Haematologica, Volume 102, issue 11 by Haematologica

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), Ziggy Kennell (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 (Utrecht); 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 2017 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 Ottobre 2017.

haematologica calendar of events

Journal of the European Hematology Association Published by the Ferrata Storti Foundation

Turkish Society of Hematology - EHA Joint Symposium November 1 - 4, 2017 Antalya, Turkey 28th Congress of the Hellenic Society of Haematology Hellenic Society of Haematology Chairs: P Panayotidis, E Terpos November 2-4, 2017 Athina, Greece International 6th ESLHO Symposium: New developments in MRD diagnostics European Scientific foundation for Laboratory Hemato Oncology (ESLHO) Chairs: M Brüggemann, J Trka, O Ottmann, K Döhner, B Durie, A Orfao, C Pott, M Ladetto, T Szczepanski November 9-10, 2017 Leiden, The Netherlands Transfusion-Transmitted Infectious Diseases and Blood Safety European School of Transfusion Medicine (ESTM) Chairs: S Sauleda, M. Schmidt. November 10-12, 2017 Barcelona, Spain

EHA-SWG Scientific Meeting on Shaping the Future of Mesenchymal Stromal Cells Therapy Chairs: W Fibbe, F Dazzi November 23-25, 2017 Amsterdam, The Netherlands EHA-SWG Scientific Meeting on Integrated Diagnosis Strategies in Oncohematology for the management of cytopenias and leukocytosis Chair: MC Béné February 8-10, 2018 Barcelona, Spain EuroClonality Workshop: Clonality assessment in Pathology European Scientific foundation for Laboratory Hemato Oncology (ESLHO) Chairs: PJTA Groenen, F Fend, AW Langerak February 19-21, 2018 Nijmegen, The Netherlands EHA-SWG Scientific Meeting on New Molecular Insights and Innovative Management Approaches for Acute Lymphoblastic Leukemia Chair: N Gökbuget April 12-14, 2018 Location TBC

14th International Conference on Thalassaemia and other Haemoglobinopathies & 16th TIF Conference for Patients and Parents Thalassaemia International Federation (TIF) Chairs: D Loukopoulos, A Taher, J Porter, MD Cappellini November 17-19, 2017 Menemeni, Greece Argentinian Society of Hematology - EHA Joint Education Day November 17-18, 2017 Mar del Plata, Argentina

Calendar of Events updated on October 2, 2017

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation

Table of Contents Volume 102, Issue 11: November 2017 Cover Figure Image generated by www.somersault1824.com.

Editorial 1807

Understanding the extracellular matrix in acute myeloid leukemia Valerio Izzi et al.

Guideline Article 1810

Haploidentical hematopoietic cell transplantation for adult acute myeloid leukemia: a position statement from the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplantation Catherine J Lee et al.

Articles Red Cell Biology & Its Disorders

1823

Clinical risks and healthcare utilization of hematopoietic cell transplantation for sickle cell disease in the USA using merged databases Staci D. Arnold et al.

Coagulation & Its Disorders

1833

The ADAMTS131239-1253 peptide is a dominant HLA-DR1-restricted CD4+ T-cell epitope Laurent Gilardin et al.

Chronic Myeloid Leukemia

1842

Treatment outcome in a population-based, ‘real-world’ cohort of patients with chronic myeloid leukemia Inge G.P. Geelen et al.

Acute Myeloid Leukemia

1850

Senescence is a Spi1-induced anti-proliferative mechanism in primary hematopoietic cells Laure Delestré et al.

1861

Bortezomib as a new therapeutic approach for blastic plasmacytoid dendritic cell neoplasm Laure Philippe et al.

Acute Lymphoblastic Leukemia

1869

Antigen receptor sequencing of paired bone marrow samples shows homogeneous distribution of acute lymphoblastic leukemia subclones Prisca M.J. Theunissen et al.

Chronic Lymphocytic Leukemia

1878

Targeting metabolism and survival in chronic lymphocytic leukemia and Richter syndrome cells by a novel NF-kB inhibitor Tiziana Vaisitti et al.

1890

SYK inhibition thwarts the BAFF - B-cell receptor crosstalk and thereby antagonizes Mcl-1 in chronic lymphocytic leukemia Cody Paiva et al.

Haematologica 2017; vol. 102 no. 11 - November 2017 http://www.haematologica.org/

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation 1901

Toll-like receptor 9 stimulation can induce IkBÎś expression and IgM secretion in chronic lymphocytic leukemia cells Eleonora Fonte et al.

Non-Hodgkin Lymphoma

1913

Efficacy and safety of subcutaneous and intravenous rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone in first-line diffuse large B-cell lymphoma: the randomized MabEase study Pieternella Lugtenburg et al.

1923

CUDC-907 in relapsed/refractory diffuse large B-cell lymphoma, including patients with MYC-alterations: results from an expanded phase I trial Yasuhiro Oki et al.

1931

Italian real-life experience with brentuximab vedotin: results of a large observational study of 40 cases of relapsed/refractory systemic anaplastic large cell lymphoma Alessandro Broccoli et al.

Cell Therapy & Immunotherapy

1936

CD56bright natural killer regulatory cells in filgrastim primed donor blood or marrow products regulate chronic graft-versus-host disease: the Canadian Blood and Marrow Transplant Group randomized 0601 study results Amina Kariminia et al.

1947

Human leukocyte antigen-E mismatch is associated with better hematopoietic stem cell transplantation outcome in acute leukemia patients Chrysanthi Tsamadou et al.

Complications in Hematology

1956

HIF1A is a critical downstream mediator for hemophagocytic lymphohistiocytosis Rui Huang et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/102/11.toc

e427

Erythropoietin stimulates murine and human fibroblast growth factor-23, revealing novel roles for bone and bone marrow Erica L. Clinkenbeard et al. http://www.haematologica.org/content/102/11/e427

e431

Increase of von Willebrand factor with aging in type 1 von Willebrand disease: fact or fiction? Mariachiara Borghi et al. http://www.haematologica.org/content/102/11/e431

e434

MLL-TET1 fusion protein promotes immortalization of myeloid progenitor cells and leukemia development Hyeng-Soo Kim et al. http://www.haematologica.org/content/102/11/e434

e438

The role of constitutive activation of FMS-related tyrosine kinase-3 and NRas/KRas mutational status in infants with KMT2A-rearranged acute lymphoblastic leukemia Henning Fedders et al. http://www.haematologica.org/content/102/11/e438

Haematologica 2017; vol. 102 no. 11 - November 2017 http://www.haematologica.org/

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation

e443

Mutational status of IGHV is the most reliable prognostic marker in trisomy 12 chronic lymphocytic leukemia Pietro Bulian et al. http://www.haematologica.org/content/102/11/e443

e447

The Bruton tyrosine kinase inhibitor CC-292 shows activity in mantle cell lymphoma and synergizes with lenalidomide and NIK inhibitors depending on nuclear factor-kB mutational status Anna Vidal-Crespo et al. http://www.haematologica.org/content/102/11/e447

e452

CXCL13 levels are elevated in patients with Waldenstrรถm macroglobulinemia, and are predictive of major response to ibrutinib Josephine M.Vos et al. http://www.haematologica.org/content/102/11/e452

e456

Circulating microRNA expressions can predict the outcome of lenalidomide plus low-dose dexamethasone treatment in patients with refractory/relapsed multiple myeloma Seung-Hyun Jung et al. http://www.haematologica.org/content/102/11/e456

e460

Mass spectrometry-based identification of a naturally presented receptor tyrosine kinase-like orphan receptor 1-derived epitope recognized by CD8+ cytotoxic T cells Falk Heidenreich et al. http://www.haematologica.org/content/102/11/e460

Comment This comment is available online only at www.haematologica.org/content/102/11.toc

e465

Late effects of blood and marrow transplantation Rahul Banerjee et al. http://www.haematologica.org/content/102/11/e465

Case Reports Case Reports are available online only at www.haematologica.org/content/102/11.toc

e466

Pseudo-monoclonal gammopathy: a report of four cases Majd D. Jawad et al. http://www.haematologica.org/content/102/11/e466

e470

ARID1A mutation in blastic plasmacytoid dendritic cell neoplasm Lei Wang et al. http://www.haematologica.org/content/102/11/e470

Haematologica 2017; vol. 102 no. 11 - November 2017 http://www.haematologica.org/

EDITORIALS Understanding the extracellular matrix in acute myeloid leukemia Valerio Izzi,1 Ritva Heljasvaara1,2 and Taina Pihlajaniemi1 1

Centre of Excellence in Cell-Extracellular Matrix Research and Biocenter Oulu, Faculty of Biochemistry and Molecular Medicine, University of Oulu, Finland and 2Centre for Cancer Biomarkers (CCBIO), Department of Biomedicine, University of Bergen, Norway E-mail: valerio.izzi@oulu.fi doi:10.3324/haematol.2017.174847

espite the continuous progress in therapy, acute myeloid leukemia (AML) remains the leading cause of death among hematologic neoplasms.1 Understanding how leukemic clones arise and how they interact with the microenvironment of the bone marrow, especially within the hematopoietic stem cell (HSC) niches, is thus necessary to provide the means to early diagnosis and intervention. In this context, much has been published about the influence of niche stromal cell-derived extracellular matrix (ECM) on leukemic stem cells (LSCs, the apex of the neoplastic progeny which is AML itself),2,3 but considerably less is known about whether LSCs directly produce their 'own' ECM to hijack niche functions and gain a growth advantage over HSCs. Hence, we investigated the expression of ECM genes in both LSCs (and, more generally, leukemic precursor cells) and mature AML cells as compared to normal HSCs and white blood cells from healthy donors. We found that a common set of 80 ECM genes (either up- or down-regulated with respect to healthy donors) characterizes all leukemic cells, independently from specific cytogenetic alterations or mutations in driver genes, and constitutes the ECM signature of the whole leukemic process from the LSCs to the final circulating blasts4 (Figure 1). Our results, together with those of a few others, suggest a much more active role for LSCs in the production of a modified ECM in the niche, one that favors disease progression. These data will help us to decipher the relationships between AML and the microenvironment, and will hopefully result in new diagnostic and therapeutic possibilities in the future. In 1889, Stephen Paget proposed the visionary â&#x20AC;&#x153;seed and soilâ&#x20AC;? theory, a conjecture which acknowledged the importance of a permissive microenvironment (the bad soil) in the growth of neoplastic cells (the bad seeds).5 Approximately 100 years later, the works by Schofield and Kimble & White changed the prevailing vision of stem cells (SCs), showing for the first time that elements other than SCs control and influence the growth and fate of these cells.6 Today we define the stem cell niches as specialized anatomical regions in which a complex network of stromal cells and ECM interact with each other and with the SCs as a 'dynamic duo',7 and we know that SCs and cancer are two sides of the same coin, with the normal SCs and their niches on one side and the cancer SCs (CSCs) and their niches on the other.2,3 The ECM provides several types of microenvironmental cues that sustain SCs. Due to its ability to seize and present cytokines and growth factors, provide an adhesive substrate for the cells, and to generate and integrate the mechanical signals needed to control cell proliferation and differentiation, the ECM is, in fact, a critical determinant of the properties and behavior of the cells embedded into it.8 HSCs and haematologica | 2017; 102(11)

their niches are no exceptions to this rule, and a substantial amount of data now point to the importance of sensing and correctly interacting with the proper ECM in the bone marrow for the HSCs to lodge in their niches and accomplish their functions.9 It is all the more surprising, then, to notice that, even though the very definition of CSCs came from an AML study,10 research on LSCs has taken a rather unidirectional road. While findings on genetic and epigenetic networks controlling LSCs have accumulated rapidly, and the importance of stromal elements and altered ECM in the growth and dissemination of leukemic cells is nowadays well accounted for,2,3,5 there seems to be a paucity of results to explain how far the leukemic cells themselves contribute to the changes occurring in the niche. For example, it is widely accepted that the expression of the major cellular ECM receptors, the integrins, on LSC allows fibronectin sensing, lodging these cells to the bone marrow niche and triggering pro-survival signal cascades that ultimately lead to post-therapy persistence of the leukemic clones.11 Likewise, the expression of another cellular receptor, CD44, on LSC and AML cells allows the interaction of these cells with hyaluronan, osteopontin, fibronectin and selectins, conferring a superior engrafting ability to the malignant clones and a crucial survival mechanism.3,11 But what about the direct interference of LSCs with the niche ECM? Considering how important a 'permissive' ECM is for LSCs, we speculated that transcriptional programs enabling leukemic cells to directly alter the composition of the surrounding ECM would give them a growth advantage as they could quickly shape and hijack the microenvironment even before establishing a detrimental co-operation with stromal cells. Following this idea, we investigated the expression of ECM genes in LSC and HSC and, in parallel, in AML cells versus normal white cells. We found a core set of 80 ECM genes that were differentially regulated in leukemic cells with respect to their normal counterparts.4 Notably, our results not only largely recapitulated previous non-systematic findings, but also gave them a wider context. Thus, in cells with high CD44 expression, we observed a significant upregulation of ECM proteins which would directly interact with CD44 itself, including structural substrates, such as collagen IV and XVIII and laminin beta 2 (COL4A5, COL18A1 and LAMB2, respectively), and proteinases implicated in ECM assembly, remodeling and growth factor activation, such as matrix metalloproteinase 2, ADAM metallopeptidase domain 17, bone morphogenetic protein 1 and cathepsin G (MMP2, ADAM17, BMP1 and CTSG, respectively).4 Building on these data, we found that ECM gene expression in leukemic cells is not univocal and that the differential 1807

Editorials

Figure 1. The extracellular matrix (ECM) signature of leukemic stem cells (LSCs) and acute myeloid leukemia (AML). Compared to normal hematopoietic stem cells (HSC), committed precursors or differentiated blood cells, the expression of ECM genes in LSC, leukemia precursor cells (LPC) or AML blasts differs significantly, and spans all the major categories of ECM components, enzymes and secreted factors represented in the Matrisome DB database. Note that some ECM genes are not present in the Matrisome DB and are reported as “Not available”.

expression of a restricted set of these genes distinguishes two subtypes of leukemic precursor cells not previously reported: one subtype, clustering with normal HSCs with respect to the overall ECM expression profile (we named them “early leukemic” group), and another with a clearly distinguishable profile (we termed these cells “definitive leukemic”). When these two profiles were used to classify AML patients, we found that individuals with the “definitive leukemic” profile had much lower overall, diseasefree and event-free survival independently of karyotype aberrations or typical driver mutations, for example, in FLT3, NPM1 or IDH1 genes.4 Recently, Foroushani et al. provided another important confirmation of the view that AML actively creates its own ECM. They employed a sophisticated network analysis to identify the architecture of active gene regulatory networks in AML and found that the regulation of ECM modules is the most significant feature of AML transcriptional profile.12 Of particular interest is the 1808

downregulation of matrix metalloproteinase 9 (MMP9) observed in both Foroushani et al.’s and our studies; a finding in contrast with the general idea that MMP9 is highly expressed in AML.13 Yet, our systematic approach suggests a possible explanation to this. In fact, we found that MMP9 belongs to a specific subnetwork of ECM elements interacting with CD44, and that the expression of CD44 characterizes patients with lower survival.4 If we then consider that a previous report suggested that MMP9 levels correlate inversely with patients’ risk of death,14 we can easily envisage a regulatory mechanism that, in this CD44 subnetwork, transcribes genes that facilitate the spreading of AML while suppressing genes (such as MMP9) that would halt it. Altogether, these findings lay the foundation for a more systematic analysis of the direct ECM-modifying activities of LSC and AML cells, and support a much more active role for these cells in the regulation of the niche ECM than had been previously supposed. haematologica | 2017; 102(11)

Editorials

References 1. Adult Acute Myeloid Leukemia Treatment (PDQÂŽ)â&#x20AC;&#x201C;Health Professional Version 2017 [updated 20.1.2017; cited 15.6.2017]. Available from: https://www.cancer.gov/types/leukemia/hp/adultaml-treatment-pdq. 2. Plaks V, Kong N, Werb Z. The cancer stem cell niche: how essential is the niche in regulating stemness of tumor cells? Cell Stem Cell. 2015;16(3):225-238. 3. Schepers K, Campbell TB, Passegue E. Normal and leukemic stem cell niches: insights and therapeutic opportunities. Cell Stem Cell. 2015;16(3):254-267. 4. Izzi V, Lakkala J, Devarajan R, et al. An extracellular matrix signature in leukemia precursor cells and acute myeloid leukemia. Haematologica. 2017;102(7):e245-e248. 5. Langley RR, Fidler IJ. The seed and soil hypothesis revisited--the role of tumor-stroma interactions in metastasis to different organs. Int J Cancer. 2011;128(11):2527-2535. 6. Albert Hubbard EJ. The C. elegans germ line: a model for stem cell biology. Dev Dyn. 2007;236(12):3343-3357. 7. Voog J, Jones DL. Stem cells and the niche: a dynamic duo. Cell Stem Cell. 2010;6(2):103-115.

haematologica | 2017; 102(11)

8. Ahmed M, Ffrench-Constant C. Extracellular Matrix Regulation of Stem Cell Behavior. Curr Stem Cell Rep. 2016;2:197-206. 9. Krause DS. Regulation of hematopoietic stem cell fate. Oncogene. 2002;21(21):3262-3269. 10. Lapidot T, Sirard C, Vormoor J, et al. A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature. 1994;367(6464):645-648. 11. Konopleva MY, Jordan CT. Leukemia stem cells and microenvironment: biology and therapeutic targeting. J Clin Oncol. 2011;29(5):591-599. 12. Foroushani A, Agrahari R, Docking R, et al. Large-scale gene network analysis reveals the significance of extracellular matrix pathway and homeobox genes in acute myeloid leukemia: an introduction to the Pigengene package and its applications. BMC Med Genomics. 2017;10(1):16. 13. Travaglino E, Benatti C, Malcovati L, et al. Biological and clinical relevance of matrix metalloproteinases 2 and 9 in acute myeloid leukaemias and myelodysplastic syndromes. Eur J Haematol. 2008;80(3):216-226. 14. Schmohl J, Santovito D, Guenther T, et al. Expression of surface-associated 82kDa-proMMP-9 in primary acute leukemia blast cells inversely correlates with patients' risk. Exp Hematol. 2016;44(5):358-362.e5.

