with an altered set of co-factors the DNA binding specificity and repressive ability of the fusion protein is expanded,19 a characteristic that is believed to play a key role in the leukemogenic process. PML/RARA has been shown to interact with key transcription factors of the granulocytic differentiation program, such as PU.120 and C/EBPα downregulation has been observed in PML/RARA expressing cells,21,22 both events possibly participating in the transformation process. Interestingly, TRIB1 has been shown to co-localize with RARA/RXR leading to negative regulation of the transcriptional activity of the complex,23 although the precise mechanism mediating this inhibition is still unclear. Given previous studies and clinical data implicating PML/RARA, MYC and TRIB1 in AML and APL leukemogenesis, we aimed to investigate the co-operation of these oncogenes in the development of leukemia. Using a retroviral bone marrow (BM) transduction and BM transplantation (BMT) approach to over-express MYC, TRIB1, or both in wild type (Wt) or PML/RARA BM from transgenic animals, we characterized the resultant neoplasms arising in the different groups. We found that MYC and TRIB1 cooperated in AML, but that MYC alone was able to drive APL development in the presence of PML/RARA. These data indicate that PML/RARA and TRIB1 could share redundant functions. We showed that TRIB1 (and TRIB2, but not a mutant of TRIB1) prevents differentiation of PML/RARA-expressing cells in the presence of ATRA, which normally derepresses the fusion protein to allow terminal maturation. We also showed that TRIB1 overexpression leads to sustained decreased C/EBPα protein levels in these cells, providing an explanation for the inability of these cells to respond to ATRA. Indeed, leukemias generated in secondary recipient animals transplanted with PML/RARA+MYC+TRIB1+ leukemia were unresponsive to ATRA, in contrast to the ATRA responsiveness of PML/RARA+MYC+ leukemia.
Methods Expression constructs and retrovirus production The HA-tagged human MYC cDNA24 was subcloned into MSCV-IRES-mCherry vector. Mouse Trib1 cDNA25 was subcloned into MSCV-IRES-GFP (MigR1) vector. Mutant Trib1 (DC4) expressed in MigR1 was previously published.26 Mouse Trib2 cDNA was subcloned into MSCV-IRES-GFP (MigR1) and MSCVIRES-NGFR (NGFR).15,17 For mouse BM transduction experiments, BOSC23 packaging cell line was transfected with pCL-Eco and retroviral expression plasmids, as previously described.9 For NB4 transduction experiments, HEK293T packaging cell line was transfected with retroviral (pCGP, VSV-G) packaging vectors and retroviral expression plasmids. Viral supernatants (sups) were harvested at 24-48 h post transfection.
Cell culture and transduction NB4 cells were cultured in RPMI supplemented with 10% fetal 1229
K. Keeshan et al.
bovine serum (FBS). Cells were transduced by spinoculation with virus and 4 Îźg/mL Polybrene at 1290g for 90 mins at room temperature (RT). Transduced cells were sorted by flow cytometry 48 h post transduction for GFP expression (MigR1, TRIB1, DC4) or using anti-biotin beads (NGFR, TRIB2). Sorted cells were plated in presence of 1 uM ATRA at a density of 0.05x106 cells/mL.
harvested and 1x106 white blood cells/mL plated into pre-stimulation media (Myelocult M5300 Stemcell Technologies, 15% FBS, 10% of IL-3 and IL-6 conditioned medium, 0.4 mM of LGlutamine and 10 ng/mL of murine recombinant SCF). Two spinoculations were performed (at 4 h and 24 h after harvest), and cells were injected retro-orbitally into lethally-irradiated 6-16-week old recipients (3x105-1x106 cells/mouse).
Bone marrow harvest, retroviral transduction and transplantation
Cytomorphology
Donor animals (6-12 weeks old) were injected intraperitoneally with 5-Fluorouracil (5 FU, 150 mg/kg animal) to enrich for hemopoietic stem and progenitor cells and push them into cycle for the facilitation of retrovirus transduction. Leg and pelvic bones were
Cytospins were prepared by harvesting 25,000 cells and slides were stained with the Kwik-Diff staining kit (Thermo Scientific) as per the manufacturerâ&#x20AC;&#x2122;s instructions. Chromatin condensation and granularity was used to define differentiation on a Leica
A
B
C
D
Figure 1. MYC and TRIB1 co-operate to initiate acute myeloid leukemia (AML), whereas MYC alone combined with PML/RARA is able to drive acute promyelocytic leukemia (APL). (A) Schematic representation of the in vivo experimental strategy used to assess the selective advantage of MYC and/or TRIB1 in transforming wild-type (Wt) or PML/RARA bone marrow (BM). Lethally-irradiated recipients were reconstituted with transduced bone marrow, animals were monitored for disease development and the arising neoplasm characterized. (B) Summary table combining all Wt leukemias and all PML/RARA leukemias. As seen, no differences were observed in the mean or median time to disease in AML or APL initiation. (C) Panels show the flow cytometry staining and gating strategy used. (D) Phenotypic analysis of resultant leukemia in representative examples of a Wt-MYC+TRIB1+, PML/RARA-MYC+ and PML/RARA-MYC+TRIB1+ leukemias. Panels show representative Wt-MYC+TRIB1+ (top left), PML/RARA-MYC+ (top right), and PML/RARA-MYC+TRIB1+ (bottom left) leukemias staining for CD45.1, CD45.2, Gr-1 and c-Kit through the indicated gate of parental population; cytomorphology of spleen, sternum, and liver from representative Wt-MYC+TRIB1+ (top left), PML/RARA-MYC+ (top right) and PML/RARA-MYC+TRIB1+ (bottom left) leukemias showing leukemic cell organ infiltration.
1230
haematologica | 2016; 101(10)
MYC, TRIB1 and PML/RARA co-operation in APL and AML
DM2000 and photographs taken on an Olympus DP70. A sample of slides were blinded and reviewed by Dr Mike Leech, Consultant Hematologist with expertise in diagnostic morphology and there was 100% correlation with our findings. Paraffin embedded sections were stained with hematoxylin & eosin (H&E). Photographs were taken on a Nikon Eclipse 80i microscope with a Nikon Digital Sight camera using NIS-Elements F2.30 or F4.30 software at a resolution of 2560 Å~ 1920.
Flow cytometry A total of 10,000-50,000 cells were assessed by staining with CD45.1, CD45.2, Gr1 (Ly6G), and c-Kit on a LSRFortessa (BD Biosciences) or with CD15 (MMA) and CD11c (3.9) (eBioscience) on a FACSCanto II (BD Biosciences). Dead cells were excluded by DAPI staining (Sigma). Analysis was performed on FlowJo (Treestar).
RNA extraction and real time q-PCR Total RNA was extracted using RNeasy Mini Kit (QIAGEN) and reverse transcribed with the High Capacity cDNA Reverse Transcription Kit (Life Technologies). QPCR was performed using Fast SYBR® Green Master Mix (Life Technologies) on a 7900HT Fast Real-Time PCR System (Life Technologies). hENOX2, hABL and hβ-2 microglobulin were used as internal controls and averaged; relative mRNA levels were calculated using the 2-δδCT method, and absolute mRNA levels calculated using the 2-δCT method. Experiments were performed on technical triplicates and biological duplicate samples. Details of primers are available in the Online Supplementary Appendix.
Immunoblotting Lysates from cultured NB4 cells (50,000-150,000) +/- ATRA were prepared by direct lysis in 2X SDS sample buffer. Antibodies used were: anti-C/EBPα (Santa Cruz sc-61), anti-Actin (Sigma Aldrich A5441).
Animals Mice were bred and maintained at University of California San Francisco (UCSF) (USA) and Cardiff University (UK) and were cared for in accordance with Institutional Animal Care and Use Committee guidelines. The recipients FVB/n CD45.1/45.2 mice were generated by crossing FVB/n (CD45.1) to FVB/n CD45.2 congenic animals. hMRP8-PML/RARA mice have been previously described.27 Fisher’s exact test was used to assess whether Wt and PML/RARA leukemias were equivalent in their likelihood to express TRIB1.
In vivo ATRA treatment of recipients transplanted with PML/RARA leukemic cells After one passage into recipient animals, resurrected cryopreserved leukemic cells (2x106) of the PML/RARA-MYC+ and PML/RARA-MYC+TRIB1+ phenotype were intravenously injected into unirradiated or sublethally irradiated (500 cGy) FVB/n mice (n=4-5 per group in 2 independent experiments). Ten days post injection, 10 mg 21-day release ATRA or placebo pellets (Innovative Research of America) were implanted into the dorsal neck scruff. Mice were sacrificed when moribund or upon veterinary advice. Statistical significance was calculated using a log rank test.
Table 1. Summary table of the outcome of the in vivo leukemogenesis experiment, specifying the phenotype of the leukemia or cause of death in the first column, disease latency (time to death column), mean time to disease, and input (donor) bone marrow (BM) specified in the last column.
Leukemic deaths
Non-leukemic deaths
Phenotype of arising leukemia
Time to death (days)
Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ Wt-MYC+TRIB1+* Wt-MYC+TRIB1+ Wt-MYC+TRIB1+ PML/RARA-MYC+ PML/RARA-MYC+ PML/RARA-MYC+ PML/RARA-MYC+ PML/RARA-MYC+* PML/RARA-MYC+TRIB1+ PML/RARA-MYC+TRIB1+ Myeloid leukemia, no phenotypic data Abnormal tumor Recipient-derived thymic tumor Recipient-derived thymic tumor No disease No disease No disease No disease
88 92 109 119 126 154 167 176 217 250 106 153 162 176 176 109 121 88 27 153 162 260 260 267 272
Mean time to AML +/SD (days)
149.8 +/- 53.6
154.6 +/- 28.9
115 +/- 8.5
N/A
Input BM Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 PML/RARA-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 PML/RARA-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1 PML/RARA-MYC-TRIB1 Wt-MYC-TRIB1 PML/RARA-MYC-TRIB1 PML/RARA-MYC-TRIB1 Wt-MYC-TRIB1 Wt: PML/RARA (1:1)-MYC-TRIB1
AML: acute myeloid leukemia; SD: standard deviation; N/A: not applicable..*These two table entries represent one individual animal which developed leukemias of both the indicated phenotypes.
haematologica | 2016; 101(10)
1231
K. Keeshan et al.
Bioinformatics analysis GSE6891 and GSE12662 datasets reflect results on the Affymetrix Human Genome U133 Plus 2.0 Array. Three TRIB1 probes are present: 202241_at, 239818_x_at, 235641_at. Signal strength was more than 10-fold higher for 202241_at as compared to the other probes and the results presented reflect this probe. P values were obtained using Microsoft EXCEL t-test (two-tailed, unequal variance). The MYC gene is represented in GSE6891 by probe 202431_s_at. Correlations of MYC and TRIB1 expression were assessed using Microsoft EXCEL, including calculation of R2 value.
Results and Discussion In order to investigate if MYC and TRIB1 function cooperatively in the context of myeloid leukemia, we utilized a retroviral transduction system to concurrently over-express MYC and Trib1 in 5-FU-treated murine BM from Wt or PML/RARA transgenic animals. The PML/RARA mouse model expresses a human PML/RARA cDNA from the hMRP8 promoter cassette, which drives transgene expression in myeloid cells.27 Three groups of donor cells (PML/RARA, Wt, or a combination of Wt + PML/RARA mixed in a 1:1 ratio) were transduced and
A
C
1232
transplanted into lethally-irradiated recipients (n=5 PML/RARA donor group, n=4 Wt donor group, n=15 combined donor group) (Figure 1A). In the combined donor experiments, injected cells were comprised of all possible combinations, including cells that lacked oncogene integration. With this approach, we anticipated that the cells expressing all three oncogenes (PML/RARA, MYC and TRIB1) would be most able to initiate leukemia. At disease manifestation (or on day 260-272 at experiment termination), BM from these animals was harvested and analyzed by flow cytometry to look at chimerism and characterization of the arising neoplasm. Every leukemia which arose from the Wt donor BM expressed both MYC and TRIB1, confirming that these oncogenes co-operate and coexpress to drive AML transformation in this transplant model (Table 1, red color-coded in column 2). In contrast, leukemias originating from PML/RARA-expressing donor BM displayed a different phenotype, with 5 leukemias expressing MYC only (Table 1, green color-coded in column 2), and only 2 leukemias expressing both MYC and TRIB1 (Table 1, blue color-coded in column 2). With regard to the non-leukemic mice in our study, at the time of experiment termination, 4 of 24 recipient mice had failed to develop disease and 3 others were identified as bearing recipient-derived malignancies likely due to irradi-
B
Figure 2. TRIB1 expression is decreased in human acute promyelocytic leukemia (APL) and does not correlate with MYC expression. (A) TRIB1 transcript levels in 21 cases of human APL are compared to levels in 440 cases of human non-APL acute myeloid leukemia (AML). Midlines represent median values, boxes represent 25th-75th percentiles, whiskers represent range of all values. TRIB1 levels are lower in APL than in non-APL AML (P<0.001). (Dataset reported values in log2 scale of GSE6891 have been transformed to linear values for clarity.) (B) TRIB1 values rise with maturation from human normal CD34 (n=5) to normal promyelocytes (n=5), but are low in human APL (n=14). Individual values are shown. [Intensity signal on arrays were scaled differently in the GSE6891 and GSE12662 datasets resulting in the different relative expression scales in (A) and (B)]. (C) Correlation plots of TRIB1 expression and MYC expression in GSE6891 for 461 AMLs, including APL, as well as for AMLs with isolated trisomy 8 (+8, n=20), AMLs with normal karyotype (NK, n=187), and APL alone (n=21). Trendlines are shown but R2 <0.5 for each of the comparisons. [Dataset log2 scale GSE6891 values were graphed for these comparisons; in contrast to (A), the values have not been linear transformed.]
haematologica | 2016; 101(10)
MYC, TRIB1 and PML/RARA co-operation in APL and AML
A
C
E
B
D
F
Figure 3. TRIB1 overexpression blocks the ATRA-mediated differentiation of NB4 cells. (A) NB4 cells were transduced with control MigR1, MigR1-TRIB1, and MigR1DC4 (Trib1 mutant unable to degrade C/EBPα), sorted and treated with 1 μM ATRA for six days and cell morphology assessed by manual differential counts. (B) Representative cell morphology at day 0 and day 6. (C) Flow cytometry analysis of CD15 and CD11c expression in vehicle control (VC top) and ATRA treated transduced NB4 cells at day 6. (D) QPCR analysis of: exogenous Trib1 (top) and endogenous TRIB1 (bottom) in MigR1 and MigR1-TRIB1 transduced NB4 cells treated with ATRA at the indicated time points. Top graph depicts absolute gene expression relative to internal controls. Bottom graphs depict relative gene expression normalized to MigR1-control at day 0. (E) QPCR analysis of G-CSFR in MigR1, MigR1-TRIB1, and MigR1-DC4 transduced NB4 cells treated with ATRA at the indicated time points. Graphs depict relative gene expression normalized to MigR1-control at day 0. Significance determined by one-way ANOVA and Bonferroni post test. *P<0.05; **P<0.01. (F) Western blot analysis for C/EBPα in direct lysis samples prepared from NB4 cells transduced with control MigR1, MigR1-TRIB1, and MigR1DC4 and treated with 1 μM ATRA for indicated time points. Actin was used as a loading control. h: hours.
haematologica | 2016; 101(10)
1233
K. Keeshan et al.
ation. In the presence of PML/RARA, the dominant leukemic phenotype was MYC co-expression alone, as opposed to the Wt setting where MYC and TRIB1 were always co-expressed to drive leukemia. It is highly unlikely that this result was due to chance. If the selective pressure to express TRIB1 had been equivalent in Wt and PML/RARA-expressing BM, then we would have expected equal proportions of Wt and PML/RARA leukemias to express TRIB1. In fact, the observed result of TRIB1 expression in 10 of 10 Wt leukemias but only 2 of 7 PML/RARA leukemias indicates that selection for TRIB1 was indeed abrogated by the presence of PML/RARA (P=0.003). Of note, this was not a result of different TRIB1 levels in the Wt and PML/RARA leukemias as similar transduction levels were obtained in the groups (data not shown). To our surprise, transduced PML/RARA marrows did not outcompete transduced wild-type marrows in the combined donor experiments, further supporting the idea that PML/RARA and TRIB1 did not co-operate in leukemia initiation. When looking at all the Wt or PML/RARA leukemias combined (Figure 1B), we do not observe statistically significant differences in terms of numbers or time to disease, although a wider range of latencies was observed for the Wt leukemias. The median
latency to disease observed in Wt and PML/RARA leukemias was 140 and 153 days, respectively, suggesting that additional events are likely necessary to complete the transformation process, which could further explain why some recipients did not develop disease within the timeframe of the experiment. Morphologically, Wt and PML/RARA leukemias resembled each other and were characterized by the expansion of immature myeloid cells. Flow cytometry analysis and histology staining from a representative Wt-MYC+TRIB1+, PML/RARA-MYC+ and a PML/RARA-MYC+TRIB1+ leukemia are shown in Figure 1D, showing similar leukemic cell organ infiltration and morphology of the indicated leukemias. Flow cytometric immunophenotyping confirmed an immature myeloid character, with some variation in levels of expression of KIT and Gr1. Thus, MYC and TRIB1 co-expression were able to drive myeloid leukemia, similar to leukemias driven by PML/RARA and MYC co-expression. We hypothesized that if TRIB1 levels play a pathogenic role in some AMLs, but not in APL, that APL might show lower TRIB1 expression than other AMLs. Using a published dataset (GSE689128) we compared TRIB1 transcript levels in 21 APLs to those seen in 440 non-APL AMLs. On average, APLs expressed less than 50% the level of TRIB1
A
B
Figure 4. TRIB1 overexpression abrogates ATRA response in vivo. (A) Schematic representation of the in vivo experimental strategy to assess ATRA response in PML/RARA+MYC+ versus PML/RARA+MYC+TRIB1+ leukemias (n=10 each group). Recipient animals were either unirradiated or sublethally-irradiated and transplanted with PML/RARA-MYC+ or PML/RARA-MYC+TRIB1+ leukemias. (B) Kaplan-Meier survival curves of transplanted animals treated with placebo or ATRA pellet implantation. ATRA is able to extend survival of PML/RARA+MYC+ but not PML/RARA+MYC+TRIB1+ secondary transplanted mice. Using the log rank (Mantel-Cox) test for each group, PML/RARA-MYC+: placebo versus ATRA P<0.0001 (longer survival with ATRA; PML/RARA-MYC+TRIB1+: placebo versus ATRA; P<0.05 (shorter survival with ATRA).
1234
haematologica | 2016; 101(10)
MYC, TRIB1 and PML/RARA co-operation in APL and AML
present in non-APL AMLs (P<0.001) (Figure 2A). In order to assess if this simply reflected the normal expression levels in promyelocytes as compared to less mature myeloid blasts, we examined TRIB1 levels in a published dataset (GSE1266229) that includes normal CD34 cells, normal promyelocytes and APLs. TRIB1 increases 5-fold with maturation from normal CD34 cells to normal promyelocytes, indicating that the low levels in APL as compared to other non-APL AMLs is not simply a reflection of cell maturation. Furthermore, in this dataset, APLs expressed much lower levels of TRIB1 than do normal promyelocytes (>4 fold decrease; P<0.003) (Figure 2B). While recognizing that mRNA levels may not reflect protein levels, these results are compatible with the idea that TRIB1 does not contribute to the pathogenesis of APL. We also examined the published dataset of 461 AMLs (including 21 APLs) to evaluate whether there is a correlation between increased TRIB1 transcripts and increased MYC transcripts, which would suggest co-selection for expression. Although our data in mice show that MYC and TRIB1 can co-operate in AML, in the human dataset there was no significant correlation between TRIB1 levels and MYC levels in AML as a whole, as well as no significant correlation in the normal karyotype (n=187), +8 (n=20), and APL (n=21) subsets (R2 <0.5 for these four comparisons) (Figure 2C). Our in vivo leukemogenesis experiments showed that TRIB1 overexpression does not co-operate with PML/RARA for disease initiation but rather appears to phenocopy the PML/RARA-induced disease in the context of MYC co-expression. These results raise the possibility that PML/RARA and TRIB1 share overlapping functions or have redundant functions in leukemic transformation, thereby explaining the lack of selective advantage for cells to express PML/RARA, MYC and TRIB1 concomitantly. C/EBPα is a key myeloid transcription factor shown to be a target of both PML/RARA and TRIB family members, potentially representing the common target of the two oncogenes in our model. To test this hypothesis, we used an established in vitro system in which PML/RARAexpressing APL cells are induced to terminally differentiate by exposure to all-trans retinoic acid (ATRA). As a master regulator of granulocyte maturation, activation of C/EBPα, C/EBPβ30 and its downstream targets is mandatory to bypass the maturation block in these cells, which ATRA can circumvent by directly targeting PML/RARA. Therefore, we hypothesized that over-expressing TRIB proteins in PML/RARA-expressing cells might abrogate the differentiating effect of ATRA due to C/EBPα inhibition or degradation. NB4 cells transduced with control, TRIB1 or TRIB1 mutant (DC4) unable to bind and degrade C/EBPα26 expressing retroviruses were cultured in presence of ATRA for six days to induce myeloid differentiation. NB4 cells differentiated after six days in ATRA, as seen by loss of nucleoli, maturation/condensation of the chromatin and granularity, and CD15 and CD11c surface marker expression, indicating that the maturation block had been bypassed following targeting of PML/RARA by ATRA (Figure 3A-C; MigR1 controls). TRIB1 overexpression was able to block ATRA-induced differentiation and this was lost upon overexpression of mutant TRIB1 (DC4) (Figure 3A-C). These data indicate that TRIB1 abrogates ATRA-induced terminal differentiation and suggest the mechanism is via C/EBP degradation. Using primers specific for endogenous human TRIB1, we observed that ATRA treatment (which degrades haematologica | 2016; 101(10)
PML/RARA) up-regulated TRIB1 in NB4 cells (Figure 3D). In combination with the finding (see above) that TRIB1 transcript levels are lower in human APL than they are in normal human promyelocytes, it appears possible PML/RARA may suppress TRIB1 expression. Although we have not investigated the mechanism for low endogenous TRIB1 levels in APL cells, PML/RARA suppression of TRIB1 may be permissive for increased C/EBP activity in response to ATRA and therefore permissive for maturation beyond the promyelocyte stage. NB4 cell transduction with TRIB1 expressing vectors resulted in significant expression of exogenous TRIB1 (detected using mouse TRIB1 specific primers) and, of note, there was no increase in endogenous TRIB1 in response to ATRA when exogenous TRIB1 was present (Figure 3D). To further investigate TRIB-mediated block of ATRAinduced differentiation via C/EBPα, we assessed the gene expression of a C/EBPα-dependent target important for granulocytic maturation, G-CSF receptor (G-CSFR).31 G-CSFR gene expression levels increased over time in the control groups, as expected, which was blocked upon expression of TRIB1 and lost upon overexpression of mutant TRIB1 (DC4) (Figure 3E), demonstrating that expression of TRIB1 protein inhibits a C/EBPα-dependent target gene important for driving the myeloid maturation program. To confirm that the inhibitory phenotypic and transcriptional activities of TRIB proteins are mediated via C/EBP protein inhibition, we assessed C/EBPα protein levels. Expression of TRIB1 attenuated ATRA-induced C/EBPα protein expression and this was lost upon overexpression of mutant TRIB1 (DC4) (Figure 3F). We also assessed C/EBPβ protein levels and found that TRIB1 attenuated ATRA-induced C/EBPβ protein levels but not with the same impact as on C/EBPα protein expression (Online Supplementary Figure S1A). Similarly, the overexpression of TRIB2 in NB4 cells blocked ATRA-induced differentiation assessed by morphology and G-CSFR gene expression, and abrogated the ATRA-mediated induction of C/EBPα, C/EBPβ and PU.1 transcription factors (Online Supplementary Figure S1B-D). These data provide strong evidence that TRIB1 or TRIB2 overexpression prevents ATRA-induced differentiation of APL cells via the inhibition of C/EBPα. To further study the clinical relevance of TRIB1 expression in APL, we investigated ATRA (or placebo) response in vivo in recipient animals inoculated with PML/RARAMYC+ compared to PML/RARA-MYC+TRIB1+ leukemia. Leukemias generated in our initial leukemogenesis experiment (Figure 1) were tested (see Figure 4A for details of the experimental strategy) in two independent experiments. Results of these experiments were concordant and the combined results are shown (Figure 4B). Significantly, the response of the leukemias to ATRA treatment was markedly different. ATRA treatment significantly delays disease progression and extends median survival from 31 to 53 days (P<0.0001) in recipients of PML/RARAMYC+ leukemias compared to placebo treated controls, as expected. However, ATRA did not extend survival in mice injected with PML/RAR-MYC+TRIB1+ leukemia. These results show that TRIB1 expression abrogates ATRA response in vivo confirming our in vitro findings. Our data indicate that TRIB1 expression provides critical oncogenic functions preventing APL cells from responding to ATRA therapy, presumably via repression of C/EBPα. To conclude, in this study we aimed to investigate the 1235
K. Keeshan et al.
co-operation of MYC and TRIB1 in the pathogenesis of AML and APL, two genes located on a contiguous fragment of chromosome 8 in humans. Given the common gain of the MYC/TRIB1-containing segment in AMLs, we hypothesized that TRIB1 could also play an important role in leukemic transformation, including for APL. Using a retroviral transduction model of in vivo leukemogenesis, we found that both TRIB1 and MYC oncogenes co-operate to initiate AML, but that in most cases, MYC is sufficient to accelerate APL leukemogenesis. C/EBPα downregulation is an overlapping function of PML/RARA and TRIB1, and this common feature explains the lack of selective leukemic outgrowth for clones expressing both these oncogenes. The responsiveness to ATRA treatment was severely impaired by the expression of TRIB1, a phenotype observed both in vitro and in vivo. Overall, our results provide critical information about the role of TRIB1 in
References 1. Estey E, Döhner H. Acute myeloid leukaemia. Lancet. 2006;368(9550):18941907. 2. Douer D. The epidemiology of acute promyelocytic leukaemia. Best Pract Res Clin Haematol. 2003;16(3):357-367. 3. Coombs CC, Tavakkoli M, Tallman MS. Acute promyelocytic leukemia: where did we start, where are we now, and the future. Blood Cancer J. 2015;5:e304. 4. Arteaga MF, Mikesch J-H, Fung T-K, So CWE. Epigenetics in acute promyelocytic leukaemia pathogenesis and treatment response: a TRAnsition to targeted therapies. Br J Cancer. 2015;112(3):413-418. 5. Cauchy P, James SR, Zacarias-Cabeza J, et al. Chronic FLT3-ITD Signaling in Acute Myeloid Leukemia Is Connected to a Specific Chromatin Signature. Cell Rep. 2015;12(5):821–836. 6. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012; 150(2):264278. 7. Grimwade D, Walker H, Oliver F, et al. The importance of diagnostic cytogenetics on outcome in AML: analysis of 1,612 patients entered into the MRC AML 10 trial. The Medical Research Council Adult and Children's Leukaemia Working Parties. Blood. 1998;92(7):2322-2333. 8. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials. Blood. 2010;116(3):354-365. 9. Jones L, Wei G, Sevcikova S, et al. Gain of MYC underlies recurrent trisomy of the MYC chromosome in acute promyelocytic leukemia. J Exp Med. 2010;207(12):25812594. 10. Hoffman B, Amanullah A, Shafarenko M, Liebermann DA. The proto-oncogene cmyc in hematopoietic development and
1236
11.
12.
13. 14. 15.
16.
17.
18.
19.
20.