1809

GUIDELINE ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1810-1822

Catherine J Lee,1 Bipin N Savani,2 Mohamad Mohty,3 Myriam Labopin,3 Annalisa Ruggeri,3 Christoph Schmid,4 Frédéric Baron,5 Jordi Esteve,6 Norbert C Gorin,7 Sebastian Giebel,8 Fabio Ciceri9 and Arnon Nagler3,10

Utah Blood and Marrow Transplant Program, Huntsman Cancer Institute, University of Utah, Salt Lake City, UT, USA; 2Vanderbilt University Medical Center, Nashville, TN, USA; 3 Department of Hematology, Saint-Antoine Hospital, INSERM, Paris, France; 4Klinikum Augsburg, Department of Hematology and Oncology, University of Munich, Augsburg, Germany; 5Department of Medicine, Division of Hematology, University of Liège, Belgium; 6Department of Hematology, Hospital Clinic, IDIBAPS, Barcelona, Spain; 7 Department of Hematology, Saint-Antoine Hospital, APHP and University UPMC, Paris, France; 8Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Poland; 9Hematology, IRCCS San Raffaele Scientific Institute, Milan, Italy and 10Hematology Division, Chaim Sheba Medical Center, Tel Hashomer, Israel 1

ABSTRACT

Correspondence: bipin.savani@vanderbilt.edu

Received: July 9, 2017. Accepted: September 5, 2017. Pre-published: September 7, 2017. doi:10.3324/haematol.2017.176107 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1810 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1810

llogeneic blood or marrow hematopoietic cell transplantation continues to be the most potent anti-leukemic treatment for adult patients with standard, high-risk, or chemo-refractory acute myeloid leukemia. Until recently, this procedure was generally limited to those recipients who had an available matched-sibling donor or matched-unrelated donor. Technical advances in graft cell processing and manipulation, control of bidirectional T cell alloreactivity, graft-versus-host disease prophylaxis, and other supportive measures in haploidentical transplantation now enable nearly all patients with acute myeloid leukemia to benefit from the graft-versus-leukemia effect with substantial reduction in procedure-related mortality. Over recent years, haploidentical donors have been increasingly adopted as a valid donor source in allogeneic hematopoietic cell transplantation for acute myeloid leukemia in the absence of an HLA-matched donor. Among centers of the European Society for Blood and Marrow Transplantation, the use of haploidentical related donor transplantation has increased by 250% since 2010, and 291% since 2005. On behalf of the Acute Leukemia Working Party of the European Society for Blood and Marrow Transplantation, we summarize recent utilization trends in haploidentical transplantation for acute myeloid leukemia and describe the transformative changes in haploidentical hematopoietic cell transplantation techniques over the past decade, which have led to the current widespread use of this procedure. Furthermore, we review the efficacy of haploidentical hematopoietic cell transplantation for acute myeloid leukemia from available studies, including preliminary comparative studies, and bring attention to remaining unanswered questions and directions for future research. We conclude this report with our recommendations for the role of haploidentical hematopoietic cell transplantation in acute myeloid leukemia.

haematologica | 2017; 102(11)

Haploidentical HCT in AML

Introduction Allogeneic hematopoietic cell transplantation (HCT) is a potentially curative therapy for acute myeloid leukemia (AML), with 3- to 5-year overall survival (OS) rates ranging from 23% to 88%1-3 depending on the AML risk profile, stage, and the presence or absence of minimal residual disease (MRD). AML continues to be the primary indication for allogeneic HCT, and the number of these procedures performed for AML among the European Society for Blood and Marrow Transplantation (EBMT) centers has steadily increased over the past decade, with the most recent report showing a record number of 6,189 allogeneic HCT performed in 2015 compared to 3,946 in 2010.4,5 In addition to the development of reduced-intensity (RI) conditioning regimens, thereby extending the use of allogeneic HCT for AML patients above the age of 60 or for those with co-morbidities, the significant growth of allogeneic HCT for AML is a result of the increased availability of alternative donors, particularly haploidentical family donors. While a HLA-matched sibling donor (MSD) remains the preferred donor choice for optimal transplant outcomes, in reality, approximately only 30% of patients from Western countries have such a donor, therefore 70% of patients require an unrelated donor source.6,7 Interestingly, a recent analysis using population data from the USA has challenged this well-accepted sibling match probability and describes a variation in rates ranging between 13% to 51%.8 Perhaps more alarming is the finding of the effect of a 40-year decline in USA birth rates on decreased availability of a MSD for transplant-eligible patients. As such, the current generation of young adults (18 to 44 years) will be 1.5 times less likely to find a MSD during the peak need for HCT (at around 61 years of age) compared to their current adult counterparts (aged 45 to 64 years).8 It is expected that a similar evolution in MSD accessibility is occurring in Western Europe as total fertility rates remain low.9 These changes highlight the upcoming demand for and utilization of alternative donor sources. Alternative donor options include HLA-matched unrelated donors (MUD), partially HLA-mismatched unrelated donors (MMUD), single or double umbilical cord blood units (sUCB or dUCB), and haploidentical (haplo) family donors. While MUD have traditionally been considered to be the next preferred donor following a MSD, the success of finding an 8/8 HLA MUD depends on race. While Caucasians have an approximately 75% likelihood of finding an 8/8 MUD, the probability falls to less than 20% for patients of African descent or other ethnic minorities.7,10 Furthermore, differences in laws for donor selection and recruitment among different countries limit or delay the acquisition of a MUD.10 The use of an unrelated donor or UCB product with a mismatch at one or two HLA loci expands the accessibility of HCT to the vast majority of patients, however, this is at the cost of an increased risk of poor transplant outcomes and/or increased expense, particularly with the use of UCB cells. Over recent years, haploidentical donors have been increasingly adopted as a valid source of donor cells for allogeneic HCT of AML in the absence of an HLAmatched donor.4,11 A haploidentical related donor is defined by the sharing of one haplotype (or a single identical copy of chromosome 6) with the patient containing the HLA region involving class I and class II histocompathaematologica | 2017; 102(11)

ibility genes. However, a haploidentical family donor may be greater than half-matched and have common alleles on the unshared haplotype (mismatched related donor). The most recent EBMT activity survey report described haploidentical donors as a family member with two or more loci mismatch within the loci HLA-A, -B, -C, -DRB1 andDQB1.4 Among centers of the EBMT, the use of haploidentical transplantation (haploHCT) for malignant and non-malignant disorders has surged by 250% since the year 2010, and by 291% since 2005. In 2010, 802 haploHCT were performed, and this number increased to 1,571 in 2013, followed by 2,012 haploHCT in 2015.4,11 The highest utilization for haploHCT in 2015 was seen in myeloid malignancies (n=1,008), and the majority of these patients had a diagnosis of AML (n=735), with an equal proportion of patients in first complete remission (CR1) or more advanced disease (Figure 1). In contrast, the utilization of unrelated umbilical cord blood transplantation (UCBT) has sharply declined for myeloid and lymphoid malignancies.4 This apparent preference for haploidentical donors is a result of improvements in conditioning regimens combined with new strategies to diminish the risk of graft-versus-host disease (GvHD) associated with one haplotype mismatched donors that have resulted in favorable clinical outcomes comparable to HLA-matched allogeneic HCT, compounded with the nearly universal and immediate availability of the donor and ease of recurrent stem cell collections for repeat cellular infusions. The ability to have rapid access to a haploidentical donor is a crucial benefit for patients with high-risk AML, as a delay in transplantation due to donor issues can result in poor outcomes. On behalf of the Acute Leukemia Working Party (ALWP) of the EBMT, herein we aim to first describe the early strategies used in haploidentical transplantation and the pivotal developments that have made its use universal and available to nearly all patients requiring hematopoietic cell transplantation. We then summarize the evidence from available studies, evaluating its efficacy in AML, including preliminary non-comparative and comparative studies of haploHCT with other alternative donor transplants, and lastly, discuss future directions for research.

Early experiences in haploidentical transplantation Initial experiences with unmodified bone marrow (BM) haploidentical HCT in acute leukemias generated poor outcomes as a consequence of intense bi-directional T cell alloreactivity associated with HLA-mismatches. Limited success was primarily related to delayed engraftment, graft failure, and acute graft-versus-host disease (aGvHD).12-15 In order to overcome these challenges, several alternative strategies were developed. In 1993, investigators from the University of Perugia pioneered a strategy of T cell-depletion (TCD) with ex vivo CD34+ cell selection and in vivo antithymocyte globulins (ATG) administration as sole prophylaxis for GvHD, accompanied by infusion of a large number of CD34+ cells following intensive myeloablative and immunosuppressive conditioning, with the rationale that this strategy would help promote engraftment and decrease graft failure (Figure 2A). â&#x20AC;&#x153;Mega-dosesâ&#x20AC;? of stem cells were obtained by combining TCD BM with granulocyte colony-stimulating factor (G-CSF) mobilized peripher1811

C.J. Lee et al.

al blood stem cells (PBSC).16-18 With additional modifications, 95% of patients with acute leukemia (AL) achieved primary engraftment, and aGvHD and chronic GvHD (cGvHD) were minimal. With more than 15 years of follow-up, the relapse incidence (RI) was 17% in patients with AML who were transplanted in any complete remission (CR), while the 17-year disease-free survival (DFS) rate was 43%.19,20 In addition to a highly myeloablative regimen, the emergence of natural killer (NK) cell alloreactions following transplantation may explain the decreased incidence of relapse and improved survival.21-24 Despite the success of the anti-leukemic effects of this strategy, TCD haploHCT was associated with high transplant-related mortality (TRM) of up to 40% mainly due to a delay in immune recovery and life-threatening infections.18,19 Findings from an EBMT retrospective analysis of 173 adults with de novo AL who received a TCD haploidentical HCT in Europe showed similar outcomes, including high engraftment rate, negligible GvHD, and high TRM.25 To circumvent the pitfalls associated with TCD haploHCT, other forms of T-cell cellular therapy were exploited, including selective T-cell-depletion,26,27 adoptive transfer of donor T cells following transplant,28 selective T-cell addbacks29-32 and gene-modified donor T cells.33

Post-transplant cyclophosphamide: a pivotal point in haploidentical transplantation The rationale behind the use of post-transplantation cyclophosphamide (PTCy) stems from early preclinical studies demonstrating its role in targeting alloreactive T cells and reducing GvHD when given within a narrow window following allografting.34-39 Furthermore, the finding of preserved hematopoietic stem and progenitor cells (and in later work, regulatory T cells40) when exposed to cyclophosphamide owing to the high expression of aldehyde dehydrogenase,41,42 gave rise to the first-in-human clinical trial at Johns Hopkins Hospital in 1999. Thirteen patients with high-risk hematologic malignancies underwent T cell-replete (TCR) haploidentical bone marrow transplantation (haploBMT) using a non-myeloablative (NMA) conditioning regimen consisting of fludarabine and low-dose total body irradiation (TBI), as well as 50 mg/kg of cyclophosphamide on day + 3 post-transplant, followed by tacrolimus and mycophenolic mofetil (MMF) on day + 4 for GvHD prophylaxis. Owing to a high rate of graft rejection (2 out of the first 3 patients), cyclophosphamide 14.5 mg/kg was introduced into the conditioning regimen. This adaptation resulted in an 80% primary

Figure 1. Trends in haploidentical HCT in Europe between 1990-2015. (A) Increasing use to haploidentical family HCT from cord blood HCT. (B) Increasing use of haploidentical HCT by main disease group. (C) Similar increase in rates of haploidentical HCT for AML early disease and AML advanced disease. (D) Haploidentical HCT by cell source; bone marrow (BM) versus peripheral blood (PB). Adapted from Passweg et al.4 used under theCreative Commons License. AML: acute myeloid leukemia; HSCT: hematopoietic stem cell transplantation; NMD: non-malignant disorders.

1812

haematologica | 2017; 102(11)

Haploidentical HCT in AML

engraftment rate (8 out of 10 patients), giving proof of concept to move into next phase studies.43 As the initial phase I study ultimately had a high cumulative incidence of graft failure and severe GvHD at 6 months post-transplant, Luznik et al.44 modified the regimen by adding an additional dose of cyclophosphamide 50mg/kg on day + 4 post-transplant (Figure 2B). In a collaborative phase 2 trial between Hopkins and Seattle, 68 patients with AML (n= 27) received the revised regimen, and results yielded an 87% engraftment rate, one-year non-relapse mortality (NRM) and relapse of 15% and 51%, respectively, and two-year OS and event-free survival (EFS) of 36% and 26%, respectively. Additionally, the cumulative incidences (CI) of grades II-IV and grade

III-IV aGvHD by day 200 were 34% and 6%, respectively. A trend towards a lower incidence of extensive cGvHD with the use of 2 doses of PTCy as compared to one dose was noted (5% vs. 25%, P=0.05). In an updated analysis of 210 recipients of NMA haploBMT, the Hopkins group reported similar outcomes.45 Due to the early reports of success with unmanipulated haploidentical HCT and pioneering of PTCy for prevention of GvHD, other centers, mainly in Western Europe and the USA, have favored the use of TCR grafts over TCD haploHCT.10,46,47 Ciurea et al.46 reported significantly improved 1-year NRM (16% vs. 42%, P=0.02), OS (64% vs. 30%, P=0.02) and progression-free survival (PFS) (50% vs. 21%, P=0.02) in 65 consecutive patients treated with a myeloablative TCR haploBMT with PTCy (n=32), compared to a TCD PBSC graft with ATG followed by infusion of CD34+ selected cells and no other post-transplantation immunosuppression (n=33). Engraftment rate and grade II-IV aGvHD were not significantly different, whereas cGvHD was significantly lower in patients treated with a TCR graft. In conclusion, given the ease of donor acquisition and administration of PTCy-based protocols alongside the favorable results seen in patients with high-risk hematologic malignancies, more investigation into the role of haploHCT in the early steps of decisional algorithms for the treatment of acute leukemias is ongoing.

Comparative donor studies of haploidentical transplantion for acute myeloid leukemia

Haploidentical versus matched sibling or unrelated donor transplantation

Figure 2. Commonly used platforms used in haploidentical-related transplantation.111 (A) University of Perugia: myeloablative conditioning and T cell-depletion with â&#x20AC;&#x153;megadoseâ&#x20AC;? CD34+ cell allografts. (B) Johns Hopkins: non-myeloablative conditioning with high-dose, post-transplantation cyclophosphamide. (C) Peking University: myeloablative conditioning and in vivo T cell modulation (GIAC protocol). Panel B was adapted from Luznik et al.44 used under the Creative Commons License. Ara-C: cytarabine; ATG: anti-thymocyte globulin; BM: bone marrow; Bu: busulfan; CSA: ciclosporin-A; Cy: cyclophosphamide; Flu: fludarabine; GCSF: granulocyte colony-stimulating factor; M-CCNU: semustine; MMF: mycophenolate mofetil; MTX: methotrexate; PBSC: peripheral-blood stem cell; TBI: total body irradiation.

haematologica | 2017; 102(11)

At present there are no prospective randomized comparisons of transplantations using a haploidentical donor versus a MSD or MUD for AML. Based on several non-randomized comparative studies evaluating transplantation outcomes following haploidentical transplantation with post-transplant cyclophosphamide or other in vivo T celldepletion methods,10,48-57 the combined data suggest similar outcomes for haploHCT compared with MSD and MUD HCT. Table 1 summarizes the available comparative studies of haploHCT with PTCy platform versus MSD or MUD HCT. In one of the earliest studies of haploHCT with PTCy, Bashey et al.49 demonstrated equivalent primary outcomes of 271 patients with a variety of hematologic malignancies (~ 34% AML), who contemporaneously underwent a T cell-replete haploidentical MSD or MUD transplant. However, post-relapse survival at 12 months was unexpectedly lower compared to a well-matched MSD or MUD HCT (17% vs. 67% vs. 63%, P<0.001). In an updated cohort of 475 patients (~ 36% AML) and median follow-up of 45 months, these investigators again reported non-significant differences between haplo, MSD, and MUD transplants in DFS (54% vs. 56% vs. 50%), OS (57% vs. 72% vs. 59%), CI of NRM (17% vs. 14% vs. 16%), and relapse (29% vs. 30% vs. 34%) at 2 years after transplantation. The CIs of grade II-IV aGvHD were not significantly different between haplo and MUD HCT, however, haploHCT was associated with a significantly higher incidence of aGvHD compared to MSD (P=0.005 for grade II-IV). The 2-year CI of moderate-severe cGvHD was also significantly lower in haploHCT than in MSD or MUD HCT recipients (31% vs. 44% vs. 47%, P=<0.05), 1813

C.J. Lee et al.

and showed a similar trend for patients receiving a PBSC graft only.50 In another contemporaneously treated cohort of 227 patients with AML/myelodysplastic syndrome (MDS), Di Stasi et al.51 reported superimposable survival curves between haplo and 10/10 HLA MUD HCT, a nonsignificant improvement in outcomes with MSD HCT and a similar CI of GvHD across all donor groups. In the largest study carried out in the USA focusing on AML, Ciurea et al.48 utilized the Center for International Blood and Marrow Transplant Research (CIBMTR) registry data and reported comparable 3-year OS following haploidentical (n=192) and 8/8 HLA matched MUD HCT (n=1982) in patients with AML in various disease stages (CR1, CR2, and more advanced) who received either a myeloablative (MA) (45% vs. 50%, P=0.38) or RI (46% vs. 44%, P=0.71) conditioning regimen. Further subset analysis revealed no differences in 3-year NRM (14% vs. 20%, P=0.14) or relapse (44% vs. 39%, P=0.37) by donor type in the MA cohort, however, there was a significant decrease

in 3-year NRM (9% vs. 23%, P=0.0001) and increase in relapse (58% vs. 42%, P=0.006) in the RI group. In both cohorts, 3-month grade II-IV and grade III-IV aGvHD, and 3-year cGvHD were lower after haploidentical compared with MUD transplants (MA: grade II-IV aGvHD: 16% vs. 33%, P<0.0001; grade III-IV aGvHD: 7% vs. 13%, P=0.02; cGvHD: 30% vs. 53%, P<0.0001; RI: grade II-IV aGvHD: 19% vs. 28%, P=0.05; grade III-IV aGvHD: 2% vs. 11%, P<0.0001; cGvHD: 33% vs. 52%, P=0.002). In this study, the majority of recipients of haploHCT received a BM graft, whereas PBSC were predominantly utilized in MUD HCT. Owing to the limitations inherent in an observational registry study, the investigators could not assess the impact of the donor source of stem cells on clinical outcomes. To address this question, Rashidi et al.52 reported results from a single-center retrospective analysis of 140 patients who underwent a haploHCT (n=52) or MUD HCT (n=88) with PBSC. This group showed a significantly faster neutrophil and platelet recovery in the MUD

Table 1. Comparative studies of haploidentical HCT with PTCy versus matched donor HCT.

Author

Bashey et al.49

Disease, Condition Graft Donor no of pts. Intensity (%) Source (%) Type (%) (AML %) All, 271 (34)

Di Stasi et al. 51 AML/MDS, 227 (85)

MA (50) RI (50)

Engraft aGvHD (%) II-IV (d)

BM (17) Haplo (20) PB (83) Both (<1) MSD (43)

97.5

MUD (37)

Haplo (14)

MSD (38)

MUD (48)

BM (37) PB (63)

Ciurea et al.48

AML, 825 (100)

BM (19) PB (81)

Haplo (11) MUD (89)

93 96

Bashey et al.50

All, 475 (36)

MA (49) RI (51)

BM (21) PB (79)

Haplo (24)

MSD (38)

How et al.53

AML, 99

MA (72) RI (28)

BM (99) PB (1)

Haplo (24) MSD (32) UD (43)b

Rashidi et al.55

AML, 83

MA (42) RI (58)

PB (100)

Haplo (75) MD (25)c

11%

4% (2)

64% (2)

60% (2)

33% (2)

7% (2)

11% (P=0.062) 12% (P<0.05)

76% (P=NS) 67% (P=NS)

53% (P=NS) 52% (P=NS)

34% (P=NS) 34% (P=NS)

13% (P=NS) 16% (P=NS)

11% (3)

66% (3)

30% (3)

33% (1)

24% (1)

31%

56%a (P=0.646)

36%

28%

20%

72% (P=0.02) 59% (P=NS) 42 (1.5) 37 (P=0.17) 36% (2) 28% 29% (P=0.75) 53% (1) 58% (P=0.31)

90 97

Haplo (37) MUD (63)

31%

NRM (year)

44% (P=0.032) 47% (P=0.004) 10% (1.5) 9% (P=0.91) 10% (1) 10% 15% (P=0.61) 6% (1) 5% (P=0.86)

Haplo (8) MUD (92)

PB (100)

29% (d100) 0% (d100)

Relapse (year)

21% 9% (P=0.005) (P=0.054) 98 48% 18% (P=NS) (P=NS) 100 40% (d180) 25% (d180) 90 36% 25% (P=0.51) (P=0.79) 83 58% (d100) 28% (d100) 91 36% 23% 91 57% 30% (P=0.11) (P=0.74) 87 40% (d180) NR 100 19% (P=0.07)

BM (23) PB (77)

MA (44) RI (56)

8% (P=NS) 11% (P=NS)

DFS (year)

45 (3) 50 (P=0.38) 46 (3) 44 (P=0.71) 57% (2)

AML, 140

27% (P=NS) 39% (P=NS)

OS (year)

21% (P=0.125) 30% (3) 53% (P<0.0001) 34% (3) 52% (P=0.002) 31% (2)

AML, 1349 (100)

Rashidi et al.52

30% (d180) 11% (d180)

cGvHD (year)

29% 6% (P=0.709) (P=0.044) 16% (d90) 7% (d90) 33% 13% (P <0.0001) (P=0.02) 19% (d90) 2% (d90) 28% 11% (P =0.05) (P<0.0001) 41% (d180) 17% (d180)

Ciurea et al.48

MUD (38)

aGvHD III-IV (d)

27% (P=0.12) NR

54% (2) 56% (P=NS) 50% (P=NS) NR

23% 35% (P=0.75) (P=0.099) 44% (3) 14% (3) 39% 20% (P=0.37) (P=0.14) 58% (3) 9% (3) 42% 23% (P=0.006) (P=0.0001) 29% (2) 17% (2) 30% (P=NS) 34% (P=NS) 29% (1.5) 43% (P=0.08) 33% (1) 28% 48% (P=0.40) 31% (1) 26% (P=0.70)

14% (P=NS) 16% (P=NS) 27% (1.5) 27% (P=0.54) 26% (2) 42% 29% (P=0.49) 22% (1) 16% (P=0.30)

a Represents combined MSD and MUD transplant group. bRepresents MUD (n=35), partially mismatched (n=6); mismatched (n=2). cRepresents combined MSD and MUD transplant group. aGvHD, acute graft-versus-host disease; AML: acute myeloid leukemia; BM: bone marrow; cGvHD: chronic graft-versus-host-disease; d: day; DFS: disease-free survival; haplo: haploidentical; MA: myeloablative; MD: matched donor; MDS: myelodysplastic syndrome; MSD: matched sibling donor; MUD: matched unrelated donor; NRM: non-relapse mortality; NR: not reported; NS: not significant; OS: overall survival; PB: peripheral blood; RI: reduced-intensity; UD: unrelated donor; yr: year.

1814

haematologica | 2017; 102(11)

Haploidentical HCT in AML

group, but no statistically significant difference in OS, NRM, aGvHD or cGvHD at 1.5 years. Lastly, the refined disease risk index (DRI) developed by Armand and colleagues58,59 in order to help stratify outcomes based upon disease risk and stage has been used to compare the effects of the graft-versus-tumor response mediated by NMA haploBMT with PTCy against historical outcomes in the setting of HLA-matched donor HCT following RI conditioning.60 Risk-stratified disease and their associated survival outcomes appeared similar between the two groups. For example, 3-year PFS estimates in the low-, intermediate-, and high/very high-risk patient groups following NMA haploBMT with PTCy were 65%, 37%, and 22%, respectively, and 66%, 31%, and 15% in the original DRI study cohort of recipients of RI HLA-matched donor transplantation.60 The viability of T cell-replete haploidentical HCT with post-transplantation cyclophosphamide in patients with active AML has also been described. How et al.53 compared outcomes of 99 patients who received either a MSD (n=32), unrelated donor (all unrelated, n=43; MUD, n=35), or a haploidentical related (n=24) donor transplantation for active AML, defined by â&#x2030;Ľ5% blasts in the pre-transplantation BM, persistent cytogenetics, or extramedullary disease. With a median follow-up of 18 months, no statistically significant differences between MSD, unrelated donor, and haploidentical donor HCT in 1- and 2-year OS were identified (1 yr: 28% vs. 41% vs. 45%; 2 yr: 28% vs. 29% vs. 36%, P=0.75), EFS (1 yr: 27% vs. 28% vs. 39%; 2 yr: 18% vs. 22% vs. 23%, P=0.93), TRM (1 yr: 42% vs. 23% vs. 26%; 2 yr: 42% vs. 29% vs. 26%, P=0.49), or 1year relapse (28% vs. 48% vs. 33%, P=0.40). Similarly, the CI of grades III-IV aGvHD at day 100 (23% vs. 30% vs. 28%, P=0.74) and severe cGvHD at 1 year (10% vs. 15% vs. 10%, P=0.61) were comparable. Although not evaluated in a comparative donor study, RI T cell-replete haploHCT incorporating donor change and utilizing PTCy for postgrafting immunosuppression has also been successfully used for patients with AL relapsing after a first autologous or allogeneic transplantation.61 These results preliminarily support the decision to use a haploidentical related donor source in transplantation of patients with active AML or relapsed AML after first transplantation, as both of these patient populations have an urgent indication to proceed to transplantation and may have a readily available haploidentical family donor. The ALWP of the EBMT have also published results of several large multi-center comparative studies using EBMT registry data (Table 2). In the first retrospective comparative analysis of 10,679 patients with AL who received allogeneic HCT from a MSD or a haploidentical donor, Ringden et al.62 sought to determine whether a stronger graft-versus-leukemia (GvL) effect is exerted with T cell-deplete or T cell-replete haploidentical transplantation due to the presence of mismatched major HLA antigens on leukemic cells. The investigators determined no difference in the probability of relapse between recipients of haploidentical and MSD grafts. In a more recent study, Salvatore et al.56 compared outcomes of T cell-replete haploHCT (n=185) to those from MSD HCT (n=2,469) among 2,654 adults with intermediate-/high-risk AML in first CR. GvHD prophylaxis consisted of PTCy in 74% of patients and ATG in 26%. In multivariate analyses of patients with intermediate-risk AML, haploHCT was associated with reduced 2-year leukemia-free survival haematologica | 2017; 102(11)

(LFS), OS and GvHD-free, relapse-free survival (GRFS), and higher NRM as compared to MSD HCT. In high-risk AML patients, 2-year RI was lower in haploHCT, however, no other differences were observed in NRM, LFS, OS, and GRFS.56 In a separate registry study which focused on 6,545 patients with poor-risk AML in CR1, Versluis et al.57 reported similar 2-year OS following MSD (n=3,511) with 10/10 MUD (n=1,959) and haploHCT (n=193) (hazard ratio [HR], 0.99 and 1.12, respectively), whereas both 9/10 MUD (n =549) and UCB (n=333) grafts were associated with inferior OS (HR, 1.23, P=0.005; and HR, 1.54, P<0.001, respectively). Although the RI was decreased for 10/10 MUD (HR, 0.74, P<0.001) and haplo (HR, 0.60, P=0.001) compared with MSD HCT, NRM was significantly higher. Lastly, Piemontese et al.54 described clinical outcomes from T cell-replete haploHCT versus allogeneic transplants from 10/10 HLA matched and 9/10 HLA mismatched unrelated donors (MMUD) for adult patients with de novo AL in CR1/CR2. In this cohort, 265 patients (AML, n=176) received a haploHCT, 2,490 patients (AML, n=1,645) received a 10/10 MUD, and 813 patients (AML, n=510) received a MMUD transplant. Post-transplant cyclophosphamide was used as GvHD prophylaxis in 40% of haploHCT. Among patients with AML, 3-year LFS, OS, and NRM were significantly improved in 10/10 MUD compared to haploHCT, but there was no difference in GRFS, grade II-IV aGvHD or cGvHD. Further, no differences were found in GvHD or survival outcomes between 9/10 MMUD HCT and haploHCT. Based on the collective data, outcomes from haploidentical transplantation are encouraging, however, a larger cohort, longer follow-up period, and prospective comparative donor analyses are needed in order to firmly establish its place in the hierarchy of alternative donors. At this time, the ALWPEBMT supports a 10/10 MUD as the best donor option in the absence of a MSD, and further supports the use of a haploidentical donor or 9/10 MMUD as equally viable alternatives in the absence of a fully matched donor, or in the case of the need for an urgent transplant.

Haploidentical versus UCBT Single- and multi-center studies have also shown the value of single or double UCBT for AML in the setting of an urgent need for transplant and lack of an HLA-matched sibling or an unacceptable unrelated donor.63-68 Therefore, early comparative studies of alternative donor sources focused on examining differences in clinical outcomes with the use of haploidentical or UCB as sources of stem cells (Figure 3).65,69-72 In the earliest retrospective comparative study, the Eurocord group, in collaboration with the ALWP-EBMT, reported outcomes on 220 adult recipients who received T cell-deplete haploHCT with PBSC (n=154) or unrelated single or double UCBT (n=66) for AML. The 2-year incidences of relapse, TRM and LFS were not statistically different after haploHCT or UCBT, however, UCBT was associated with delayed neutrophil recovery and a higher incidence of aGvHD.69 In another large EBMT observational study of 918 AML patients (haplo, n=158; UCBT, n=558), Ruggeri et al.71 demonstrated similar findings of a comparable RI (HR=0.95, P=0.76), NRM (HR=1.16, P=0.47), and LFS (HR=0.78, P=0.78) between unmanipulated haploHCT and UCBT. While grade II-IV and grade III-IV aGvHD were similar between the two groups, the CI of cGvHD was less in the UCBT cohort (HR=0.63, P=0.008). 1815

C.J. Lee et al.

In order to study the reproducibility of the results found in retrospective analyses, the USA Blood and Marrow Transplantation Clinical Trials Network (BMT CTN) con-

ducted two parallel multicenter prospective clinical trials focused on outcomes associated with unmanipulated related haplo-BM graft with PTCy (n=50) and dUCBT (n=50)

Table 2. Published ALWP-EBMT studies of haploidentical transplantation in adults with AML.

Reference (year) 69

Rocha et al. (2005)

Ciceri et al.25 (2008)

Gorin et al.110 (2015)

Ruggeri et al.71 (2015)

Piemontese et al.10 (2015)

Ringden et al.62 (2016)

Rubio et al.88 (2016)

Sun et al.82 (2016)

Ruggeri et al.89 (2016)

Piemontese et al.54 (2017)

Ruggeri et al.83 (2017)

Versluis et al.57 (2017)

Study Objective

Conclusions

Retrospective comparative analysis of outcomes of 364 adult patients with AML/ALL receiving either UCBT versus T cell-depleted PB haploHCT between 1998-2002. Retrospective analysis of outcomes of 266 adult patients with de novo acute AML/ALL receiving T cell-depleted PB haploHCT between 1995-2004. Matched pair analysis of outcomes of 188 T cell-replete haploHCT and 356 autologous transplants (ASCT) in adult patients with acute leukemia between 2007-2012.

In AML, no difference between groups in relapse, TRM, and LFS. UCBT had increased grade II-IV aGvHD.

Retrospective comparative analysis of outcomes of 1,446 adult patients with de novo AML/ALL receiving either UCBT versus unmanipulated haploHCT between 2007-2012. Retrospective analysis of outcomes of 229 adult patients with de novo AML/ALL in CR or non-remission who received an unmanipulated haploHCT between 2007-2011. Retrospective comparative analysis of relapse and survival outcomes of 10,679 adult patients with AML/ALL receiving MSD HCT versus T cell-replete or deplete haploHCT between 2007-2012. Retrospective comparative analysis of outcomes of 696 adult patients with AML/ALL receiving T cell-replete haploHCT with RI versus MA conditioning regimen between 2001-2012. Retrospective one-to-one matched pair comparative study of outcomes following GIAC-based haploHCT and MA 10/10 MUD HCT in 174 patients with de novo intermediate-risk (based on cytogenetics) AML in CR1. Retrospective comparative analysis of outcomes of 451 adult patients with AML/ALL in CR1/CR2 receiving T cell-replete haploHCT with PTCy with PB versus BM stem cells between 2010-2014. Retrospective comparative analysis of outcomes of 3,568 adult patients with de novo AML/ALL in CR1/CR2 who received T cell-replete haploHCT versus 10/10 MUD or 9/10 MMUD HCT between 2007-2013. Retrospective comparative analysis of outcomes of 308 adult patients with AML in CR1/CR2 who underwent T cell-replete haploHCT using PTCy versus ATG-based GvHD prophylaxis between 2007-2014. Retrospective comparative analysis of outcomes of 6,545 adult patients with poor-risk AML in CR1 receiving an allogeneic HCT using MRD versus 10/10 or 9/10 MUD, UCB, T cell-replete haplo-identical donor between 2000-2014.

Salvatore et al.56 (2017)

Retrospective comparative analysis of outcomes of 2,654 adult patients with int-AML or high-risk AML in CR1 receiving T cell-replete haploHCT versus MSD HCT.

Canaani et al.96 (2017)

Retrospective comparative analysis of outcomes in 837 adult patients with AML who received ABO-matched versus ABO-mismatched haploHCT between 2005-2014.

Engraftment occurred in 91% of the patents; 2-year LFS was 48% ± 10% for patients with AML in CR1; 21% ± 5% in ≥ CR2; and 1% ± 1% in non-remission; GvHD was minimal. Haploidentical centers were divided into “expert” vs. “regular” based on the number of haploHCT performed. NRM was higher among all haploHCT compared to ASCT. LFS and OS were higher following ASCT compared to haploHCT in regular centers and similar In multivariate analysis of the AML group, UCBT was associated with lower cGvHD. No significant differences in relapse, NRM, and LFS between the two groups. Engraftment occurred in 93% of the patents; For the total group: 3-year relapse, LFS, OS and NRM was 42%, 30%, 37%, and 28%. 100-day CI of Grade II-IV aGvHD (32%) and 3-year CI of cGvHD (34%) were similar to historical outcomes of MSD HCT. No difference in relapse between MSD and haploHCT groups while NRM and LFS was superior in the MSD group.

Multivariable analysis in AML showed no difference in NRM, cGvHD, LFS, and OS. There was a trend toward high relapse with RI conditioning. Similar 5-year LFS, OS, relapse, NRM, grade III-IV aGvHD, and cGvHD.

Multivariable analysis showed that PB was associated with increased risk of grade II-IV aGvHD. Otherwise, no significant difference in other GvHD or survival outcomes. Weighted Cox model showed significantly higher LFS and OS in transplants from 10/10 MUD compared to haploHCT but no difference between 9/10 MMUD and haploHCT. Acute and chronic GvHD were not impacted by the donor type. HaploHCT with a PTCy-based prophylaxis produced superior LFS, GRFS, and lower incidence of GvHD and NRM compared to ATG-based prophylaxis. Multivariable analysis confirmed no differential impact on OS and RFS following MRD, 10/10 MUD, or haplo HCT, and significantly worse OS with 9/10 MUD and UCB grafts. Relapse was decreased for 10/10 MUD compared to MRD and haploHCT, and NRM was significantly higher for all alternative donors compared to MRD HCT. Int-AML: multivariate analysis showed reduced LFS, OS, GRFS, and higher NRM after haploHCT. High-risk AML: increased grade II-IV aGvHD and lower relapse incidence with haploHCT, and similar NRM, LFS, and OS between both donor groups. Major ABO mismatching was associated with inferior day 100 engraftment, and bi-directional mismatching had increased risk of grade II-IV aGvHD. Otherwise, NRM, relapse, LFS, OS, and cGvHD were similar.

aGvHD: acute graft-versus-host disease; AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; ASCT,: autologous stem cell transplantation; ATG: antithymocyte globulin; BM: bone marrow; cGvHD: chronic graft-versus-host-disease; CI: cumulative incidence; CR: complete remission; GRFS: GvHD-free, relapse-free survival; GvHD: graft-versus-host disease; haplo: haploidentical; HCT: hematopoietic cell transplantation; int-AML: intermediate AML; LFS: leukemia-free survival; MA: myeloablative; MMUD: mismatched unrelated donor; MRD: minimal residual disease; MSD: matched sibling donor; MUD,: matched unrelated donor; NRM: non-relapse mortality; OS: overall survival; PB: peripheral blood; PTCy: post-transplant cyclophosphamide; RFS: relapse-free survival; RI: reducedintensity; TRM: transplant-related mortality; UCB: umbilical cord blood; UCBT: umbilical cord blood transplantation.

1816

haematologica | 2017; 102(11)

Haploidentical HCT in AML

following identical RI conditioning regimens for patients with high-risk leukemia or lymphoma (AML: haplo, n=22; dUCB, n=29). Results from this trial successfully replicated those from single-center studies and showed no significant differences between the two modalities.65 Furthermore, both cohorts had comparable survival rates to patients with high-risk hematologic malignancies who underwent MUD HCT with blood or marrow after RI conditioning.73 An ongoing phase III multicenter randomized trial (clinicaltrials.gov Identifier: 01597778) is attempting to clarify the relative efficacies of double unrelated cord and haploidentical related BMT, and the estimated completion date of the trial is June 2019.