21.
leukemogenesis and responsiveness to ATRA treatment. Our data support a role for PML/RARA in altering C/EBPα and C/EBPβ in APL leukemogenesis and provide strong genetic evidence for a key role of C/EBP family members in myeloid leukemias beyond those with mutations in or methylation of the CEBPA gene. Funding Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award Number R01CA095274 (to SCK), by the Howat Foundation, Children with Cancer UK and Bloodwise (LLR 13011) (to KK) and Bloodwise (to NO). Acknowledgments We thank Dr. Huimin Geng for assistance with statistical analysis and Dr Mike Leech for morphology assessment.
leukemogenesis. Oncogene. 2002; 21(21): 3414-3421. Ye M, Zhang H, Amabile G, et al. C/EBPa controls acquisition and maintenance of adult haematopoietic stem cell quiescence. Nat Cell Biol. 2013;15(4):385-394. Röthlisberger B, Heizmann M, Bargetzi MJ, Huber AR. TRIB1 overexpression in acute myeloid leukemia. Cancer Genet Cytogenet. 2007;176(1):58-60. Liang KL, Rishi L, Keeshan K. Tribbles in acute leukemia. Blood. 2013;121(21):42654270 Lohan F, Keeshan K. The functionally diverse roles of tribbles. Biochem Soc Trans. 2013;41(4):1096-1100. Keeshan K, He Y, Wouters BJ, et al. Tribbles homolog 2 inactivates C/EBPalpha and causes acute myelogenous leukemia. Cancer Cell. 2006;10(5):401-411. Dedhia PH, Keeshan K, Uljon S, et al. Differential ability of Tribbles family members to promote degradation of C/EBPalpha and induce acute myelogenous leukemia. Blood. 2010;116(8):1321-1328. Keeshan K, Bailis W, Dedhia PH, et al. Transformation by Tribbles homolog 2 (Trib2) requires both the Trib2 kinase domain and COP1 binding. Blood. 2010; 116(23):4948-4957. Melnick A, Licht JD. Deconstructing a disease: RARalpha, its fusion partners, and their roles in the pathogenesis of acute promyelocytic leukemia. Blood. 1999; 93(10):3167-3215. Saeed S, Logie C, Stunnenberg HG, Martens JHA. Genome-wide functions of PML-RAR in acute promyelocytic leukaemia. Br J Cancer. 2011;104(4):554558. Wang K, Wang P, Shi J, et al. PML/RARalpha targets promoter regions containing PU.1 consensus and RARE half sites in acute promyelocytic leukemia. Cancer Cell. 2010;17(2):186-197. Guibal FC, Alberich-Jorda M, Hirai H, et al. Identification of a myeloid committed progenitor as the cancer-initiating cell in acute promyelocytic leukemia. Blood. 2009;
114(27):5415-5425. 22. Gaillard C, Tokuyasu TA, Rosen G, et al. Transcription and methylation analyses of preleukemic promyelocytes indicate a dual role for PML/RARA in leukemia initiation. Haematologica. 2015;100(8):10641075. 23. Imajo M, Nishida E. Human Tribbles homolog 1 functions as a negative regulator of retinoic acid receptor. Genes Cells. 2010; 15(10):1089-1097. 24. Hemann MT, Bric A, Teruya-Feldstein J, et al. Evasion of the p53 tumour surveillance network by tumour-derived MYC mutants. Nature. 2005;436(7052):807-811. 25. Jin G, Yamazaki Y, Takuwa M, et al. Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogenesis. Blood. 2007; 109(9):3998-4005. 26. Yokoyama T, Kanno Y, Yamazaki Y, Takahara T, Miyata S, Nakamura T. Trib1 links the MEK1/ERK pathway in myeloid leukemogenesis. Blood. 2010; 116(15):2768-2775. 27. Brown D, Kogan S, Lagasse E, et al. A PMLRARalpha transgene initiates murine acute promyelocytic leukemia. Proc Natl Acad Sci USA. 1997;94(6):2551-2556. 28. Verhaak RGW, Wouters BJ, Erpelinck CAJ, et al. Prediction of molecular subtypes in acute myeloid leukemia based on gene expression profiling. Haematologica. 2009; 94(1):131-134. 29. Payton JE, Grieselhuber NR, Chang L-W, et al. High throughput digital quantification of mRNA abundance in primary human acute myeloid leukemia samples. J Clin Invest. 2009;119(6):1714-1726. 30. Duprez E, Wagner K, Koch H, Tenen DG. C/EBPbeta: a major PML-RARA-responsive gene in retinoic acid-induced differentiation of APL cells. EMBO J. 2003; 22(21):5806-5816. 31. Smith LT, Hohaus S, Gonzalez DA, Dziennis SE, Tenen DG. PU.1 (Spi-1) and C/EBP alpha regulate the granulocyte colony-stimulating factor receptor promoter in myeloid cells. Blood. 1996;88(4):12341247.
haematologica | 2016; 101(10)
ARTICLE
Hodgkin Lymphoma
Definition of bulky disease in early stage Hodgkin lymphoma in computed tomography era: prognostic significance of measurements in the coronal and transverse planes
EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
Anita Kumar,1 Irene A. Burger,2 Zhigang Zhang,3 Esther N. Drill,3 Jocelyn C. Migliacci,1 Andrea Ng,4 Ann LaCasce,5 Darci Wall,6 Thomas E. Witzig,7 Kay Ristow,7 Joachim Yahalom,8 Craig H. Moskowitz,1 and Andrew D. Zelenetz1
Lymphoma Service, Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 2Department Medical Radiology, University Hospital Zurich, Switzerland; 3Biostatistics and Epidemiology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA; 4Department of Radiation Oncology, Brigham and Womenâ&#x20AC;&#x2122;s Hospital and Dana-Farber Cancer Institute, Boston, MA, USA; 5Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; 6Department of Hematology, Mayo Clinic, Rochester, MN, USA; 7Department of Hematology, Mayo Clinic, Rochester, MN, USA; and 8Department of Radiation Oncology, Memorial Sloan-Kettering Cancer Center, New York, NY, USA
1
Haematologica 2016 Volume 101(10):1237-1243
ABSTRACT
D
isease bulk is an important prognostic factor in early stage Hodgkin lymphoma, but its definition is unclear in the computed tomography era. This retrospective analysis investigated the prognostic significance of bulky disease measured in transverse and coronal planes on computed tomography imaging. Early stage Hodgkin lymphoma patients (n=185) treated with chemotherapy with or without radiotherapy from 2000-2010 were included. The longest diameter of the largest lymph node mass was measured in transverse and coronal axes on pre-treatment imaging. The optimal cut off for disease bulk was maximal diameter greater than 7 cm measured in either the transverse or coronal plane. Thirty patients with maximal transverse diameter of 7 cm or under were found to have bulk in coronal axis. The 4-year overall survival was 96.5% (CI: 93.3%, 100%) and 4-year relapse-free survival was 86.8% (CI: 81.9%, 92.1%) for all patients. Relapse-free survival at four years for bulky patients was 80.5% (CI: 73%, 88.9%) compared to 94.4% (CI: 89.1%, 100%) for non-bulky; Cox HR 4.21 (CI: 1.43, 12.38) (P=0.004). In bulky patients, relapse-free survival was not impacted in patients treated with chemoradiotherapy; however, it was significantly lower in patients treated with chemotherapy alone. In an independent validation cohort of 38 patients treated with chemotherapy alone, patients with bulky disease had an inferior relapse-free survival [at 4 years, 71.1% (CI: 52.1%, 97%) vs. 94.1% (CI: 83.6%, 100%), Cox HR 5.27 (CI: 0.62, 45.16); P=0.09]. Presence of bulky disease on multidimensional computed tomography imaging is a significant prognostic factor in early stage Hodgkin lymphoma. Coronal reformations may be included for routine Hodgkin lymphoma staging evaluation. In future, our definition of disease bulk may be useful in identifying patients who are most appropriate for chemotherapy alone.
Introduction The presence of bulky disease at presentation has long been considered a poor prognostic factor in early stage Hodgkin lymphoma (ESHL).1,2 Historically, bulk in the mediastinum was focused upon and defined using radiographic criteria from a haematologica | 2016; 101(10)
Correspondence: kumara2@mskcc.org
Received: January 14, 2016. Accepted: July 5, 2016. Pre-published: July 6, 2016. doi:10.3324/haematol.2016.141846
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1237
Š2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is required for any other use.
1237
A. Kumar et al.
standing posterior-anterior (PA) chest radiograph (CXR). In the 1989 Cotswolds revision of the Ann Arbor Staging system, bulk in the mediastinum was defined as “when the maximum width is equal or greater than one-third of the internal transverse diameter of the thorax at the level of T5/6” on a PA CXR and bulk at an alternate site was defined as any mass measuring 10 cm or more by any imaging study.3 Today, the presence of bulky disease in ESHL remains an unfavorable prognostic feature in all modern risk classification systems, but is variably defined as more than one-third of the mediastinal mass ratio (MMR), more than one-third of the mediastinal thoracic ratio (MTR), or any mass over 10 cm.4-7 Although definitions of disease bulk were originally developed in the CXR-era, computed tomography (CT) imaging is now the standard staging imaging modality in lymphoma.8 In the CT era, the definition of disease bulk remains elusive. Various retrospective studies have defined CT criteria for bulky disease associated with increased risk of relapse in ESHL, ranging from greater than 5 to 10 cm for the maximal mediastinal mass diameter in transverse plane.9-11 The recent “Lugano Classification” for initial staging in lymphoma retains the historical definition of bulk with a single nodal mass of 10 cm or more or mediastinal bulk more than one-third of the transthoracic diameter. However, with uncertainty remaining regarding an evidence-based definition of bulk in the CT-era, the authors also suggest elimination of the modifier “X” and propose recording the longest measurement by CT scan, presumably in the transverse plane.8 However, there remains a critical need for a bulky disease cut-off point in ESHL as increasingly patients are being
A
treated with chemotherapy alone to avoid the late effects of radiotherapy, an approach that has generally excluded patients with disease bulk.12,13 For example, the recently published UK RAPID study included stage IA and IIA patients without mediastinal bulk defined as maximal mediastinal diameter of 33% or more of the internal thoracic diameter at T5-T6.12 Importantly, the transverse diameter of a lymph node mass may not reflect the longest dimension of an oblique or sagittal positioned mass in HL. On CT, lymph node masses can be measured in the transverse, coronal, and sagittal planes; however, measurements in axes other than the transverse are rarely reported in HL, and 3-dimensional or volumetric assessments of tumor bulk are difficult to routinely assess. In this study, we aimed to assess the prognostic significance of the longest diameter of the largest nodal mass measured in either the transverse and coronal planes using CT imaging.
Methods Patients For the training cohort of this retrospective study, we identified pediatric and adult patients with stage I-II classical HL treated at Memorial Sloan Kettering (MSK) Cancer Center between January 2000 and December 2010. Patients were included if they had pretreatment CT scan performed within 30 days of therapy initiation and received doxorubicin-containing chemotherapy with or without radiation at MSK. For the independent validation cohort, adult patients with classical HL, stage I-II, with baseline CT scan performed within 30
B
Figure 1. Representative images of the longest diameters measured using calipers of a right cervical mass in (A) transverse plane, 2.6 cm and (B) coronal plane, 12.1 cm.
1238
haematologica | 2016; 101(10)
Disease bulk in Hodgkin lymphoma
days of therapy, and treated with chemotherapy alone at Dana Farber Cancer Institute (DFCI) or Mayo Clinic (Mayo) between January 2000 and December 2010 were included. These retrospective analyses were conducted on waivers of authorization approved by the institutional review boards at each institution.
Radiological assessment Measurements for the training cohort were performed by a single board-certified radiologist (IB) who was blinded to patients’ clinical history, including treatment and outcome. All available pre-treatment imaging studies, including diagnostic computed tomography (dCT), FDG-PET and FDG-PET/CT were reviewed. CT scans were performed on GE Lightspeed helical scanners. MSK patients were scanned on dedicated PET/CT systems [Discovery STE, LS or 690 (GE Medical Systems) or Biograph 16 (Siemens Medical Systems)]. Images were reconstructed at 5- or 7.5-mm intervals in the Picture Archiving and Communication System (PACS) (Centricity; GE Medical Systems, Milwaukee, WI, USA). CT scans from outside institutions were digitized on PACS,
Table 1. Patient’s characteristics of early stage Hodgkin lymphoma, n=185.
N Median age (range), years Female sex Age < 45 Age ≥ 45 Pediatric, age < 18 Histological subtype Nodular sclerosing Mixed cellularity Lymphocyte rich Classical HL, NOS CD20+ classical Hodgkin lymphoma Stage IA Stage IB Stage IIA Stage IIB Pericardial or pleural effusion Extranodal disease Lung involvement Chest wall involvement Thyroid involvement ESR >50 if A, ESR > 30 if B, n=167 # Nodal Sites >2 defined by GHSG Unfavorable risk by GHSG*, n=181 Treatment Combined modality therapy Chemotherapy alone Regimen ABVD regimen Stanford V Other†
% 41 (9-85)
105 142 43 21
57 77 23 11
144 22 2 17 16 10 0 122 53 26 26 13 8 4 65 89 128
78 12 1 9 9 5 0 66 29 14 14 7 4 2 39 48 71
115 70
62 38
107 34 44
58 18 24
HL: Hodgkin lymphoma; NOS: not otherwise specified; GHSG: German Hodgkin Study Group; ESR: erythrocyte sedimentation rate. *GHSG unfavorable risk criteria: bulky mediastinal mass (≥one-third maximum transverse thoracic diameter on posterior-anterior chest X-ray or ≥10 cm by CT imaging in axial plane), ESR ≥50 mm/h or ESR ≥30 mm/h in patients with “B” symptoms, extranodal involvement, or 3 or more involved lymph node sites (reference). ABVD: doxorubicin, bleomycin, vinblastine, dacarbazine. † COPP/ABV, AVG, BEACOPP, CHOPE, CHOP, EVA, rituximab+chemotherapy, and various pediatric protocols such as ABVE-PC.
haematologica | 2016; 101(10)
reviewed and included only if of acceptable quality. Most measurements were obtained using dCT (134 of 185, 72%). Others were from low-dose CT performed in conjunction with FDG-PET imaging (51 of 185, 28%). Using calipers with measurements in centimeters, the longest diameter of the largest individual or conglomerate lymph node mass was measured in the transverse plane and coronal plane (see Figure 1). This measurement could lie obliquely or in any orientation, to ensure ascertainment of the maximal diameter. Coronal images were analyzed and re-formatted using the open-source software OsiriX.14 The quality of the coronal reconstruction was dependent on the slice thickness, ranging from 1.2 to 7.5 mm. The quality of coronal reconstructions was assessed using a subjective scale: 0-poor, 1-fair, 2-good. For the validation cohorts, the measurements were performed in a similar manner as described for the training cohort. At DFCI, the measurements were performed by a single medical oncologist (AK) and at Mayo by a single board-certified radiologist (DW), both blinded to patient outcome. Coronal reconstructions were created using OsiriX and TeraRecon (TeraRecon Inc.) software at DFCI (n=25) and at Mayo (n=13), respectively. Measurements were made from dCT at baseline (23 of 38, 61%) and low-dose CT in conjunction with FDG-PET imaging (15 of 38, 39%).
Statistical analysis Survival analyses were performed using the Kaplan-Meier method in SPSS 22 and R 3.2.3. Relapse-free survival (RFS) and overall survival (OS) were defined as the time from initiation of treatment until progression of disease or relapse (RFS) or until death from any cause (OS), respectively. Log rank tests were used to compare survival differences. P≤0.05 was considered statistically significant. The Pearson correlation test assessed the degree of correlation between the maximal transverse and coronal diameters. With few deaths, only the association between disease bulk and RFS was assessed. To identify the optimal cut off for the transverse and coronal maximal diameters to predict outcome, we identified a range of potential cut-off points (in cm) based upon the distribution of the data (between the 10th and 90th percentiles), and examined their significance levels using log rank tests correlating with
Figure 2. Correlation between maximal transverse and coronal measurements with lines for the 7 cm maximal transverse and coronal cut-offs for disease bulk.
1239
A. Kumar et al.
RFS. The cut-off point resulting in the maximal significance level was identified as the optimal cut off for the transverse and coronal dimensions, and this P-value was adjusted by the maximal χ2 method since multiple tests were performed.15 Further, we aimed to identify the most predictive “combined” criterion. By using concordance probability analysis, we identified the set of maximal diameter cut offs (i.e. transverse > X OR coronal > Y) that most strongly predicted RFS.16 The methods for determining bulky disease cut off, including maximal χ2 method and concordance probability analysis, are described in detail in the Online Supplementary Appendix.
Results In total, there were 185 patients who met inclusion criteria for this study. Of the 185 images examined, 158 (85%) were characterized by the interpreting radiologist as good quality for coronal measurement assessment and 27 fair quality (15%). Patients’ characteristics are shown in Table 1. The median age of patients was 41 years (range 9 to 85) with slight female predominance (57%). Of the histological subtypes of classical HL, the predominant was the nodular sclerosing subtype (78%). Overall, 95% of patients presented with stage II disease by Ann Arbor classification. An effusion (pericardial or pleural) or extranodal disease (lung, chest wall, or thyroid involvement) was present in a minority of patients. Using the German Hodgkin Study Group system, 71% of patients met criteria for unfavorable risk disease.5 All patients were treated with standard or novel doxorubicin-based chemotherapy regimens (Table 1). The vast majority of patients were treated with ABVD or Stanford V, and others were treated with various doxorubicin-containing regimens, including clinical trial protocols. A proportion of CD20+ HL patients were treated with rituximab-containing regimens (6 of 16). One hundred and fifteen patients (62%) were treated with combined modality therapy (CMT) and 70 patients (38%) with chemotherapy alone.
A
Outcomes Of the 185 patients, 23 patients relapsed and 6 patients died after initial therapy: 4 deaths due to progressive HL after multiple lines of salvage treatment, one death from secondary myelodysplastic syndromes (MDS), and one death of unknown cause. The median follow up for all patients was five years. The 4-year OS was 96.5% (CI: 93.2%, 100%) and the 4-year RFS was 86.8% (CI: 81.8%, 92.1%).
Disease bulk Training cohort: in most cases the longest diameter in the transaxial and coronal plane was measured from the same lymph node mass; however, in 13 cases the longest diameters were from different sites. In univariate analyses, both transverse maximal diameter and coronal maximal diameter were significantly associated with RFS (Online Supplementary Table S1). Using log rank tests to compare RFS, the optimal cut-off point for transverse maximal diameter was 7 cm (P=0.01, after adjustment by the maximal χ2 method P=0.02). The optimal cut-off point for coronal max diameter was 10.5 cm (P=0.005, after adjustment P=0.007). Using the concordance probability technique to determine optimal combined criterion, the criteria of transverse max diameter more than 7.0 cm OR coronal max diameter more than 7.0 cm was the best predictor for progression (Online Supplementary Appendix). Figure 2 demonstrates that the transverse and coronal maximal diameters are highly correlated (correlation coefficient R=0.856; P<0.001). Using a cut-off point of more than 7cm in maximal transverse dimension or maximal coronal dimension (>7cm in MTD or MCD), approximately half of the patients are defined as bulky (101 of 187, 54%). Seventythree patients had disease bulk more than 7.0 cm by transverse criteria and 10 of these patients had disease bulk in the transverse plane alone without disease bulk in the coronal plane. Ninety-one patients had disease bulk more than 7.0 cm by coronal criteria with 28 patients identified as having disease bulk more than 7.0 cm on coronal imag-
B
Figure 3. Relapse-free survival (RFS) by presence of bulky disease. (A) RFS for non-bulky versus bulky disease (transverse or coronal max diameter > 7cm). (B) RFS for coronal bulk alone (coronal max measurement > 7cm, transverse max, measurement ≤ 7cm) compared to traditional definition of bulk (transverse max, measurement >7cm).
1240
haematologica | 2016; 101(10)
Disease bulk in Hodgkin lymphoma
ing that was not observed in the transaxial plane. Of the 101 patients with bulky disease (>7cm in MTD or MCD), 83 patients had evidence of mediastinal disease bulk, 17 with other sites of supradiaphragmatic bulk, and one with infradiaphragmatic disease bulk (Table 2). Univariate analyses for RFS demonstrated that disease bulk defined as more than 7cm was associated with increased risk for relapse, with similar findings for transverse versus coronal plane measurements and for mediastinal versus non-mediastinal disease (Table 2). The presence of bulky disease, defined as more than 7cm in maximal transverse dimension or maximal coronal dimension (>7cm in MTD or MCD), significantly correlated with RFS. At four years, relapse-free survival for bulky patients was 80.5% (CI: 73%, 88.9%) compared to 94.4% (CI: 89.1%, 100%) for non-bulky; Cox HR 4.21 (CI: 1.43, 12.38) (P=0.004) (Figure 3A). There was no significant difference in OS between patients with bulky versus nonbulky disease [Cox HR 4.63 (CI 0.54, 40.04); P=0.126]. In an outcomes analysis of bulky patients (Figure 3B) there is no apparent difference between RFS for patients uniquely identified in this study with coronal bulk alone (coronal > 7cm and transverse ≤7cm) versus traditional definition of bulk using transverse measurement (transverse > 7cm) [Cox HR 0.91 (CI: 0.33, 2.51); P=0.846]. Analysis of bulk stratified by therapy (CMT vs. chemotherapy alone) demonstrated that patients with bulky disease treated with chemotherapy alone had a particularly unfavorable prognosis with 4-year RFS of 55.2% (CI: 40.2%, 75.7%) (Figure 4). The definition of bulk as more than 7 cm in MTD or MCD was a powerful prognostic factor in patients treated with chemotherapy alone. However, for patients who received CMT, this definition of disease bulk was not prognostic; rather the traditional definition of bulky disease (>10cm in transverse plane) delineated a group with increased risk of relapse (P=0.001) (Online Supplementary Figure S1).
years for this cohort. Nineteen patients (50%) had disease bulk and 5 patients (13%) were identified to have disease bulk on coronal measurement, but not by transverse measurement. Among the 38 patients, there were 6 relapse events and 2 deaths, one undifferentiated sarcoma eight years post treatment in a 53-year old and one bleomycin-related respiratory failure in a 67-year old; both deaths occurred in patients with non-bulky disease. Disease bulk was associated with a trend toward inferior RFS: in bulky patients, 4-year RFS was 71.1% (CI: 52.1%, 97%) versus 4-year RFS of 94.1% (CI: 83.6%, 100%) in non-bulky patients [Cox HR 5.27 (CI: 0.62, 45.16); P=0.09] (Figure 5).
Discussion The importance of re-examining the definition of bulk is based upon improved multidimensional CT data quality and evolving treatment strategies, including reduction and elimination of radiotherapy. Previous definitions of dis-
Validation cohort The novel MSK definition of disease bulk (>7cm in MTD or MCD) was prognostically relevant among patients who received chemotherapy alone, and, therefore, this prognostic factor was examined in an independent validation set of 38 patients with stage II disease treated with chemotherapy alone (36 with 4-6 cycles of ABVD, 1 with 4 cycles of BCVPP, and 1 with 3 cycles of MOPP and 3 cycles of ABVD). The median follow up was four
Figure 4. Relapse-free survival by presence of bulky disease (transverse or coronal max, diameter > 7cm) and treatment [chemotherapy alone (Chemo) vs. combined modality therapy (CMT)].
Table 2. Relapse-free survival (RFS) at four years for bulky disease subgroups defined as more than 7cm in transverse or coronal plane. The corresponding hazard ratios (HR) are reported for the bulky disease subgroups when compared to the complementary non-bulky disease subgroup.
Bulky disease measurement Transverse > 7cm Transverse > 7cm only (coronal ≤7cm) Coronal > 7cm Coronal > 7cm only (transverse ≤7cm) Transverse or coronal > 7cm Mediastinal Non-mediastinal Supradiaphragmatic Infradiaphragmatic Non-bulky (transverse AND coronal ≤7cm) haematologica | 2016; 101(10)
N
4-year RFS
HR
73 10 91 28 101 83 18 17 1 84
0.80 (0.71 – 0.90) 0.80 (0.59 – 1.00) 0.81 (0.73 – 0.89) 0.82 (0.69 – 0.98) 0.81 (0.73 – 0.89) 0.82 (0.74 – 0.91) 0.72 (0.54 – 0.96)
2.56 (1.09, 5.84) 1.75 (0.41, 7.46) 3.09 (1.22, 7.83) 1.59 (0.59, 4.27) 4.21 (1.43, 12.38)
0.94 (0.89 – 1.0) 1241
A. Kumar et al.
ease bulk were formulated in the era when combined modality therapy was the standard of care for all ESHL patients.17 Our proposed definition of disease bulk (>7cm in transverse or coronal plane) has heightened prognostic relevance for patients treated with chemotherapy alone. To our knowledge, this is the first study to analyze the prognostic significance of bulky disease defined in both transverse and coronal planes in the era of CT imaging. This study demonstrates that, when measured in the coronal plane, bulky disease is significantly associated with inferior outcome. The early stage risk classification systems (including GHSG, EORTC, NCIC, and NCCN) use MMR and MTR to define mediastinal bulky disease; however, in modern clinical practice and in clinical trials, bulky disease is often defined by the maximal transverse diameter of the largest lymph node mass more than 10 cm, likely because this measurement is a quick and reproducible method of assessing disease bulk on CT imaging. Other studies have suggested alternate definitions of mediastinal bulky disease, including lymph node masses greater than 5 cm, 6 cm, or 7.5 cm.9,10,17,18 Herein the optimal criterion for bulky disease of more than 7 cm in MTD or MCD falls within the range of previously reported CT bulk definitions. Our data show that bulky disease in any location, in the mediastinum or an alternate site, is of prognostic value. Coronal and sagittal reformations in CT studies provide additional diagnostic and clinical information in various clinical settings.19,20 Evaluation of coronal images in HL is important for disease bulk positioned obliquely in the mediastinum or for conglomerate lymph node mass extending along the cranio-caudal axis, such as involvement of contiguous cervical, supraclavicular, and infraclavicular lymph nodes. We found additional patients (approx. 30%) with bulky disease were identified using CT coronal reformations that would not have otherwise been identified using transaxial imaging alone, identifying additional patients at increased risk of relapse. Importantly, patients with coronal-only bulk had similar outcomes when compared to patients with bulky disease defined traditionally using transverse measurements. These data have the potential to add to the â&#x20AC;&#x153;Lugano classificationâ&#x20AC;? system by clarifying the importance of measuring disease bulk in the coronal axis in addition to the transverse axis. OsiriX software was used to create coronal reformats from transaxial CT images in this study; however, coronal and sagittal reformations are a routine component of modern multi-planar CT protocols. Obtaining these measurements on modern pre-treatment staging CT scans can easily be performed in routine practice. The standard of care for patients with ESHL and bulky mediastinal disease is CMT.21-23 In the modern treatment of HL, there is interest in limiting or eliminating the use of radiotherapy in select patients to decrease long-term radiotherapy-related toxicities. Many clinical trials have aimed to eliminate radiotherapy in select ESHL patients, such as the United Kingdom RAPID, NCIC HD6, MSK, EORTC/LYSA/FIL H10, and SWOG ESHL trials, and have been limited to patients with non-bulky disease.12,24-26 In the current retrospective analysis, we analyzed risk factors for relapse in patients treated with chemotherapy alone versus CMT, and identified a novel definition of disease bulk (>7cm in MTD or MCD) that is prognostically most relevant for patients treated with chemotherapy alone. This novel definition was evaluated in an independent 1242
Figure 5. Relapse-free survival by presence of bulky disease (transverse or coronal max, diameter > 7cm) in validation cohort patients (n=38) treated with chemotherapy alone.