The Beijing experience An alternative strategy for prevention of GvHD after T cell-replete haploidentical donor transplantation has incorporated pre-transplant ATG. The Peking University group in China pioneered an approach that used a combination of G-CSF-priming of the donor, intensified immunosuppression, ATG, and combination of T cell-replete BM plus peripheral blood as the stem cell source (GIAC protocol) (Figure 2C).74,75 An early trial in patients with acute leukemia, including 108 AML patients, suggested encouraging GvL effects with universal engraftment, low incidence of relapse following transplantation (13 out of 108 AML patients) and 3-year relapse probabilities of 11.9% and 20.2% in the standard- and high-risk AML groups, resulting in DFS rates of 71% and 56%, respectively. While TRM at D+100 was favorable in both risk groups, the 3year TRM in standard-risk and high-risk AML groups was 19.4% and 29.4%. The CI of grade II-IV and grade III-IV aGvHD were 45.8% and 13.4% at D+100, respectively, while the 3-year CI of total cGvHD and extensive cGvHD were 53.9% and 22.6%.75 An updated trial including 756 patients with AL over a time period of 9 years confirmed their previous findings.76 A subsequent comparative study in patients with AML who received the GIAC haploHCT protocol revealed a similar CI of acute and chronic GvHD, TRM, 5-year relapse and OS rates when compared to MUD HCT, but a significantly reduced incidence of 5-year relapse (14.2% vs. 34%, P=0.008) compared to MSD HCT. A superior GvL effect for high-risk leukemia was also observed in haploHCT, as 5-year relapse rates were 15.4%, 28.2%, and 49.9% in haplo, MUD (P=0.07), and MSD HCT, respectively (P=0.002).77 Results from Wang et al.78 also suggested a superior GvL effect by haploHCT compared to a matched sibling HCT in patients with high-risk AL (50 AML out of 117), whereas other studies indicated no significant difference.79,80 In three of the four studies, grade II-IV aGvHD was significantly more frequent after haploHCT compared to MSD HCT. In the only prospective study comparing post-transplantation outcomes in 450 patients with intermediate- or high-risk AML in CR1 who received a haplo or MSD HCT, Wang et al.81 demonstrated a similar CI of relapse (15% vs. 15%, P=0.98), 3-year DFS (74% vs. 78%, P=0.34), NRM (13% vs. 8%, P=0.13), and OS (79% vs. 82%, P=0.36). The CI of 100-day aGvHD and 1-year cGvHD, including severe cGvHD, was significantly higher in the haploHCT group. Owing to the lack of randomization, this comparative study suggests haploidentical HCT as a valid alternative option for this patient population for whom no matched sibling donor is available. haematologica | 2017; 102(11)

Due to the reported high leukemia-free survival rates associated with the Beijing strategy, the ALWP of the EBMT performed a retrospective one-to-one matched pair comparative study of outcomes following GIAC-based haploidentical HCT and myeloablative (non-TBI based) 10/10 MUD HCT in patients with de novo intermediaterisk (based on cytogenetics) AML in CR1.82 Subjects were matched in age, time to transplant, and number of induction courses to reach CR1. Similar outcomes were observed between haploHCT and MUD HCT in terms of 5-year LFS (73.5% vs. 60.3%, P=0.15), OS (78.2% vs. 63.6%, P=0.15), relapse (12.7% vs. 24%, P=0.08), NRM (13.8% vs. 15.7%, P=0.96), grade III-IV aGvHD (9.2% vs. 9.4%, P=1), and cGvHD (42.5% vs. 34.9%, P=0.39). Based on this analysis, the authors concluded that the Beijing protocol is a feasible alternative to allogeneic transplantation with a 10/10 MUD. Following several publications showing very low incidences of GvHD after ATG-based intensive immunosuppression established in the GIAC haploHCT protocol, Ruggeri et al.83 compared this GvHD prophylaxis regimen to the PTCy platform in the setting of unmanipulated haploHCT for patients with various-risk AML in CR1 or CR2. A total of 308 patients were studied (PTCy, n=193; ATG, n=115), and both groups were well matched in regards to recipient and donor age, AML disease risk, disease status at transplant, and conditioning intensity. Notably, a BM stem cell source was used more frequently in the PTCy group (60.1% vs. 39.9%, P=0.01), and that cohort also had shorter follow up (18 vs. 36 months, P<0.001). At day 100, similar outcomes in grade II-V aGvHD were observed between patients receiving PTCy versus ATG (31% vs. 21%, P=0.07), however, grade III-IV aGvHD was significantly lower in the PTCy group (4.7% vs. 12.5%, P=0.01). The incidence of 2-year cGvHD did not differ between the two groups (33.7% vs. 28.3%, P=0.33). Multivariate analysis of NRM, LFS, OS, and GRFS also significantly favored the PTCy regimen. Although different haploHCT methods have not been prospectively compared in a randomized fashion, the available cumulative evidence demonstrates the feasibility of haploidentical transplantation and the benefit of having a readily accessible donor, regardless of the platform used.

Ongoing research in T cell-replete haploidentical transplantation Since the demonstration of the safety and efficacy of NMA haploHCT with PTCy, there has been increasing research interest in optimizing clinical outcomes for different patient populations through modifications of the original platform. For example, some groups have explored optimizing the anti-leukemia effects of haploHCT, particularly in high-risk or advanced AML, by intensifying the conditioning regimen or substituting BM with PBSC as the stem cell graft source, due to the concern of high relapse rates associated with NMA haploHCT and PTCy. In the former setting, several single-center non-comparative studies have reported a low risk of acute and chronic GvHD and encouraging rates of TRM and OS with myeloablative conditioning.84-87 These observations were recently validated by the first large retrospective comparative analysis performed by the ALWP-EBMT showing similar OS, LFS, NRM, and cGvHD between MA and RI conditioning regi1817

C.J. Lee et al.

Table 3. Comparative studies of haploidentical HCT versus umbilical cord blood transplantation.

Author

Disease, no of pts. (AML, n)

Rocha et al.69

AML, 220

Brunstein et al.65^ All, 100 (AML, 51) Ruggeri et al.71 AML, 918

El-Cheikh et al.72 All, 150 (AML, 40)

Condition

Graft Source Intensity (%) (n)

Donor Engraft aGvHD Type (%) II-IV (yr) (no. of patients)

NR (haplo)

Haplo (154) UCB (66)

RI (100)

BM (50)

MA (54%) RI (46%)

BM (171) PB (78) Both (14) BM (NR) PB (NR)

Haplo (50) UCB (50) Haplo (360) UCB (558)

RI (100)

Haplo (69) UCB (81)

5 =/- 5% 23 +/- 5% <0.0001

aGvHD III-IV (yr)

cGvHD (yr)

OS (yr)

DFS (yr)

24±4% (2) 30±6% (P=0.39)

94 32% (d100) 0% (d100) 13% (1) 62% (1) 48% (1) 96 40% 21% 25% 54% 46% 91 27% (*) 11% (*) 29% (*) 38% (*) 32% (2) 84 31% 12% 24% 42% 38% (P=0.003) (P=0.10) (P=0.41) (P=0.19) (P=0.269) (P=0.102) 94 34% (*) 5% (*) 6% (*) 69% (2) 65% (2) 90 50% 33% 12% 45% 36% (P=0.08) (P<0.0001) (P=0.001) (P=0.10) (P=0.01)

Relapse (yr)

NRM (yr)

18±3% (2) 58±4% 24±5% 46±2% (P=0.44) (P=0.23) 45% (1) 31% 41% (*) 32% (P=0.008)α 18% (*) 38% (P=0.03)

7% (1) 24% 27% (*) 30% (P=0.356) 18% (2) 23% (P=0.49)

The data are from 2 separate but parallel multicenter phase 2 trials with identical objectives, eligibility, and clinical endpoints. The clinical outcomes should not be compared directly. * Year not reported. αMultivariate analysis of relapse was not statistically different between the haplo and UCB groups (HR 0.95, P=0.76). aGvHD: acute graft-versus-host disease; AML: acute myeloid leukemia; ALL: acute lymphocytic leukemia; BM: bone marrow; cGvHD, chronic graft-versus-host-disease; DFS: disease-free survival; haplo: haploidentical; MA: myeloablative; NRM: non-relapse mortality; NR: not reported; OS: overall survival; PB: peripheral blood; RI: reduced-intensity; UCB: umbilical cord blood; yr, year. ^

mens in T cell-replete haploHCT, in particular for patients with AML in CR1.88 Multivariable analyses revealed a trend towards higher relapse incidence, with RI versus MA conditioning (HR 1.34, P=0.09), and when taken collectively the data supported the use of either high or low intensity conditioning haploHCT with PTCy in the first-line treatment of high-risk AML.88 In this study, there was an increased risk of grade II-IV aGvHD and cGvHD independent of the conditioning regimen intensity and of the use of PTCy with the use of a PBSC graft compared to BM, but no difference was seen with regard to the incidence of NRM and other survival outcomes.88 Ruggeri et al.89 also described the use of PBSC as the sole factor associated with an increased risk of grade II-IV aGvHD (HR 2.2, 95% CI 1.27-3.9, P=0.005) in patients with AL, the majority of whom were transplanted with a MA regimen for AML in CR1. Otherwise, the type of stem cell graft (PBSC vs. BM) proved to have no significant difference on grade III-IV aGvHD, cGvHD, relapse, or survival. In line with the attempts to exploit a PBSC source, Peccatori et al. developed a calcineurin inhibitor-free GvHD prophylaxis based on rapamycin, mycophenolate mofetil and ATG, with the aim of promoting a fast post-transplant immune recovery with a preferential accumulation of regulatory T cells.90 Recently, this sirolimus platform has been modified with the substitution of ATG by PTCy, which showed a significant reduction in cGvHD.91 In the NMA haploHCT setting, both Castagna et al.92 and O’ Donnell et al.93 reported comparable outcomes in acute and chronic GvHD, engraftment rates, NRM, and OS after haplo-BM or haplo-PBSC transplantation; however, the incidence of relapse at 1 to 3 years was significantly lower after haplo-PBSC transplant compared with haplo-BM transplants in the latter study. Other groups have also demonstrated the feasibility of NMA haploHCT with either PBSC or BM stem cells in older adults.94,95 The significance of ABO incompatibility on outcomes after haploHCT for AML have recently been published by the ALWP-EBMT, and preliminarily demonstrate a significantly increased risk of grade II-IV aGvHD with bi-directional ABO mismatching and a lower OS rate in patients with major ABO mismatching transplanted 1818

with BM grafts.96 Lastly, the impact of haploHCT for specific high-risk AML cytogenetic and molecular risk groups as well as the role of post-transplant cellular therapies are of interest. The significance of pre-transplant MRD as a poor prognostic and predictive factor of outcomes after allogeneic HCT in AML has been reported.3,97-99 For example, the Seattle group published inferior 3-year OS and relapse outcomes among AML patients receiving a myeloablative matched donor HCT with pre-transplant MRD-positive (morphologic remission) compared to MRD-negative (morphologic remission), and further demonstrated comparable outcomes to patients with active disease at the time of HCT.3 Several other groups have studied the significance of pre-transplant MRD on a more granular level and demonstrated that the level of pre-transplant MRD may differentially impact post-transplantation outcomes100,101 The significance of pre-transplant MRD has also been described in the setting of haploHCT.102,103 Wang et al.102 retrospectively evaluated outcomes of 255 patients with AML in CR1 or CR2. Multivariate analysis indicated failure of CR after 2 courses of induction therapy as the strongest independent prognostic factor for relapse and LFS. In subgroup analysis, positive pre-transplant MRD as compared to negative MRD also resulted in worse LFS at 3 years (76% vs. 52%, P=0.041) and CI of relapse at 2 years (10% vs. 35%, P=0.002). These results must be interpreted cautiously due to the limited patient sample (negative MRD, n=110; positive MRD, n=20). Conversely, other groups have reported no significant influence of MRD status (positive vs. negative) prior to haploHCT on PFS104 or relapse103 for patients with AML in CR1/CR2, and further hypothesize that while detectable MRD before HCT is a strong unfavorable prognostic factor, its adverse impact may be overcome by the potentially stronger GvL effects of unmanipulated haploHCT.103

Perspectives Over the last two decades, the international BMT community have witnessed incredible advances in HLA-typhaematologica | 2017; 102(11)

Haploidentical HCT in AML

Figure 3. Recommended donor choice algorithm for adults with intermediate or high-risk AML with an indication for allogeneic HCT. AML: acute myeloid leukemia; BM: bone marrow; CR: complete remission; alloHCT: allogeneic hematopoietic cell transplantation; haploHCT: haploidentical hematopoietic cell transplantation; MMUD: mismatched unrelated donor; MRD: matched related donor; MUD: matched unrelated donor; PB: peripheral blood; UCB: umbilical cord blood; UCBT: umbilical cord blood transplantation. RI: reduced-intensity; MA: myeloablative.

ing and alternative donor transplantation strategies, such that in the present day nearly all transplant-eligible patients with AML will have an available donor. Unmanipulated haploidentical related transplantation with post-transplant cyclophosphamide has emerged as a potentially powerful strategy for the cure of AML and is the dominant haploHCT platform in Europe.105 Other significant advantages of haploHCT with PTCy include its associated low non-relapse mortality and GvHD, ease of donor accessibility often leading to minimal length of time to transplantation, and low acquisition costs. The costeffectiveness associated with haploHCT with PTCy may have the most appeal in developing countries, where economic resources are more limited.105 Unmanipulated haploidentical transplantation with post-transplantation immunosuppression also shows promise in decreasing post-transplant infections and death due to infections, however, further data on immune reconstitution, infections and their related complications (i.e., hemorrhagic cystitis) among different haploHCT strategies are warranted.106 The incidence of post-transplant cardiomyopathy secondary to GvHD prophylaxis with high-dose PTCy appears non-significant in the absence of severe infection,107 however, further research evaluating predictive factors for cardiomyopathy following PTCy based HCT are necessary. While there has been questioning of donor-derived malignancies (DDM) associated with PTCy, a recent retrospective study by the Hopkins group showed an extremely low proportion of patients with a DDM (4 out of 789) over a 10-year period, suggesting that haematologica | 2017; 102(11)

PTCy does not appear to increase the risk of DDM.108 However, the authors acknowledge the short follow-up period of their study and report the need for continued close monitoring of DDMs over a longer follow-up time. Another key issue arising in the setting of unmanipulated haploHCT is the selection of the â&#x20AC;&#x153;bestâ&#x20AC;? donor, as some patients will have multiple haploidentical donor candidates and donor selection may significantly impact GvHD, relapse, TRM, and survival outcomes. Owing to improved approaches of unmanipulated haploHCT with PTCy or ATG-based GvHD prophylaxis, the effects of HLA disparity have vanished, nonetheless, other donorrelated variables should be considered. These include the selection of donors for whom there are no recipient donor-specific antibodies (DSA); alternatively, measures to remove DSA should be undertaken in the patient; the selection of a younger, male donor over an older, female donor due to the potential for superior survival, decreased risk of grade II-IV aGvHD and age-related clonal hematopoiesis leading to subsequent malignancies; and the selection of an ABO compatible donor, followed by a minor ABO mismatched and then a major ABO mismatched donor. Other factors to consider include donor and recipient cytomegalovirus (CMV) serostatus, NK cell alloreactivity and KIR haplotype matching, and non-inherited maternal HLA antigens (NIMA) mismatching.109 However, more research is needed as the significance of each of these factors may change depending on the haploHCT protocol or platform used, and indeed, may vanish, due to the emergence of new variables as 1819

C.J. Lee et al.

haploHCT becomes increasingly utilized. In conclusion, the growing body of literature has consistently demonstrated comparable outcomes of haploidentical donor HCT as compared to an UCB, matched sibling and unrelated donor transplantation for patients with AML. However, the available studies are nonrandomized, underpowered, and lack long-term follow-up data. Accordingly, the ALWP-EBMT endorses haploidentical transplantation as a valid post-remission therapy for highrisk AML in the absence of a matched donor or in the case of the need for an urgent transplant procedure (Figure 3). Further prospective studies randomizing haploHCT to UCBT (in the USA) or to MUD or MMUD HCT (in Europe) are ongoing, and will help to establish its position in the hierarchy of alternative donors.

References 1. Zuckerman T, Rowe JM. Transplantation in acute myeloid leukemia. Hematol Oncol Clin North Am. 2014;28(6):983-994. 2. Cornelissen JJ, Gratwohl A, Schlenk RF, et al. The European LeukemiaNet AML Working Party consensus statement on allogeneic HSCT for patients with AML in remission: an integrated-risk adapted approach. Nat Rev Clin Oncol. 2012;9(10):579-590. 3. Araki D, Wood BL, Othus M, et al. Allogeneic hematopoietic cell transplantation for acute myeloid leukemia: time to move toward a minimal residual diseasebased definition of complete remission? J Clin Oncol. 2016;34(4):329-336. 4. Passweg JR, Baldomero H, Bader P, et al. Use of haploidentical stem cell transplantation continues to increase: the 2015 European Society for Blood and Marrow Transplant activity survey report. Bone Marrow Transplant. 2017;52(6):811-817. 5. Passweg JR, Baldomero H, Gratwohl A, et al. The EBMT activity survey: 1990-2010. Bone Marrow Transplant. 2012;47(7):906923. 6. Tiercy JM. How to select the best available related or unrelated donor of hematopoietic stem cells? Haematologica. 2016;101(6):680687. 7. Gragert L, Eapen M, Williams E, et al. HLA match likelihoods for hematopoietic stemcell grafts in the U.S. registry. N Engl J Med. 2014;371(4):339-348. 8. Besse K, Maiers M, Confer D, Albrecht M. On modeling human leukocyte antigenidentical sibling match probability for allogeneic hematopoietic cell transplantation: estimating the need for an unrelated donor source. Biol Blood Marrow Transplant. 2016;22(3):410-417. 9. Employment, Social Affairs & Inclusion Eurostat Demography Report. European Commission. Luxembourg: Publications Office of the European Union; 2015 [updated May 2015; cited 8 August 2017]. Available from: http://ec.europa.eu/eurostat/web/population-demography-migration-projections/overview 10. Piemontese S, Ciceri F, Labopin M, et al. A survey on unmanipulated haploidentical hematopoietic stem cell transplantation in adults with acute leukemia. Leukemia.

1820

Position statement from the ALWP- EBMT • Haploidentical donor transplantation is a valid option for patients with AML lacking a matched sibling or unrelated donor. • In certain clinical situations, especially in the case of a need for an urgent transplant procedure and lack of a MDS, a readily available haploidentical donor may be considered over initiating an unrelated donor search. • The evidence for the superiority of haploidentical vs. MMUD vs. UCBT is insufficient, but there is the potential for a cost benefit with regard to haploHCT. • There is insufficient evidence for the superiority of one haploidentical HCT platform over another. Economic factors, together with individual center experience, may be decisive.

2015;29(5):1069-1075. 11. Passweg JR, Baldomero H, Bader P, et al. Hematopoietic SCT in Europe 2013: recent trends in the use of alternative donors showing more haploidentical donors but fewer cord blood transplants. Bone Marrow Transplant. 2015;50(4):476-482. 12. Beatty PG, Clift RA, Mickelson EM, et al. Marrow transplantation from related donors other than HLA-identical siblings. N Engl J Med. 1985;313(13):765-771. 13. Anasetti C, Amos D, Beatty PG, Appelbaum FR, Bensinger W, Buckner CD, et al. Effect of HLA compatibility on engraftment of bone marrow transplants in patients with leukemia or lymphoma. N Engl J Med. 1989;320(4):197-204. 14. Anasetti C, Beatty PG, Storb R, et al. Effect of HLA incompatibility on graft-versus-host disease, relapse, and survival after marrow transplantation for patients with leukemia or lymphoma. Hum Immunol. 1990;29(2):79-91. 15. Powles RL, Morgenstern GR, Kay HE, et al. Mismatched family donors for bone-marrow transplantation as treatment for acute leukaemia. Lancet. 1983;1(8325):612-615. 16. Aversa F, Tabilio A, Terenzi A, et al. Successful engraftment of T-cell-depleted haploidentical "three-loci" incompatible transplants in leukemia patients by addition of recombinant human granulocyte colonystimulating factor-mobilized peripheral blood progenitor cells to bone marrow inoculum. Blood. 1994;84(11):3948-3955. 17. Aversa F, Tabilio A, Velardi A, et al. Treatment of high-risk acute leukemia with T-cell-depleted stem cells from related donors with one fully mismatched HLA haplotype. N Engl J Med. 1998;339(17): 1186-1193. 18. Aversa F, Terenzi A, Tabilio A, et al. Full haplotype-mismatched hematopoietic stem-cell transplantation: a phase II study in patients with acute leukemia at high risk of relapse. J Clin Oncol. 2005;23(15):3447-3454. 19. Aversa F. Setting the standard in T-celldepleted haploidentical transplantation and beyond. Best Pract Res Clin Haematol. 2011;24(3):325-329. 20. Mancusi A, Ruggeri L, Velardi A. Haploidentical hematopoietic transplantation for the cure of leukemia: from its biology to clinical translation. Blood. 2016;128 (23):2616-2623.

21. Ruggeri L, Capanni M, Casucci M, et al. Role of natural killer cell alloreactivity in HLAmismatched hematopoietic stem cell transplantation. Blood. 1999;94(1):333-339. 22. Ruggeri L, Capanni M, Urbani E, et al. Effectiveness of donor natural killer cell alloreactivity in mismatched hematopoietic transplants. Science. 2002;295(5562):20972100. 23. Ruggeri L, Mancusi A, Capanni M, et al. Donor natural killer cell allorecognition of missing self in haploidentical hematopoietic transplantation for acute myeloid leukemia: challenging its predictive value. Blood. 2007;110(1):433-440. 24. Velardi A, Ruggeri L, Mancusi A, Aversa F, Christiansen FT. Natural killer cell allorecognition of missing self in allogeneic hematopoietic transplantation: a tool for immunotherapy of leukemia. Curr Opin Immunol. 2009;21(5):525-530. 25. Ciceri F, Labopin M, Aversa F, et al. A survey of fully haploidentical hematopoietic stem cell transplantation in adults with high-risk acute leukemia: a risk factor analysis of outcomes for patients in remission at transplantation. Blood. 2008;112(9):3574-3581. 26. Andre-Schmutz I, Le Deist F, Hacein-BeyAbina S, et al. Immune reconstitution without graft-versus-host disease after haemopoietic stem-cell transplantation: a phase 1/2 study. Lancet. 2002;360 (9327):130-137. 27. Amrolia PJ, Muccioli-Casadei G, Huls H, et al. Adoptive immunotherapy with allodepleted donor T-cells improves immune reconstitution after haploidentical stem cell transplantation. Blood. 2006;108(6):17971808. 28. Triplett BM, Shook DR, Eldridge P, et al. Rapid memory T-cell reconstitution recapitulating CD45RA-depleted haploidentical transplant graft content in patients with hematologic malignancies. Bone Marrow Transplant. 2015;50(7):968-977. 29. Hoffmann P, Ermann J, Edinger M, Fathman CG, Strober S. Donor-type CD4(+)CD25(+) regulatory T cells suppress lethal acute graftversus-host disease after allogeneic bone marrow transplantation. J Exp Med. 2002;196(3):389-399. 30. 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.

haematologica | 2017; 102(11)

Haploidentical HCT in AML 31. Di Ianni M, Falzetti F, Carotti A, et al. Tregs prevent GVHD and promote immune reconstitution in HLA-haploidentical transplantation. Blood. 2011;117(14):3921-3928. 32. Martelli MF, Di Ianni M, Ruggeri L, et al. HLA-haploidentical transplantation with regulatory and conventional T-cell adoptive immunotherapy prevents acute leukemia relapse. Blood. 2014;124(4):638-644. 33. Ciceri F, Bonini C, Stanghellini MT, et al. Infusion of suicide-gene-engineered donor lymphocytes after family haploidentical haemopoietic stem-cell transplantation for leukaemia (the TK007 trial): a non-randomised phase I-II study. Lancet Oncol. 2009;10(5):489-500. 34. Berenbaum MC, Brown IN. Prolongation of homograft survival in mice with single doses of cyclophosphamide. Nature. 1963;200:84. 35. Berenbaum MC, Brown IN. Dose-response relationships for agents inhibiting the immune response. Immunology. 1964;7:6571. 36. Santos GW, Owens AH. Production of graftversus-host disease in the rat and its treatment with cytotoxic agents. Nature. 1966;210(5032):139-140. 37. Mayumi H, Himeno K, Shin T, Nomoto K. Drug-induced tolerance to allografts in mice. VI. Tolerance induction in H-2-haplotypeidentical strain combinations in mice. Transplantation. 1985;40(2):188-194. 38. 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. 39. Luznik L, Engstrom LW, Iannone R, Fuchs EJ. Posttransplantation cyclophosphamide facilitates engraftment of major histocompatibility complex-identical allogeneic marrow in mice conditioned with low-dose total body irradiation. Biol Blood Marrow Transplant. 2002;8(3):131-138. 40. Kanakry CG, Ganguly S, Zahurak M, et al. Aldehyde dehydrogenase expression drives human regulatory T cell resistance to posttransplantation cyclophosphamide. Sci Transl Med. 2013;5(211):211ra157. 41. Kastan MB, Schlaffer E, Russo JE, Colvin OM, Civin CI, Hilton J. Direct demonstration of elevated aldehyde dehydrogenase in human hematopoietic progenitor cells. Blood. 1990;75(10):1947-1950. 42. Jones RJ, Barber JP, Vala MS, et al. Assessment of aldehyde dehydrogenase in viable cells. Blood. 1995;85(10):2742-2746. 43. O'Donnell PV, Luznik L, Jones RJ, et al. Nonmyeloablative bone marrow transplantation from partially HLA-mismatched related donors using posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2002;8(7):377-386. 44. Luznik L, O'Donnell PV, Symons HJ, et al. HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and highdose, posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2008;14(6):641-650. 45. Munchel A, Kesserwan C, Symons HJ, et al. Nonmyeloablative, HLA-haploidentical bone marrow transplantation with high dose, post-transplantation cyclophosphamide. Pediatr Rep. 2011;3 Suppl 2:e15. 46. Ciurea SO, Mulanovich V, Saliba RM, et al. Improved early outcomes using a T cell replete graft compared with T cell depleted haploidentical hematopoietic stem cell

haematologica | 2017; 102(11)

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

transplantation. Biol Blood Marrow Transplant. 2012;18(12):1835-1844. Devillier R, Bramanti S, Furst S, et al. Treplete haploidentical allogeneic transplantation using post-transplantation cyclophosphamide in advanced AML and myelodysplastic syndromes. Bone Marrow Transplant. 2016;51(2):194-198. Ciurea SO, Zhang MJ, Bacigalupo AA, et al. Haploidentical transplant with posttransplant cyclophosphamide vs matched unrelated donor transplant for acute myeloid leukemia. Blood. 2015;126(8):1033-1040. Bashey A, Zhang X, Sizemore CA, et al. Tcell-replete HLA-haploidentical hematopoietic transplantation for hematologic malignancies using post-transplantation cyclophosphamide results in outcomes equivalent to those of contemporaneous HLA-matched related and unrelated donor transplantation. J Clin Oncol. 2013;31(10): 1310-1316. Bashey A, Zhang X, Jackson K, et al. Comparison of outcomes of hematopoietic cell transplants from T-replete haploidentical donors using post-transplantation cyclophosphamide with 10 of 10 HLA-A, -B, -C, -DRB1, and -DQB1 allele-matched unrelated donors and HLA-identical sibling donors: a multivariable analysis including disease risk index. Biol Blood Marrow Transplant. 2016;22(1):125-133. Di Stasi A, Milton DR, Poon LM, et al. Similar transplantation outcomes for acute myeloid leukemia and myelodysplastic syndrome patients with haploidentical versus 10/10 human leukocyte antigen-matched unrelated and related donors. Biol Blood Marrow Transplant. 2014;20(12):1975-1981. Rashidi A, DiPersio JF, Westervelt P, et al. Comparison of outcomes after peripheral blood haploidentical versus matched unrelated donor allogeneic hematopoietic cell transplantation in patients with acute myeloid leukemia: a retrospective singlecenter review. Biol Blood Marrow Transplant. 2016;22(9):1696-1701. How J, Slade M, Vu K, et al. T Cell-Replete T cell-replete peripheral blood haploidentical hematopoietic cell transplantation with post-transplantation cyclophosphamide results in outcomes similar to transplantation from traditionally matched donors in active disease acute myeloid leukemia. Biol Blood Marrow Transplant. 2017;23(4):648653. Piemontese S, Ciceri F, Labopin M, et al. A comparison between allogeneic stem cell transplantation from unmanipulated haploidentical and unrelated donors in acute leukemia. J Hematol Oncol. 2017;10(1):24. Rashidi A, Slade M, DiPersio JF, Westervelt P, Vij R, Romee R. Post-transplant high-dose cyclophosphamide after HLA-matched vs haploidentical hematopoietic cell transplantation for AML. Bone Marrow Transplant. 2016;51(12):1561-1564. Salvatore D, Labopin M, Ruggeri A, et al. Outcomes of non T cell-depleted haploidentical HSCT versus HSCT from matched sibling donors in patients with acute myeloid leukemia in first complete remission, an ALWP-EBMT study. 22nd European Hematology Association Congress; 2017; Madrid, Spain; 2017. Versluis J, Labopin M, Ruggeri A, et al. Alternative donors for allogeneic hematopoietic stem cell transplantation in poor-risk AML in CR1. Blood Advances. 2017;1:477-485. Armand P, Gibson CJ, Cutler C, et al. A dis-

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

ease 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. McCurdy SR, Kanakry JA, Showel MM, et al. Risk-stratified outcomes of nonmyeloablative HLA-haploidentical BMT with highdose posttransplantation cyclophosphamide. Blood. 2015;125(19):3024-3031. Tischer J, Engel N, Fritsch S, et al. Second haematopoietic SCT using HLA-haploidentical donors in patients with relapse of acute leukaemia after a first allogeneic transplantation. Bone Marrow Transplant. 2014;49(7):895-901. Ringden O, Labopin M, Ciceri F, et al. Is there a stronger graft-versus-leukemia effect using HLA-haploidentical donors compared with HLA-identical siblings? Leukemia. 2016;30(2):447-455. Ballen KK, Spitzer TR, Yeap BY, et al. Double unrelated reduced-intensity umbilical cord blood transplantation in adults. Biol Blood Marrow Transplant. 2007;13(1):82-89. Brunstein CG, Barker JN, Weisdorf DJ, et al. Umbilical cord blood transplantation after nonmyeloablative conditioning: impact on transplantation outcomes in 110 adults with hematologic disease. Blood. 2007;110(8): 3064-3070. Brunstein CG, Fuchs EJ, Carter SL, et al. Alternative donor transplantation after reduced intensity conditioning: results of parallel phase 2 trials using partially HLAmismatched related bone marrow or unrelated double umbilical cord blood grafts. Blood. 2011;118(2):282-288. Laughlin MJ, Barker J, Bambach B, et al. Hematopoietic engraftment and survival in adult recipients of umbilical-cord blood from unrelated donors. N Engl J Med. 2001;344(24):1815-1822. Barker JN, Weisdorf DJ, DeFor TE, et al. Transplantation of 2 partially HLA-matched umbilical cord blood units to enhance engraftment in adults with hematologic malignancy. Blood. 2005;105(3):1343-1347. Eapen M, Rocha V, Sanz G, et al. Effect of graft source on unrelated donor haemopoietic stem-cell transplantation in adults with acute leukaemia: a retrospective analysis. Lancet Oncol. 2010;11(7):653-660. Rocha V, Aversa F, Labopin M, et al. Outcomes of unrelated cord blood and haploidentical stem cell transplantation in adults with acute leukaemia. Blood. 2005;106(11): 301. Raiola AM, Dominietto A, di Grazia C, et al. Unmanipulated haploidentical transplants compared with other alternative donors and matched sibling grafts. Biol Blood Marrow Transplant. 2014;20(10):1573-1579. Ruggeri A, Labopin M, Sanz G, et al. Comparison of outcomes after unrelated cord blood and unmanipulated haploidentical stem cell transplantation in adults with acute leukemia. Leukemia. 2015;29(9):18911900. El-Cheikh J, Crocchiolo R, Furst S, et al. Unrelated cord blood compared with haploidentical grafts in patients with hematological malignancies. Cancer. 2015;121(11): 1809-1816. Giralt S, Logan B, Rizzo D, et al. Reducedintensity conditioning for unrelated donor progenitor cell transplantation: long-term follow-up of the first 285 reported to the national marrow donor program. Biol Blood

1821

C.J. Lee et al. Marrow Transplant. 2007;13(7):844-852. 74. Huang XJ, Liu DH, Liu KY, et al. Haploidentical hematopoietic stem cell transplantation without in vitro T-cell depletion for the treatment of hematological malignancies. Bone Marrow Transplant. 2006;38(4):291-297. 75. Huang XJ, Liu DH, Liu KY, et al. Treatment of acute leukemia with unmanipulated HLA-mismatched/haploidentical blood and bone marrow transplantation. Biol Blood Marrow Transplant. 2009;15(2):257-265. 76. Wang Y, Liu DH, Liu KY, et al. Long-term follow-up of haploidentical hematopoietic stem cell transplantation without in vitro T cell depletion for the treatment of leukemia: nine years of experience at a single center. Cancer. 2013;119(5):978-985. 77. Luo Y, Xiao H, Lai X, et al. T-cell-replete haploidentical HSCT with low-dose anti-Tlymphocyte globulin compared with matched sibling HSCT and unrelated HSCT. Blood. 2014;124(17):2735-2743. 78. Wang Y, Liu DH, Xu LP, et al. Superior graftversus-leukemia effect associated with transplantation of haploidentical compared with HLA-identical sibling donor grafts for high-risk acute leukemia: an historic comparison. Biol Blood Marrow Transplant. 2011;17(6):821-830. 79. Chen XH, Gao L, Zhang X, et al. HLA-haploidentical blood and bone marrow transplantation with anti-thymocyte globulin: long-term comparison with HLA-identical sibling transplantation. Blood Cells Mol Dis. 2009;43(1):98-104. 80. Lu DP, Dong L, Wu T, et al. Conditioning including antithymocyte globulin followed by unmanipulated HLA-mismatched/haploidentical blood and marrow transplantation can achieve comparable outcomes with HLA-identical sibling transplantation. Blood. 2006;107(8):3065-3073. 81. Wang Y, Liu QF, Xu LP, et al. Haploidentical vs identical-sibling transplant for AML in remission: a multicenter, prospective study. Blood. 2015;125(25):3956-3962. 82. Sun Y, Beohou E, Labopin M, et al. Unmanipulated haploidentical versus matched unrelated donor allogeneic stem cell transplantation in adult patients with acute myelogenous leukemia in first remission: a retrospective pair-matched comparative study of the Beijing approach with the EBMT database. Haematologica. 2016;101 (8):e352-354. 83. Ruggeri A, Sun Y, Labopin M, et al. Posttransplant cyclophosphamide versus antithymocyte globulin as graft- versus-host disease prophylaxis in haploidentical transplant. Haematologica. 2017;102(2):401-410. 84. Raiola AM, Dominietto A, Ghiso A, et al. Unmanipulated haploidentical bone marrow transplantation and posttransplantation cyclophosphamide for hematologic malignancies after myeloablative conditioning. Biol Blood Marrow Transplant. 2013;19(1): 117-122. 85. Bacigalupo A, Dominietto A, Ghiso A, et al. Unmanipulated haploidentical bone marrow transplantation and post-transplant cyclophosphamide for hematologic malignanices following a myeloablative conditioning: an update. Bone Marrow Transplant. 2015;50 Suppl 2:S37-39. 86. Solomon SR, Sizemore CA, Sanacore M, et al. Haploidentical transplantation using T cell replete peripheral blood stem cells and myeloablative conditioning in patients with

1822

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

high-risk hematologic malignancies who lack conventional donors is well tolerated and produces excellent relapse-free survival: results of a prospective phase II trial. Biol Blood Marrow Transplant. 2012;18(12): 1859-1866. Solomon SR, Sizemore CA, Sanacore M, et al. Total body irradiation-based myeloablative haploidentical stem cell transplantation is a safe and effective alternative to unrelated donor transplantation in patients without matched sibling donors. Biol Blood Marrow Transplant. 2015;21(7):1299-1307. Rubio MT, Savani BN, Labopin M, et al. Impact of conditioning intensity in T-replete haplo-identical stem cell transplantation for acute leukemia: a report from the acute leukemia working party of the EBMT. J Hematol Oncol. 2016;9:25. Ruggeri A, Labopin M, Bacigalupo A, Gulbas Z, Koc Y, D B. Use of bone marrow or peripheral blood stem cell grafts in non T cell depleted haploidentical transplants using post-transplant cyclophosphamide, an ALWP-EBMT analysis. 58th American Society of Hematology; 2016; San Diego, CA; 2016. Peccatori J, Forcina A, Clerici D, et al. Sirolimus-based graft-versus-host disease prophylaxis promotes the in vivo expansion of regulatory T cells and permits peripheral blood stem cell transplantation from haploidentical donors. Leukemia. 2015;29(2):396-405. Cieri N, Greco R, Crucitti L, et al. Post-transplantation cyclophosphamide and sirolimus after conditioning and peripheral blood stem cells. Biol Blood Marrow Transplant. 2015;21(8):1506-1514. Castagna L, Crocchiolo R, Furst S, et al. Bone marrow compared with peripheral blood stem cells for haploidentical transplantation with a nonmyeloablative conditioning regimen and post-transplantation cyclophosphamide. Biol Blood Marrow Transplant. 2014;20(5):724-729. O'Donnell PV, Eapen M, Horowitz MM, et al. Comparable outcomes with marrow or peripheral blood as stem cell sources for hematopoietic cell transplantation from haploidentical donors after non-ablative conditioning: a matched-pair analysis. Bone Marrow Transplant. 2016;51(12):1599-1601. Kasamon YL, Bolanos-Meade J, Prince GT, et al. Outcomes of nonmyeloablative HLAhaploidentical blood or marrow transplantation with high-dose post-transplantation cyclophosphamide in older adults. J Clin Oncol. 2015;33(28):3152-3161. Blaise D, Furst S, Crocchiolo R, et al. Haploidentical T cell-replete transplantation with post-transplantation cyclophosphamide for patients in or above the sixth decade of age compared with allogeneic hematopoietic stem cell transplantation from a human leukocyte antigen-matched related or unrelated donor. Biol Blood Marrow Transplant. 2016;22(1):119-124. Canaani J, Savani BN, Labopin M, et al. Impact of ABO incompatibility on patients' outcome after haploidentical hematopoietic stem cell transplantation for acute myeloid leukemia - a report from the Acute Leukemia Working Party of the EBMT. Haematologica. 2017;102(6):1066-1074. Walter RB, Gooley TA, Wood BL, et al. Impact of pretransplantation minimal residual disease, as detected by multiparametric flow cytometry, on outcome of myeloabla-