cohort of patients treated with chemotherapy alone. Unfortunately, in this validation cohort, the sample size was small and there were few relapse events, such that the RFS difference in this validation cohort was not statistically significant but showed a trend towards this (P=0.09). However, findings in the validation cohort were consistent with the training set and demonstrate a clinically meaningful difference with most relapse events (5 of 6) occurring among patients with disease bulk. These data provide preliminary evidence that suggest a potential for caution against eliminating radiotherapy in patients with disease more than 7 cm in MTD or MCD as our retrospective analysis suggests that chemotherapy alone is associated with poor outcomes in bulky patients [in the training set: 4-year RFS of 55.2% (CI: 40.2%, 75.7%); in the validation set: 4-year RFS 71.1% (CI: 52.1%, 97%)]. A corollary of these data is that patients without disease bulk per our definition appear to have excellent relapse-free survival when treated with chemotherapy alone [in the training set: 4-year RFS of 97.4% (CI: 92.4%, 100%); in the validation set: 4-year RFS 94% (CI: 83.6%, 100%)]. Full course chemotherapy, as was commonly administered for patients included in this retrospective analysis, is likely not necessary for these non-bulky patients who have an excellent prognosis and would be appropriate candidates for the RAPID or CALGB/Alliance 50604 risk-adapted treatment paradigms with abbreviated chemotherapy regimens (i.e. 3-4 cycles of ABVD) for PET-negative patients.12,13 Finally, the training data also suggest that patients with very bulky disease with masses more than 10 cm in the transverse plane have a high risk of relapse even when treated with standard CMT, suggesting a role for more intensive chemotherapy such as escalated BEACOPP or incorporation of novel agents, such as brentuximab vedotin or nivolumab, in this patient population. A limitation of this study is that interim and end-oftreatment PET scans were not routinely available. Preliminary data from the British Columbia Cancer Agency reports excellent outcomes for bulky HL patients haematologica | 2016; 101(10)
Disease bulk in Hodgkin lymphoma
who achieve a negative PET scan after full-course ABVD without additional consolidative RT, and there is an ongoing CALGB/Alliance 50801 study exploring interim PETbased risk-adapted therapy in HL patients with disease bulk.27 These studies will further clarify whether achievement of an interim or end-of-treatment PET may overcome the negative prognostic value of disease bulk at initial presentation. In addition, the current study was inadequately powered to assess association with OS due to few deaths. Furthermore, due to a limited number of patients and events, subset analyses were underpowered and predominantly univariate analyses were performed. We hope future studies in larger data sets can provide external validation. In conclusion, our study demonstrates that measurement of bulky disease on coronal reformations in addition
References 1. Mauch P, Goodman R, Hellman S. The significance of mediastinal involvement in early stage Hodgkin's disease. Cancer. 1978;42(3):1039-1045. 2. Schomberg PJ, Evans RG, O'Connell MJ, et al. Prognostic significance of mediastinal mass in adult Hodgkin's disease. Cancer. 1984;53(2):324-328. 3. Lister TA, Crowther D, Sutcliffe SB, et al. Report of a committee convened to discuss the evaluation and staging of patients with Hodgkin's disease: Cotswolds meeting. J Clin Oncol. 1989;7(11):1630-1636. 4. Eghbali H, Raemaekers J, Carde P, Group EL. The EORTC strategy in the treatment of Hodgkin's lymphoma. Eur J Haematol. 2005;(Suppl):135-40. 5. Engert A, Schiller P, Josting A, et al. Involved-field radiotherapy is equally effective and less toxic compared with extended-field radiotherapy after four cycles of chemotherapy in patients with early-stage unfavorable Hodgkin's lymphoma: results of the HD8 trial of the German Hodgkin's Lymphoma Study Group. J Clin Oncol. 2003;21(19):3601-3608. 6. Meyer RM, Gospodarowicz MK, Connors JM, et al. Randomized comparison of ABVD chemotherapy with a strategy that includes radiation therapy in patients with limited-stage Hodgkin's lymphoma: National Cancer Institute of Canada Clinical Trials Group and the Eastern Cooperative Oncology Group. J Clin Oncol. 2005;23(21):4634-4642. 7. Hoppe RT, Advani RH, Ai WZ, et al. Hodgkin lymphoma, version 2.2012 featured updates to the NCCN guidelines. J Natl Compr Canc Netw. 2012;10(5):589597. 8. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32(27):3059-3068. 9. Mendenhall NP, Cantor AB, Barre DM, Lynch JW Jr, Million RR. The role of prognostic factors in treatment selection for early-stage Hodgkin's disease. Am J Clin
haematologica | 2016; 101(10)
to standard transaxial measurements identifies additional patients with disease bulk, and our data suggest that these patients are potentially at increased risk of relapse when treated with chemotherapy alone versus combined modality therapy. Routine review of transverse and coronal reformats on staging CT examinations in ESHL is feasible and has clinical impact. Acknowledgments The authors would like to thank collaborators at Dana Farber Cancer Institute and Mayo Clinic for their important contributions toward validating this work. Funding This research was supported by the Lymphoma Research Foundation and NIH.
Oncol. 1994;17(3):189-195. 10. North LB, Fuller LM, Hagemeister FB, Rodgers RW, Butler JJ, Shullenberger CC. Importance of initial mediastinal adenopathy in Hodgkin disease. Am J Roentgenology. 1982;138(2):229-235. 11. Bradley AJ, Carrington BM, Lawrance JA, Ryder WD, Radford JA. Assessment and significance of mediastinal bulk in Hodgkin's disease: comparison between computed tomography and chest radiography. J Clin Oncol. 1999;17(8):2493-2498. 12. Radford J, Illidge T, Counsell N, et al. Results of a Trial of PET-Directed Therapy for Early-Stage Hodgkin's Lymphoma. N Engl J Med. 2015;372(17):1598-1607. 13. Straus D, Pitcher BN, Kostakoglu L, al. E. Initial Results of US Intergroup Trial of response-adapted chemotherapy or chemotherapy/radiation therapy based on PET for non-bulky Stage I and II Hodgkin lymphoma (HL) (CALGB/Alliance 50604). ASH Annual Meeting Abstracts. Blood. 2015;126(23):578. 14. Rosset A, Spadola L, Pysher L, Ratib O. Informatics in radiology (infoRAD): navigating the fifth dimension: innovative interface for multidimensional multimodality image navigation. Radiographics. 2006; 26(1):299-308. 15. Mazumdar M, Glassman JR. Categorizing a prognostic variable: review of methods, code for easy implementation and applications to decision-making about cancer treatments. Stat Med. 2000;19(1):113-132. 16. Gonen M, Heller G. Concordance probability and discriminatory power in proportional hazards regression. Biometrika. 2005;92(4):965-970. 17. Klimm B, Goergen H, Fuchs M, et al. Impact of risk factors on outcomes in earlystage Hodgkin's lymphoma: an analysis of international staging definitions. Ann Oncol. 2013;24(12):3070-6. 18. Picardi M, De Renzo A, Pane F, et al. Randomized comparison of consolidation radiation versus observation in bulky Hodgkin's lymphoma with postchemotherapy negative positron emission tomography scans. Leuk Lymphoma. 2007;48(9):1721-1727. 19. Sandrasegaran K, Rydberg J, Tann M, Hawes DR, Kopecky KK, Maglinte DD.
20.
21.
22.
23.
24.
25.
26.
27.
Benefits of routine use of coronal and sagittal reformations in multi-slice CT examination of the abdomen and pelvis. Clin Radiol. 2007;62(4):340-347. Wei SC, Ulmer S, Lev MH, Pomerantz SR, Gonzalez RG, Henson JW. Value of coronal reformations in the CT evaluation of acute head trauma. Am J Neuroradiol. 2010; 31(2):334-339. Behar RA, Horning SJ, Hoppe RT. Hodgkin's disease with bulky mediastinal involvement: effective management with combined modality therapy. Int J Radiat Oncol Biol Phys. 1993;25(5):771-776. Hughes-Davies L, Tarbell NJ, Coleman CN, et al. Stage IA-IIB Hodgkin's disease: management and outcome of extensive thoracic involvement. Int J Radiat Oncol Biol Phys. 1997;39(2):361-369. Leopold KA, Canellos GP, Rosenthal D, Shulman LN, Weinstein H, Mauch P. Stage IA-IIB Hodgkin's disease: staging and treatment of patients with large mediastinal adenopathy. J Clin Oncol. 1989;7(8):10591065. Raemaekers JM, Andre MP, Federico M, et al. Omitting radiotherapy in early positron emission tomography-negative stage I/II Hodgkin lymphoma is associated with an increased risk of early relapse: Clinical results of the preplanned interim analysis of the randomized EORTC/LYSA/FIL H10 trial. J Clin Oncol. 2014;32(12):1188-1194. Meyer RM, Gospodarowicz MK, Connors JM, et al. ABVD alone versus radiation-based therapy in limited-stage Hodgkin's lymphoma. N Engl J Med. 2012;366(5):399-408. Straus DJ, Portlock CS, Qin J, et al. Results of a prospective randomized clinical trial of doxorubicin, bleomycin, vinblastine, and dacarbazine (ABVD) followed by radiation therapy (RT) versus ABVD alone for stages I, II, and IIIA nonbulky Hodgkin disease. Blood. 2004;104(12):3483-3489. Savage KJ, Connors JM, Villa DR, al. E. Advanced Stage Classical Hodgkin Lymphoma Patients with a Negative PETScan Following Treatment with ABVD Have Excellent Outcomes without the Need for Consolidative Radiotherapy Regardless of Disease Bulk at Presentation Blood (ASH Annual Meeting Abstracts). 2015;126.
1243
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Non-Hodgkin Lymphoma
Ferrata Storti Foundation
Haematologica 2016 Volume 101(10):1244-1250
Non-Hodgkin lymphoma in the developing world: review of 4539 cases from the International Non-Hodgkin Lymphoma Classification Project
Anamarija M. Perry,1 Jacques Diebold,2 Bharat N. Nathwani,3 Kenneth A. MacLennan,4 Hans K. Müller-Hermelink,5 Martin Bast,6 Eugene Boilesen,7 James O. Armitage,1 and Dennis D. Weisenburger3
Department of Pathology, University of Manitoba, Winnipeg, Manitoba, Canada; Department of Anatomic Pathology and Cytology, Hotel-Dieu, University Denis Diderot, Paris, France; 3Department of Pathology, City of Hope National Medical Center, Duarte, CA, USA; 4Section of Pathology and Leeds Institute of Molecular Medicine, St. James University Hospital, Leeds, UK; 5Institute of Pathology, University of Würzburg, Germany; 6 Department of Internal Medicine, University of Nebraska Medical Center, Omaha, NE, USA; and 7Center for Collaboration on Research Design and Analysis, College of Public Health, University of Nebraska Medical Center, Omaha, NE, USA 1 2
ABSTRACT
T Correspondence: dweisenburger@coh.org
Received: May 4, 2016. Accepted: June 23, 2016. Pre-published: June 27, 2016. doi:10.3324/haematol.2016.148809
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1244
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1244
he distribution of non-Hodgkin lymphoma subtypes varies around the world, but a large systematic comparative study has never been done. In this study, we evaluated the clinical features and relative frequencies of non-Hodgkin lymphoma subtypes in five developing regions of the world and compared the findings to the developed world. Five expert hematopathologists classified 4848 consecutive cases of lymphoma from 26 centers in 24 countries using the World Health Organization classification, and 4539 (93.6%) were confirmed to be non-Hodgkin lymphoma, with a significantly greater number of males than females in the developing regions compared to the developed world (P<0.05). The median age at diagnosis was significantly lower for both low- and high-grade B-cell lymphoma in the developing regions. The developing regions had a significantly lower frequency of B-cell lymphoma (86.6%) and a higher frequency of T- and natural killer-cell lymphoma (13.4%) compared to the developed world (90.7% and 9.3%, respectively). Also, the developing regions had significantly more cases of high-grade B-cell lymphoma (59.6%) and fewer cases of low-grade B-cell lymphoma (22.7%) compared to the developed world (39.2% and 32.7%, respectively). Among the B-cell lymphomas, diffuse large B-cell lymphoma was the most common subtype (42.5%) in the developing regions. Burkitt lymphoma (2.2%), precursor B- and T-lymphoblastic leukemia/lymphoma (1.1% and 2.9%, respectively) and extranodal natural killer/T-cell lymphoma (2.2%) were also significantly increased in the developing regions. These findings suggest that differences in etiologic and host risk factors are likely responsible, and more detailed epidemiological studies are needed to better understand these differences.
Introduction The relative frequencies of various subtypes of non-Hodgkin lymphoma (NHL) vary significantly in different geographic regions of the world,1-7 and environmental and lifestyle factors, as well as host genetic makeup, appear to play an important role in the development of NHL.8-10 The International NHL Classification Project was initiated in 1995 with a goal to investigate the geographic differences in NHL subtype distribution and clinical features.1,2 Between 1995 and 2012, five expert haematologica | 2016; 101(10)
Non-Hodgkin lymphoma in the developing world
hematopathologists visited 26 sites in 24 countries on five continents (Table 1), and found significant geographic differences in the relative frequencies and clinical features of various NHL subtypes.1-7 The aim of this study was to aggregate this data by region and evaluate the relative frequencies of NHL subtypes in five developing regions of the world. Moreover, we compare the findings in the developing regions to those in the developed world which, to our knowledge, has never been done before.
Methods International NHL Classification Project Twenty six institutions form 24 countries in seven regions including North America (NA), Western Europe (WEU), Southeastern Europe (SEEU), Central and South America (CSA), North Africa/the Middle East and India (NAF/ME/IN), Southern Africa (SAF), and the Far East (FE) participated in the study (Table 1). Each institution was instructed to collect 200 consecutive, newly-diagnosed and untreated cases of NHL representative of their local region or country. Leukemias were excluded from the study unless a tissue biopsy other than bone marrow was performed prior to therapy. Approval for this study was obtained from the Institutional Review Board of the University of Nebraska Medical Center and at each of the participating institutions as required by institutional policies. This study was conducted in accordance with the Declaration of Helsinki. At each site, hematoxylin and eosin-stained slides, immunostains, pathology reports, clinical data, and the results of ancillary studies were organized for review. A panel of five expert hematopathologists (JD, KAM, HKM-H, BNN, and DDW) then reviewed all of the cases using the 2001 World Health Organization classification.3,11 A consensus diagnosis was reached when at least four of the experts agreed on a diagnosis. For cases in which a consensus diagnosis could not be reached, a specific diagnostic algorithm was developed for each case and agreed upon by the experts. Additional requested clinical data and material, either paraffin blocks or unstained slides, were then sent to one of the experts who performed additional ancillary testing (immunostains, in situ stains, molecular studies, etc.) and assigned a consensus diagnosis to each case based on the algorithm. For this analysis, only information on age and sex is included because the clinical data collected at the various institutions was often incomplete and quite variable. Cases of composite lymphoma were classified according to the low-grade component, and mature B-cell NHL was further subdivided into low-grade (LG) and high-grade (HG) subgroups.3 Due to the small numbers, the rare T-cell subtypes were all grouped together under the category of peripheral T-cell lymphoma, other types. The data from the two developed regions (NA and WEU) were combined and compared to the data from the five developing regions, and the information from each developing region was also compared to the combined data from the other four developing regions.
Statistical analysis Data analysis was done using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA). Comparisons of medians for continuous variables were conducted using the Wilcoxon rank sum test. Comparisons of categorical variables were done using χ2 or Fisher's exact tests; the latter was used when the χ2 test may not have been valid due to small numbers. The P-values for pairwise comparisons were adjusted using the Hochberg step-up Bonferroni method, and adjusted P-values of less than 0.05 were considered to be statistically significant. haematologica | 2016; 101(10)
Table 1. Geographic regions and sites included in the study. North America Omaha, United States of America Vancouver, Canada Western Europe London, United Kingdom Würzburg, Germany Lyon, France Locarno, Switzerland Cape Town, South Africa* Southeastern Europe Zagreb, Croatia Cluj-Napoca, Romania Skopje, Macedonia Central/South America Guatemala City, Guatemala Lima, Peru Sao Paolo, Brazil Santiago, Chile Buenos Aires, Argentina North Africa/Middle East/India Algiers, Algeria Cairo, Egypt Kuwait City, Kuwait Riyadh, Saudi Arabia Mumbai, India Southern Africa** Cape Town, South Africa Johannesburg, South Africa Harare, Zimbabwe Far East Shanghai, China Hong Kong, China Bangkok, Thailand Jakarta, Indonesia *Whites only;: **non-whites only.
Results In this study, a total of 4848 cases were collected for expert review and 4539 (93.6%) were confirmed to be NHL, whereas the other 309 (6.4%) had diagnoses other than NHL and were excluded from further analysis. Among the excluded cases, 89 were Hodgkin lymphomas, 117 had diagnoses other than lymphoma, and 105 were unclassifiable cases. The number of reclassified cases in the developing regions (7.5%) was significantly higher than in the developed world (2.2%; P<0.05). The relative distribution of NHL subtypes in the developing and developed regions are shown in Table 2. Of the 3560 cases in the developing regions, 3082 (86.6%) were B-cell lymphomas and 478 (13.4%) were T- and natural killer (NK)-cell lymphomas. The developing regions had a significantly lower relative frequency of B-NHL and a higher frequency of T- and NK-cell NHL compared to the developed world (90.7% and 9.3%, respectively). Also, the developing regions had a significantly higher relative frequency of HG B-NHL (59.6%) and a lower frequency of LG B-NHL (22.7%) compared to the developed world (39.2% and 32.7%, respectively; Table 3). Comparison of the individual developing regions with the developed world showed that all of the developing regions had higher relative frequencies of HG B-NHL. Moreover, all of the developing regions, except SEEU, had a lower frequency of LG B-NHL compared to the developed world. When 1245
A.M. Perry et al.
compared to the rest of the developing regions, NAF/ME/IN and the FE had higher relative frequencies of HG B-NHL whereas SEEU and CSA had lower frequencies. Southeastern Europe had a significantly higher frequency of LG B-NHL whereas NAF/ME/IN had a lower frequency compared to the rest of the developing world.
Cases of pediatric NHL (age <19 years) comprised only 3.3% of the study cases, too few for meaningful analysis, and these are listed in the Online Supplementary Table S1. Among B-cell lymphomas (Table 2), diffuse large B-cell lymphoma (DLBCL) was the most common subtype in the developing world (42.5%), as well as in each of the
Table 2. Relative frequencies of non-Hodgkin lymphoma subtypes by region.
SEEU % (N)
CSA % (N)
NAF/ME/IN % (N)
SAF % (N)
FE % (N)
Developing Regions % (N)
Developed Regions % (N)
Chronic lymphocytic leukemia/small lymphocytic lymphoma
38.8! (231) 15.8! (94) 11.3*! (67)
39.0*! (357) 20.7*! (189) 3.1*! (28)
47.2*! (420) 12.4*! (110) 7.0 (62)
36.3 *! (177) 18.1! (88) 8.4 (41)
48.6*! (327) 9.4*! (63) 2.7*! (18)
42.5! (1512) 15.3! (544) 6.1 (216)
28.9 (283) 25.5 (250) 7.0 (69)
Marginal zone B-cell lymphoma, MALT type
6.6 (39)
7.0* (64)
2.7*! (24)
2.5*! (12)
6.8 (46)
5.2! (185)
8.8 (86)
Mantle cell lymphoma
5.9* (35) 3.7 (22) 1.5 (9) 2.4 (14) 0.3*! (2) 0.7 (4) 1.8 (11) 0.5 (3) 1.2 (7) 0.7 (4) 91.1* (542)
5.0 (46) 2.6 (24) 3.0! (27) 1.5 (14) 0.5*! (5) 2.4*! (22) 0.5 (5) 0.2! (2) 0.7 (6) 0.9 (8) 87.1 (797)
2.2*! (20) 2.0 (18) 2.7! (24) 2.2 (20) 1.6 (14) 1.1 (10) 0.9 (8) 0.1! (1) 1.1 (10) 3.6*! (32) 86.9 (773)
1.8! (9) 1.8 (9) 1.6 (8) 1.8 (9) 8.2*! (40) 0.2 (1) 0.2 (1) 0.8 (4) 0.6 (3) 3.3! (16) 85.8! (418)
3.6! (24) 2.2 (15) 1.8 (12) 1.9 (13) 1.0 (7) 0.4 (3) 1.5 (10) 0.3 (2) 0.7 (5) 1.0 (7) 82.0*! (552)
3.8! (134) 2.5 (88) 2.2! (80) 2.0 (70) 1.9 (68) 1.1! (40) 1.0 (35) 0.3! (12) 0.9 (31) 1.9! (67) 86.6! (3082)
7.8 (76) 3.0 (29) 0.8 (8) 2.1 (21) 2.5 (24) 0.3 (3) 0.4 (4) 1.4 (14) 1.2 (12) 0.9 (9) 90.7 (888)
2.7 (16) 1.5 (9) 0.8 * (5) 1.0 (6) 1.5 (9) 0.7 (4) 0.0 (0) 0.7 (4) 8.9* (53) 595
2.5 (23) 1.5* (14) 3.0! (27) 1.5 (14) 1.4 (13) 0.4 (4) 1.1*! (10) 1.4 (13) 12.9 (118) 915
2.7 (24) 4.0! (36) 1.1* (10) 3.3* (29) 0.7 (6) 0.7 (6) 0.0 (0) 0.7 (6) 13.1 (117) 890
6.4*! (31) 3.7! (18) 0.4* (2) 1.8 (9) 0.4 (2) 0.4 (2) 0.0 (0) 1.0 (5) 14.2! (69) 487
4.6 (31) 3.7! (25) 5.2*! (35) 1.9 (13) 2.5* (17) 0.0 (0) 0.0 (0) 0.0 (0) 18.0*! (121) 673
3.5 (125) 2.9! (102) 2.2! (79) 2.0 (71) 1.3 (47) 0.4 (16) 0.3 (10) 0.8 (28) 13.4! (478) 3560
2.6 (25) 1.3 (13) 0.3 (3) 2.2 (22) 1.2 (12) 1.0 (10) 0.1 (1) 0.5 (5) 9.3 (91) 979
B-cell lymphomas Diffuse large B-cell lymphoma Follicular lymphoma, all grades
Marginal zone lymphoma, nodal/splenic Burkitt lymphoma Primary mediastinal B-cell lymphoma High-grade B-cell lymphoma, Burkitt-like Precursor B-lymphoblastic leukemia/lymphoma Plasmacytoma Lymphoplasmacytic lymphoma Unclassifiable low-grade B-cell lymphoma Unclassifiable high-grade B-cell lymphoma Subtotal T- and NK-cell lymphomas Peripheral T-cell lymphoma, not otherwise specified Precursor T-lymphoblastic leukemia/lymphoma Extranodal NK/T-cell lymphoma, nasal type Anaplastic large T-cell lymphoma, ALK+ and ALKAngioimmunoblastic T-cell lymphoma Mycosis fungoides Adult T-cell leukemia/lymphoma Peripheral T-cell lymphoma, other types Subtotal Total
SEEU: Southeastern Europe; CSA: Central/South America; NAF/ME/IN: North Africa, Middle East and India; SAF: Southern Africa; FE: Far East. *Significantly different from the other developing regions combined. !Significantly different from the developed regions (North America and Western Europe). NK: natural killer; ALK: anaplastic lymphoma kinase.
1246
haematologica | 2016; 101(10)
Non-Hodgkin lymphoma in the developing world
developing regions. Follicular lymphoma (FL) was the second most common subtype in the developing world (15.3%) and was the most common in CSA (20.7%). Among the other B-cell lymphomas, marginal zone lymphoma (MZL) of mucosa-associated lymphoid tissue (MALT) type (7%) was more common in CSA, chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL; 11.3%) and mantle cell lymphoma (MCL; 5.9%) in SEEU, precursor B-lymphoblastic leukemia/lymphoma (2.4%) in CSA, and Burkitt-like lymphoma (8.2%) in SAF. When individual lymphoma subtypes in the developing regions were compared to the developed world (Table 2), DLBCL (42.5%) was significantly more frequent and FL (15.3%) was less frequent than in the developed world (28.9% and 25.5%, respectively). Furthermore, when FL was separated into LG and HG subtypes (Table 3), the developing regions had a significantly lower relative frequency of LG FL (9.5%) compared to the developed world (18.9%). Analysis of DLBCL and FL frequencies in the individual developing regions (Table 2) showed that all of the regions had higher relative frequencies of DLBCL and lower frequencies of FL compared to the developed world. When compared to the other developing regions, NAF/ME/IN and the FE had significantly higher relative frequencies of DLBCL and lower frequencies of FL. Burkitt lymphoma (2.2%) and precursor B-lymphoblastic leukemia/lymphoma (1.1%) were also significantly more common in the developing regions compared to the developed world (0.8% and 0.3%, respectively). Among the individual regions, CSA and NAF/ME/IN had a significantly higher frequency of Burkitt lymphoma compared to the developed world. Furthermore, CSA had a higher relative frequency of precursor B-lymphoblastic leukemia/lymphoma than the rest of the developing regions or the developed world. The developing regions had significantly lower frequencies of MZL of MALT type (5.2%) and MCL (3.8%) compared to the developed world (8.8% and 7.8%, respectively). The relative frequency of MZL of MALT type was significantly lower in the NAF/ME/IN and SAF compared to the rest of the developing regions. As for MCL, NAF/ME/IN, SAF and the FE had lower relative frequencies compared to the developed world. Southeastern Europe had a higher frequency of CLL/SLL compared to the other developing regions and the developed world. Finally, the developing regions had a higher relative frequency of unclassifiable high-grade B-cell lymphoma (1.9%) compared to the developed world (0.9%), mainly due to increases in NAF/ME/IN and SAF.
Among the T- and NK-cell lymphomas (Table 2), peripheral T-cell lymphoma (PTCL), not otherwise specified, was the most common subtype in SEEU (2.7%) and SAF (6.4%). Extranodal NK/T-cell lymphoma was the most common subtype in CSA (3%) and the FE (5.2%), whereas the most common subtype of PTCL in NAF/ME/IN was anaplastic large T-cell lymphoma. Furthermore, angioimmunoblastic T-cell lymphoma was relatively common in the FE (2.5%), and adult T-cell leukemia/lymphoma was seen only in CSA (1.1%). When compared to the developed world, the developing regions had significantly higher relative frequencies of extranodal NK/T-cell lymphoma (2.2%) and precursor T-lymphoblastic leukemia/lymphoma (2.9%). The rare subtypes and unclassifiable cases of T-cell lymphoma were too few for meaningful analysis and are listed in the Online Supplementary Table S2. The distribution of NHL by sex and age is shown in Table 4. The ages of patients in the developing regions ranged from less than 1 to 105 years old, and 57.1% were male. In the developed world, the age range was 4 to 100 years old, and 51.1% were male. There was a significantly greater number of males in the developing regions compared to the developed world. When the individual regions were compared to the developed world, NAF/ME/IN and the FE had a significantly greater number of males. The median ages of patients with LG B-NHL (61 years) and HG B-NHL (53 years) were significantly lower in the developing regions than in the developed world (63 and 65 years, respectively). The median age of patients with HG B-NHL was also significantly lower in each of the developing regions compared to the developed world. Furthermore, in NAF/ME/IN and SAF, LG B-NHL patients also had a significantly lower median age compared to the developed world. Among the individual regions, SEEU and CSA had a significantly higher median age for HG BNHL patients, whereas SAF had a lower median age than the other developing regions. Patients with HG FL in the developing regions also had a significantly lower median age (58 years) than those in the developed world (65 years), whereas there were no differences for LG FL. There were also no significant differences in the median ages of T- and NK-cell NHL patients between the developing and developed world.
Discussion This study included 4848 cases initially diagnosed as NHL from 24 countries in the developing and developed
Table 3. Relative frequencies of mature B-cell non-Hodgkin lymphomas according to grade.