tive hematopoietic cell transplantation for acute myeloid leukemia. J Clin Oncol. 2011;29(9):1190-1197. 98. Walter RB, Buckley SA, Pagel JM, et al. Significance of minimal residual disease before myeloablative allogeneic hematopoietic cell transplantation for AML in first and second complete remission. Blood. 2013;122 (10):1813-1821. 99. Ivey A, Hills RK, Simpson MA, et al. Assessment of minimal residual disease in standard-risk AML. N Engl J Med. 2016;374(5):422-433. 100. Anthias C, Dignan FL, Morilla R, et al. Pretransplant MRD predicts outcome following reduced-intensity and myeloablative allogeneic hemopoietic SCT in AML. Bone Marrow Transplant. 2014;49(5):679-683. 101. Buccisano F, Maurillo L, Piciocchi A et al. Variable outcome of allogeneic stem cell transplant according to the different levels of pre-transplant minimal residual disease, in adult patients with acute myeloid leukemia. Blood. 2015;126(23):3230. 102. Wang Y, Liu DH, Liu KY, et al. Impact of pretransplantation risk factors on post transplantation outcome of patients with acute myeloid leukemia in remission after haploidentical hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2013;19(2):283-290. 103. Zhao XS, Qin YZ, Liu YR, et al. The impact of minimal residual disease prior to unmanipulated haploidentical hematopoietic stem cell transplantation in patients with acute myeloid leukemia in complete remission. Leuk Lymphoma. 2017;58(5):11351143. 104. Bachegowda LS, Saliba RM, Ramlal R, et al. Predictive model for survival in patients with AML/MDS receiving haploidentical stem cell transplantation. Blood. 2017;129(22):3031-3033. 105. Apperley J, Niederwieser D, Huang XJ, et al. Haploidentical hematopoietic stem cell transplantation: a global overview comparing Asia, the European Union, and the United States. Biol Blood Marrow Transplant. 2016;22(1):23-26. 106. Aversa F, Prezioso L, Manfra I, Galaverna F, Spolzino A, Monti A. Immunity to infections after haploidentical hematopoietic stem cell transplantation. Mediterr J Hematol Infect Dis. 2016;8(1):e2016057. 107. Lin CJ, Vader JM, Slade M, DiPersio JF, Westervelt P, Romee R. Cardiomyopathy in patients after posttransplant cyclophosphamide-based hematopoietic cell transplantation. Cancer. 2017;123(10):1800-1809. 108. Majzner RG, Mogri H, Varadhan R, et al. Post-transplantation cyclophosphamide after bone marrow transplantation is not associated with an increased risk of donorderived malignancy. Biol Blood Marrow Transplant. 2017;23(4):612-617. 109. Chang YJ, Luznik L, Fuchs EJ, Huang XJ. How do we choose the best donor for Tcell-replete, HLA-haploidentical transplantation? J Hematol Oncol. 2016;9:35. 110. Gorin NC, Labopin M, Piemontese S, et al. T-cell-replete haploidentical transplantation versus autologous stem cell transplantation in adult acute leukemia: a matched pair analysis. Haematologica. 2015;100(4): 558564. 111. Kanakry CG, Fuchs EJ, Luznik L. Modern approaches to HLA-haploidentical blood or marrow transplantation. Nat Rev Clin Oncol. 2016;13(1):10-24.

haematologica | 2017; 102(11)

ARTICLE

Red Cell Biology & Its Disorders

Clinical risks and healthcare utilization of hematopoietic cell transplantation for sickle cell disease in the USA using merged databases

Staci D. Arnold,1 Ruta Brazauskas,2,3 Naya He,2 Yimei Li,4 Richard Aplenc,4 Zhezhen Jin,41 Matt Hall,5 Yoshiko Atsuta,6,7 Jignesh Dalal,8 Theresa Hahn,9 Nandita Khera,10 Carmem Bonfim,11 Navneet S. Majhail,12 Miguel Angel Diaz,13 Cesar O. Freytes,14 William A. Wood,15 Bipin N. Savani,16 Rammurti T. Kamble,17 Susan Parsons,18 Ibrahim Ahmed,8 Keith Sullivan,19 Sara Beattie,20 Christopher Dandoy,21 Reinhold Munker,22 Susana Marino,23 Menachem Bitan,24 Hisham Abdel-Azim,25 Mahmoud Aljurf,26 Richard F. Olsson,27,28 Sarita Joshi,29 Dave Buchbinder,30 Michael J. Eckrich,31 Shahrukh Hashmi,26,32 Hillard Lazarus,33 David I. Marks,34 Amir Steinberg,35 Ayman Saad,36 Usama Gergis,37 Lakshmanan Krishnamurti,1 Allistair Abraham,38 Hemalatha G. Rangarajan,29 Mark Walters,39 Joseph Lipscomb,40 Wael Saber2,* and Prakash Satwani5,*

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1823-1832

*Co-senior Authors

1 Emory University Hospital, Atlanta, GA, USA; 2CIBMTR (Center for International Blood and Marrow Transplant Research), Department of Medicine, Medical College of Wisconsin, Milwaukee, WI, USA; 3Division of Biostatistics, Institute for Health and Society, Medical College of Wisconsin, Milwaukee, WI, USA; 4University of Pennsylvania, Philadelphia, PA, USA; 5Division of Pediatric Hematology, Oncology and Stem Cell Transplantation, Department of Pediatrics, Columbia University Medical Center, New York, NY, USA; 6 Japanese Data Center for Hematopoietic Cell Transplantation, Nagoya, Japan; 7Nagoya University Graduate School of Medicine, Japan; 8Rainbow Babies & Children’s Hospital, Cleveland, OH, USA; 9Department of Medicine, Roswell Park Cancer Institute, Buffalo, NY, USA; 10Department of Hematology/Oncology, Mayo Clinic, Phoenix, AZ, USA; 11Hospital de Clinicas-Federal University of Parana, Curitiba, Brazil; 12Blood & Marrow Transplant Program, Cleveland Clinic Taussig Cancer Institute, Cleveland, OH, USA; 13Department of Hematology/Oncology, Hospital Infantil Universitario Nino Jesus, Madrid, Spain; 14Texas Transplant Institute, San Antonio, TX, USA; 15Division of Hematology/Oncology, Department of Medicine, University of North Carolina, Chapel Hill, NC, USA; 16Division of Hematology/Oncology, Department of Medicine, Vanderbilt University Medical Center, Nashville, TN, USA; 17Divsion of Hematology and Oncology, Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX, USA; 18Tufts Medical Center, Boston, MA, USA; 19Duke University Medical Center, Durham, NC, USA; 20University of Ottawa, Canada; 21Cincinnati Children’s Hospital Medical Center, OH, USA; 22Section of Hematology/Oncology, Department of Internal Medicine, Louisiana State University Health Shreveport, LA, USA; 23University of Chicago Hospitals, Chicago, IL, USA; 24Department of Pediatric Hematology/Oncology, Tel-Aviv Sourasky Medical Center, Israel; 25Division of Hematology, Oncology and Blood & Marrow Transplantation, Children’s Hospital of Los Angeles, University of Southern California Keck School of Medicine, CA, USA; 26 Department of Oncology, King Faisal Specialist Hospital Center & Research, Riydah, Saudi Arabia; 27Division of Therapeutic Immunology, Department of Laboratory Medicine, Karolinksa Institutet, Stockholm, Sweden; 28Centre for Clinical Research Sormland, Uppsala University, Sweden; 29Pediatric Hematology, Oncology and BMT, Nationwide Children’s Hospital and Ohio State University Wexner, Columbus, OH, USA; 30Division of Pediatrics Hematology, Children’s Hospital of Orange County, Orange, CA, USA; 31Levine Children’s Hospital, Charlotte, NC, USA; 32Department of Internal Medicine, Mayo Clinic, Minneapolis, MN, USA; 33Seidman Cancer Center, University Hospitals Case Medical Center, Cleveland, OH, USA; 34Adult Bone Marrow Transplant, University Hospitals Bristol NHS Trust, Bristol, United Kingdom; 35Department of Hematology-Oncology, Mount Saini Hospital, New York, NY, USA; 36Division of Hematology/Oncology, Department of Medicine, University of Alabama at Birmingham, Birmingham, AL, USA; 37Hematologic Malignancies & Bone Marrow Transplant, Department of Medical Oncology, New York Presbyterian Hospital/Weill Cornell Medical College, New York, NY, USA; 38Division of Blood and Marrow Transplantation, Center for Cancer and Blood Disorders, Children’s National Medical Center, Washington, DC, USA; 39Children’s Hospital & Research Center Oakland, Oakland, NY, USA; 40Health Policy and Management, Rollins School of Public Health, Winship Cancer Institute, Emory University, Atlanta, GA, USA and 41Department of Biostatistics, Mailman School of Public Health, Columbia University, New York, NY, USA.

ABSTRACT

dvances in allogeneic hematopoietic cell transplantation for sickle cell disease have improved outcomes, but there is limited analysis of healthcare utilization in this setting. We hypothesized that, compared to late transplantation, early transplantation (at age <10 years) haematologica | 2017; 102(11)

Correspondence: wsaber@mcw.edu

Received: March 23, 2017. Accepted: August 10, 2017. Pre-published: August 17, 2017. doi:10.3324/haematol.2017.169581 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1823 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1823

S.D. Arnold et al.

improves outcomes and decreases healthcare utilization. We performed a retrospective study of children transplanted for sickle cell disease in the USA during 2000-2013 using two large databases. Univariate and Cox models were used to estimate associations of demographics, sickle cell disease severity, and transplant-related variables with mortality and chronic graft-versus-host disease, while Wilcoxon, Kruskal-Wallis, or linear trend tests were applied for the estimates of healthcare utilization. Among 161 patients with a 2-year overall survival rate of 90% (95% confidence interval [CI] 85-95%) mortality was significantly higher in those who underwent late transplantation versus early (hazard ratio (HR) 21, 95% CI 2.8-160.8, P=0.003) and unrelated compared to matched sibling donor transplantation (HR 5.9, 95% CI 1.7-20.2, P=0.005). Chronic graftversus host disease was significantly more frequent among those translanted late (HR 1.9, 95% CI 1.0-3.5, P=0.034) and those who received an unrelated graft (HR 2.5, 95% CI 1.2-5.4; P=0.017). Merged data for 176 patients showed that the median total adjusted transplant cost per patient was $467,747 (range: $344,029-$799,219). Healthcare utilization was lower among recipients of matched sibling donor grafts and those with low severity disease compared to those with other types of donor and disease severity types (P<0.001 and P=0.022, respectively); no association was demonstrated with late transplantation (P=0.775). Among patients with 2-year pre- and post-transplant data (n=41), early transplantation was associated with significant reductions in admissions (P<0.001), length of stay (P<0.001), and cost (P=0.008). Early transplant outcomes need to be studied prospectively in young children without severe disease and an available matched sibling to provide conclusive evidence for the superiority of this approach. Reduced post-transplant healthcare utilization inpatient care indicates that transplantation may provide a sustained decrease in healthcare costs over time.

Introduction Sickle cell disease (SCD) affects approximately 100,000 people in the United States of America (USA) with 2,000 new cases detected via newborn screening annually. There is a lack of clinical predictors to estimate overall outcomes of SCD-associated morbidities, including painful crises and organ dysfunction, which respond variably to medical management, have a devastating impact on quality of life, and can lead to premature death.1 As a result, many people with SCD are left with sequelae of the disease and its complications. Allogeneic hematopoietic cell transplantation (alloHCT) remains the only established curative option for these individuals. Despite mounting evidence of rising alloHCT success rates over time, such that the 5-year disease-free survival in children with SCD is now 92%, many still regard alloHCT as an experimental therapy, only for patients with severe disease.2,3 The indications for alloHCT remain unclear for non-transplant providers when compared to the benefits of medical management.4 In addition, a recent retrospective study from Belgium suggested that patients with SCD managed medically with hydroxyurea may have a better survival than those treated with alloHCT.5 However, short-term improvements in outcome with medical therapy must be balanced against a disease with an unpredictable clinical course and substantial impact on healthcare utilization. USA individuals with SCD account for an estimated $1.6 billion per year in healthcare costs.6 SCD ranked fifth among the top ten diagnoses of hospital stays among Medicaid super-utilizers.7 The substantial healthcare utilization and cost of SCD-related morbidity suggests that a greater focus on curative approaches for this disease is needed. AlloHCT, when successful, can be curative, but also carries the risks of death and substantial morbidity from chronic graft-versus-host disease (GvHD). In addition, the initial cost of alloHCT represents a significant financial burden of approximately $400,000 in the transplant year.8 This research investigates alloHCT for pediatric SCD using a comprehensive, systematic database analysis 1824

exploring patient-, disease-, and transplant-related variables that may reduce healthcare utilization over time while sustaining excellent clinical outcomes. The findings may provide transplant and non-transplant physicians with additional information to help choose between recommending medical therapy and alloHCT.

Methods Data sources Outcomes analysis The Center for International Blood and Marrow Transplant Research (CIBMTR) database contains alloHCT data for recipients and their donors. Data are collected prior to and at various intervals post-alloHCT. Upon CIBMTR registration, a weighted randomization scheme selects a subset of patients for more detailed data collection in comprehensive research forms (CRF) which provide more specific transplant-related data (SCD complications, pre-transplant therapy, etc.) (Online Supplementary Figure A1).

Healthcare utilization analysis CIBMTR data on all alloHCT recipients are submitted as transplant essential data (TED) (Online Supplementary Figure A2). TED forms record donor and recipient demographic, clinical, and transplant data but lack specific CRF data. The Pediatric Health Information System (PHIS; Childrenâ&#x20AC;&#x2122;s Hospital Association, Overland Park, KS, USA) records the corresponding inpatient healthcare utilization data. PHIS, a confidential database of 43 member hospitals in the USA (Online Supplementary Figure A3), has participating hospitals submit de-identified data with an encrypted medical record number for identification of readmissions at the same hospital. Institutional and patient-specific information, including patientâ&#x20AC;&#x2122;s age, date of service, visit codes, length of stay (LOS), adjusted costs, and daily billing data, are collected. The PHIS has been merged for similar research purposes including a number of recent scientific publications.9

Merging and validating datasets Patients in the PHIS database who underwent alloHCT for SCD during the study period were identified utilizing International Classification of Diseases version 9 (ICD9) and alloHCT diagnohaematologica | 2017; 102(11)

Transplant risks and utilization for SCD Table 1. Characteristics of USA pediatric patients (age ≤21) receiving first allogeneic hematopoietic cell transplant for sickle cell disease.

Variable

Number of patients Patient-related Age, median, years Age at transplant, years <10 ≥10 Gender Male Female Race African-American Other/missing Karnofsky/Lansky score prior to transplant, % >90 ≤90 Missing Disease-related Chronic transfusion No Yes Missing Hydroxyurea No Yes Missing Sickle cell related complications Stroke Acute chest syndrome Recurrent vaso-occlusive pain Transplant-related Time from diagnosis to transplant (months) Transplant indication Stroke Acute chest syndrome Recurrent vaso-occlusive pain Excessive transfusion requirements Other*/missing Conditioning regimen Myeloablative Reduced intensity Non-myeloablative Graft source Bone marrow Peripheral blood Cord blood Donor/recipient CMV match -/-/+ +/+/+ Missing Donor type Cord blood Related Unrelated HLA identical sibling Well-matched unrelated Other unrelated Missing haematologica | 2017; 102(11)

Outcomes analysis (CRF) N (%)

HCU analysis (TED/PHIS) N (%)

161

183

10 (<1-21)

9 (<1-20)

86 (53) 75 (47)

100 (55) 83 (45)

81 (50) 80 (50)

101 (55) 82 (45)

142 (88) 19 (22)

91 (57) 52 (32) 18 (11)

155 (85) 28 (15)

164 (90) 2 (1) 17 (9)

82 (51) 72 (45) 7 (4)

Not recorded Not recorded -

62 (39) 85 (53) 89 (55)

29 (16) 46 (25) 134 (73)

112 (7-242) 47 (29) 19 (12) 35 (22) 17 (11) 43 (26)

109 (10-227) Not recorded -

96 (60) 54 (33) 11(7)

121 (66) 58 (32) 4 (2)

53 (33) 103 (64) 5 (3)

97 (60) 9 (6) 55 (34)

144 (79) 4 (2) 35 (19)

47 (29) 25 (16) 31 (19) 41 (25) 17 (11)

55 (30) 20 (11) 22 (12) 51 (28) 35 (19)

55 (34) 22 (40) 33 (60) 67 (42) 27 (17) 11 (7) 1 (<1)

35 (19) 15 (43) 20 (57) 126 (69) 10 (5) 7 (4) 5 (3)

continued on next page 1825

S.D. Arnold et al. continued from the previous page Year of transplant 2000-2006 2007-2013 GvHD prophylaxis FK506 ± MMF or MTX CSA ± MMF or MTX Others**/missing Median follow-up of survivors (range), months

42 (26) 119 (74)

54 (30) 129 (70)

57 (36) 89 (55) 15 (9) 49 (3-138)

40 (22) 134 (73) 11 (5) 49 (11-145)

Well matched- 10/10 HLA match; other unrelated – mismatched or <10/10 HLA match; CRF: comprehensive research form; HCU: healthcare utilization; TED: transplant essential data; PHIS: pediatric health information system; CsA: cyclosporine, FK506: tacrolimus; CMV: cytomegalovirus; CY: cyclophosphamide; ATG: antithymocyte globulin;FLUD-fludarabine; BU: busulfan; MEL: melphalan; MMF-mycophenolate mofetil; MTX: methotrexate *abnormal transcranial Doppler(TCD)-magnetic resonance imaging angiography with narrowing of supraclinoid portions of the internal carotid arteries bilatecally (n=1); abnormal TCD (n=1); acute chest syndrome (ACS); pain; transfusions (n=1); best long-term, lifelong option for patient (n=1); both ACS and pain crisis (n=1); cardiomyopathy/pulmonary stenosis (n=1); cerebral vasculopathy (n=1); combination of ACS and pain crisis (chronically ill) (n=1); cranial vasculopathy therefore stroke prevention (n=1); cure sickle cell (n=1); develop allo antibodies, increasing hgb, decreased response to hydroxyurea (n=1); elevated transcranial Doppler (n=1); extensive complications from sickle cell (n=1); family wanted to move back to Nigeria where there is not modern care nor safe transfusion(n=1); fever, ileus and mild ACS (n=1); improved quality of life (n=1); increased frequency of pain crisis, at significant risk end organ damage anddysfunction in adulthood (n=1); liver transplant (n=1); matched sibling and hiistory of pain crisis (n=1); osteonecrosis/requiring hip replacement neuropathy-vision loss (n=1); pain;avascular necrosis; magnetic resonance imaging changes to correct SCD (n=1); parents wanted a cure for their child’s SCD (n=1); presence of silent infarcts on magnetic resonance imaging (n=1); rare disease type (n=1); SCD (n=1); **cor+methotrexate (n=1), methotrexate (n=1)

sis-related group (DRG) codes (282.6 and 1803, respectively) as well as PHIS procedure codes. These patients were identified within the CIBMTR using a probabilistic algorithm. This process occurred under the guidance of the CIBMTR via the National Marrow Donor Program institutional review board. TED data were merged with PHIS data to determine risk factors and clinical outcomes associated with healthcare utilization (Online Supplementary Figure A4). A target of 85% merge accuracy was set based on the available database population and previously published reports.9-11 Once linked, the merge accuracy was assessed by performing institutional level validation under an existing pilot institutional review board process.8 This validation confirmed 100% patient identification in this subset. SCD-related complications identified in the PHIS were validated against CRF data, where available, and showed concordance.