B-cell Lymphomas
SEEU % (N)
CSA % (N)
NAF/ME/IN % (N)
SAF % (N)
FE % (N)
Developing Regions % (N)
Developed Regions % (N)
Low-grade B-cell lymphoma
33.9* (184) 48.7*! (264) 7.6! (41) 9.8 (53)
22.0! (175) 54.3*! (433) 14.6*! (116) 9.2 (73)
18.5*! (143) 67.3*! (520) 6.7*! (52) 7.5 (58)
18.9! (79) 60.0! (251) 13.2*! (55) 7.9 (33)
21.7! (120) 66.8 *! (369) 5.3*! (29) 6.2 (34)
22.7! (701) 59.6! (1837) 9.5! (293) 8.1 (251)
32.7 (290) 39.2 (348) 18.9 (168) 9.2 (82)
High-grade B-cell lymphoma Follicular lymphoma, low-grade Follicular lymphoma, high-grade
SEEU: Southeastern Europe; CSA: Central/South America; NAF/ME/IN: North Africa, Middle East and India; SAF: Southern Africa; FE: Far East. *Significantly different from the other developing regions combined. !Significantly different from the developed regions (North America and Western Europe).
haematologica | 2016; 101(10)
1247
A.M. Perry et al.
world and, to our knowledge, represents the largest epidemiological study of this kind ever done. We found that the developing regions had a significantly higher frequency of misclassified cases (7.5%) than the developed world (2.2%). The field of lymphoma pathology is challenging and rapidly changing.12 Lymphomas are currently classified according to the 2008 WHO Classification,13 which mandates correlation of the morphological findings with the clinical features and often requires complex ancillary studies to make a diagnosis. In resource-poor countries, pathologists often lack training and experience in hematopathology, as well as the technology to perform the necessary ancillary studies.14,15 Moreover, the developing regions had a higher number of unclassifiable cases, which mainly reflects problems associated with specimen handling and tissue processing that result in poor-quality slides.12 Our findings suggest that training in hematopathology, as well as technical improvements, are needed in the developing world to decrease the number of misdiagnosed cases. The distribution of NHL subtypes in the five developing regions was markedly different compared to the developed world (Table 2). Overall, the developing regions had a significantly lower relative frequency of B-cell NHL and a higher frequency of T- and NK-cell NHL. Among the B-cell lymphomas, a significantly higher relative frequency of HG B-NHL and a lower frequency of LG B-NHL were observed in the developing regions compared to the developed world. The most common lymphoma in the developing regions was DLBCL, whereas the relative frequency of FL was low. Notably, a high frequency of highgrade Burkitt-like lymphoma was observed in SAF, which is likely due to the human immunodeficiency virus (HIV) epidemic occurring there.16 Other studies from developing countries have reported a similar distribution pattern of Bcell lymphoma, with a high relative frequency of DLBCL and a low frequency of FL.17-20 These differences are likely multifactorial in origin. In North America, whites have the highest incidence of FL of all races.8 Studies that have looked at the incidence of different lymphoma subtypes in foreign-born Asians and Asians born in the United States (US) have shown a higher incidence of FL in US-born
Asians, supporting a role for environmental and lifestyle factors in the development of FL.21,22 Among lifestyle factors that have been studied, a high intake of red meat and saturated fat has been associated with an increased risk of FL.23-25 This could explain, at least in part, the high incidence of FL in the developed world since developing countries have lower meat consumption per capita than the developed world.26 Moreover, studies from Japan27 and Taiwan28 have reported a significant increase in the relative frequency of FL in recent times. Therefore, meat and saturated fat consumption may be surrogate markers for western lifestyle. Additional studies are needed to better understand these differences The developing regions also had higher relative frequencies of other aggressive B-cell lymphomas including Burkitt lymphoma and unclassifiable HG B-NHL, as well as both B- and T-lymphoblastic leukemia/lymphoma. In our study, Burkitt lymphoma was more frequent in all of the developing regions compared to the developed world, but most prominently in CSA and NAF/ME/IN. Burkitt lymphoma has three clinical forms â&#x20AC;&#x201C; endemic, immunodeficiency-associated and sporadic forms. Epstein-Barr virus (EBV) infection is implicated in virtually all cases of endemic Burkitt lymphoma, but is also seen in a significant proportion of the other types of Burkitt lymphoma.13 The occurrence of EBV in Burkitt lymphoma is higher in developing countries, where EBV seroconversion occurs at an early age.29 A study of US children found a significantly higher seroprevalence of EBV infection in children of lower socioeconomic status.30 Furthermore, the seroprevalence of EBV infection in Chinese children is higher compared to US children.30,31 These findings may explain, at least in part, the higher frequency of Burkitt lymphoma in the developing regions. Reasons for the higher relative frequency of B- and T-lymphoblastic leukemia/lymphoma in the developing world are unclear. Other studies from developing countries have reported similar findings,32-34 and further large epidemiological studies are needed to better understand these differences. Our study also found a significantly higher relative frequency of T- and NK-cell lymphomas in the developing regions, with the highest frequencies observed in SAF and the FE. In western countries, T-cell lymphomas account
Table 4. Sex distribution and median ages by region for non-Hodgkin lymphoma.
Sex Male Female Median age, years B-cell lymphomas Low-grade High-grade Follicular lymphoma, low-grade Follicular lymphoma, high-grade T- and NK-cell lymphomas
SEEU % (N)
CSA % (N)
NAF/ME/IN % (N)
SAF % (N)
FE % (N)
Developing Regions % (N)
Developed Regions % (N)
50.1* (298) 49.9* (297)
50.8* (450) 49.2* (436)
64.0*! (564) 36.0*! (317)
57.5 (277) 42.5 (205)
62.3*! (417) 37.7*! (252)
57.1! (2006) 42.9! (1507)
51.1 (482) 48.9 (461)
62.0 59.0*! 55.0
63.0 58.0*! 57.0
58.0! 52.0! 53.5
58.0! 43.0*! 53.5
59.0 51.0! 52.0
61.0! 53.0! 55.0
63.0 65.0 58.0
57.0
60.0
49.0!
63.0
55.5
58.0!
65.0
49.0
46.0
38.5
41.0
49.0
45.0
52.5
SEEU: Southeastern Europe; CSA: Central/South America; NAF/ME/IN: North Africa, Middle East and India; SAF: Southern Africa; FE: Far East. *Significantly different from the other developing regions combined. !Significantly different from the developed regions (North America and Western Europe). NK: natural killer.
1248
haematologica | 2016; 101(10)
Non-Hodgkin lymphoma in the developing world
for 5-10% of all NHL, whereas in the Asian countries, 15-20% of all NHL are classified as T- or NK-cell lymphomas.35 In the US, the incidence of PTCL increased by 280% between 1992 and 2005, and the cause for this increase is largely unknown.36 However, pathologists are diagnosing T-cell lymphomas more accurately today due to changes in lymphoma classification over the last 20 years as well as a wider availability of ancillary studies, which could be contributing to the increasing incidence of PTCL. Interestingly, in the US, the incidence of PTCL is the highest in blacks.8,36,37 which could also explain the high incidence of this lymphoma in SAF where most of the patients were black or mixed-race Africans.5 Further large epidemiological and genetic studies are needed to explain these differences. Extranodal NK/T-cell lymphoma, nasal type, was significantly increased in CSA and the FE in our study. An increased frequency of extranodal NK/T-cell lymphoma in Asia and the native populations of CSA has been well documented.13,19,34 Extranodal NK/T-cell lymphoma has a very strong association with EBV infection and studies suggest that genetics plays a significant role in the development of this lymphoma.8 Asians who live in the US have a significantly increased incidence of this lymphoma compared to other races.8,36.37 Moreover, studies have shown a similar incidence of this lymphoma among US-born and foreignborn Asians.21,38 Our study found that extranodal NK/T-cell lymphoma was particularly increased in the native populations of Guatemala, Chile and Peru, which share a similar genetic background with Asians.3 We also observed a significant difference in the sex distribution of NHL in the developing regions, with a significantly higher number of males compared to the developed world. Even though NHL overall is more common in males,39 there are likely additional factors in the developing world that contribute to the increased number of males. Sex inequality in health care is common in some developing countries. Women may have less access to
References 6. 1. A clinical evaluation of the International Lymphoma Study Group classification of non-Hodgkin's lymphoma. The NonHodgkin's Lymphoma Classification Project. Blood. 1997;89(11):3909-3918. 2. Anderson JR, Armitage JO, Weisenburger DD. Epidemiology of the non-Hodgkin's lymphomas: distributions of the major subtypes differ by geographic locations. NonHodgkin's Lymphoma Classification Project. Ann Oncol. 1998;9(7):717-720. 3. Laurini JA, Perry AM, Boilesen E, et al. Classification of non-Hodgkin lymphoma in Central and South America: a review of 1028 cases. Blood. 2012;120(24):4795-4801. 4. Dotlic S, Perry AM, Petrusevska G, et al. Classification of non-Hodgkin lymphoma in South-eastern Europe: review of 632 cases from the International non-Hodgkin Lymphoma Classification Project. Br J Haematol. 2015;171(3):366-372. 5. Perry AM, Perner Y, Diebold J, et al. NonHodgkin lymphoma in Southern Africa: review of 487 cases from the International
haematologica | 2016; 101(10)
7.
8.
9.
10.
medical care or may seek medical care less frequently and, therefore, it is possible that lymphomas are underdiagnosed in females in these countries.40,41 Further study is needed to confirm and explain these findings. The overall median age of LG and HG B-NHL patients in the developing regions was significantly lower than in the developed world. A lower median age for HG B-NHL was seen in all five of the developing regions, and NAF/ME/IN and SAF also had a significantly lower median age for LG B-NHL patients. Other studies from the developing world have also reported lower median ages in B-NHL patients.19,20,34,42 Our finding that B-NHL occurs at an earlier age seems to correlate with a lower gross domestic product in the developing world,43 and suggests that socioeconomic factors may play a role in lymphomagenesis. The much lower median age of patients with HG B-NHL observed in SAF can be attributed to HIV infection, which predominantly affects young blacks.44 In conclusion, this study is the largest systematic study of the distribution of NHL subtypes in the developing world, and the first to compare its findings to the developed world. Our data from the developed world is very similar to data recently reported from the National Cancer Database in the United States.45 We were unable to calculate the incidence rates of the different NHL subtypes in the developing regions due to the lack of centralized and comprehensive population-based cancer registries in many of the countries. However, we did find significant differences in the relative frequencies of many NHL subtypes between the developing and developed world, as well as differences in age and sex. The reasons for these differences are likely multifactorial, and large epidemiological studies are needed to confirm and better explain our findings. Acknowledgments We wish to thank all of the clinicians and pathologists who participated in this study (Online Supplementary Table S3).
Non-Hodgkin Lymphoma Classification Project. Br J Haematol. 2016;172(5):716-723. Perry AM, Diebold J, Nathwani BN, et al. Relative frequency of non-Hodgkin lymphoma subtypes in selected centres in North Africa, the Middle East and India: a review of 971 cases. Br J Haematol. 2015;172(5):699-708. Perry AM, Diebold J, Nathwani BN, et al. Non-Hodgkin lymphoma in the Far East: review of 730 cases from the International non-Hodgkin Lymphoma Classification Project. Ann Hematol. 2016;95(2):245-251. Morton LM, Wang SS, Devesa SS, Hartge P, Weisenburger DD, Linet MS. Lymphoma incidence patterns by WHO subtype in the United States, 1992-2001. Blood. 2006;107(1):265-276. Cerhan JR, Linet MS, Maskarinec G, et al, ed. Medical history, lifestyle, family history, and occupational risk factors for nonHodgkin lymphoma subtypes: pooled analysis from the International Lymphoma Epidemiology Consortium. J Natl Cancer Inst Monogr. 2014;(48):76-86. Chiu BC, Hou N. Epidemiology and etiology of non-Hodgkin's lymphoma. Cancer
Treat Res. 2015;165:1-25. 11. Jaffe ES, Harris NL, Stein H, Vardiman JW. ed. World Health Organization Classification of Tumours: Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press; 2001. 12. Wilkins BS. Pitfalls in lymphoma pathology: avoiding errors in diagnosis of lymphoid tissues. J Clin Pathol. 2011;64(6):466476. 13. Swerdlow SH, Campo E, Harris NL, et al, eds. World Health Organization Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC; 2008. 14. Chang C, Huang SW, Su IJ, Chang KC. Hematopathologic discrepancies between referral and review diagnoses: a gap between general pathologists and hematopathologists. Leuk Lymphoma. 2014;55(5):1023-1030. 15. Cainelli F, Tanko MN, Vento S. The challenge of lymphomas in sub-Saharan Africa. Lancet Oncol. 2010;11(7):610-611. 16. Wiggill TM, Mayne ES, Willem P. Challenges in lymphoma diagnosis in HIV
1249
A.M. Perry et al.
17.
18.
19.
20.
21.
22.
23.
24.
25.
1250
positive patients in the South African setting. Transfus Apher Sci. 2013;49(2):157162. World Health Organization Classification of malignant lymphomas in Japan: incidence of recently recognized entities. Lymphoma Study Group of Japanese Pathologists. Pathol Int. 2000;50(9):696702. Abdel-Fattah MM and Yassine OG. NonHodgkin's lymphomas in Alexandria, Egypt; incidence rates and trend study (1995-2004). Eur J Cancer Prev. 2007;16(5):479-485. Sun J, Yang Q, Lu Z, et al. Distribution of lymphoid neoplasms in China: analysis of 4,638 cases according to the World Health Organization classification. Am J Clin Pathol. 2012;138(3):429-434. Arora N, Manipadam MT, Nair S. Frequency and distribution of lymphoma types in a tertiary care hospital in South India: analysis of 5115 cases using the World Health Organization 2008 classification and comparison with world literature. Leuk Lymphoma. 2013;54(5):1004-1011. Herrinton LJ, Goldoft M, Schwartz SM, Weiss NS. The incidence of non-Hodgkin's lymphoma and its histologic subtypes in Asian migrants to the United States and their descendants. Cancer Causes Control. 1996;7(2):224-230. Clarke CA, Glaser SL, Gomez SL, et al. Lymphoid malignancies in U.S. Asians: incidence rate differences by birthplace and acculturation. Cancer Epidemiol Biomarkers Prev. 2011;20(6):1064-1077. Cross AJ, Ward MH, Schenk M, et al. Meat and meat-mutagen intake and risk of nonHodgkin lymphoma: results from a NCISEER case-control study. Carcinogenesis. 2006;27(2):293-297. Aschebrook-Kilfoy B, Ollberding NJ, Kolar C, et al. Meat intake and risk of nonHodgkin lymphoma. Cancer Causes Control. 2012;23(10):1681-1692. Ollberding NJ, Aschebrook-Kilfoy B, Caces DB, et al. Phytanic acid and the risk of nonHodgkin lymphoma. Carcinogenesis. 2013;
34(1):170-175. 26. Food and Agriculture Organization of the United Nations. The State of Food and Agriculture 2009. Available from: http://www.fao.org/docrep/012/i0680e/i06 80e.pdf. [Last accessed: January 10, 2016]. 27. Miyazato H, Nakatsuka S, Miyanaga I, et al. Follicular lymphoma in Osaka, Japan: histological features and chronological change. Int J Hematol. 2002;76(4):333-337. 28. Chuang SS. Significant increase in the relative frequency of follicular lymphoma in Taiwan in the early 21st century. J Clin Pathol. 2008;61(7):879-880. 29. Anwar N, Kingma DW, Bloch AR, et al. The investigation of Epstein-Barr viral sequences in 41 cases of Burkitt's lymphoma from Egypt: epidemiologic correlations. Cancer. 1995;76(7):1245-1252. 30. Dowd JB, Palermo T, Brite J, McDade TW, Aiello A. Seroprevalence of Epstein-Barr virus infection in U.S. children ages 6-19, 2003-2010. PLoS One. 2013;8(5):e64921. 31. Xiong G, Zhang B, Huang MY, et al. Epstein-Barr virus (EBV) infection in Chinese children: a retrospective study of age-specific prevalence. PLoS One. 2014;9(6):e99857. 32. Sukpanichnant S. Analysis of 1983 cases of malignant lymphoma in Thailand according to the World Health Organization classification. Hum Pathol. 2004;35(2):224-230. 33. Naresh KN, Advani S, Adde M, et al. Report of an International Network of Cancer Treatment and Research workshop on non-Hodgkin's lymphoma in developing countries. Blood Cells Mol Dis. 2004; 33(3):330-337. 34. Yang QP, Zhang WY, Yu JB, et al. Subtype distribution of lymphomas in Southwest China: analysis of 6,382 cases using WHO classification in a single institution. Diagn Pathol. 2011;6:77-1596-6-77. 35. Vose J, Armitage J, Weisenburger D, International T-Cell Lymphoma Project. International peripheral T-cell and natural killer/T-cell lymphoma study: pathology findings and clinical outcomes. J Clin Oncol. 2008;26(25):4124-4130.
36. Abouyabis AN, Shenoy PJ, Lechowicz MJ, Flowers CR. Incidence and outcomes of the peripheral T-cell lymphoma subtypes in the United States. Leuk Lymphoma. 2008;49(11):2099-2107. 37. Adams SV, Newcomb PA, Shustov AR. Racial patterns of peripheral T-cell lymphoma incidence and survival in the United States. J Clin Oncol. 2016;34(9):963-971. 38. Pan JW, Cook LS, Schwartz SM, Weis NS. Incidence of leukemia in Asian migrants to the United States and their descendants. Cancer Causes Control. 2002;13(9):791795. 39. National Cancer Institute. Surveillance, Epidemiology, and End Results (SEER) Program. SEER Stat Fact Sheets: NonHodgkin Lymphoma. Available from: http://seer.cancer.gov/statfacts/html/nhl.ht ml. [Last accessed: January 4, 2016]. 40. Gijsbers van Wijk CM, van Vliet KP, Kolk AM. Gender perspectives and quality of care: towards appropriate and adequate health care for women. Soc Sci Med. 1996;43(5):707-720. 41. Balarajan Y, Selvaraj S, Subramanian SV. Health care and equity in India. Lancet. 2011;377(9764):505-515. 42. Goldman L, Ezzat S, Mokhtar N, et al. Viral and non-viral risk factors for nonHodgkin's lymphoma in Egypt: heterogeneity by histological and immunological subtypes. Cancer Causes Control. 2009;20(6):981-987. 43. The World Bank Database. GDP per capita. Available from: http:// d a t a . w o r l d b a n k . o r g / i n d i c a t o r / N Y. DP.PCAP.CD. [Last accessed: January 4, 2016]. 44. UNAIDS. Countries. Available from: http://www.unaids.org/en/regionscountries/countries. [Last accessed: July 10, 2015]. 45. Al-Hamadani M, Habermann TM, Cerhan JR, Macon WR, Maurer MJ, Go RS. NonHodgkin lymphoma subtype distribution, geodemographic patterns, and survival in the US: A longitudinal analysis of the National Cancer Data Base from 1998 to 2011. Am J Hematol. 2015;90(9):790-795.
haematologica | 2016; 101(10)
ARTICLE
Cell Therapy & Immunotherapy
A phase I study of CD25/regulatory T-cell-depleted donor lymphocyte infusion for relapse after allogeneic stem cell transplantation
EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
Sarah Nikiforow,1,2 Haesook T. Kim,3,4 Heather Daley,1 Carol Reynolds,1 Kyle Thomas Jones,1 Philippe Armand,1,2 Vincent T. Ho,1,2 Edwin P. Alyea III,1,2 Corey S. Cutler,1,2 Jerome Ritz,1,2 Joseph H. Antin,1,2 Robert J. Soiffer,1,2 and John Koreth1,2
Division of Hematologic Malignancies, Department of Medical Oncology, Dana-Farber Cancer Institute; 2Harvard Medical School; 3Department of Biostatistics and Computational Biology, Dana-Farber Cancer Institute; and 4Harvard School of Public Health, Boston, MA, USA
1
Haematologica 2016 Volume 101(10):1251-1259
ABSTRACT
D
onor lymphocyte infusions are used to treat relapse after allogeneic hematopoietic stem cell transplantation, but responses are inadequate. In addition to effector cells, infusions contain CD25+ regulatory T cells (Treg) that may suppress graft-versus-tumor responses. We undertook a phase I study of donor lymphocyte infusions depleted of CD25+ T cells in patients with hematologic malignancies who had relapsed after transplantation. Twenty-one subjects received CD25/Treg-depleted infusions following removal of CD25+ cells using antibody-conjugated magnetic beads. Sixteen subjects received prior cytoreductive therapy. Four were in complete remission at the time of infusion. Two dose levels were administered: 1x107 (n=6) and 3x107 CD3+ cells/kg (n=15). A median 2.3 log-depletion of CD4+CD25+FOXP3+ Treg was achieved. Seven subjects (33%) developed clinically significant graft-versus-host disease by 1 year, including one patient who died. At dose level 1, five subjects had progressive disease and one had stable disease. At dose level 2, nine subjects (60%) achieved or maintained responses (8 complete responses, 1 partial response), including seven with active disease at the time of infusion. A shorter period between relapse and infusion was associated with response at dose level 2 (P=0.016). The 1-year survival rate was 53% among patients treated with dose level 2. Four of eight subjects with acute myeloid leukemia remained in remission at 1 year. When compared to unmodified donor lymphocyte infusions in 14 contemporaneous patients meeting study eligibility, CD25/Treg depletion was associated with a better response rate and improved event-free survival. Circulating naĂŻve and central memory CD4+ T cells increased after CD25/Treg-depleted infusion, but no immunophenotypic signature for response was noted. CD25/Tregdepleted donor infusion appears feasible and capable of inducing graft-versus-tumor responses without excessive graft-versus-host disease. (ClinicalTrials.gov NCT#00675831)
Correspondence: sarah_nikiforow@dfci.harvard.edu
Received: January 4, 2016. Accepted: June 15, 2016. Pre-published: June 27, 2016. doi:10.3324/haematol.2015.141176
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1251
Introduction Patients with a hematologic malignancy who relapse after allogeneic hematopoietic stem cell transplantation (HSCT) have a dismal prognosis. Re-establishing disease control through donor lymphocyte infusion (DLI) is one accepted approach, as this can invoke graft-versus-tumor effects mediated by recognition of minor histocompatibility antigens and tumor antigens on malignant cells.1 DLI has been successful in relapsed chronic myeloid leukemia with greater than 70% sustained response rates.2,3 Its efficacy in other diseases is limited by two factors. First, the potency and durability of the graft-versus-tumor effects vary, with response rates in diseases other than chronic myeloid leukemia being much lower.4-6 Initial responses to DLI can be as low as 15-29% in acute myeloid leukemia and 5-27% in acute lymhaematologica | 2016; 101(10)
Š2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1251
S. Nikiforow et al.
phoid leukemia, with a median time of 10 months to leukemia recurrence.3,7,8 Second, DLI is often accompanied by toxicity from graft-versus-host disease (GvHD). The incidence of GvHD varies depending on the DLI cell dose, degree of HLA mismatch, and use of concurrent immunosuppression but ranges from 25% to 70% after DLI from fully HLA-matched related or unrelated donors.9-12 Strategies to enhance the efficacy of graft-versus-tumor effects after DLI have included ex vivo activation of effector cells with CD3/CD28-coated beads; incubation with interferon-γ, interleukin-2, and anti-CD3 to produce “cytokine-induced killer cells”; or incubation with interleukin-2 alone. These approaches have resulted in disappointing efficacy and/or unacceptable rates of GvHD.13-16 CD4+CD25+FoxP3+ regulatory T cells (Treg) account for ~5% to 10% of circulating CD4+ T cells in healthy individuals, can dominantly suppress auto-reactive T, B and natural killer effectors, and control innate and adaptive immune responses.17,18 In vitro Treg suppress proliferation of CD4+ and CD8+ T cells after polyclonal or antigenic stimulation.19 In vivo Treg can impede immune responses to solid and liquid tumors.20-23 They also play a role in alleviating GvHD as demonstrated by the inverse correlation between circulating Treg and the onset and severity of GvHD.24-28 Surprisingly, adoptive transfer or in vivo expansion of Treg improves GvHD without jeopardizing graftversus-tumor effects.29-32 We hypothesized that, elimination of CD25+ cells, including CD25-expressing Treg, from DLI products by selective depletion of CD25+ cells might boost anti-tumor efficacy, with the risk that potency gained could be offset by worsened GvHD. The feasibility of large-scale CD25/Treg depletion from apheresis products has previously been demonstrated using bead-bound anti-CD25 antibodies and magnetic separation.33 We evaluated the feasibility and safety of administering CD25/Treg-depleted DLI to subjects with a hematologic malignancy who had relapsed after allogeneic HSCT, a population of patients for whom there are few effective therapeutic options.
Methods This phase I, dose-escalation trial was approved by DanaFarber/Harvard Cancer Center Institutional Review Board (ClinicalTrials.gov NCT#00675831). Its primary objectives were to determine the feasibility and safety of depleting CD25+ cells from leukapheresis products. The secondary objective was to assess response to the depleted DLI.
Patients’ characteristics
Eligible subjects were ≥18 years old, with relapsed hematologic malignancies (other than those with stable-phase chronic myeloid leukemia), ≥2 months after HLA –A, -B, -C and -DRB1-matched donor allogeneic HSCT with donor total leukocyte chimerism ≥20%, and disease involving ≤50% bone marrow cellularity and/or lymph nodes ≤5 cm. The subjects had been off systemic immune suppression for ≥2 weeks, without active GvHD, and had not received chemotherapy (except hydroxyurea) within 4 weeks or immunotherapy within 8 weeks of the DLI.
Donor leukocyte infusion Original stem cell donors underwent one or two leukaphereses without growth-factor stimulation. If twice the CD3+ cell/kg tar1252
get dose for infusion was not met, an unmanipulated DLI was administered. Planned dose levels were 1x107 and 3x107 CD3+ cells/kg; a lower dose of 1x106/kg was also envisaged if toxicities were encountered. CD25/Treg depletion was achieved using the CliniMACS CD25 Reagent System according to the manufacturer’s instructions (Miltenyi Biotec, Cambridge, MA, USA; IDE 13423). Release criteria included CD3+ cell dose, depletion of CD4+CD25high cells to ≤0.5%, and ≥70% viability. Standard methods were used for the immunophenotypic analyses of DLI products and recipients’ peripheral blood before and after CD25/Tregdepleted DLI (Online Supplementary Methods).
Study design and definitions CD25/Treg-depleted DLI entailed a single infusion of fresh cells immediately following selection. Recipients not requiring chemotherapy within 8 weeks after DLI were evaluable for doselimiting toxicities: grade III-IV acute GvHD, severe pancytopenia, or DLI-related CTCAE ≥ grade 3 toxicities unrelated to progressive disease or GvHD. Acute GvHD was assessed using consensus criteria, while chronic GvHD was graded “limited” or “extensive”.34,35 A complete response consisted of resolution of histological/radiological evidence of disease and chromosomal abnormalities (acute leukemia). A partial response was defined by ≥50% reduction, stable disease by <50% reduction or <50% increase, and progressive disease by ≥50% increase of lymph node or bone marrow disease burden.
Exploratory contemporaneous comparator cohort In an exploratory analysis, a contemporaneous comparator cohort of patients from the Dana-Farber Harvard Cancer Center was identified. This cohort comprised 14 patients with relapsed hematologic malignancy (other than those with stable phase chronic myeloid leukemia) after HLA-matched HSCT, with ≥20% donor chimerism, who received unmanipulated DLI with 2-4x107 CD3+ cells/kg (dose level 2), and who were not on study secondary to the patients’ preference or logistical challenges.
Statistical analyses Descriptive statistics were used for patient- and transplant-related characteristics. Fisher exact test, the χ2 test, or Wilcoxon-ranksum tests were used for cohort comparisons. Overall survival and event-free survival were estimated using the Kaplan-Meier method. Differences in survival curves between cohorts were tested using the log-rank test. Cumulative incidences of non-relapse mortality and relapse/progression were constructed in a framework of competing risks and differences evaluated using the Gray test.36 Repeated measures analysis was performed for effect of time and response on immunophenotypic profiles. Immunophenotypic parameters were log- or square-root transformed to meet the normality assumption prior to modeling. Multiplicity was adjusted within each model. P-values were twosided with a significance level of 0.05. Analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC, USA) and R version 2.15.12 (the CRAN project).
Results Feasibility Between May, 2008 and October, 2011, 24 subjects were enrolled and evaluable for feasibility. Two subjects received unmanipulated DLI because of an insufficient CD3+ cell collection from the donor after two leukaphereses. One subject developed GvHD shortly after enrollment, so DLI was not pursued. Of the 21 CD25/Treghaematologica | 2016; 101(10)
CD25/Treg-depleted DLI for relapse after SCT
depleted products administered, 20 were collected in a single leukapheresis. The median CD3+ recovery after CD25 depletion was 73.3% (range, 46.1 - 107.7%) (Table 1). The median log-depletion of CD4+CD25high cells was 2.46 (range, 0.83 - 3.37), while that of CD4+CD25+FoxP3+ Treg was 2.28 (range, 0.73-3.33) (Figure 1). A similar log-depletion (median 2.32) of Treg, based on CD4+CD25+CD127low surface expression, was achieved. At dose level 2, the median number of residual Treg infused was 4.24 x 103/kg (range, 0.64 x 103/kg to 95.66 x 103/kg).
leukemia, had received prior unmanipulated DLI without achieving a sustained response. Four (19%) subjects were in complete response and seven (33%) in partial response prior to DLI. Five (24%) had morphological disease without receipt of cytoreductive therapy since relapse, and five (24%) had progressive/persistent disease despite therapy. The median total leukocyte donor chimerism was 99%, (range, 32% -100%) (Online Supplementary Table S1). The median follow-up among survivors was 44 months (range, 26 – 59 months).
Subjects receiving CD25/Treg-depleted donor lymphocyte infusions The median age of the 21 subjects evaluable for safety and efficacy was 42 years (range, 19-71 years) (Table 2). The majority (n=15, 71%) had relapsed acute myeloid leukemia, acute lymphoblastic leukemia, or myelodysplastic syndrome. Thirteen subjects (62%) received cells from matched siblings, eight from matched unrelated donors. All had received peripheral blood stem cell grafts; one was CD34-selected. Eleven subjects (52%) received myeloablative conditioning; ten received reduced-intensity conditioning. Two subjects (9.5%) had prior grade I acute GvHD and four subjects (19.0%) had prior chronic GvHD after HSCT but did not have symptoms at the time of the DLI. The median time from HSCT to relapse was 6.5 months (range, 3 – 30 months), while the median time from relapse to DLI was 2.8 months (range, 1 – 15 months). Sixteen (76%) subjects received cytoreductive therapy after discontinuation of immune suppression at the discretion of the treating clinician, all more than 4 weeks prior to DLI. The median number of therapies after relapse in these 16 subjects was one (range, 1 - 2). Two subjects, one with Hodgkin lymphoma and one with acute myeloid
Table 1. Efficiency of processing and CD25/Treg depletion.
N. of CD3+ T Cells Initial Apheresis After Depletion (x107/kg) (x107/kg) Median Range
15.5 7.3-34.0
11.2 7.5-25.9
% Recovery 73.3% 46.1-107.7%
% CD4+CD25high Cells of Lymphocytes % Initial Apheresis % After Depletion Log-Depletion Median Range
3.54% 0.83-12.74%
0.022% 0.001-0.40%
2.46 0.83-3.37
% CD4+CD25+FOXP3+ Treg of Lymphocytes % Initial apheresis % After depletion Log-depletion Median Range
2.35% 0.60-8.14%
0.013% 0.001-0.32
2.28 0.73-3.33
N. of CD4+CD25+FoxP3+ Treg infused Dose level 1 Dose level 2 (x103/kg) (x103/kg) Median Range
0.94 0.18-8.87
4.24 0.64-95.66
Figure 1. Flow cytometric analysis of CD4+CD25+FoxP3+ cells confirms Treg depletion. The initial apheresis product (top plots) and target fraction after CD25 depletion using the CliniMACS system (bottom plots) were subjected to extracellular staining for CD4 and CD25 expression and intracellular staining for FoxP3 protein. As shown for a representative DLI product, gating on CD45+CD4+ cells demonstrated reduction in cells expressing high levels of CD25 (left plots). Gating on CD4+CD25high cells demonstrated reduction in cells expressing FoxP3 (right plots) after the depletion process.
haematologica | 2016; 101(10)
1253
S. Nikiforow et al.
Safety and toxicities
Efficacy
No significant infusion reactions occurred. Five subjects were enrolled (6 treated) at dose level 1 (1x107 CD3+ cells/kg recipient). Given that no significant adverse events occurred at this dose, escalation to dose level 2 (3x107 CD3+ cells/kg) was allowed for the subsequent 15 subjects. Three subjects with acute myleloid leukemia and two with acute lymphoblastic leukemia required cytotoxic therapy for disease progression within 8 weeks of DLI (n=2, dose level 1; n=3, dose level 2). The maximum-tolerated dose was not reached. One severe adverse event was observed in the first 8 weeks: namely, grade III acute GvHD (stage 2 gastrointestinal) 25 days after DLI at dose level 2, eventually leading to death (Table 3). There were two cases of grade I acute GvHD at dose level 2: one in the first 8 weeks, and one between 8 weeks and 1 year. Three subjects developed extensive chronic GvHD within 8 weeks following DLI: one at dose level 1 and two at dose level 2. After 8 weeks, three additional cases of chronic GvHD developed, yielding observed rates of significant GvHD of 19% (n=4) at 8 weeks and of 33% (n=7) at 1 year. All other adverse events observed, namely respiratory distress, transaminase elevations, cytopenias, or death were deemed related to GvHD, disease progression, subsequent cytoreductive therapy, or unrelated rather than secondary to CD25/Treg-depleted DLI.
21 13 (61.9%) 42 (19-71)
Twenty-one subjects were evaluated for response at 8 weeks. Only those with evidence of relapse/disease progression received additional therapy following DLI. At dose level 1, four (67%) of six subjects were alive 8 weeks after DLI, but all had persistent or progressive disease (Table 4). At 1 year, two (33%) of those six subjects remained alive. At dose level 2, 13 (87%) of 15 subjects were alive 8 weeks after DLI. One succumbed to acute GvHD and one to disease progression. Eight subjects at dose level 2 had a complete response (53%), for an overall response (complete responses + partial responses) rate of 60%. Responses were observed in five subjects with acute myeloid leukemia, one with acute lymphoblastic leukemia, one with acute lymphoblastic leukemia, two with Hodgkin lymphoma and one with nonHodgkin lymphoma (Online Supplementary Table S1). At 1 year, eight (53%) subjects receiving dose level 2 were alive, and four remained in complete remission. At dose level 2, the planned phase II dose, estimated 1-year overall and event-free survival rates were 53% (95% CI, 26-74) and 27% (95% CI, 8-50), respectively. The 1-year cumulative incidences of relapse and non-relapse mortality were 67% (95% CI, 35-86) and 7% (95% CI, 0.4-27), respectively (Figure 2 and Table 5). The median survival time was 8.7 months (range, 1.3 months - not reached) for all subjects and 12.7 months (range, 1.3 months - not reached) at dose level 2. Salvage therapies after CD25/Treg-depleted DLI included second CD25/Treg-depleted DLI outside of the current protocol (n=1), unmanipulated DLI (n=1), subsequent HSCT (n=2), and antibody-drug conjugate therapy. The only clinical characteristic associated with response was time from relapse to DLI, with a median time of 2.7 months (range, 1.4 - 4.7 months) in responders and 5.4 months (range, 1.9 - 14.5 months) in non-responders
11 (52.4%) 4 (19.0%) 4 (19.0%) 2 (9.5%)
Table 3. Toxicity after CD25/Treg-depleted DLI.
Table 2. Demographics of the study subjects.
Subjects Enrolled Receiving CD25/Treg-depleted DLI Male Median age, years (range) Malignancy Acute myeloid leukemia/myelodysplastic syndrome Acute lymphoblastic leukemia Hodgkin lymphoma Non-Hodgkin lymphoma Peripheral blood allogeneic HSCT Donor source HLA-matched sibling donor HLA-matched unrelated donor Conditioning intensity Myeloablative Reduced-intensity Prior acute GvHD Grade I Grades II-IV Prior chronic GvHD Limited Extensive Median time HSCT to relapse, months (range) Median time relapse to DLI, months (range) Median time HSCT to DLI, months (range) Received cytoreductive therapy prior to DLI Median n. of therapies (range) Prior unmanipulated DLI Disease status prior to study DLI Complete remission Partial response Untreated Progressive disease after cytoreduction Median total donor chimerism (range) 1254
n=24
13 (61.9%) 8 (38.1%) 11 (52.4%) 10 (47.5%) 2 (9.5%) 0 1 (4.8%) 3 (14.3%) 6.5 (2.7-30.2) 2.8 (0.6-14.7) 10.4 (5.1-38) 16 (76.2%) 1 (1-2) 2 (9.5%) 4 (19.0%) 7 (33.3%) 5 (23.8%) 5 (23.8%) 99% (32-100%)
Dose level 1 1x107 CD3+/kg Toxicity potentially related to DLI n=6 Acute GvHD Onset within 8 weeks
0
Onset at 8 weeks to 1 year Chronic GvHD Onset within 8 weeks Onset at 8 weeks to 1 year
0 1 - Extensive 1 - Limited
Significant GvHD at 8 weeks# Significant GvHD at 1 year
1 (16.7%) 2 (33.3%)
Dose level 2 3x107 CD3+/kg n=15 1 - Grade I 1 - Grade III->V* 1 - Grade I 2 - Extensive 1 - Limited 1 - Extensive 3 (20.0%) 5 (33.3%)
Toxicity related to GvHD, disease, or chemotherapy within 8 weeks Death Grade 4 respiratory distress Grade 3 transaminase elevations Grade 4 neutropenia Grade 3-4 thrombocytopenia
2 1 1 0 2
2 (1-GvHD DLT*) 0 2 1 3
1
0
Unrelated toxicity Grade 3 pain
*Denotes dose-limiting toxicity; # Significant GvHD connotes grades II-IV acute GvHD or chronic GvHD- N.B. One subject enrolled at dose level 2 actually received cells at dose level 1. Subjects were analyzed based on cell dose received.
haematologica | 2016; 101(10)
CD25/Treg-depleted DLI for relapse after SCT
(P=0.016). Development of GvHD after DLI was not a prerequisite for response; four of nine subjects with an initial response and two of four with sustained complete responses never developed GvHD. Entering DLI in remission was also not necessary: six subjects with disease at the time of DLI achieved a complete response at 8 weeks and three sustained those responses for at least 1 year. All four subjects remaining in complete response at 1 year had prior acute myeloid leukemia. Neither of the two subjects who had undergone prior unmanipulated DLI responded to CD25/Treg-depleted DLI. None of the other clinical factors, including time from HSCT to relapse or DLI, conditioning intensity of prior HSCT, donor chimerism, or receipt of disease-directed chemotherapy prior to DLI, correlated with response.
a comparator cohort. This cohort included two subjects initially enrolled on study who received unmanipulated DLI given the numbers of donor T cells collected. The age and gender distribution of the comparator cohort was similar to that of the study population, without significant differences in type of malignancy, donor source, prior GvHD incidence, transplant characteristics, time from HSCT to relapse or DLI, receipt of disease-directed therapy prior to DLI, or chimerism at the time of DLI (Online Supplementary Table S2). Five patients (36%) were in complete remission, two in
Table 4. Response to CD25/Treg-depleted DLI.
Disease response
Comparison to unmanipulated donor lymphocyte infusion In an exploratory analysis, we collected baseline data on expected toxicity (i.e., subsequent GvHD and non-relapse mortality) and efficacy of unmodified DLI at our center in a contemporaneous cohort of patients. Fifty patients whose donor chimerism was â&#x2030;Ľ20% received unmanipulated DLI for relapse between November 2006 and May 2011 after undergoing an 8/8 HLA-matched HSCT. Of these 50, 36 received 1x107 CD3+ cells/kg and 14 received 2-4x107 CD3+ cells/kg. To provide an exploratory baseline for subjects receiving dose level 2 of CD25/Treg-depleted DLI, we included the 14 patients receiving 2-4x107 CD3+ cells/kg as
Status at 8 weeks Complete response Partial response Stable disease Progressive disease Alive Dead Status at 1 year Complete response Alive Dead
A
B
Dose level 1 1x107 CD3+/kg n=6
Dose level 2 3x107 CD3+/kg n=15
0 0 1 (16.7%) 5 (83.3%) 4 (66.7%) 2 (33.3%)
8 (53.3%) 1 (6.7%) 0 6 (40.0%) 13 (86.7%) 2 (13.3%)
0 2 (33.3%) 4 (66.6%)
4 (26.7%) 8 (53.3%) 7(47.7%)
C
Figure 2. Clinical outcomes after CD25/Treg-depleted DLI at dose level 2. (A) Kaplan Meier curves demonstrating overall survival after infusion for recipients of CD25/Treg-depleted DLI at 3x107 CD3+/kg (solid line) versus unmanipulated DLI at cell doses of 2.2-3.7x107 CD3+/kg (dotted line). (B) Kaplan Meier curves demonstrating improved event-free survival after CD25/Treg-depleted DLI (solid line) versus unmanipulated DLI (dotted line). (C) Relapse and non-relapse mortality: plot displaying cumulative incidences of disease relapse and non-relapse mortality after CD25/Treg-depleted DLI (solid and fine dotted lines, respectively) versus unmanipulated DLI (hashed and coarse dotted lines, respectively).
haematologica | 2016; 101(10)
1255
S. Nikiforow et al.
partial remission, and seven (50%) had untreated/progressive disease at the time of the unmanipulated DLI. The median DLI cell dose received was 3x107 CD3+ cells/kg (range, 2.2 x 107 - 3.7 x107 cells/kg). Acknowledging caveats of comparing small, heterogeneous cohorts, specifically limited statistical power and selection bias, the response rate after unmanipulated DLI was 14% at 8-12 weeks, lower than the 60% response rate seen after CD25/Treg-depleted DLI at dose level 2 (P=0.02).
There were two complete responses, both in recipients with acute myeloid leukemia, which lasted 5.1 and 6.1 months (Online Supplementary Table S3). The 1-year eventfree survival rate of this cohort was 0% as compared with 27% in the CD25/Treg-depleted DLI cohort (P=0.0075), the 1-year cumulative incidence rate of relapse was 100% as compared to 67% (P=0.002), and the 1-year overall survival rate was 36% as compared to 53% in the CD25/Tregdepleted DLI cohort (P=0.08). Cumulative incidences of
Table 5. Outcomes of Treg-depleted DLI compared to unmanipulated DLI.
Responses at 8-12 weeks Complete + partial responses Persistent/progressive disease Responses in patients with AML at DLI Complete + partial responses Persistent/progressive disease Cumulative incidence (95% CI) 1-year overall survival 1-year event-free survival 1-year relapse 1-year non-relapse mortality 6-month grade II-IV acute GvHD 6-month chronic GvHD
CD25/Treg-depleted DLI 3x107 CD3+/kg n=15
Unmanipulated DLI 2-4x107 CD3+/kg n=14
P value
9 (60%) 6 (40%) n=8 5 (62.5%) 3 (37.5%)
2 (14.3%) 12 (85.7%) n=7 2 (28.6%) 5 (71.4%)
0.02
53% (26-74) 27% (8-50) 67% (35-86) 6.7% (0.4-27) 6.7% (3.8-27) 27% (7.6-51)
36% (13-59) 0% 100% 0% 0% 0%
0.08 0.0075 0.002 NS NS 0.04
NS
AML: acute myeloid leukemia; NS: not significant. Statistically significant values are shown in bold.
A
B
1256
Figure 3. NaĂŻve and central memory cells increase within CD4+ and CD8+ T-cell compartments after CD25/Treg-depleted DLI. Peripheral blood mononuclear cells drawn at indicated times were stained for extracellular markers and analyzed by flow cytometry. (A) Changes in median distribution of CD4+ T cells among all subjects: the percentage of CD3+CD4+ T cells in each subset just prior to (0 months) or in the first 2 months after DLI is indicated. Panels reflect percentage of naĂŻve cells (CD45RO-CD62L+), central memory cells effector (CD45RO+CD62L+), + memory cells (CD45RO CD62L-), and terminal effectors (CD45ROCD62L-). *Indicates P<0.05 for the comparison between baseline (month 0) and the indicated time point. (B) Changes in median distribution of CD8+ T cells. Individual CD8+ subsets and significant differences among all subjects after CD25/Tregdepleted DLI are displayed as detailed above.
haematologica | 2016; 101(10)
CD25/Treg-depleted DLI for relapse after SCT
non-relapse mortality at 1 year and acute GvHD at 6 months were not different. Given the early deaths from disease seen after unmanipulated DLI, the difference in incidence of chronic GvHD noted was not deemed clinically relevant (Table 5 and Figure 2).
Immunological effects of CD25/Treg-depleted donor lymphocyte infusions In the first 2 months after CD25/Treg-depleted DLI, there were no significant changes from prior to DLI in white blood cell count, absolute lymphocyte count, percentages and numbers of circulating peripheral CD3+, CD4+, and CD8+ T cells, Treg and CD20+ B cells, or in Treg to conventional T-cell (Tcon) ratios when analyzing all subjects (Online Supplementary Figure S1 and data not shown). Median CD8:CD4 ratios increased at 1 month (median, 0.74 to 1.06; P=0.013), but this difference was not maintained at 2 months. There were significant drops from baseline in circulating natural killer cells (median, 201.5 to 111.6 to 86.0/μL; P=0.027, 1 month; P=0.002, 2 months) and dendritic cells (median, 301.1 to 99.7 to 17.4/ μL; P=0.029, 1 month; P=0.0002, 2 months), which were no different between those who responded to DLI and those who did not. Thus, CD25/Treg-depleted DLI did not notably facilitate numerical expansion of early total lymphocytes, Tcon, B cells, or natural killer cells. There were qualitative shifts within the naïve and memory compartments of circulating CD4+ T cells after CD25/Treg-depleted DLI (Figure 3A). The percentage of CD4+ naïve T cells increased by 1 month (median, 11.0% to 24.1%; P=0.008) and remained above baseline at 2 months (median, 21.7%; P=0.0005). The percentage of CD4+ terminal effectors decreased by 1 month (median, 12.8% to 8.6%; P=0.053) and remained below baseline at 2 months (median, 2.9%; P=0.003). Increases in CD4+ central memory cells at 2 months (P=0.004) and decreases in CD4+ effector memory cells at 1 and 2 months after DLI (P=0.04 and P=0.0007, respectively) were statisticially significant. Identical shifts were seen in CD8+ T cells: namely, increases in naïve cells (P=0.0006 and P=0.0004), decreases in terminal effectors (P=0.04 and P=0.03), increases in central memory cells (P=0.004 and P=0.004), and decreases in effector memory cells (P=0.01 and P=0.004) at 1 and 2 months after DLI, respectively (Figure 3B). In these analyses after DLI, no differences between responder and non-responder populations were noted, perhaps because of the limited sample size. Similar analyses were not available for bone marrow cell subsets or for patients receiving unmanipulated DLI. When examining immunological profiles prior to DLI, there were no correlations between response and circulating lymphocyte subset values including CD4+ or CD8+ Tcell counts, CD4+ Tcon or Treg counts, or CD8:CD4 ratios (data not shown). The ratio of peripheral blood Treg to Tcon prior to DLI, the Treg:Tcon ratio in the DLI product itself, and the number of residual Treg infused had no correlation with disease response (P=0.76, P=0.41, and P=0.24, respectively); this was the case regardless of whether all subjects or only those treated with dose level 2 were analyzed.
Discussion Given the historically poor efficacy of unmanipulated DLI in treating relapse of hematologic malignancies other than chronic myeloid leukemia after HSCT, we investigated haematologica | 2016; 101(10)
whether ex vivo depletion of CD25+ cells including suppressive Treg from DLI products was feasible, safe, and could enhance anti-tumor activity without increasing GvHD. Twenty-one of 23 (91.3%) donors yielded sufficient CD3+ cells for selection and DLI, 20 (87.0%) in one leukapheresis. Given that only two donors did not yield sufficient T cells for CD25 depletion, it is unlikely that the results discussed below are a result of selection bias for “optimal” donors from among the donor pool. The median CD3+ T-cell recovery was 73% after depletion. Thus, it may be possible to target higher doses in future studies, at least for related donors for whom multiple leukaphereses are more practical. This approach proved feasible with a median 2.3 logdepletion of FoxP3+ Treg, which is comparable to efficiencies obtained in prior studies.33 We refer to CD25-depleted products as “Treg-depleted”. CD25 (interleukin-2 receptor-α) is also expressed on activated CD4+ Tcon and CD8+ T cells, so the selection process could affect these populations as well. Depletion of CD25+ donor Tcon cells already activated in vivo at the time of leukapheresis might potentially decrease non-specific alloreactivity but would not be expected to improve graftversus-tumor effects. We, therefore, postulate that our observed results, particularly as regards efficacy, are primarily a result of Treg depletion, a concept that has support from animal models.23 Extensive phenotyping of the DLI product with characterization of CD25+ cells depleted/retained during processing was not pursued in this phase I study. This limitation could, however, be addressed in future trials which should include characterization of Treg phenotype (e.g., HLA-DR, CTLA-4, Lag-3, CD45RA, Ki-67 expression), suppressive activity, and markers of thymic activity. Regarding safety, there were no infusion reactions. Subsequent cytopenias and adverse events were no more frequent or severe than expected after unmanipulated DLI. Apart from one case of acute GvHD progressing to death, CD25/Treg-depleted DLI was well-tolerated, and overall GvHD rates were not excessive. Comparison to a limited contemporaneous cohort receiving unmanipulated DLI at a study-equivalent dose at our center showed no significant difference in the incidence of acute GvHD, although small sample sizes limit the statistical robustness of this comparison. However, our 33% observed rate of GvHD requiring systemic therapy after 3x107 CD3+/kg CD25-Treg-depleted DLI compares favorably with other studies of unmanipulated DLI at similar doses, e.g. 45% in a large retrospective analysis by Bar et al., as does our 1-year non-relapse mortality rate of 7%.10 Importantly, prior GvHD does not appear a contraindication to receiving CD25/Treg-depleted DLI, although active GvHD at the time of enrollment was an exclusion criterion. Regarding efficacy, over 70% of study subjects had acute leukemia or myelodysplastic syndrome, diseases with historically poor response rates and remission durability. The 8-week response rate of 60% at dose level 2 (3x107 CD3+ cells/kg) and the 27% 1-year event-free survival rate, while still leaving much room for improvement, were higher than seen in a contemporaneous group of patients receiving unmanipulated DLI in our exploratory analysis. However, the implications of these comparisons are limited by small sample size, clinical heterogeneity and potential selection bias, and should be confirmed in subsequent prospective trials. The 1-year survival of 53% in the CD25/Treg-depleted study group was in the range 1257
S. Nikiforow et al.
of 17-67% survival seen at similar cell doses in other studies, although historical comparisons are also limited by variations in disease distribution, disease status, prior cytoreduction, chimerism, mobilization of the DLI product, and year of DLI.9,10,37 Six of the nine responses were in study subjects with acute leukemia, and four of those six responses were sustained for a year. The observed response rate of 63% and 1year overall survival of 50% in patients with acute myeloid leukemia/myelodysplastic syndrome at dose level 2 was surprising, particularly as all but one patient had active disease at the time of DLI. After unmanipulated DLI, albeit in a small group, of seven patients with acute myeloid leukemia only two had complete responses, both of whom relapsed within 6 months or less. Findings after CD25/Tregdepleted DLI compare favorably with 2-year survival rates of 56% for DLI in complete response and 15% for DLI with active disease in an European Group for Blood and Marrow Transplantation study devoted to relapse of acute myeloid leukemia after HSCT.7 Responses in our study were not obviously dependent on occurrence of GvHD after DLI, indicating the potential for immunologically separating graft-versus-tumor effects from GvHD, even in subjects with relapsed acute myeloid leukemia, through manipulation of the composition of the DLI. The numbers and profiles of circulating lymphocyte numbers either before or after CD25/Treg-depleted DLI did not correlate with disease response in this study population. However, significant increases in CD4+ and CD8+ na誰ve and central memory cells and decreases in CD4+ and CD8+ effector memory and terminal effector cells were seen. Recovery of na誰ve and central memory cells typically indicates improved thymopoiesis and T-cell diversity and has been associated with decreased relapse and increased survival after HSCT.38,39 Whether this normalization of na誰ve, memory, and effector-cell distribution is augmented by CD25/Treg depletion could not be determined in this single-arm study. We did not see a correlation between responses and pre-existing CD8+ T-cell counts as previously noted in bone marrow after CD4+ DLI in myeloma and chronic myeloid leukemia, in which donor lymphoid expansion and reversal of residual exhausted T cells were demonstrated, respectively.40,41 However, the current analysis was restricted to early time points only after CD25/Treg-depleted DLI and to peripheral blood samples, not bone marrow. Identifying a cellular mechanism specific to responses after CD25/Tregdepleted or other DLI products will require prospective monitoring of bone marrow and sites of disease rather than just peripheral blood. Characterization of markers of activation and exhaustion such as PD-1 as well as detection of T-cell frequencies against various tumor antigens could be considered. The impacts of this study are two-fold. First, we confirm that CD25/Treg-depleted DLI is feasible and safe and may be modestly more effective than unmanipulated DLI, particularly in acute myeloid leukemia. Larger numbers of
References 1. Horowitz M, Gale R, Sondel P, et al. Graftversus-leukemia reactions after bone marrow
1258
patients will be needed to confirm a disease-specific effect.42 Previously, Maury et al. reported two responses to one infusion of Treg-depleted DLI and four more responses after a second infusion in 17 subjects with relapse after HSCT. In that study, 16 subjects had previously failed standard DLI, and responses were only seen following development of GvHD. Response rates were increased after chemotherapyinduced lymphodepletion was introduced.43 In our study, we targeted a different population, namely subjects early after relapse (19 of 21 without prior DLI), without pre-DLI lymphodepletion, and observed a 60% response rate and responses that lasted at least 1 year in 50% of initial complete responders (n=4). Whether the durability of response can be prolonged by multiple infusions is unknown. The likelihood of clinical response was associated with shorter time from relapse to DLI. Our responses were not stringently linked with GvHD and broaden the application of CD25/Treg-depleted DLI to use earlier in relapse after HSCT. Secondly, our data support the concept that manipulating the composition of stem cell and other cellular therapies has the potential to improve efficacy and reduce toxicity.44,45 In this vein, trials of ex vivo depletion of na誰ve T cells from stem cell grafts to reduce GvHD and preserve anti-tumor and anti-infectious activity are underway.46,47 Analogous to the introduction of immune check-point inhibitors to boost activity of otherwise ineffective lymphocytes, modulation of the stimulatory versus suppressive T-cell balance within cellular therapy products and their recipients holds promise.48,49 Significant caveats in interpreting suggestions of efficacy are the small sample size and heterogeneity in this phase I study. Full details for each patient are laid out in the Online Supplementary Tables, but retrospective comparison to a small, diverse comparator group in our secondary analysis cannot take the place of a randomized study in which disease burden, molecular risk subtype, prior therapies, and time from chemotherapy to DLI are rigorously controlled between recipients of unmanipulated and CD25/Treg-depleted DLI. The comparative benefit of CD25-Treg depletion in specific disease groups such as acute myeloid leukemia and the role of lymphodepleting chemotherapy just prior to DLI administration will be addressed in future trials.50 In conclusion, depletion of CD25+ Treg from DLI is feasible, well-tolerated, and may increase efficacy when compared to unmanipulated DLI. Future avenues of research include increasing cell doses or giving multiple infusions of CD25/Treg-depleted donor lymphocytes, prospective randomized evaluation in relapsed acute leukemia versus unmanipulated DLI, and combining CD25/Treg-depleted DLI with novel immunomodulatory agents. Characterization of lymphocytes within active DLI products and at sites of disease response may lead to cellular manipulations capable of dissociating graft-versustumor from GvHD activity in the currently bleak situation of relapse after HSCT.
transplantation. Blood. 1990;75(3):555-562. 2. Kolb H, Schattenberg A, Goldman J, et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. Blood. 1995;86(5):2041-2050.