Determination of the severity of sickle cell disease TED/PHIS variables were used to determine SCD severity. Younger patients (age <10 years) without disease sequelae were considered low risk. Younger patients with disease sequelae or older patients (age ≥10 years) without disease sequelae were considered moderate risk. Older patients with disease sequelae or patients of any age with stroke were considered high risk. Disease sequelae were defined as any episode of acute chest syndrome, and/or three or more vaso-occlusive crises requiring hospitalization in 1 year.12

Variables and outcomes Outcomes analysis The study population consisted of children 21 years or younger who had undergone alloHCT for SCD in the USA between 20002013 and for whom CRF data were available. The CRF provided information on clinical risk factors and outcomes including overall survival, graft failure, grade II-IV acute GvHD, chronic GvHD, and GvHD-related event free survival (GREFS). GREFS was defined as the survival free of graft failure, chronic GvHD, or death and was used to better assess the post-alloHCT morbidity and associations of clinical risk factors with outcomes.

Healthcare utilization analysis The total adjusted cost reported to the PHIS is based on a fixed hospital-wide ratio of cost to charge adjusted by geographical location. Adjusted costs for each service unit or department (clini1826

cal, pharmacy, imaging, etc.) were reported using service-specific ratios of cost to charge. Charges in the PHIS database were adjusted for the wage and price index (published annually in the Federal Register) and reported from the hospital perspective. Total adjusted costs were determined for all inpatient admissions for each patient and include direct medical costs, excluding provider fees, incurred. Indirect costs, outpatient costs, and costs incurred at non-PHIS hospitals were not captured. Adjusted cost data only were analyzed as the primary outcome of interest because charges and reimbursements vary across each institution and state. Patients without available adjusted cost data were excluded (n=7). Zero-dollar research or study-related costs, reflecting largely workup, medication, or laboratory-related account credits, were included in the analysis. Additional healthcare utilization outcomes included number of admissions and LOS. PHIS data were used for descriptive analyses of the selected cohort of patients throughout all the study periods. Healthcare utilization for the initial alloHCT admission (conditioning to first recorded discharge) and alloHCT year (conditioning to day +365) was described and considered separately in the analyses. Likewise, the pre-alloHCT period (2 years preceding transplant through to the day of transplant conditioning) and the post-alloHCT period (2 years from day +366 onward) were analyzed separately (Online Supplementary Figure B).

Analysis of allogeneic hematopoietic cell transplant year Factors influencing healthcare utilization during the alloHCT year were analyzed using the following TED clinical variables age at transplant, gender, performance status, recipient cytomegalovirus status, income level, insurance, distance from center, SCD complications, donor type, graft source, conditioning regimens, and transplant year. A secondary analysis of disease severity and healthcare utilization was also performed.

Pre- and post-allogeneic hematopoietic cell transplantation comparison To standardize costs for comparisons, the total adjusted cost per 30 hospital days was calculated for each patient with both pre- and post-alloHCT inpatient admissions and used as the primary healthcare utilization outcome. The change in an individual patient's healthcare utilization pre- and post-alloHCT was compared. This change in healthcare utilization was also analyzed by disease severity. haematologica | 2017; 102(11)

Transplant risks and utilization for SCD

Table 2. Cox regression model of outcomes with patient- and transplant- related variables as reported in CRF data (n=161).

Variable Age (≥10 years) Gender (female) Performance status (<90%) SCD complications (>2) Hydroxyurea Chronic transfusion CMV+ recipient MUD vs. MSD CBT vs. MSD CSA vs. FK506 prophylaxis AlloHCT after 2006 NMA/RIC vs. MA

Mortality (n=16) HR (95% CI)

Graft failure (n=3) HR (95% CI)

aGVHD (n=33) HR (95% CI)

cGVHD (n=44) HR (95% CI)

GREFS (n=54) HR (95% CI)

*21.21 (2.80-160.76) 3.09 (1.00-9.57) 1.03 (0.34-3.06) 1.09 (0.38-3.13) 2.71 (0.92-7.93) 1.96 (0.55-6.94) 1.59 (0.56-4.46) *5.88 (1.71-20.19) 0.92 (0.20-4.1) *0.33 (0.12-0.91) 2.90 (0.66-12.78) 1.52 (0.57-4.06)

*1.61 (1.17-2.21) 1.21 (0.88-1.66) 1.11 (0.78-1.58) 1.09 (0.78-1.54) 1.21 (0.89-1.69) 0.8 (0.57-1.12) 0.92 (0.67-1.26) 1.93 (1.22-3.06) *0.50 (0.34-0.74) 1.02 (0.72-1.43) 1.11 (0.77-1.58) 1.27 (0.92-1.76)

*2.63 (1.07-6.45) 1.87 (0.78-4.46) 1.74 (0.75-4.01) 0.79 (0.33-1.91) 1.72 (0.73-4.03) 1.13 (0.46-2.76) *2.75 (1.06-7.09) *4.36 (1.43-13.34) 1.56 (0.47-5.11) 0.68 (0.29-1.6) 3.75 (0.88-16.03) 1.17 (0.51-2.71)

*1.92 (1.05-3.50) 1.49 (0.82-2.73) 1.67 (0.91-3.06) 1.89 (0.96-3.74) 1.62 (0.87-3.02) 1.11 (0.58-2.13) 1.31 (0.72-2.39) *2.53 (1.18-5.41) 0.94 (0.43-2.05) *0.48 (0.26-0.88) *2.81 (1.18-6.65) 1.26 (0.70-2.28)

*2.2 (1.26-3.82) 1.65 (0.95-2.85) 1.57 (0.90-2.74) 1.64 (0.90-3.01) *1.77 (1.00-3.12) 1.21 (0.66-2.2) 1.37 (0.80-2.36) *3.00 (1.51-5.96) 1.07 (0.52-2.16) *0.49 (0.28-0.86) *2.25 (1.10-4.61) 1.03 (0.60-1.77)

CRF: comprehensive research form; aGvHD: grade 3-4 acute graft-versus-host disease; cGvHD: chronic graft-versus-host disease; GREFS: the survival free of graft failure, chronic graft-versus-host disease, or death; SCD: sickle cell disease; CMV: cytomegalovirus; MUD: matched unrelated donor; MSD: matched sibling donor; CSA: cyclosporina; alloHCT: allogeneic hematopoietic cell transplant; NMA: non-myeloablative; RIC: reduced intensity conditioning; MA: myeloablative conditioning. *P≤0.05

Statistical analysis Outcomes analysis After documenting descriptive statistics, Cox proportional hazards modeling of CRF level data determined the impact of risk factors on alloHCT outcomes. Due to low event rates and small sample sizes, only bivariate analyses involving one explanatory variable at a time were performed. The cumulative incidences of acute GvHD and chronic GvHD were calculated using a competing risk framework.

Healthcare utilization analysis TED/PHIS data were evaluated to identify whether demographic factors, SCD severity, or alloHCT variables correlate with healthcare utilization changes during the alloHCT year using Wilcoxon, Kruskal-Wallis, or linear trend tests. Healthcare utilization pre- and post-alloHCT was compared using the Wilcoxon signed rank test for continuous variables and the McNemar test for binary variables. Finally, the impact of SCD severity on healthcare utilization across these two time periods was examined by Poisson regression for number of visits and LOS. Total adjusted cost were analyzed by linear regression. All analyses were performed using SAS version 9.3 statistical software (Cary, NC, USA).

(64%). The most common source of a graft for transplantation was a matched sibling donor (MSD) (42%) and the majority of patients received myeloablative conditioning (60%).

Transplant outcomes The 2-year overall survival was 90% [95% confidence interval (CI): 85-95%]: 96% (95% CI: 89-100%) for cord blood transplant (CBT), 94% (95% CI: 86-98%) for MSD transplants, and 74% (95% CI: 54-90%) for transplants from well-matched unrelated donors (MUD) (P=0.002) (Online Supplementary Table B and Online Supplementary Figure C). All 16 deaths occurred among children with prealloHCT complications of SCD and were due to organ failure (37.5%), infections (25%), GvHD (6.2%), and other/unknown causes (31.2%) (Online Supplementary Table C). The majority of deaths (62.5%) occurred during the alloHCT year; six patients died after day +365 (2 from organ failure, 1 from infection, 1 from sickle cell-associated vasculopathy, and 2 from a missing/other unspecified cause). The cumulative incidence of acute GvHD at day 100 was 14% (95% CI: 9-20%), and chronic GvHD developed in 31% (95% CI: 23-38%) at 2 years. Of those with chronic GvHD, 64% had extensive disease with over half having a MSD (n=9) or MUD (n=11). The 2-year GREFS was 64% (95% CI: 56-71%).

Results Bivariate analysis

Outcomes analysis Demographics CRF data were available for 161 patients with a median age of 10 years (range <1-21) of whom 50% were female (Table 1; Online Supplementary Table A). The majority (84%) of patients had the HbSS genotype. The most commonly documented transplant indication was stroke (29%) followed by recurrent vaso-occlusive crises (22%). However, 39% had stroke as a documented SCD complication, 55% vaso-occlusive crises, and 53% acute chest syndrome. The majority of patients reported use of medical therapy, hydroxyurea (45%) and chronic transfusions haematologica | 2017; 102(11)

Age ≥10 years and use of a MUD showed significant negative associations with outcomes (Table 2). The use of cyclosporine-A prophylaxis and year of alloHCT exhibited significant associations with chronic GvHD-related outcomes (Table 2).

Healthcare utilization analysis Demographics Combined TED/PHIS data were available for 183 patients with a median age of 9 years (range: <1-20) of whom 45% were female (Table 1). With regards to SCD complications, 73% had vaso-occlusive crises, 25% had 1827

S.D. Arnold et al.

Figure 1. Comparison of 2-year pre- and post-allogeneic hematopoietic cell transplant inpatient healthcare utilization per 30 days as reported in the Pediatric Health Information System (n=41). PrealloHCT: the 2 years preceding transplant through to the day of transplant conditioning; Post-alloHCT: day +366 onward; alloHCT: hematopoietic cell transplant; PHIS: Pediatric Health Information System: visits: inpatient admissions; LOS: length of stay.

acute chest syndrome, and 16% had a stroke prior to alloHCT. This translated into 19.7% with low severity disease, 36.6% with moderate severity, and 43.7% with high severity disease (Online Supplementary Table D). The majority of patients received MSD alloHCT (69%) and a myeloablative regimen (66%). Complete cost data were available for 176 of the 183 patients.

Allogeneic hematopoietic cell transplantation admission The median total adjusted cost per patient was $380,320 [interquartile range (IQR): $297,710-$563,462] with a median LOS of 39.5 days (IQR: 31-53) (Table 3). The highest costs were associated with room and nursing charges followed by pharmacy costs (Online Supplementary Figure D).

decreased to $33,112 (IQR: $14,291-$161,959) postalloHCT (Table 3). The highest adjusted costs during these periods were associated with room and nursing charges followed by laboratory costs (Online Supplementary Figure D). Only 41 patients were admitted to the hospital during both pre- and post-transplant periods: healthcare utilization was significantly reduced with total adjusted cost per 30 days decreasing from $9,393 (IQR: $4,595-$31,291) to $1,873 (IQR: $571-$6,504) (P=0.008) (Figure 1). Of these, 30 patients had high severity disease and had a significant reduction in healthcare utilization (admissions P<0.001, LOS P<0.001, cost P=0.002) (Online Supplementary Table E).

Discussion Allogeneic hematopoietic cell transplantation year The median total adjusted cost per patient was $467,747 (IQR: $344,029-$799,219) with a median LOS of 52 days (IQR: 38-72) (Table 3). Again, the highest costs were associated with room and nursing charges followed by pharmacy costs (Online Supplementary Figure D). During the alloHCT year, age â&#x2030;Ľ10 years and total adjusted cost per 30 days were not significantly associated (P=0.775) (Table 4). Total adjusted cost was also not associated with income level, insurance type, or distance from transplant center (P=0.417, 0.918, and 0.253, respectively). The total adjusted cost per 30 days was lower for MSD transplants than for CBT or MUD transplants (P<0.001). CBT was associated with a higher total adjusted cost per 30 days compared to bone marrow transplants (P=0.004). Increased total adjusted cost per 30 days were associated with prior stroke (P=0.004) and vaso-occlusive crises (P=0.009) but not acute chest syndrome (P=0.291). Overall, total adjusted cost per 30 days increased with SCD severity (P=0.022).

Pre- and post-allogeneic hematopoietic cell transplantation Two-year inpatient healthcare utilization data were available for 134 pre-alloHCT patients and 45 postalloHCT. The median total adjusted cost per patient was $56,416 (IQR: $32,848-$103,270) pre-alloHCT; this 1828

Substantial advances in alloHCT for SCD have been made in the past two decades; nevertheless, the eligibility of patients, especially those without severe disease, remains controversial. This study provides additional insight into eligibility as alloHCT outcomes were favorably linked to age and donor type suggesting that early alloHCT, before the age of 10 years, and MSD alloHCT have optimal outcomes with the latter also showing a significant healthcare utilization advantage. These outcomes are in keeping with a recent report from an international study of adult and pediatric MSD recipients describing lower event-free survival rates with increasing age at transplantation and support the recent expert panel recommendation of early alloHCT, prior to the onset of SCD complications, for children with SCD and an available MSD.13,14 MSD alloHCT is not, however, a viable option for many patients, as fewer than 25% will have a suitable HLAmatched donor, necessitating MUD alloHCT for severe disease. Among patients undergoing MUD alloHCT, higher GvHD risk could be hypothesized as the etiology of poorer outcomes and increased healthcare utilization of older recipients.13,15-17 Although the GvHD-related mortality rate in this study is lower than that in published reports, the low percentage of deaths after day +365 may mitigate haematologica | 2017; 102(11)

Transplant risks and utilization for SCD

Table 3. Inpatient healthcare utilization analysis as reported over time in the PHIS.

Time period

Pre-alloHCT (n=134) AlloHCT admission (n=176) AlloHCT year (n=176) Post-alloHCT (n=45)

Median number of inpatient admissions per patient (IQR)

Median length of stay per patient [days (IQR)]

Median total adjusted cost per patient (IQR)

3 (1-6)

9 (4-18)

40 (32-53)

2 (1-4)

52 (39-73)

$56,416 (32,848-103,270) $380,320 (297,711-563,462) $467,747 (344,029-799,220)

2 (1-4)

8 (3-17)

$33,112 (14,291-161,960)

PHIS: Pediatric Health Information System; alloHCT: allogeneic hematopoietic cell transplant; Pre-alloHCT - the 2 years preceding transplantation through to the day of transplant conditioning; AlloHCT admission: conditioning to first recorded discharge; AlloHCT year: conditioning to day +365, Post-alloHCT: day +366 onward; IQR: interquartile range.

the mortality associated with chronic GvHD. However, the 26% mortality among the MUD alloHCT patients remains well over the 5% mortality threshold accepted by most parents and adolescents for the cure of SCD, as recently published.18 Efforts are still needed to elucidate more precisely the cause of death (unknown cause in 31%) and transplant-related mortality, especially among patients who died after the alloHCT year, in order to improve transplant procedures to prevent these complications. Our findings support the current practice of restricting MUD alloHCT to individuals with severe disease.19 The results of the analysis of donor type were consistent with published reports and indicated that the overall survival of patients treated with CBT is similar to that of recipients of MSD bone marrow; however, healthcare utilization was higher with CBT.20 Many of the drawbacks of CBT have been associated with insufficient cell dose. Ongoing research into cord blood expansion may mitigate this limitation and the associated healthcare utilization by reducing the delay in engraftment.21-23 However, this cohort of both related and unrelated CBT showed no statistical difference in GREFS or other GvHD outcomes, with a lower cost compared to MUD, suggesting that more analysis is needed to determine the optimal donor type if a MSD is unavailable. Donor type and SCD severity had a significant impact on both outcomes and healthcare utilization, such that patients with high severity disease and MUD had poorer outcomes and increased healthcare utilization. The correlation of healthcare utilization and disease severity, not age, is unclear but may suggest that healthcare utilization is linked to management of persistent SCD complications after alloHCT. Although some end-organ disease is reversible after alloHCT, complications of chronic lung disease and pain can persist in the first year postalloHCT.24 These complications correlate with increased healthcare utilization in individuals with SCD.25 In addition, severe disease remains an indication for MUD alloHCT research trials suggesting this as a possible confounder in healthcare utilization analysis. MUD outcomes associated with degree of HLA matching, supportive care measures, and infection must also be considered in more detail.16 Understanding and mitigating risk factors associated with poor outcomes and increased healthcare utilizahaematologica | 2017; 102(11)

tion following MUD alloHCT is needed because improvements in unrelated alloHCT clinical outcomes have the potential to have the greatest clinical and financial impact. The healthcare utilization analysis also described variations over time with a subset of patients having a significant reduction in healthcare utilization pre- and postalloHCT. However, the current sample size and/or 2-year time period may not be sufficient to document a change in inpatient healthcare utilization for the entire population of patients. Donor type variations (e.g., CBT) and disease severity within our small sample size likely have a role; more robust analysis is ongoing to understand this phenomenon better. In addition, outpatient healthcare utilization was not described which may account for the substantial reduction of data available for pre- and postalloHCT analysis. However, previous publications on healthcare utilization in this population indicated that outpatient costs remained flat pre- and post-alloHCT.8

Limitations This is a retrospective study and analysis is therefore limited to variables and data collected by the CIBMTR at the time of alloHCT. This limitation is somewhat mitigated by using multiple sources of data (PHIS and CIBMTR, both TED and CRF forms) to increase sample size and data availability or quality. Retrospective studies also do not allow control of exposures (pre-alloHCT treatment, conditioning, etc.) which may influence outcomes. Bivariate and multivariate analyses can elucidate cofounders; however, due to sample size and low event rates, multivariate analysis was not performed in this study. However, univariate analysis was used to document the impact of these exposures on outcomes. The retrospective nature of this database study also does not allow for comparison to controls with SCD who have not been transplanted or incorporation of prospective metrics including quality of life in the analysis. Previous studies have documented the quality of life improvements gained after alloHCT for SCD.26,27 Other studies have described significant quality of life differences between non-transplant interventions such as chronic transfusion and hydroxyurea.1,28 Outcomes and cost of care for children have been well documented, with a recent analysis of management with hydroxyurea based 1829

S.D. Arnold et al.

on Medicaid claims data suggesting that lifetime costs of care may be influenced by the age of hydroxyurea initiation.5,29 These data largely demonstrate the substantial lifetime costs associated with caring for individuals with SCD and the need to compare costs across the various available treatment options.6 Collectively these findings suggest a multicenter case control study incorporating quality of life is needed to truly understand the full impact of alloHCT on outcomes and healthcare utilization. Prospective studies, including STRIDE2 (NCT02766465), are underway to fill this gap in knowledge of alloHCT for SCD. The study includes a population that is heterogeneous for donor type, graft source, and conditioning regimen. Analysis of this complex group reflects the current clinical paradigm and its potential to influence healthcare utilization; however, diversity does introduce confounders. A recent study showed that conditioning regimen does not influence outcomes, and graft source only influences overall survival, while age and year of transplant influence both overall and event-free survival.13 Survival outcomes of this cohort are similar to this and other previously published estimates; these findings suggest that the heterogeneity of our cohort had limited influence on the outcomes analysis. The study focuses on a pediatric population which is unique in that children are largely less affected by the disease than their adult counterparts and alloHCT has quite a different long-term impact. Certainly, efforts to offer early MSD alloHCT will have a significant effect on adult care; in the meantime, a large number of adults live with SCD. Therefore, efforts to better understand the cost of “late alloHCT” are needed. Specifically, the combination of favorable outcomes of CBT and advances in cord blood expansion technology may make this a more viable option for adults.21-23 Although, this study excluded data on haploidentical transplants, promising clinical outcomes to date suggest this is as another viable means of expanding the donor pool for adults and children with SCD.24 Ultimately, as the field advances with changes in conditioning regimens, modified donor source options, and the advent of gene therapy, ongoing analysis of outcomes and cost will be needed. Studies of alloHCT for SCD are also limited by the absence of sufficient data on late effects. The risk of impaired fertility and long-term quality of life are not well described. Many studies have documented the impact of alloHCT on sperm production and ovarian failure in this population, but the actual fertility risk remains unclear particularly in light of the use of newer reduced intensity regimens that are less gonadotoxic.3,17,30 However, a recent study of patients’ and parents’ attitudes toward alloHCT documented that 56% were willing to accept infertility18 suggesting the potential for cure may far outweigh the risk of this complication. At the same time, another recent study of alloHCT not only reported the impact on fertility but also on sexual function and other patient-reported outcomes. In a cohort of individuals studied at least 10 years after transplantation for malignant disease, reports of sexual problems, restrictions in social function, memory and attention concerns, denial of life and health insurance were significant.31 Future studies should determine late outcomes after transplantation for SCD, collecting data not only on end-organ complications but also patientreported outcomes including organ function and quality of life evaluations for both alloHCT recipients and controls. 1830

Finally, this study involved a USA population and its findings cannot be easily extrapolated to a global population. However, recent publications have documented the impact of alloHCT throughout the developing world. AlloHCT outcomes in India and Mexico have been described as nearly equivalent to those in the developed world but are performed at substantially lower costs.32,33 International experience suggests that further study of USA transplant approaches to learn potential cost efficiencies from the global experience will make alloHCT more viable for SCD and other diseases throughout the world.