3. Porter DL, Collins Jr RH, Shpilberg O, et al. Long-term follow-up of patients who achieved complete remission after donor leukocyte infusions. Biol Blood Marrow Transplant. 1999;5(4):253-261.
haematologica | 2016; 101(10)
CD25/Treg-depleted DLI for relapse after SCT
4. Collins RH, Shpilberg O, Drobyski WR, et al. Donor leukocyte infusions in 140 patients with relapsed malignancy after allogeneic bone marrow transplantation. J Clin Oncol. 1997;15(2):433-444. 5. Beitinjaneh AM, Saliba R, Bashir Q, et al. Durable responses after donor lymphocyte infusion for patients with residual multiple myeloma following non-myeloablative allogeneic stem cell transplant. Leuk Lymphoma. 2012;53(8):1525-1529. 6. Campregher PV, Gooley T, Scott BL, et al. Results of donor lymphocyte infusions for relapsed myelodysplastic syndrome after hematopoietic cell transplantation. Bone Marrow Transplant. 2007;40(10):965-971. 7. Schmid C, Labopin M, Nagler A, et al. Donor lymphocyte infusion in the treatment of first hematological relapse after allogeneic stem-cell transplantation in adults with acute myeloid leukemia: a retrospective risk factors analysis and comparison with other strategies by the EBMT Acute Leukemia Working Party. J Clin Oncol. 2007;25(31): 4938-4945. 8. El-Jurdi N, Reljic T, Kumar A, et al. Efficacy of adoptive immunotherapy with donor lymphocyte infusion in relapsed lymphoid malignancies. Immunotherapy. 2013;5(5):457-466. 9. Tomblyn M, Lazarus HM. Donor lymphocyte infusions: the long and winding road: how should it be traveled? Bone Marrow Transplant. 2008;42(9):569-579. 10. Bar M, Sandmaier BM, Inamoto Y, et al. Donor lymphocyte infusion for relapsed hmatological malignancies after allogeneic hematopoietic cell transplantation: prognostic relevance of the initial CD3+ T cell dose. Biol Blood Marrow Transplant. 2013;19(6): 949-957. 11. Dazzi F, Szydlo RM, Cross NC, et al. Durability of responses following donor lymphocyte infusions for patients who relapse after allogeneic stem cell transplantation for chronic myeloid leukemia. Blood. 2000;96(8):2712-2716. 12. Yun HD, Waller EK. Finding the sweet spot for donor lymphocyte infusions. Biol Blood Marrow Transplant. 2013;19(4):507-508. 13. Jiang Y-Z, Barrett J. The Allogeneic CD4+ Tcell-mediated graft-versus-leukemia effect. Leuk Lymphoma. 1997;28(1-2):33-42. 14. Slavin S, Naparstek E, Nagler A, et al. Allogeneic cell therapy with donor peripheral blood cells and recombinant human interleukin-2 to treat leukemia relapse after allogeneic bone marrow transplantation. Blood. 1996;87(6):2195-2204. 15. Laport GG, Sheehan K, Baker J, et al. Adoptive immunotherapy with cytokineinduced killer cells for patients with relapsed hematologic malignancies after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2011;17(11):1679-1687. 16. Kumar AJ, Hexner EO, Frey NV, et al. Pilot study of prophylactic ex vivo costimulated donor leukocyte infusion after reduced-intensity conditioned allogeneic stem cell transplantation. Biol Blood Marrow Transplant. 2013;19(7):1094-1101. 17. Fehérvari Z, Sakaguchi S. Development and function of CD25+CD4+ regulatory T cells. Curr Opin Immunol. 2004;16(2):203-208. 18. Piccirillo CA, Shevach EM. Naturally-occurring CD4+CD25+ immunoregulatory T cells: central players in the arena of peripheral tolerance. Semin Immunol. 2004;16(2):81-88. 19. Thornton AM, Shevach EM. CD4+CD25+ Immunoregulatory T cells suppress polyclon-
haematologica | 2016; 101(10)
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30.
31.
32.
33.
34.
35.
al T cell activation in vitro by inhibiting interleukin 2 production. J Exp Med. 1998;188(2): 287-296. Turk MJ, Guevara-Patiño JA, Rizzuto GA, Engelhorn ME, Houghton AN. Concomitant tumor immunity to a poorly immunogenic melanoma is prevented by regulatory T cells. J Exp Med. 2004;200(6):771-782. Sugiyama D, Nishikawa H, Maeda Y, et al. Anti-CCR4 mAb selectively depletes effectortype FoxP3+CD4+ regulatory T cells, evoking antitumor immune responses in humans. Proc Natl Acad Sci USA. 2013;110(44):1794517950. Pastille E, Bardini K, Fleissner D, et al. Transient ablation of regulatory T cells improves antitumor immunity in colitis-associated colon cancer. Cancer Res. 2014; 74(16):4258-4269. Zhou Q, Bucher C, Munger ME, et al. Depletion of endogenous tumor-associated regulatory T cells improves the efficacy of adoptive cytotoxic T-cell immunotherapy in murine acute myeloid leukemia. Blood. 2009;114(18):3793-3802. Zorn E, Kim HT, Lee SJ, et al. Reduced frequency of FOXP3+ CD4+CD25+ regulatory T cells in patients with chronic graft-versushost disease. Blood. 2005;106(8):2903-2911. Rieger K, Loddenkemper C, Maul J, et al. Mucosal FOXP3+ regulatory T cells are numerically deficient in acute and chronic GvHD. Blood. 2006;107(4):1717-1723. Rezvani K, Mielke S, Ahmadzadeh M, et al. High donor FOXP3-positive regulatory T-cell (Treg) content is associated with a low risk of GVHD following HLA-matched allogeneic SCT. Blood. 2006;108(4):1291-1297. Blazar BR, Murphy WJ, Abedi M. Advances in graft-versus-host disease biology and therapy. Nat Rev Immunol. 2012;12(6):443-458. Fujioka T, Tamaki H, Ikegame K, et al. Frequency of CD4+FOXP3+ regulatory Tcells at early stages after HLA-mismatched allogeneic hematopoietic SCT predicts the incidence of acute GVHD. Bone Marrow Transplant. 2013;48(6):859-864. Nguyen VH, Shashidhar S, Chang DS, et al. The impact of regulatory T cells on T-cell immunity following hematopoietic cell transplantation. Blood. 2008;111(2):945-953. Kennedy-Nasser AA, Ku S, Castillo-Caro P, et al. Ultra low-dose IL-2 for GVHD prophylaxis after allogeneic hematopoietic stem cell transplantation mediates expansion of regulatory T cells without diminishing antiviral and antileukemic activity. Clin Cancer Res. 2014;20(8):2215-2225. Koreth J, Matsuoka K-i, Kim HT, et al. Interleukin-2 and regulatory T cells in graftversus-host disease. N Engl J Med. 2011; 365(22):2055-2066. Ganguly S, Ross DB, Panoskaltsis-Mortari A, et al. Donor CD4+ Foxp3+ regulatory T cells are necessary for posttransplantation cyclophosphamide-mediated protection against GVHD in mice. Blood. 2014;124 (13):2131-2141. Powell DJ, Parker LL, Rosenberg SA. Largescale depletion of CD25+ regulatory T cells from patient leukapheresis samples. J Immunother. 2005;28(4):403-411. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HLAmatched sibling donors. Transplantation. 1974;18(4):295-304. Shulman H, Sullivan KM, Weiden P, et al. Chronic graft-versus-host syndrome in man. A long-term clinicopathologic study of 20
36. 37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
Seattle patients. Am J Med. 1980;69(2):204217. Gray R. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16(3):1141-1154. Chiorean EG, DeFor TE, Weisdorf DJ, et al. Donor chimerism does not predict response to donor lymphocyte infusion for relapsed chronic myelogenous leukemia after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2004;10(3): 171-177. Komanduri KV, St. John LS, de Lima M, et al. Delayed immune reconstitution after cord blood transplantation is characterized by impaired thymopoiesis and late memory Tcell skewing. Blood. 2007;110(13):4543-4551. Jeljeli M, Guérin-El Khourouj V, Porcher R, et al. Relationship between cytomegalovirus (CMV) reactivation, CMV-driven immunity, overall immune recovery and graft-versusleukaemia effect in children. Br J Haematol. 2014;166(2):229-239. Bellucci R, Alyea EP, Weller E, et al. Immunologic effects of prophylactic donor lymphocyte infusion after allogeneic marrow transplantation for multiple myeloma. Blood. 2002;99(12):4610-4617. Bachireddy P, Hainz U, Rooney M, et al. Reversal of in situ T-cell exhaustion during effective human antileukemia responses to donor lymphocyte infusion. Blood. 2014;123 (9):1412-1421. Ustun C, Miller JS, Munn DH, Weisdorf DJ, Blazar BR. Regulatory T cells in acute myelogenous leukemia: is it time for immunomodulation? Blood. 2011;118(19): 5084-5095. Maury S, Lemoine FM, Hicheri Y, et al. CD4+CD25+ Regulatory T cell depletion improves the graft-versus-tumor effect of donor lymphocytes after allogeneic hematopoietic stem cell transplantation. Sci Transl Med. 2010;2(41):41ra52. Waller EK, Logan BR, Harris WAC, et al. Improved survival after transplantation of more donor plasmacytoid dendritic or naïve T cells from unrelated-donor marrow grafts: results from BMTCTN 0201. J Clin Oncol. 2014;32(22):2365-2372. Reshef R, Huffman AP, Gao A, et al. A survival benefit for reduced intensity allogeneic transplants from young unrelated donors compared to older sibling donors depends on the graft CD8 T-cell Content. Biol Blood Marrow Transplant. 2015;21(2):S42-S43. Zheng H, Matte-Martone C, Li H, et al. Effector memory CD4+ T cells mediate graftversus-leukemia without inducing graft-versus-host disease. Blood. 2008;111 (4):24762484. Bleakley M, Heimfeld S, Loeb KR, et al. Outcomes of acute leukemia patients transplanted with naive T cell–depleted stem cell grafts. J Clin Invest. 2015;125(7):2677-2689. Frigault MJ, Lee J, Basil M, et al. Identification of chimeric antigen receptors that mediate constitutive or inducible proliferation of T cells. Cancer Immunol Res. 2015;3(4):356367. Ansell SM, Lesokhin AM, Borrello I, et al. PD1 blockade with nivolumab in relapsed or refractory Hodgkin's lymphoma. N Engl J Med. 2015;372(4):311-319. Schroeder T, Frobel J, Cadeddu RP, et al. Salvage therapy with azacitidine increases regulatory T cells in peripheral blood of patients with AML or MDS and early relapse after allogeneic blood stem cell transplantation. Leukemia. 2013;27(9):1910-1913.
1259
ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION
Stem Cell Transplantation
Ferrata Storti Foundation
Donor and recipient sex in allogeneic stem cell transplantation: what really matters Haesook T. Kim,1 Mei-Jie Zhang,2 Ann E. Woolfrey,3 Andrew St. Martin,4 Junfang Chen,4 Wael Saber,4 Miguel-Angel Perales,5 Philippe Armand6 and Mary Eapen4
Biostatistics and Computational Biology, Dana-Farber Cancer Institute, Boston, MA; Division of Biostatistics, Medical College of Wisconsin, Milwaukee, WI; 3Department of Medicine, Fred Hutchinson Cancer Research Center, Seattle, WA; 4Center for International Blood and Marrow Transplant Research, Department of Medicine, Medical College of Wisconsin, Milwaukee, WI; 5Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, NY; and 6Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA
1
2
Haematologica 2016 Volume 101(10):1260-1266
ABSTRACT
W
Correspondence: htkimc@jimmy.harvard.edu
Received: April 9, 2016. Accepted: June 23, 2016. Pre-published: June 27, 2016. doi:10.3324/haematol.2016.147645
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1260
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1260
e investigated whether and how recipient-donor sex affects transplantation outcomes of 11,797 patients transplanted between 2008 and 2010. Thirty-seven percent were male recipients with male donors, 21% male recipients with female donors, 25% female recipients with male donors, and 17% female recipients with female donors. In multivariable analyses, male recipients had inferior overall survival and progression-free survival compared to females regardless of donor sex, with an 11% relative increase in the hazard of death (P<0.0001) and a 10% relative increase in the hazard of death or relapse (P<0.0001). The detrimental effect of male recipients varied by donor sex. For male recipients with male donors, there was a 12% relative increase in the subdistribution hazard of relapse compared with female recipients with male donors (P=0.0036) and male recipients with female donors (P=0.0037). For male recipients with female donors, there was a 19% relative increase in the subdistribution hazard of non-relapse mortality compared with male recipients with male donors (P<0.0001) and a 22% relative increase compared with female recipients with male donors (P=0.0003). In addition, male recipients with female donors showed a 21% relative increase in the subdistribution hazard of chronic graft-versus-host disease (P<0.0001) compared with female recipients with male donors. Donor sex had no effect on outcomes for female recipients. Transplantation of grafts from male and female donors was associated with inferior overall survival and progression-free survival in male recipients with differing patterns of failure. Recipient sex is an important prognostic factor independent of donor sex.
Introduction In patients undergoing allogeneic hematopoietic cell transplantation (HSCT), it has been previously reported that sex mismatch between donor and recipient affects HSCT outcome across diseases.1-11 Most studies have reported that the combination of male recipient and female donor (F→M) is associated with a higher incidence of graft-versus-host disease (GvHD)2,5,7,9 and non-relapse mortality (NRM),1,3 as well as, in some studies, a lower relapse rate.2,3 The increased GvHD rate in this setting is thought to be mediated via male recipient minor histocompatibility antigens (mHAs) targeted by female donor T cells,10,11 and likely explains the increased NRM. At the same time, the theoretical increase in graft-versus-tumor (GvT) mediated by the same mHAs may explain the decreased risk of relapse. On balance, however, the increased toxicity of F→M transplants resulted in a decreased survival, suggesting that this combination was deleterious in patients undergoing haematologica | 2016; 101(10)
Donor/recipient sex combination in HSCT
HSCT.1-3 F→M is, therefore, included as a risk factor in the (modified) EBMT risk score,12-16 and many transplant clinicians will use sex matching as a criterion in donor selection for patients undergoing HSCT. However, much less attention has been paid to recipient sex, although there have been a few reports that male recipients had a poor survival irrespective of donor sex. 2,3 Moreover, the exact sex-based determinants of HSCT outcome have not been rigorously examined in a modern transplantation cohort that is carefully stratified by disease risk. Recently, a study of disease risk conducted at a single institution found that the only significant risk factor for mortality related to sex was recipient sex, with a hazard ratio for mortality of approximately 0.9 for female compared to male recipients.17 Given this, we undertook an analysis of donor/recipient sex in a large cohort of patients transplanted between 2008 and 2010 in the United States and reported to the Center for International Blood and Marrow Transplant Research (CIBMTR). The primary goal of this study was to examine the effect of recipient sex and donor-recipient sex combinations on overall and progression-free survival (OS and PFS) after HSCT. In addition, we sought to determine whether the effect of recipient and/or donor sex on OS and PFS were mediated mainly by acute or chronic GvHD, NRM, or relapse.
ing risks framework considering relapse, NRM and death or relapse without developing chronic GvHD, respectively, as competing events. All time to events were measured from the date of stem cell infusion. The difference between cumulative incidence curves in the presence of a competing risk was tested using the Gray method.19 Multivariable regression analysis was performed using the Cox model for OS, PFS and Fine and Gray model for relapse, NRM, and chronic GvHD.20,21 Models were stratified by conditioning intensity as this variable did not meet the proportional hazards assumption. Potential prognostic factors considered in the analyses included recipient and donor sex, disease risk index (DRI),17,18 age, conditioning intensity, cytomegalovirus (CMV) serostatus of recipient and donor, graft source, donor HLA type,22 co-morbidity index (HCT-CI),23 and Karnofsky performance status at HSCT. Prior to modeling, the proportional hazards assumption and significance of interaction terms were examined. Acute GvHD was analyzed as a binary outcome using a landmark analysis at day 100 of HSCT, and multivariable analysis for acute GvHD was performed using logistic regression analysis. The threshold for statistical significance was set at 0.01 to account for multiple testing. All tests were two-sided and all analyses were performed using SAS 9.3 (SAS Institute Inc., Cary, NC, USA), and R v.3.2.2 (the CRAN project; www.cran.r-project.org).
Results Patients’ characteristics
Methods Study population The Center for International Blood and Marrow Transplant Research (CIBMTR) comprises a voluntary network of more than 450 transplantation centers worldwide that contribute detailed data on consecutive allogeneic and autologous HSCT to a centralized Statistics Center.18 Observational studies conducted by the CIBMTR are performed in compliance with all applicable federal regulations pertaining to the protection of human research participants. Protected Health Information used in the performance of such research is collected and maintained in CIBMTR’s capacity as a Public Health Authority under the HIPAA Privacy Rule. The Institutional Review Board of the National Marrow Donor Program approved this study. The study cohort consisted of patients aged 18 years or over who underwent HSCT between 2008 and 2010, excluding autologous, syngeneic, and cord transplantations. Among the 14,126 potential patients, we further excluded 2329 patients (16%) with missing disease type or pretransplant disease status, and transplantations for benign or rare disorders (including histiocytic disorders, large granular lymphocyte or natural killer cell leukemia). The remaining 11,797 patients are included in the current analysis.
Statistical analysis Patients’ baseline characteristics were reported descriptively. End points of interest were OS, PFS, relapse, NRM, as well as acute and chronic GvHD. OS was defined as the time from stem cell infusion to death from any cause. Patients who were alive were censored at the time last seen alive. PFS was defined as the time from stem cell infusion to disease relapse, progression or death from any cause, whichever occurred first. Patients who were alive without disease relapse or progression were censored at the time last seen alive and progression-free. OS and PFS were estimated using the Kaplan-Meier method and the log rank test stratified by conditioning intensity was used for comparisons of Kaplan-Meier curves. Cumulative incidence curves for non-relapse death, relapse and chronic GvHD were constructed in the compethaematologica | 2016; 101(10)
The baseline characteristics of the 11,797 patients are shown in Table 1. The median age of the entire cohort was 52 years (range 18-80). The cohort included a broad representation of diseases, disease risk, donor types, and graft sources. Forty-one percent of patients received grafts from matched sibling donor; 54% were conditioned with a myeloablative regimen. Among the 11,797 patient/donor pairs, 37% were male recipients with male donors (M→M), 21% male recipients with female donors (F→M), 25% female recipients with male donors (M→F), and 17% female recipients with female donors (F→F). Overall, 54% of pairs were sex matched: 42% among female recipients and 64% among male recipients (P<0.0001). The median follow up among survivors was 48 months (range 2-76).
Overall and progression-free survival Overall, male recipients had worse OS and PFS than females, regardless of donor sex (Figures 1A and B and Online Supplementary Table S1). The 4-year OS was 41% in male recipients (41% for M→M and 40% for F→M) and 45% in female patients (45% for M→F and 44% for F→F) (P=0.001); the corresponding 4-year PFS was 33% in male recipients (33% for M→M and 32% for F→M) and 36% in female patients (37% for M→F and 35% for F→F) (P=0.0005). The result was consistent in multivariable analysis; hazard ratio (HR) of male compared to female recipients was 1.11 (95%CI: 1.05-1.16) for OS (P<0.0001) and 1.10 (95%CI: 1.05-1.15) for PFS (P<0.0001) (Table 2A). When all possible recipient and donor sex combinations were considered in the multivariable model, the F→M group showed an inferior OS (HR 1.14, P=0.0004) and PFS (HR 1.1 P=0.0044) compared with the M→F group (reference group). Worse survival outcome was also seen in the M→M group (HR 1.1, P=0.0032 for OS; HR 1.11, P=0.0004 for PFS), but not in the F→F group (HR 1.02 for both OS and PFS; P=0.64 and P=0.58, respectively) (Table 2B). When the F→M group was compared with the 1261
H.T. Kim et al. Table 1. Baseline characteristics.
Female recipients F–›F M–›F N (%) N (%) Number of patients Age, years (median, range) Age < 40 Age 40-49 Age 50-64 Age ≥ 65 Disease ALL AML CLL CML Hodgkin lymphoma MDS Myeloproliferative neoplasms Non-Hodgkin lymphoma Multiple myeloma Disease risk indexa Low Intermediate High Very high HCT-CIb 0 1-2 3+ Missing Karnofsky performance score < 90 90-100 Missing Donor matchc MRD Non-MRD 8/8 HLA-match URD 7/8 HLA-match URD Haploidentical relative 7/8 HLA-match relative 6/8 HLA-match URD Matching unknown Graft source Bone marrow Peripheral blood Conditioning regimen Myeloablative Reduced intensity GvHD prophylaxis CnI + methotrexate CnI + mycophenolate T-cell depletion Post-transplant Cy Other CMV serostatus R- /DR-/D+ R+/DR+/D+ Unknown
2061 (17) 51 (18-76) 458 (22) 480 (23) 964 (47) 159 (8)
Male recipients M–›M F–›M N (%) N (%)
2899 (25) 52 (18-75) 654 (23) 652 (22) 1339 (46) 254 (9)
4349 (37) 54 (18-81) 941 (22) 768 (18) 2059 (47) 581 (13)
260 (13) 926 (45) 76 (4) 82 (4) 70 (3) 233 (11) 75 (4) 260 (13) 79 (4)
361 1317 113 112 94 330 95 360 117
570 1570 354 177 128 541 146 707 156
226 (11) 1371 (67) 401 (19) 63 (3)
All N (%)
2488 (21) 53 (18-78) 569 (23) 487 (20) 1161 (47) 271 (11)
11797 (100) 52 (18-80) 2622 (22) 2387 (20) 5523 (47) 1265 (11)
(13) (36) (8) (4) (3) (12) (3) (16) (4)
331 (13) 910 (37) 184 (7) 97 (4) 85 (3) 297 (12) 77 (3) 400 (16) 107 (4)
1522 (13) 4723 (40) 727 (6) 468 (4) 377 (3) 1401 (12) 393 (3) 1727 (15) 459 (4)
342 (12) 1866 (64) 575 (20) 116 (4)
682 (16) 2668 (61) 820 (19) 179 (4)
363 (15) 1495 (60) 518 (21) 112 (5)
1613 (14) 7400 (63) 2314 (20) 470 (4)
761 (37) 593 (29) 681 (33) 26 (1)
1143 (39) 764 (26) 956 (33) 36 (1)
1802 (41) 1224 (28) 1260 (29) 63 (1)
1035 (42) 735 (30) 690 (28) 28 (1)
4741 (40) 3316 (28) 3587 (30) 153 (1)
713 (35) 1266 (61) 82 (4)
1027 (35) 1754 (61) 118 (4)
1481 (34) 2703 (62) 165 (4)
819 (33) 1553 (62) 116 (5)
4040 (34) 7276 (62) 481 (4)
998 (48)
1038 (36)
1492 (34)
1293 (52)
4821 (41)
672 (33) 239 (12) 82 (4) 26 (1) 31 (2) 13 (<1)
1374 (47) 337 (12) 88 (3) 19 (<1) 28 (<1) 15 (<1)
2119 (49) 512 (12) 128 (3) 34 (<1) 43 (<1) 21 (<1)
727 (29) 293 (12) 103 (4) 27 (1) 26 (1) 19 (<1)
4892 (41) 1381 (12) 401 (3) 106 (1) 128 (1) 68 (1)
291 (14) 1770 (86)
415 (14) 2484 (86)
565 (13) 3784 (87)
285 (11) 2203 (89)
1556 (13) 10241 (87)
1160 (56) 901 (44)
1645 (57) 1254 (43)
2262 (52) 2087 (48)
1341 (54) 1147 (46)
6408 (54) 5389 (46)
1156 (56) 471 (23) 47 (2) 100 (5) 287 (14)
1697 (59) 653 (23) 47 (2) 79 (3) 423 (15)
2449 (56) 1053 (24) 73 (2) 128 (3) 646 (15)
1426 (57) 600 (24) 42 (2) 88 (4) 332 (13)
6728 (57) 2777 (24) 209 (2) 395 (3) 1688 (14)
472 (23) 233 (11) 560 (27) 756 (37) 40 (2)
688 (24) 271 (9) 1021 (35) 859 (30) 60 (2)
1369 (31) 484 (11) 1273 (29) 1143 (26) 80 (2)
634 (25) 374 (15) 564 (23) 860 (35) 56 (2)
3163 (27) 1362 (12) 3418 (29) 3618 (31) 236 (2)
(12) (45) (4) (4) (3) (11) (3) (12) (4)
continued on the next page
1262
haematologica | 2016; 101(10)
Donor/recipient sex combination in HSCT continued from the previous page
Year of transplant 2008 2009 2010 Median follow up among survivors, months (range)
635 (31) 691 (34) 735 (36)
884 (30) 981 (34) 1034 (36)
1311 (30) 1459 (34) 1579 (36)
758 (30) 840 (34) 890 (36)
3588 (30) 3971 (34) 4238 (36)
48 (12, 76)
48 (2.5, 75)
48 (3, 76)
48 (2.3, 76)
48 (2.3-76)
Numbers are frequencies with percentage in parenthesis except medians and ranges. Percentages may not add up to 100 because of rounding. aClassified according to Armand18 et al. bClassified according to Sorror22 et al. cClassified according to Lee23 et al.; haploidentical category also includes 5/8 and 6/8 matched relatives. M–>M: male donor with male recipient; F–>M: female donor with male recipient; F–>M: female donor with male recipient; F–>F: female donor with female recipient. ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; CLL: chronic lymphocytic leukemia; CML: chronic myelogenous leukemia; MDS: myelodysplastic syndrome; HCT-CI: HCT comorbidity index; MRD: matched related donor; URD: unrelated donor; GvHD: graft-versus-host disease; CnI: calcineurin inhibitor; Cy: cyclophosphamide; CMV: cytomegalovirus.
A
B
OS
PFS
Years from transplantation
Years from transplantation
D
C
NRM
Relapse
Years from transplantation
Years from transplantation
E cGvHD
Figure 1. Transplantation outcome by recipient and donor sex. (A) Overall survival (OS). (B) Progression-free survival (PFS). (C) Cumulative incidence of relapse. (D) Cumulative incidence of non-relapse mortality (NRM). (E) Cumulative incidence of chronic GvHD.
M→M group, the HR was 1.03 (95%CI: 0.97-1.1, P=0.31) for OS and 0.99 (95%CI: 0.93-1.05, P=0.77) for PFS (Table 2B). Results were consistent when the analysis was repeated by disease (myeloid vs. lymphoid) (data not shown) or when the analysis was restricted to matched related or well matched unrelated (data not shown).
Relapse and non-relapse mortality Relapse and NRM were slightly worse in male recipients haematologica | 2016; 101(10)
compared with female recipients (Table 2A and Online Supplementary Table S1), but not significant at the 0.01 level. When sex combinations were considered, the 4-year cumulative incidence of NRM was 26% in the F→M group, 23% in each of the M→F, F→M, M→M groups (P=0.045 for the 4 group comparison, P=0.0075 for F→M vs. the other 3 groups combined) (Figure 1D, Online Supplementary Table S1) and the 4-year cumulative incidence of relapse was 44% in the M→M group, 40% in 1263
H.T. Kim et al. Table 2A. Multivariable regression analysis* for overall survival, progression-free survival, relapse, non-relapse mortality and chronic graft-versushost disease.
Recipient
OS
PFS
Relapse
NRM
Sex HR 95% CI P HR 95% CI P HR 95% CI P Male vs. female 1.11 1.05 1.16 <0.0001 1.10 1.05 1.15 <0.0001 1.06 1.002 1.12 0.04 Table 2B. Multivariable regression analysis* for OS, PFS, relapse, NRM, and cGVHD.
sHR 95% CI P 1.09 1.01 1.17 0.032
cGvHD sHR 95% CI P 1.02 0.96 1.08 0.47
Donor and recipient Sex F–>F vs. M–>F M–>M vs. M–>F F–>M vs. M–>F F–>M vs. M–>M
HR 1.02 1.10 1.14 1.03
95% CI P HR 95% CI P sHR 95% CI P 0.94 1.10 0.64 1.02 0.95 1.10 0.58 1.02 0.93 1.11 0.66 1.03 1.17 0.0032 1.11 1.05 1.18 0.0004 1.12 1.04 1.20 0.0036 1.06 1.22 0.0004 1.10 1.03 1.18 0.0044 0.99 0.91 1.08 0.89 0.97 1.10 0.31 0.99 0.93 1.05 0.77 0.89 0.82 0.96 0.0037
sHR 1.02 1.03 1.22 1.19
95% CI P 0.90 1.15 0.78 0.93 1.13 0.62 1.10 1.37 0.0003 1.08 1.32 0.0005
sHR 1.11 0.99 1.21 1.23
95% CI P 1.01 1.21 0.023 0.92 1.06 0.76 1.12 1.32 <0.0001 1.13 1.33 <0.0001
*Models are stratified by conditioning intensity. Cox model was used for overall survival (OS) and progression-free survival (PFS). Fine and Gray model was used for cumulative incidence of relapse, non-relapse mortality (NRM) and chrnoic graft- versus-host disease (cGvHD).Variables included in each model are listed in Table 1, except year of transplant.
M→F, 42% in F→F, and 42% in F→M (P=0.03 for the 4 group comparison, P=0.009 for M→M vs. the other 3 groups combined) (Figure 1C, Online Supplementary Table S1). These results were consistent in multivariable analysis with a subdistribution hazard ratio (sHR) of NRM 1.22 for F→M compared with M→F (P=0.0003) and 1.19 for F→M compared with M→M (P=0.0005); sHR of relapse 1.12 for M→M compared with M→F (P=0.0036) and 1.12 for M→M compared with F→M (P=0.0037) (sHR of relapse 0.89 for F→M compared with M→M) (Table 2B).
This was confirmed in multivariable analysis with an sHR of 1.21 (P<0.0001) for F→M compared with M→F, 1.23 (P<0.0001) compared with M→M (Table 2B). Among female donors, the cumulative incidence of chronic GvHD in F→F was somewhat higher compared with M→F (sHR 1.11, P=0.023). Since graft source is a significant prognostic factor for chronic GvHD, we analyzed chronic GvHD by sex match and graft source. Again the F→M group had a higher incidence of chronic GvHD compared to the other three groups, regardless of the graft source (Online Supplementary Table S2).
Acute graft-versus-host disease Out of 11,797 patients, for 168 patients (1.4% male and 1.5% female recipients) information regarding acute GvHD was missing. Of the remaining 11,629 patients, 5003 patients (43%) developed grade II-IV acute GvHD, and 2086 (18%) developed grade III-IV acute GvHD. Of the 5003 patients with grade II-IV acute GvHD, for 2988 (60%) the onset date of acute GvHD was not available. This large amount of missing informative data precluded an analysis of acute GvHD in time-to-event analysis. In order to circumvent the informative missing data regarding onset date as well as the confounding factor of early death without developing acute GvHD, we performed a landmark analysis on the frequency of acute GvHD restricted to patients who were alive at day 100 of HSCT (n=10,184). Of the 10,184 patients included in the landmark analysis, 4413 (43%) developed grade III-IV acute GvHD. No combination of recipient and donor sex was significantly associated with an increased frequency of grade II-IV or III-IV acute GvHD (Table 3A). However, in multivariable logistic regression analysis, F→M had slightly higher odds of developing grade II-IV acute GvHD (OR 1.17, 95CI:1.04-1.32, P=0.01).
Chronic graft-versus-host disease Among all the patients, 467 patients had missing onset date of chronic GvHD; the rate of missing information was similar across all recipient and donor sex combinations (approx. 4%). Among all patients with available information (n=11,330), when the cumulative incidence of chronic GvHD was compared by recipient and donor sex, there was a significant increase in the F→M group. The 1-year cumulative incidence rate in this group was 42% (P<0.0001), compared with 37% in M→F, 38% (P=0.14) in F→F, and 36% (P=0.79) in M→M (Table 3B and Figure 1E). 1264
Discussion By analyzing a large cohort of patients undergoing allogeneic transplants from multiple centers in the modern transplantation era, we show that male recipients have worse OS and PFS compared to female recipients regardless of donor sex, with approximately a 10% relative increase in the hazard of death or death/relapse in multivariable analyses. The basis for the detrimental effect of male recipient sex appears to vary by donor sex. In F→M transplants, there is an increase in NRM, likely attributable to an increase in chronic GvHD. Despite this increase in NRM and chronic GvHD, we could not identify a decreased incidence of relapse in the F→M group compared with M→F, and the ultimate effect of this sex combination was a decrement in OS and PFS. When the F→M group was compared with the M→M group, there was no difference in OS and PFS, but a decrease in relapse and an increase in NRM and chronic GvHD. These results in F→M are largely consistent with previous studies.2,3,5,7,9 Randolph et al.2 proposed that F→M pairs have the lowest risk for relapse and the greatest odds for GvHD compared to other recipient and donor sex combination groups, suggesting a selective graft-versus-leukemia (GvL) effect in this cohort. However, the worst OS was seen in F→M, implying that the NRM was significantly higher in F→M than the other three groups in that cohort, which outweighed the benefit in terms of relapse. In addition, we observed that M→M pairs have an increased incidence of relapse, compared to all of the other recipient and donor sex combination groups. There is no obvious biological basis for this result from our knowledge of Y chromosome mHAs unless Y chromosome itself attributes to this effect, as it has been seen in the general haematologica | 2016; 101(10)
Donor/recipient sex combination in HSCT
Table 3A. Frequency of acute graft-versus-host disease (aGvHD): landmark analysis at day 100 post hematopoietic stem cell transplantation.
M–>F F–>F M–>M F –>M Pc
II-IV aGvHD %
N
Na
2468 1786 3771 2159
1058 783 1602 970
Donor and patient sex
42.9 43.8 42.5 44.9
III-IV aGvHD Nb
%
383 292 603 356
15.5 16.3 16 16.5
0.29
0.81
Frequency of grade II-IV aGVHD; bfrequency of grade III-IV aGVHD; cboth P-values (0.29 and 0.81) are for the 4 group comparison. P-values for all pairwise comparisons were >0.1.
a
Table 3B. Chronic graft-versus-host disease outcome by recipient and donor sex combination.
Donor and patient sex M–>F F–>F M–>M F–>M
N
1-year (95%CI)
Cumulative incidence 2-year (95%CI)
P
2780 1979 4189 2382
37% (35-39) 38% (36-40) 36% (35-38) 42% (40-44)
41% (39-43) 44% (42-44) 41% (39-42) 46% (44-48)
ref 0.14 0.79 <0.0001
population that males have a shorter life expectancy than females. The higher incidence of relapse seen in M→M resembles an increased relapse rate seen in matched sibling donors as compared to matched unrelated donors. We note, however, that the final outcome of these effects, namely an inferior OS and PFS in male recipients of HSCT compared to female recipients, regardless of donor sex, is consistent with that described in another large and fully disease risk-annotated HSCT cohort.18 Nonetheless, one may argue that male recipients have higher risk baseline characteristics compared to female recipients. However, the distributions of DRI and HCT-CI are largely compatible across all recipient and donor sex combinations (Table 1), suggesting that female donors were not particularly used for high-risk male recipients. Given this, and the very large size of the present cohort, it seems highly unlikely that this is a statistical artifact. Instead, we hope that our findings can generate new biological studies of sex itself and sex-related mHAs to explain this phenomenon. As to sex mismatch for female recipients, donor sex has no effect on OS, PFS, relapse, and NRM, which is consistent with the result reported in Randolph et al.2 but different from a previous EBMT report.3 In their study, Gahrton et al.3 reported that the F→F group had a significantly lower NRM compared with M→F and had the best OS among all recipient and donor sex combinations. We acknowledge several important limitations of this work. First, it is retrospective in nature and therefore, like other retrospective studies, is subject to possible confounding factors, even though a sex-based selection bias seems unlikely. Second, about 4% of patients had missing onset date of chronic GvHD; as this percentage was similar across all sex combinations, the results should not be significantly affected. For acute GvHD, because the essential transplant data collection form does not mandate capturing the onset date of acute GvHD, a large portion of patients did not have this information, which precluded a time-to-event data analysis of this outcome. However, since most acute GvHD occur early after HSCT, and given haematologica | 2016; 101(10)
that nearly all patients (>99.97%) were followed for at least six months, the landmark analysis we performed is unbiased and the results for group comparison should be consistent whether binary data analysis or time-to-event analysis is used, even though the overall incidence rates of acute GvHD may be slightly underestimated. The effect of donor parity and donor age on transplantation outcomes was not explored in the current analyses. Data on parity for sibling donors was not collected consistently during the study period, which prevented us from exploring any effects of parity in the setting of HLA-matched related donor HSCT. A recent report from our group on optimal donor characteristics for unrelated donor transplantation showed higher mortality risk with increasing donor age, higher NRM and fewer relapse with parous female donors.24 However, parity was not associated with survival, as any advantage from lower relapse risk was negated by higher NRM.24 The age of the sibling donor is tightly correlated with recipient age and was not examined further. In summary, for female recipients of allograft, donor sex has no detectable effect on HSCT outcomes. In contrast, for male recipients, female donors are associated with a decreased incidence of relapse, an increased incidence of NRM and chronic GvHD, while male donors are associated with an increased incidence of relapse. As a result, both OS and PFS are significantly worse for all male recipients, regardless of donor sex. Furthermore, because OS and PFS are similar between F→M and M→M, one could argue that male recipients fare better with male donors considering the chronic GvHD related quality of life (QoL). However, it is not obvious whether QoL would be more severely affected by GvHD or by management of relapse, so that it may be premature to make recommendations about the preferred donor sex for male recipients until additional studies that include QoL are conducted. Finally, it is important to recognize that the absolute differences in clinical outcomes across recipient and donor sex combinations are small (within 1265
H.T. Kim et al.
5% across all sex combinations) and much smaller than those attributable to important prognostic factors such as disease risk index and donor-recipient HLA match. Nonetheless, our results do have a direct and important bearing on the choice of HSCT donors. In our cohort, the donor sex distribution seemed to be skewed toward male donors for male recipients (64% vs. 42% sex matched in female recipients), which may reflect the commonly held view that a male donor is preferable for a male recipient, which was perhaps derived from the clinical reports in previous years. However, based on our findings, recipient sex rather than donor sex appears to
References 1. Gratwohl A, Hermans J, Niederwieser D, van Biezen A, van Houwelingen HC, Apperley J. Female donors influence transplant-related mortality and relapse incidence in male recipients of sibling blood and marrow transplants. Hematol J. 2001;2(6):363-370. 2. Randolph SS, Gooley TA, Warren EH, Appelbaum FR, Riddell SR. Female donors contribute to a selective graft-versusleukemia effect in male recipients of HLAmatched, related hematopoietic stem cell transplants. Blood. 2004;103(1):347-352. 3. Gahrton G, Iacobelli S, Apperley J, et al. The impact of donor gender on outcome of allogeneic hematopoietic stem cell transplantation for multiple myeloma: reduced relapse risk in female to male transplants. Bone Marrow Transplant. 2005;35(6):609617. 4. Gallardo D, Pérez-García A, de la Cámara R, et al. Clinical outcome after sex-mismatched allogeneic stem cell transplantation from human lymphocyte antigenidentical sibling donors: influence of stem cell source. Leukemia. 2006;20(8):14611464. 5. Loren AW, Bunin GR, Boudreau C, et al. Impact of donor and recipient sex and parity on outcomes of HLA-identical sibling allogeneic hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2006;12(7):758-769. 6. Gahrton G. Risk assessment in haematopoietic stem cell transplantation: impact of donor-recipient sex combination in allogeneic transplantation. Best Pract Res Clin Haematol. 2007;20(2):219-229. 7. Flowers ME, Inamoto Y, Carpenter PA, et al. Comparative analysis of risk factors for
1266
8.
9.
10. 11.
12.
13.
14.
be the predominant prognostic driver. Consequently, donor sex should not be considered in the donor selection algorithm until we have a better insight into the biology of sex-based alloreactivity. Funding This work was supported by U19 AI 29530 from NIAID, PO1 HL 070149 from NHLBI, U24-CA76518 from the NCI/ NHLBI/ NIAID, HHSH234200637015C (HRSA/DHHS) and P01 CA142106 from the NCI. The views expressed in this article do not reflect the official policy or position of the National Institutes of Health or any other agency of the U.S. Government.
acute graft-versus-host disease and for chronic graft-versus-host disease according to National Institutes of Health consensus criteria. Blood. 2011;117(11):3214-3219. Olsson RF, Logan BR, Chaudhury S, et al. Primary graft failure after myeloablative allogeneic hematopoietic cell transplantation for hematologic malignancies. Leukemia. 2015;29(8):1754-1762. Arai S, Arora M, Wang T, et al. Increasing incidence of chronic graft-versus-host disease in allogeneic transplantation: a report from the Center for International Blood and Marrow Transplant Research. Biol Blood Marrow Transplant. 2015;21(2):266274. Goulmy E, Termijtelen A, Bradley BA, van Rood JJ. Alloimmunity to human H-Y. Lancet. 1976;2(7996):1206. Miklos DB, Kim HT, Miller KH, et al. Antibody responses to H-Y minor histocompatibility antigens correlate with chronic graft-versus-host disease and disease remission. Blood. 2005;105(7):2973-2978. Gratwohl A, Hermans J, Goldman JM, et al. Risk assessment for patients with chronic myeloid leukaemia before allogeneic blood or marrow transplantation. Chronic Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Lancet. 1998; 352(9134): 1087-1092. Gratwohl A, Stern M, Apperley J, et al. The EBMT risk score predicts outcome after allogeneic HSCT in all haematological disease categories and is independent of stem cell source or conditioning intensity. Bone Marrow Transplant. 2008;41:S3. Terwey TH, Hemmati PG, Martus P, et al. A modified EBMT risk score and the hematopoietic cell transplantation-specific comorbidity index for pre-transplant risk assessment in adult acute lymphoblastic
15.
16. 17.
18.
19. 20. 21. 22.
23.
24.
leukemia. Haematologica. 2010;95(5):810818. Gratwohl A, Stern M, Brand R, et al. European Group for Blood and Marrow Transplantation and the European Leukemia Net. Risk score for outcome after allogeneic hematopoietic stem cell transplantation: a retrospective analysis. Cancer. 2009;115(20):4715-4726. Gratwohl A. The EBMT score. Bone Marrow Transplant. 2012;47(6):749-756. Armand P, Gibson CJ, Cutler C, et al. A disease risk index for patients undergoing allogeneic stem cell transplantation. Blood. 2012;120(4):905-913. Armand P, Kim HT, Logan BR, et al. Validation and refinement of the Disease Risk Index for allogeneic stem cell transplantation. Blood. 2014;123(23):3664-3671. Gray RJ. A class of K-sample tests for comparing the cumulative incidence of a competing risk. Ann Stat. 1988;16:1140-1154. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of competing risk. J Am Stat Assoc. 1999;94:496-509. Zhou B, Latouche A, Rocha V, Fine J. Competing risks regression for stratified data. Biometrics. 2011;67(2):661-670. Lee SJ, Klein J, Haagenson M, et al. Highresolution donor-recipeint HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110(13):4567-4583. Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. Kollman C, Spellman SR, Zhang MJ, et al. The effect of donor characteristics on survival after unrelated donor transplantation for hematologic malignancy. Blood. 2016; 127(2):260-267.
haematologica | 2016; 101(10)
ARTICLE
Stem Cell transplantation
Human leukocyte antigen supertype matching after myeloablative hematopoietic cell transplantation with 7/8 matched unrelated donor allografts: a report from the Center for International Blood and Marrow Transplant Research Aleksandr Lazaryan,1 Tao Wang,2 Stephen R. Spellman,3 Hai-Lin Wang,2 Joseph Pidala,4 Taiga Nishihori,4 Medhat Askar,5 Richard Olsson,6 Machteld Oudshoorn,7 Hisham Abdel-Azim,8 Agnes Yong,9 Manish Gandhi,10 Christopher Dandoy,11 Bipin Savani,12 Gregory Hale,13 Kristin Page,14 Menachem Bitan,15 Ran Reshef,16 William Drobyski,17 Steven GE Marsh,18 Kirk Schultz,19 Carlheinz R. Müller,20 Marcelo A. Fernandez-Viña,21 Michael R. Verneris,1 Mary M. Horowitz,2 Mukta Arora,1 Daniel J. Weisdorf,1 and Stephanie J. Lee22
EUROPEAN HEMATOLOGY ASSOCIATION
Ferrata Storti Foundation
Haematologica 2016 Volume 101(10):1267-1274
1 University of Minnesota Medical Center, Fairview, Minneapolis, MN, USA; 2Center for International Blood and Marrow Transplant Research, Milwaukee, WI, USA; 3Center for International Blood and Marrow Transplant Research, Minneapolis, MN, USA; 4H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL, USA; 5Baylor University Medical Center, Dallas, TX, USA; 6Karolinska University Hospital, Centre for Allogeneic Stem Cell Transplantation, Stockholm, Sweden; 7Leiden University Medical Centre, The Netherlands; 8 Children's Hospital of Los Angeles, CA, USA; 9Royal Adelaide Hospital / SA Pathology, Australia; 10 Mayo Clinic, Rochester, MN, USA; 11Cincinnati Children's Hospital Medical Center, OH, USA; 12 Vanderbilt University Medical Center, Nashville, TN, USA; 13All Children's Hospital, St. Petersburg, FL, USA; 14Duke University Medical Center, Pediatric Blood and Marrow Transplant, Durham, NC, USA; 15Tel-Aviv Sourasky Medical Center, Israel; 16Columbia University Medical Center, New York, NY, USA; 17Froedtert Memorial Lutheran Hospital, Milwaukee, MN, USA; 18 Anthony Nolan Research Institute & University College London Cancer Institute, Royal Free Campus, UK; 19British Columbia's Children's Hospital, Vancouver, British Columbia, Canada; 20 Zentrales Knochenmarkspender-Register Deutschland, Ulm, Germany; 21Stanford University Medical Center, CA, USA; and 22Fred Hutchinson Cancer Research Center, Seattle, WA, USA
ABSTRACT
Correspondence: alazarya@umn.edu
T
he diversity of the human leukocyte antigen (HLA) class I and II alleles can be simplified by consolidating them into fewer supertypes based on functional or predicted structural similarities in epitopebinding grooves of HLA molecules. We studied the impact of matched and mismatched HLA-A (265 versus 429), -B (230 versus 92), -C (365 versus 349), and -DRB1 (153 versus 51) supertypes on clinical outcomes of 1934 patients with acute leukemias or myelodysplasia/myeloproliferative disorders. All patients were reported to the Center for International Blood and Marrow Transplant Research following single-allele mismatched unrelated donor myeloablative conditioning hematopoietic cell transplantation. Single mismatched alleles were categorized into six HLA-A (A01, A01A03, A01A24, A02, A03, A24), six HLA-B (B07, B08, B27, B44, B58, B62), two HLA-C (C1, C2), and five HLA-DRB1 (DR1, DR3, DR4, DR5, DR9) supertypes. Supertype B mismatch was associated with increased risk of grade II-IV acute graft-versus-host disease (hazard ratio =1.78, P=0.0025) compared to supertype B match. Supertype B07-B44 mismatch was associated with a higher incidence of both grade II-IV (hazard ratio=3.11, P=0.002) and III-IV (hazard ratio=3.15, P=0.01) acute graft-versus-host disease. No significant associations were detected between supertype-matched versus -mismatched groups at other HLA loci. These data suggest that avoiding HLAB supertype mismatches can mitigate the risk of grade II-IV acute graft-versus-host disease in 7/8-mismatched unrelated donor hematopoietic cell transplantation when multiple HLA-B supertype-matched donors are available. Future studies are needed to define the mechanisms by which supertype mismatching affects outcomes after alternative donor hematopoietic cell transplantation. haematologica | 2016; 101(10)
Received: February 12, 2016. Accepted: May 25, 2016. Pre-published: May 31, 2016. doi:10.3324/haematol.2016.143271
Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/10/1267
©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.
1267
A. Lazaryan et al.
Introduction Cellular immune responses are mediated in part by cytotoxic T lymphocytes that recognize peptide molecules bound to human leukocyte antigen (HLA) on the surface of antigen-presenting cells. HLA molecules are extremely polymorphic1 and most of their variation is centered within the peptide-binding grooves that accommodate the primary anchor positions of the peptides displayed for immune recognition.2 Discovery of HLA supertypes almost a decade ago offered a simplification of diverse HLA nomenclature by consolidating individual HLA class I and II alleles into fewer supertype clusters based on functional or predicted structural similarities in epitope-binding specificities of HLA molecules.3 HLA alleles belonging to each particular supertype have either experimentally proven or predicted ability to present antigenic peptides with similar anchoring amino acids at the second (B-pocket) and C-terminal (F-pocket) positions of the peptide molecules.4 Although HLA class I and II supertypes have been increasingly studied in association with immune susceptibility to infection5-7 and cancer,8,9 the significance of individual allele mismatching within and outside of HLA class I or II supertypes remains unknown in the context of allogeneic hematopoietic cell transplantation (alloHCT). Recent encouraging outcomes in fully HLA-matched unrelated donor (MUD) HCT10,11 have contributed in part to the steady rise of MUD allografts that now outnumber related donor transplants reported annually to the Center for International Blood and Marrow Transplant Research (CIBMTR). Allografts mismatched at a single HLA-A, -B, -C, or -DRB1 locus [i.e. 7/8 mismatched unrelated donor (MMUD) HCT] were previously reported to be associated with lower overall and disease-free survival, higher treatment-related mortality, and more acute graft-versus-host disease (GVHD) compared to outcomes of 8/8 MUD allografts.12-14 Despite these risks, 7/8 MMUD grafts remain a viable option for HCT, particularly in minorities who lack suitable donors or in patients with aggressive hematologic malignancies for whom the risks of disease progression due to delays in identifying optimal donors15 is offset in part by the benefits of earlier transplantation with a 7/8 MMUD alloHCT. Although multiple strategies have been sought to identify “permissible mismatches” associated with improved outcomes of a single-allele MMUD HCT,13,16-19 the clinical significance of clustering mismatched alleles within HLA class I or II supertypes has not been established. We therefore conducted a large registry analysis of the CIBMTR database of single-allele mismatched myeloablative allografts to determine whether HLA class I or II supertype mismatching is associated with worse outcomes after 7/8 MMUD alloHCT.
Methods Study design and patient selection The study base population consisted of 2218 recipients of myeloablative conditioning followed by 7/8 HLA MUD bone marrow or peripheral blood stem cell allografts for acute myeloid leukemia, acute lymphoblastic leukemia, chronic myeloid leukemia, and myelodysplastic syndrome between 1999 and 2011. The patients’ data were reported to the National Marrow Donor 1268
Program (NMDP)/CIBMTR, and subjects were excluded if: (i) they did not consent to participate (n=55); (ii) they had fewer than 100 days of post-transplant follow up (n=11); (iii) their disease status prior to alloHCT was missing (n=32); or (iv) they had undergone ex vivo T-cell depletion (n=186). Recipients of prior HCT were excluded. Allografts performed for lymphoid malignancies and non-malignant disorders were also excluded in order to enhance the overall homogeneity of the study population. All eligible adult and pediatric study participants (n=1934) from 175 transplant centers and 16 countries provided informed consent to participate in NMDP/CIBMTR research. This was a retrospective observational study approved by NMDP/CIBMTR’s Institutional Review Board. Standard methods of NMDP/CIBMTR data analysis were used to mitigate any bias related to exclusion of non-consenting study candidates.12
HLA class I and II typing and supertype assignment High-resolution allele-level typing at HLA-A, -B, -C, and -DRB1 loci was performed through the NMDP’s high-resolution HLA typing project according to well established and validated DNAbased techniques as previously reported.20 Single allele mismatch at HLA-A, -B, -C, or -DRB1 was defined as a “7/8 match”. The assignment algorithm for HLA-A and -B supertypes (Online Supplementary Table S1) was based on an updated supertype classification with revised main HLA anchor specificities.21 This method extends the previously described nine HLA-A and -B supertype designations3 (A1, A2, A3, A24, B27, B44, B58, and B62) to 12 supertype groups (A01, A01A03, A01A24, A02, A03, A24, B07, B08, B27, B44, B58, B62), mostly due to the fact that certain HLA-A alleles were found to have peptide-binding repertoires with overlapping supertype specificities thereby resulting in newly defined A01A03 and A01A24 supertype categories. This revised classification of HLA-A and -B supertypes captured 99% of the allelic diversity of allograft recipients and their donors. The remaining 1% of unclassified HLA-A and -B alleles were grouped into supertypes using bioinformatics methods.22 Two HLA-C supertypes (C1 and C2) were derived from hierarchical cluster analysis22 with distinct amino-acid fingerprints in protein structure for HLA-C1 (Ser77) and -C2 (Asn77), which also coincide with killer Ig-like receptor binding specificities for HLA-C.23 The grouping of HLA-DRB1 alleles into supertypes was accomplished according to previously described in-silico methods on the basis of common structural and functional features of HLA class II molecules.24 The significance of alternative supertype designations and individual supertype effects was further assessed in the post-hoc exploratory analysis.
Study endpoints The primary comparison between the 7/8 supertype-matched and 7/8 supertype-mismatched allografts was conducted across major clinical endpoints including overall survival, disease-free survival, relapse, treatment-related mortality, acute GVHD, chronic GVHD, and time-to-neutrophil recovery (absolute neutrophil count ≥ 0.5x109/L). Overall survival corresponded to the time from transplantation to death from any cause and surviving patients were censored at the time of their last follow-up. Disease-free survival was defined as the time between transplantation and relapse or death from any cause; patients who remained alive and in remission were censored at the time of their last follow-up. Clinical relapse of the primary disease and treatment-related mortality were defined by established CIBMTR criteria with the latter defined as death while in continuous remission. Relapse was therefore considered a competing risk endpoint for treatmentrelated mortality, and treatment-related mortality was considered a competing risk for relapse. The onset of grades II-IV or III-IV haematologica | 2016; 101(10)
HLA supertype matching in alloHCT
acute GVHD was determined based on the Consensus criteria25 while the onset of chronic GVHD was determined based on the Seattle criteria.26 Neutrophil engraftment was defined as time-toneutrophil recovery. Death was considered a competing risk endpoint for engraftment and GVHD.
Statistical analysis Descriptive frequency estimates and comparisons for HLA alleles and supertypes as well as non-HLA study variables were obtained through the standard methods of categorical and continuous data analysis. Univariate probabilities for overall and diseasefree survival were calculated using the Kaplan-Meier estimator,27 whereas probabilities of treatment-related mortality, relapse, acute GVHD, chronic GVHD, and neutrophil engraftment were calculated as cumulative incidence rates while accounting for competing risks.28 Survival curves were compared by the log-rank test. Multivariate models for overall survival, disease-free survival, relapse, treatment-related mortality, acute GVHD, chronic GVHD and neutrophil engraftment were built using Cox proportional hazards models. All clinical variables were tested for the affirmation of the proportional hazards assumption. Variables that were found to violate this assumption were adjusted for by stratification. Final outcome-specific models were developed using a stepwise model building procedure with the threshold of α=0.05 for both entry and retention of co-variates in the model. Main variables, including HLA supertypes, were forced into the models with the interactions between the main variables and the adjusted covariates being tested at the significance level of α=0.01. Given the multiple testing, P values <0.01 were considered statistically significant.
Results HLA class I and II alleles and supertypes For 1934 recipients of 7/8 MMUD alloHCT, single-allele mismatches occurred within the HLA-A (36%), -B (17%), C (37%), and -DRB1 (11%) loci. Individual HLA-A, -B, -C, and -DRB1 allele-level mismatches were matched by corresponding HLA supertypes in 38%, 71%, 51%, and 75%, respectively (Table 1). Overall, supertype-level matching at any one of the four HLA loci was observed in 52% of study subjects.
Non-HLA characteristics Baseline patient and clinical characteristics are summarized in Table 2. In brief, the patients’ median age was 35 years (range, 1-70), and less than 20% of the study population was of non-Caucasian background. Acute myeloid leukemia and acute lymphoblastic leukemia accounted for 76% of all hematologic malignancies with over half of
patients classified as having intermediate or advanced risk disease. Peripheral blood stem cell allografts were used in 56% of all transplant procedures. Conditioning regimens for alloHCT included total body irradiation in 58% of cases, whereas anti-thymocyte globulin or alemtuzumab was incorporated into conditioning regimens in 36% of cases. The majority of GVHD prophylactic regimens included tacrolimus (62%) or cyclosporine (36%). The median follow-up of surviving patients was 54 months (range, 3-149) after alloHCT. In the crude comparisons of supertype-matched (any locus) versus –mismatched 7/8 allografts, significant differences were observed in underlying hematologic malignancies, conditioning regimens and timing of alloHCT (all P<0.01). Specifically, the supertype-matched group contained a greater proportion of total body irradiation-based conditioning regimens (62% versus 54%, P<0.001) and a smaller proportion of in vivo Tcell-depleted grafts (32% versus 40%, P<0.001). Supertypemismatched grafts were also more common in recent years (P<0.001).
HLA supertype-matched and -mismatched outcomes Univariate analyses of post-transplant outcomes based on HLA supertype matching are summarized in Table 3. Recipients of HLA supertype B-mismatched allografts (n=62) had a significantly higher cumulative incidence of grade II-IV acute GVHD than did recipients of HLA-B supertype-matched (n=174) allografts (67% versus 47%, respectively) (Figure 1, log-rank P=0.007). This association was primarily driven by an excess in grade II acute GVHD as no difference was found in the incidence of severe grade III-IV acute GVHD with supertype-B mismatching. The independent effect of HLA-B supertype matching on grade II-IV acute GVHD was confirmed by the multivariable analysis [hazard ratio (HR)=1.78; 95% confidence interval (CI), 1.23-2.59; P=0.0025] adjusting for age, gender, disease type, ABO-mismatch, graft source, and in vivo T-cell depletion (Figure 3). No other class I supertype mismatch (including supertype mismatch at any locus) was found to be significantly associated with any of the study endpoints (engraftment, chronic GVHD, relapse or death) at the pre-specified statistical threshold. HLA-B supertype mismatches involving B07-B44, B27B44, and B07-B62 were found to be the most prevalent and these individual mismatches were subsequently examined in the post-hoc analysis for their association with acute GVHD. In contrast to all other HLA-B mismatched supertypes, B07-B44 mismatched allografts were associated with a higher incidence of both grade II-IV (HR=3.11; 95% CI, 1.54-6.28, P=0.002) and III-IV acute GVHD (HR=3.15; 95% CI, 1.30-7.65, P=0.01).
Table 1. HLA class I/II allele- and supertype-level distribution.
Single allele-mismatch
N Matched (%)
HLA-A HLA-B HLA-C HLA-DRB1 Any allele
694 322 714 204 1934*
265 (38.2) 230 (71.4) 365 (51.1) 153 (75) 1000 (51.7)
HLA supertype Mismatched (%) 429 (61.8) 92 (28.6) 349 (48.9) 51 (25) 921 (48.3)
*Missing supertype assignment for 13 patients.
haematologica | 2016; 101(10)
1269
A. Lazaryan et al. Table 2. Non-HLA characteristics of the study population.
Characteristic Number of patients Number of centers Age, median (range), years Age at alloHCT, years <18 years old 19-35 years old 36-55 years old >55 years old Gender Male Female KPS prior to alloHCT < 90 90-100 Missing Race of recipient Caucasian African-American Asian / Pacific Islander Hispanic Native American CMV Donor/Recipient D-/RD-/R+ D+/RD+/R+ Disease Acute myeloid leukemia Acute lymphoblastic leukemia Chronic myeloid leukemia Myelodysplastic syndrome Disease risk¥ Early Intermediate Advanced Donor parity Male or non-parous female Parous female Missing Donor/Recipient sex match Male / male Male / female Female / male Female / female Graft source Bone marrow Peripheral blood stem cells Conditioning with TBI In vivo T-cell depletion§ GVHD prophylaxis Tacrolimus-based Cyclosporin A-based Other‡ Year of alloHCT 1999-2002 2003-2006 2007-2011 Follow up, median (range), months
All
Supertype-matched
Supertype-mismatched
1934* 175 35 (1-70)
999 161 35 (1-69)
922 148 37 (1-70)
410 (21%) 541 (28%) 758 (39%) 225 (12%)
212 (21%) 292 (29%) 386 (39%) 109 (11%)
194 (21%) 246 (27%) 367 (40%) 115 (12%)
1090 (56%) 844 (44%)
563 (56%) 436 (44%)
521 (57%) 401 (43%)
510 (26%) 1303 (67%) 121 (6%)
259 (26%) 666 (67%) 74 (7%)
248 (27%) 627 (68%) 47 (5%)
1574 (81%) 175 (9%) 62 (3%) 71 (4%) 52(3%)
824 (82%) 75 (8%) 35 (4%) 41 (4%) 7 (<1%)
740 (80%) 100 (11%) 27 (3%) 28 (3%) 7 (<1%)
533 (28%) 602 (31%) 272 (14%) 499 (26%)
294 (29%) 300 (30%) 137 (14%) 254 (25%)
235 (25%) 298 (32%) 133 (14%) 242 (26%)
870 (45%) 609 (31%) 274 (14%) 181 (9%)
427 (43%) 345 (35%) 147 (15%) 80 (8%)
436 (47%) 262 (28%) 124 (13%) 100 (11%)
813 (42%) 612 (32%) 509 (26%)
447 (45%) 310 (31%) 242 (24%)
362 (39%) 298 (32%) 262 (28%)
1389 (72%) 443 (23%) 102 (5%)
692 (69%) 246 (25%) 61 (6%)
688 (75%) 193 (21%) 41 (4%)
657 (34%) 467 (24%) 433 (22%) 377 (19%)
329 (33%) 233 (23%) 234 (23%) 203 (20%)
325 (35%) 228 (25%) 196 (21%) 173 (19%)
845 (44%) 1089 (56%) 1127 (58%) 693 (36%)
452 (45%) 547 (55%) 623 (62%) 316 (32%)
385 (42%) 537 (58%) 498 (54%) 372 (40%)
1193 (62%) 689 (36%) 52 (3%)
625 (63%) 346 (35%) 28 (3%)
560 (61%) 338 (37%) 24 (3%)
405 (21%) 633 (33%) 896 (46%) 54 (3-149)
234 (23%) 351 (35%) 414 (41%) 60 (3-149)
167 (18%) 278 (30%) 477 (52%) 48 (3-145)
P value
0.16 0.51
0.95
0.11
0.12
0.43
0.006
0.03
0.03
0.43
0.12 <0.001 <0.001 0.64
<0.001
alloHCT: allogeneic hematopoietic cell transplantation; CMV: cytomegalovirus; D: donor; R: recipient; TBI: total body irradiation; GVHD: graft-versus-host disease; KPS: Karnofsky performance score; *Including 13 cases with missing supertypes. ¥According to ASBMT 2006 definitions. §Antithymocyte globulin or alemtuzumab. ‡Mycophenolate mofetil + other (n=5); methotrexate + other (n=10); antithymocyte globulin ± corticosteroid (n=6); sirolimus (n=1); unknown (n=30).
1270
haematologica | 2016; 101(10)
HLA supertype matching in alloHCT
Table 3. Univariate probabilities of clinical outcomes between HLA supertype-matched (M) and â&#x20AC;&#x201C;mismatched (MM) 7/8 unrelated donor allografts.
Outcomes
Timing post-HCT
HLA-A ST M vs. MM
HLA-B ST M vs. MM
HLA-C ST M vs. MM
HLA-DRB1 ST M vs. MM
Any HLA ST M vs. MM
Acute GVHD II-IV
Day 100
Acute GVHD III-IV
Day 100
Any chronic GVHD
2 years
ANC recovery
Day 28
Treatment-related mortality Relapse
3 years
Disease-free survival Overall survival
3 years
54% vs. 54% P=0.95 27% vs. 32% P=0.26 47% vs. 44% P=0.39 95% vs. 94% P=0.68 45% vs. 40% P=0.19 27% vs. 28% P=0.87 27% vs. 32% P=0.23 31% vs. 38% P=0.09
47% vs.67% P=0.006 31% vs. 32% P=0.84 51% vs. 42% P=0.14 94% vs. 95% P=0.78 45% vs. 37% P=0.25 20% vs. 26% P=0.28 35% vs. 37% P=0.81 40% vs. 41% P=0.88
51% vs. 48% P=0.56 25% vs. 26% P=0.73 45% vs. 37% P=0.03 94% vs. 93% P=0.58 37% vs. 40% P=0.42 30% vs. 33% P=0.31 33% vs. 27% P=0.06 39% vs. 30% P=0.02
50% vs. 56% P=0.6 22% vs. 22% P=0.98 43% vs. 46% P=0.78 91% vs. 94% P=0.5 31% vs. 50% P=0.03 27% vs. 19% P=0.26 42% vs. 31% P=0.2 46% vs. 30% P=0.05
51% vs. 53% P=0.53 26% vs. 29% P=0.27 47% vs. 41% P=0.012 94% vs. 94% P=0.91 40% vs. 40% P=0.89 26% vs. 29% P=0.16 34% vs. 30% P=0.14 38% vs. 35% P=0.12
3 years
3 years
ST: supertype; ANC recovery: absolute neutrophil count over 0.5x109/L.
Since the impact of single-allele mismatching was most apparent in patients with early and intermediate risk disease, as demonstrated in the prior large NMDP analysis, we analyzed the effect of HLA-B supertype mismatching within the subset of patients with early and intermediate risks. Similar to our major finding, compared to HLA-Bmatched supertype grafts, HLA-B-mismatched supertype grafts were associated with an increased risk of grade II-IV acute GVHD (HR=1.84; 95% CI, 1.20-2.84, P<0.01). Although HLA-DRB1 supertype-mismatched transplants (n=51) were associated with faster neutrophil engraftment (median 12 versus 16 days, Figure 2), this early difference in engraftment kinetics was not evident by day 28 after the transplant (94% versus 90%, P=0.4). Notably, the relatively slower neutrophil engraftment among HLADRB1 supertype-matched allograft recipients had no adverse influence on treatment-related mortality or other major post-transplant outcomes. On the contrary, mismatching at HLA-DRB1 supertypes was associated with a trend towards higher treatment-related mortality (HR=1.64; 95% CI, 0.99-2.74, P=0.057) and inferior overall survival (HR=1.58; 95% CI 1.04-2.38, P=0.037) compared to that associated with HLA-DRB1 supertype-matched allografts.
Discussion In this large registry analysis of the CIBMTR database of single allele mismatched myeloablative allografts, we found a significant increase in the hazard of grade II-IV acute GVHD among HLA-B supertype-mismatched compared to HLA-B supertype-matched recipients of 7/8 MMUD allografts. Allele-level 7/8 HLA-B mismatch was proven in the past to be associated with a higher incidence of acute GVHD compared to 8/8 HLA-match [estimated 28% (95% CI, 26%-30%) incidence of grade III-IV acute GVHD].12 In our cohort, the cumulative incidence rates of grades II-IV and III-IV acute GVHD among HLA-B alleleMMUD allograft recipients were 53% (95% CI, 48%haematologica | 2016; 101(10)
59%) and 31% (95% CI, 25%-37%), respectively. This study has further extended the significance of HLA-B mismatch in regards to acute GVHD at the supertype level for 7/8 allele-mismatched allografts. This observation conforms to our primary hypothesis of adverse post-transplant outcomes with mismatched HLA supertypes and it further supports the notion of increased alloreactivity with HLA-B supertype-mismatched 7/8 MMUD transplants as opposed to supertype-level mismatches at HLA-A, -C, or DRB1 loci. There are several possible explanations for our findings. First, HLA-B alleles in humans have the highest degree of described polymorphism relative to other class I or II alleles,29 likely as a result of the evolutionary pressures from various infectious pathogens. The contribution of HLA supertypes to immune-mediated responses against a number of viral infections was well established by prior studies.5-7,30-32 It is therefore possible that early post-transplant inflammatory responses mediated by mismatched HLA-B supertypes could perpetuate alloreactive immune responses such as acute GVHD. Although addressing this hypothesis was beyond the scope of this study, this could be tested in future studies. Second, it is possible that the supertype categorization algorithm used in this study could have obscured some of the underlying true associations between class I and II supertypes with major clinical outcomes after alloHCT. Considerable diversity and a variable degree of overlap exist between major HLA class I and II supertype classifications. In this study we used the revised supertype assignment algorithm proposed by Sette and Sidney for HLA class I A- and B-supertypes.21 As opposed to other alternatives,33-37 our chosen algorithm provided successful supertype classification for the entire study population with most of the HLA-A and -B alleles classified based on experimentally established motifs in epitope-binding pockets of HLA molecules. Our supertype assignment strategy therefore ensured the most stringent selection of corresponding alleles, and by doing so it strengthened the internal validity of the study. In addition, our post-hoc exploratory analysis of alternative HLA-A, -B, 1271
A. Lazaryan et al.
Figure 1. Cumulative incidence of grade II-IV acute GVHD according to HLA-B supertypes.
Figure 2. Neutrophil recovery according to HLA-DRB1 supertypes.
and -DRB1 supertype classifications (data not shown) revealed either limited capacity of other algorithms to cluster allelic diversity of our study population into supertype categories (e.g. the algorithms by Reche,37 Harjanto,34 Hertz/Yanover,35 and Greenbaum33), or significant overlap between our supertype assignment algorithm and other classifications such as those proposed by Lund36 or Doytchinova.22 Furthermore, accounting for the mismatch vector direction (i.e. graft-versus-host or host-versus-graft) did not further influence or enrich the findings from this study. Future practical implications of HLA-B supertypematched donor selection of 7/8 HLA-B MMUD allografts can be expected to lower the incidence of grade II-IV acute GVHD for a modest fraction (5%) of all 7/8 MMUD HCT according to the donor selection practices reflected in this study. Avoidance of B07-B44 supertype mismatches should be interpreted with caution given the small number of allografts (n=9) in that subset analysis of the individ1272
ual HLA-B supertype mismatches. Nevertheless, all but one B07-B44 supertype mismatched allografts were complicated by grade II-IV acute GVHD with over half classified as severe acute GVHD. Major limitations of this study are inherent to its retrospective design and in statistical challenges of analyzing multiple endpoints across various HLA class I supertypes. Consequently, we found faster neutrophil engraftment among HLA-DRB1 supertype-mismatched allograft recipients than among HLA-DRB1 supertype-matched allograft recipients to be more controversial and difficult to explain. Although recipients of HLA-DRB1 supertype-mismatched allografts achieved neutrophil recovery on average 4 days earlier, they demonstrated a trend towards inferior overall survival. Notably, the median estimated time of neutrophil engraftment for DRB1-matched supertypes (16 days), which accounted for 75% of all 7/8 DRB1 allelemismatched allografts, was overall comparable to data reported for 8/8 MUD HCT.38 In contrast, DRB1 supertype haematologica | 2016; 101(10)
HLA supertype matching in alloHCT
Figure 3. Multivariate analysis of the impact of supertype mismatching at HLA-A, -B, -C, and -DRB1. DFS: disease-free survival; TRM: treatment-related mortality; aGVHD: acute graft-versus-host disease; cGVHD: chronic graft-versus-host disease; ANC>500: absolute neutrophil count > 0.5x109/L.
mismatches accounted for only 2.6% of all 7/8 MMUD HCT in this study thereby raising the possibility of a random effect in lieu of a less plausible cause-and-effect relationship between DRB1 supertype mismatch and neutrophil engraftment kinetics. Further studies are needed to provide definitive guidance on incorporating DRB1 supertype matching in donor selection algorithms as increased treatment-related mortality and inferior overall survival, albeit not statistically significant in this dataset, are concerning. This large observational study has provided the first evidence of “permissible” supertype-based donor selection of optimal 7/8 MMUD for myeloablative alloHCT. Pending validation in an independent dataset, our findings suggest that avoiding HLA-B supertype mismatch can serve as a novel strategy to mitigate the risk of grade II-IV acute GVHD in 7/8 MMUD HCT when multiple potential HLAB supertype-matched donors are available. This study offers new insights and testable hypotheses for future studies on the role of HLA supertypes among recipients of reduced intensity MMUD HCT and recipients of other mismatched alternative donor allografts such as umbilical cord blood or haploidentical HCT. Acknowledgments The authors gratefully acknowledge Katharina Fleischhauer, Neng Yu, Biju George, and Jason Dehn for their participation in the study protocol review, and Michael Franklin, MS, for his assistance in editing the manuscript. The CIBMTR is supported by Public Health Service Grant/Cooperative Agreement 5U24CA076518 from the National Cancer Institute (NCI), the
References 1. Robinson J, Halliwell JA, Hayhurst JD, Flicek P, Parham P, Marsh SG. The IPD and IMGT/HLA database: allele variant databases. Nucleic Acids Res. 2015;43(Database issue):D423-431. 2. Felix NJ, Allen PM. Specificity of T-cell alloreactivity. Nat Rev Immunol. 2007;7 (12):942-953.
haematologica | 2016; 101(10)
National Heart, Lung and Blood Institute (NHLBI) and the National Institute of Allergy and Infectious Diseases (NIAID); a Grant/Cooperative Agreement 5U10HL069294 from NHLBI and NCI; a contract HHSH250201200016C with Health Resources and Services Administration (HRSA/DHHS); two grants N00014-14-1-0028 and N00014-15-1-0848 from the Office of Naval Research; and grants from Alexion; Amgen, Inc.; Anonymous donation to the Medical College of Wisconsin; Be the Match Foundation; Bristol Myers Squibb Oncology; Celgene Corporation; Chimerix, Inc.; Fred Hutchinson Cancer Research Center; Gamida Cell Ltd.; Genentech, Inc.; Genzyme Corporation; Gilead Sciences, Inc.; Health Research, Inc. Roswell Park Cancer Institute; HistoGenetics, Inc.; Incyte Corporation; Jazz Pharmaceuticals, Inc.; Jeff Gordon Children’s Foundation; The Leukemia & Lymphoma Society; The Medical College of Wisconsin; Merck & Co, Inc.; Mesoblast; Millennium: The Takeda Oncology Co.; Miltenyi Biotec, Inc.; National Marrow Donor Program; Neovii Biotech NA, Inc.; Novartis Pharmaceuticals Corporation; Onyx Pharmaceuticals; Optum Healthcare Solutions, Inc.; Otsuka America Pharmaceutical, Inc.; Otsuka Pharmaceutical Co, Ltd. – Japan; Oxford Immunotec; Perkin Elmer, Inc.; Pharmacyclics; Sanofi US; Seattle Genetics; Sigma-Tau Pharmaceuticals; Spectrum Pharmaceuticals, Inc.; St. Baldrick’s Foundation; Sunesis Pharmaceuticals, Inc.; Swedish Orphan Biovitrum, Inc.; Telomere Diagnostics, Inc.; TerumoBCT; Therakos, Inc.; University of Minnesota; and Wellpoint, Inc. The views expressed in this article do not reflect the official policy or position of the National Institute of Health, the Department of the Navy, the Department of Defense, Health Resources and Services Administration (HRSA) or any other agency of the U.S. Government.
3. Sette A, Sidney J. Nine major HLA class I supertypes account for the vast preponderance of HLA-A and -B polymorphism. Immunogenetics. 1999;50(3-4):201-212. 4. Barber LD, Gillece-Castro B, Percival L, Li X, Clayberger C, Parham P. Overlap in the repertoires of peptides bound in vivo by a group of related class I HLA-B allotypes. Curr Biol. 1995;5(2):179-190. 5. Trachtenberg E, Korber B, Sollars C, et al. Advantage of rare HLA supertype in HIV
disease progression. Nat Med. 2003;9 (7):928-935. 6. Lazaryan A, Song W, Lobashevsky E, et al. The influence of human leukocyte antigen class I alleles and their population frequencies on human immunodeficiency virus type 1 control among African Americans. Hum Immunol. 2011;72(4):312-318. 7. Lazaryan A, Song W, Lobashevsky E, et al. Human leukocyte antigen class I supertypes and HIV-1 control in African
1273
A. Lazaryan et al. Americans. J Virol. 2010;84(5):2610-2617. 8. Matsueda S, Takedatsu H, Yao A, et al. Identification of peptide vaccine candidates for prostate cancer patients with HLA-A3 supertype alleles. Clin Cancer Res. 2005;11(19 Pt 1):6933-6943. 9. Matsueda S, Takedatsu H, Sasada T, et al. New peptide vaccine candidates for epithelial cancer patients with HLA-A3 supertype alleles. J Immunother. 2007;30(3):274-281. 10. Saber W, Opie S, Rizzo JD, Zhang MJ, Horowitz MM, Schriber J. Outcomes after matched unrelated donor versus identical sibling hematopoietic cell transplantation in adults with acute myelogenous leukemia. Blood. 2012;119(17):3908-3916. 11. Yakoub-Agha I, Mesnil F, Kuentz M, et al. Allogeneic marrow stem-cell transplantation from human leukocyte antigen-identical siblings versus human leukocyte antigen-allelic-matched unrelated donors (10/10) in patients with standard-risk hematologic malignancy: a prospective study from the French Society of Bone Marrow Transplantation and Cell Therapy. J Clin Oncol. 2006;24(36):5695-5702. 12. Lee SJ, Klein J, Haagenson M, et al. Highresolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110(13):4576-4583. 13. Pidala J, Lee SJ, Ahn KW, et al. Nonpermissive HLA-DPB1 mismatch increases mortality after myeloablative unrelated allogeneic hematopoietic cell transplantation. Blood. 2014;124(16):25962606. 14. Flomenberg N, Baxter-Lowe LA, Confer D, et al. Impact of HLA class I and class II high-resolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood. 2004;104(7):19231930. 15. Lazaryan A, Weisdorf D. Optimal donor selection for allogeneic hematopoietic stem cell transplantation. In: Wingard J, Gastineau D, Leather H, Snyder EL, Szczepiorkowski ZM, eds. Hematopoietic Stem Cell Transplantation: A Handbook for Clinicians: AABB, 2015. 16. Wade JA, Hurley CK, Takemoto SK, et al. HLA mismatching within or outside of
1274
17.
18.
19.
20.
21.
22.
23.
24. 25.
26.
cross-reactive groups (CREGs) is associated with similar outcomes after unrelated hematopoietic stem cell transplantation. Blood. 2007;109(9):4064-4070. Fleischhauer K, Shaw BE, Gooley T, et al. Effect of T-cell-epitope matching at HLADPB1 in recipients of unrelated-donor haemopoietic-cell transplantation: a retrospective study. Lancet Oncol. 2012;13(4):366-374. Kawase T, Morishima Y, Matsuo K, et al. High-risk HLA allele mismatch combinations responsible for severe acute graft-versus-host disease and implication for its molecular mechanism. Blood. 2007;110(7):2235-2241. Heemskerk MB, Cornelissen JJ, Roelen DL, et al. Highly diverged MHC class I mismatches are acceptable for haematopoietic stem cell transplantation. Bone Marrow Transplant. 2007;40(3):193-200. Spellman S, Setterholm M, Maiers M, et al. Advances in the selection of HLA-compatible donors: refinements in HLA typing and matching over the first 20 years of the National Marrow Donor Program Registry. Biol Bone Marrow Transplant. 2008;14(9 Suppl):37-44. Sidney J, Peters B, Frahm N, Brander C, Sette A. HLA class I supertypes: a revised and updated classification. BMC Immunol. 2008;9:1. Doytchinova IA, Guan P, Flower DR. Identifiying human MHC supertypes using bioinformatic methods. J Immunol. 2004;172(7):4314-4323. Richardson J, Reyburn HT, Luque I, ValesGomez M, Strominger JL. Definition of polymorphic residues on killer Ig-like receptor proteins which contribute to the HLA-C binding site. Eur J Immunol. 2000;30(5):1480-1485. Doytchinova IA, Flower DR. In silico identification of supertypes for class II MHCs. J Immunol. 2005;174(11):7085-7095. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on acute GVHD grading. Bone Marrow Transplant. 1995;15(6):825-828. Barker JN, Weisdorf DJ, Wagner JE. Creation of a double chimera after the transplantation of umbilical-cord blood from two partially matched unrelated donors. N Engl J Med. 2001;344(24):1870-1871.
27. Kaplan EL, Meier P. Nonparametric-estimation from incomplete observations. J Am Stat Assoc. 1958;53(282):457-481. 28. Lin DY. Non-parametric inference for cumulative incidence functions in competing risks studies. Stat Med. 1997;16(8):901910. 29. https://www.ebi.ac.uk/ipd/imgt/hla/ stats.html. IMGT HLA database. (HLA-B = 3,131 proteins, 4,242 alleles; HLA-A = 2,396 proteins, 3,399 alleles; HLA-C = 2,089 proteins, 2,950 alleles; HLA-DRB1 = 1,369 proteins, 1,883 alleles; Accessed on 04/15/2016). 30. Kuniholm MH, Anastos K, Kovacs A, et al. Relation of HLA class I and II supertypes with spontaneous clearance of hepatitis C virus. Genes Immun. 2013;14(5):330-335. 31. Vejbaesya S, Thongpradit R, Kalayanarooj S, et al. HLA class I supertype associations with clinical outcome of secondary dengue virus infections in ethnic Thais. J Infect Dis. 2015;212(6):939-947. 32. Hertz T, Nolan D, James I, et al. Mapping the landscape of host-pathogen coevolution: HLA class I binding and its relationship with evolutionary conservation in human and viral proteins. J Virol. 2011;85(3):1310-1321. 33. Greenbaum J, Sidney J, Chung J, Brander C, Peters B, Sette A. Functional classification of class II human leukocyte antigen (HLA) molecules reveals seven different supertypes and a surprising degree of repertoire sharing across supertypes. Immunogenetics. 2011;63(6):325-335. 34. Harjanto S, Ng LF, Tong JC. Clustering HLA class I superfamilies using structural interaction patterns. PloS one. 2014;9(1):e86655. 35. Hertz T, Yanover C. Identifying HLA supertypes by learning distance functions. Bioinformatics. 2007;23(2):e148-155. 36. Lund O, Nielsen M, Kesmir C, et al. Definition of supertypes for HLA molecules using clustering of specificity matrices. Immunogenetics. 2004;55(12):797-810. 37. Reche PA, Reinherz EL. Definition of MHC supertypes through clustering of MHC peptide-binding repertoires. Methods Mol Biol. 2007;409:163-173. 38. Anasetti C, Logan BR, Lee SJ, et al. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):1487-1496.
haematologica | 2016; 101(10)