Table 4. Comparison of inpatient total adjusted costs per 30 days by TED/PHIS patient and transplant variables during the alloHCT year (n=176).

Variables

Patient-related Age <10 years ≥10 years Gender Male Female Performance status <80% ≥80% CMV+ recipient Yes No History of stroke Yes No History of VOC Yes No History of ACS Yes No SCD severity High Low Transplant-related Donor type Matched sibling Well-matched unrelated Cord blood Graft source Cord blood Bone marrow Conditioning Non-myeloablative RIC (N=1) Myeloablative AlloHCT year Before 2006 After 2006

Median Total Adjusted Cost (IQR, $ per 30 days)

0.775 121,506 (89,377-180,961) 128,731 (90,120-167,746) 0.263 119,492 (83,738-172,825) 132,841 (93,741-172,900) 0.529 106,761 (93,741-119,781) 126,343 (90,086-171,574) 0.061 132,768 (104,432-190,135) 122,517 (83,541-156,185) <0.001 167,746 (102,818-233,031) 121,506 (87,291-157,945) 0.009 135,464 (93,552-187,355) 111,261 (83,928-142,718) 0.291 132,841 (93,552-186,437) 123,088 (84,321-171,574) 0.022 130,380 (93,552-201,784) 112,565 (90,052-144,859) <0.001 112,835 (85,640-145,895) 246,903 (151,522-279,149) 170,322 (102,034-217,186) 0.004 170,322 (102,034-217,186) 120,331 (88,195-153,998) 0.953 119,754 (75,011-210,103) 53,319 126,993 (93,746-160,187) 0.001 97,416 (69,488-131,898) 135,464 (102,818-190,135)

TED: transplant essential data; PHIS: Pediatric Health Information System; alloHCT: allogeneic hematopoietic cell transplant; IQR: interquartile range;VOC: vaso-occlusive crises; ACS: acute chest syndrome; RIC: reduced intensity conditioning.

haematologica | 2017; 102(11)

Transplant risks and utilization for SCD

Conclusion Performing analyses of both clinical and financial outcomes is challenging because of the lack of a single, exhaustive data source. The data merging process is invaluable in that it successfully utilizes existing datasets and combines them to create the largest and most accurate surrogate for assessment in a comprehensive manner. This provides proof of principle for this methodology and type of analysis and builds a foundation for future research in the field. Specifically, the superior clinical outcomes among children <10 years old lay the basis for prospective studies among low-risk SCD patients for whom a MSD is available. In addition, alloHCT, although costly, can provide a sustained decrease in healthcare utilization for patients over time. These results and future studies will provide guidance and insight into health policy determinations for the optimal use of this curative therapy for children with SCD. Acknowledgments Special acknowledgments to our patients and families. The numerous mentors and advisors who helped develop and support this project. Dr. Arnold was supported in part by a Robert Wood Johnson Foundation Harold Amos Medical Faculty Development Program award. Dr. Aplenc was supported by 1R01CA166581. CIBMTR Support List: the CIBMTR is supported primarily by Public Health Service Grant/Cooperative Agreement 5U24CA076518 from the National Cancer Institute (NCI), the 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

References 1. Panepinto JA, Bonner M. Health-related quality of life in sickle cell disease: past, present, and future. Pediatr Blood Cancer. 2012;59(2):377-385. 2. Gluckman E. Allogeneic transplantation strategies including haploidentical transplantation in sickle cell disease. Hematology Am Soc Hematol Educ Program. 2013;2013:370-376. 3. Walters MC, De Castro LM, Sullivan KM, et al. Indications and results of HLA-identical sibling hematopoietic cell transplantation for sickle cell disease. Biol Blood Marrow Transplant. 2016; 22(2):207-211. 4. Mikles B, Bhatia M, Oyeku SO, Green NS. Hematology provider perspectives on hematopoietic stem cell transplantation for pediatric sickle cell disease. Blood. 2012;120(21):4276-4276. 5. Lê PQ, Gulbis B, Dedeken L, et al. Survival among children and adults with sickle cell disease in Belgium: benefit from hydroxyurea treatment. Pediat Blood Cancer. 2015;62(11):1956-1961. 6. Kauf TL, Coates TD, Huazhi L, ModyPatel N, Hartzema AG. The cost of health care for children and adults with sickle cell disease. Am J Hematol. 2009;84(6):323327. 7. Jiang HJ, Weiss AJ, Barrett ML, Sheng M. Characteristics of hospital stays for super-

haematologica | 2017; 102(11)

10.

11.

12.

Resources and Services Administration (HRSA/DHHS); two Grants N00014-15-1-0848 and N00014-16-1-2020 from the Office of Naval Research; and grants from *Actinium Pharmaceuticals, Inc.; Alexion; *Amgen, Inc.; Anonymous donation to the Medical College of Wisconsin; Astellas Pharma US; AstraZeneca; Atara Biotherapeutics, Inc.; Be the Match Foundation; *Bluebird Bio, Inc.; *Bristol Myers Squibb Oncology; *Celgene Corporation; Cellular Dynamics International, Inc.; Cerus 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; Janssen Scientific Affairs, LLC; *Jazz Pharmaceuticals, Inc.; Jeff Gordon Children’s Foundation; The Leukemia & Lymphoma Society; Medac, GmbH; MedImmune; The Medical College of Wisconsin; *Merck & Co, Inc.; *Mesoblast; MesoScale Diagnostics, Inc.; *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; PCORI; Perkin Elmer, Inc.; Pfizer, Inc; *Sanofi US; *Seattle Genetics; *Spectrum Pharmaceuticals, Inc.; St. Baldrick’s Foundation; *Sunesis Pharmaceuticals, Inc.; Swedish Orphan Biovitrum, Inc.; Takeda Oncology; Telomere Diagnostics, 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. *Corporate Members.

utilizers by payer, 2012: Statistical Brief #190. Healthcare Cost and Utilization Project (HCUP) Statistical Briefs. Rockville (MD) 2015. Arnold SD, Jin Z, Sands S, Bhatia M, Kung AL, Satwani P. Allogeneic hematopoietic cell transplantation for children with sickle cell disease is beneficial and cost-effective: a single-center analysis. Biol Blood Marrow Transplant. 2015;21(7):1258-1265. Berry JG, Hall DE, Kuo DZ, et al. Hospital utilization and characteristics of patients experiencing recurrent readmissions within children's hospitals. JAMA. 2011; 305(7):682-690. Li Y, Hall M, Fisher BT, et al. Merging Children's Oncology Group data with an external administrative database using indirect patient identifiers: a report from the Children's Oncology Group. PLoS One. 2015;10(11):e0143480. Aplenc R, Fisher BT, Huang YS, et al. Merging of the National Cancer Institutefunded cooperative oncology group data with an administrative data source to develop a more effective platform for clinical trial analysis and comparative effectiveness research: a report from the Children's Oncology Group. Pharmacoepidemiol Drug Saf. 2012;21 (Suppl 2):37-43. Shenoy S. Has stem cell transplantation come of age in the treatment of sickle cell disease? Bone Marrow Transplant. 2007;40 (9):813-821.

13. Gluckman E, Cappelli B, Bernaudin F, et al. Sickle cell disease: an international survey of results of HLA-identical sibling hematopoietic stem cell transplantation. Blood. 2016;129(11):1548-1556. 14. Angelucci E, Matthes-Martin S, Baronciani D, et al. Hematopoietic stem cell transplantation in thalassemia major and sickle cell disease: indications and management recommendations from an international expert panel. Haematologica. 2014; 99(5):811-820. 15. Lee SJ, Klar N, Weeks JC, Antin JH. Predicting costs of stem-cell transplantation. J Clin Oncol. 2000;18(1):64-71. 16. Svahn BM, Remberger M, Alvin O, Karlsson H, Ringden O. Increased costs after allogeneic haematopoietic SCT are associated with major complications and re-transplantation. Bone Marrow Transplant. 2012;47(5):706-715. 17. Shenoy S, Eapen M, Wu J, et al. A multicenter phase II trial of unrelated donor reduced intensity bone marrow transplantation for children with severe sickle cell disease (SCURT): results of the blood and Marrow Transplant Clinical Trials Network (BMT CTN 0601) study. Blood. 2015;126(23):619619. 18. Meier ER, Dioguardi JV, Kamani N. Current attitudes of parents and patients toward hematopoietic stem cell transplantation for sickle cell anemia. Pediatr Blood Cancer. 2015;62(7):1277-1284.

1831

S.D. Arnold et al. 19. Bhatia M, Sheth S. Hematopoietic stem cell transplantation in sickle cell disease: patient selection and special considerations. J Blood Med. 2015;6:229-238. 20. Majhail NS, Mothukuri JM, Brunstein CG, Weisdorf DJ. Costs of hematopoietic cell transplantation: comparison of umbilical cord blood and matched related donor transplantation and the impact of posttransplant complications. Biol Blood Marrow Transplant. 2009;15(5):564-573. 21. Horwitz ME, Chao NJ, Rizzieri DA, et al. Umbilical cord blood expansion with nicotinamide provides long-term multilineage engraftment. J Clin Invest. 2014;124 (7):3121-3128. 22. Wagner JE, Brunstein CG, Boitano AE, et al. Phase I/II trial of StemRegenin-1 Expanded umbilical cord blood hematopoietic stem cells supports testing as a stand-alone graft. Cell Stem Cell. 2015;18(1):144-155. 23. Wagner JE, Jr., Eapen M, Carter S, et al. One-unit versus two-unit cord-blood transplantation for hematologic cancers. N Engl J Med. 2014;371(18):1685-1694. 24. Bolanos-Meade J, Fuchs EJ, Luznik L, et al. HLA-haploidentical bone marrow trans-

1832

25.

26.

27.

28.

plantation with posttransplant cyclophosphamide expands the donor pool for patients with sickle cell disease. Blood. 2012;120(22):4285-4291. Carroll CP, Haywood C, Jr., Lanzkron S. Prediction of onset and course of high hospital utilization in sickle cell disease. J Hosp Med. 2011;6(5):248-255. Bhatia M, Kolva E, Cimini L, et al. Healthrelated quality of life after allogeneic hematopoietic stem cell transplantation for sickle cell disease. Biol Blood Marrow Transplant. 2015;21(4):666-672. Kelly MJ, Pennarola BW, Rodday AM, Parsons SK, Journeys to Recovery Study HCS. Health-related quality of life (HRQL) in children with sickle cell disease and thalassemia following hematopoietic stem cell transplant (HSCT). Pediatr Blood Cancer. 2012;59(4):725-731. Ware RE, Davis BR, Schultz WH, et al. Hydroxycarbamide versus chronic transfusion for maintenance of transcranial Doppler flow velocities in children with sickle cell anaemia-TCD With Transfusions Changing to Hydroxyurea (TWiTCH): a multicentre, open-label, phase 3, non-infe-

29.

30.

31.

32.

33.

riority trial. Lancet. 2016;387(10019):661670. Wang WC, Oyeku SO, Luo Z, et al. Hydroxyurea is associated with lower costs of care of young children with sickle cell anemia. Pediatrics. 2013;132(4):677-683. Bernaudin F, Robin M, Ferry C, et al. Related myeloablative stem cell transplantation (SCT) to cure sickle cell anemia (SCA): update of French results. Blood. 2010:116(21):3518. Syrjala KL, Langer SL, Abrams JR, Storer BE, Martin PJ. Late effects of hematopoietic cell transplantation among 10-year adult survivors compared with case-matched controls. J Clin Oncol. 2005;23(27):65966606. Jaime-Pérez JC, Heredia-Salazar AC, Cantú-Rodríguez OG, et al. Cost structure and clinical outcome of a stem cell transplantation program in a developing country: the experience in northeast Mexico. Oncologist. 2015;20(4):386-392. Sharma SK, Choudhary D, Gupta N, et al. Cost of hematopoietic stem cell transplantation in India. Mediterr J Hematol Infect Dis. 2014;6(1):e2014046.

haematologica | 2017; 102(11)

ARTICLE

Coagulation & Its Disorders

The ADAMTS131239-1253 peptide is a dominant HLA-DR1-restricted CD4+ T-cell epitope

Laurent Gilardin,1,2,3,4,# Sandrine Delignat,1,2,3 Ivan Peyron,1,2,3 Mathieu Ing,1,2,3 Yu-Chun Lone,5 Bagirath Gangadharan,1,2,3 Baptiste Michard,1,2,3 Yousra Kherabi,1,2,3 Meenu Sharma,1,2,3 Anastas Pashov,1,2,3,6 Jean-Baptiste Latouche,7 Mohamad Hamieh,7 Olivier Toutirais,8 Pascale Loiseau,9 Lionel Galicier,10 Agnès Veyradier,11 Srini Kaveri,1,2,3,12 Bernard Maillère,13 Paul Coppo4,14 and Sébastien Lacroix-Desmazes1,2,3,12 #

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Current affiliation: Service d’Onco-Hématologie, Hôpital Saint-Louis, AP-HP, 75010 Paris, France

Institut National de la Santé et de la Recherche Médicale (INSERM), Unité Mixte de Recherche en Santé (UMR S) 1138, Centre de Recherche des Cordeliers, Equipe Immunopathology and Therapeutic Immunointervention, Paris, France; 2Sorbonne Universités, Université Pierre et Marie Curie–Paris 6, UMR S 1138, Centre de Recherche des Cordeliers, Equipe Immunopathology and Therapeutic Immunointervention, Paris, France; 3Université Paris Descartes – Paris 5, UMR S 1138, Centre de Recherche des Cordeliers, Equipe Immunopathology and Therapeutic Immunointervention, Paris, France; 4Centre National de Référence sur les Microangiopathies Thrombotiques, Hôpital Saint Antoine, AP-HP, Paris, France; 5 Institut National de la Santé et de la Recherche Médicale (INSERM), U1014, Hôpital Paul Brousse, Villejuif, France; 6Department of Immunology, Institute of Microbiology, BAS, Sofia, Bulgaria; 7Laboratoire de Génétique Moléculaire, CHU CH.NICOLLE, Rouen, France; 8Laboratoire d’Immunologie and d’Immunopathologie, CHU Caen, France; 9 Laboratoire d'Immunologie et Histocompatibilité, Hôpital Saint-Louis, AP-HP, Paris, France; 10Département d’Immunologie Clinique, Hôpital Saint-Louis, AP-HP, Paris, France; 11Service d'Hématologie Biologique, Hôpital Lariboisière, AP-HP, Paris, France; 12 International Associated Laboratory IMPACT (INSERM, France–Indian Council of Medical Research, India), National Institute of Immunohaematology, Mumbai, India; 13 Institute of Biology and Technologies, SIMOPRO, Labex LERMIT, Labex VRI, Commissariat à l’énergie Atomique (CEA) Saclay, Gif sur Yvette, France and 14Service d’Hématologie, Hôpital Saint Antoine, AP-HP, Paris, France 1

Haematologica 2017 Volume 102(11):1833-1841

Correspondence: gilardin@gmail.com ABSTRACT

cquired thrombotic thrombocytopenic purpura is a rare and severe disease characterized by auto-antibodies directed against “A Disintegrin And Metalloproteinase with Thrombospondin type 1 repeats, 13th member" (ADAMTS13), a plasma protein involved in hemostasis. Involvement of CD4+ T cells in the pathogenesis of the disease is suggested by the IgG isotype of the antibodies. However, the nature of the CD4+ T-cell epitopes remains poorly characterized. Here, we determined the HLA-DR-restricted CD4+ T-cell epitopes of ADAMTS13. Candidate T-cell epitopes were predicted in silico and binding affinities were confirmed in competitive enzyme-linked immunosorbent assays. ADAMTS13-reactive CD4+ T-cell hybridomas were generated following immunization of HLA-DR1 transgenic mice (Sure-L1 strain) and used to screen the candidate epitopes. We identified the ADAMTS131239-1253 peptide as the single immunodominant HLA-DR1restricted CD4+ T-cell epitope. This peptide is located in the CUB2 domain of ADAMTS13. It was processed by dendritic cells, stimulated CD4+ T cells from Sure-L1 mice and was recognized by CD4+ T cells from an HLA-DR1-positive patient with acute thrombotic thrombocytopenic purpura. Interestingly, the ADAMTS131239-1253 peptide demonstrated promiscuity towards HLA-DR11 and HLA-DR15. Our work paves the way towards the characterization of the ADAMTS13-specific CD4+ T-cell response in patients with thrombotic thrombocytopenic purpura using ADAMTS131239-1253-loaded HLA-DR tetramers. haematologica | 2017; 102(11)

Received: September 21, 2015. Accepted: July 21, 2017. Pre-published: July 27, 2017. doi:10.3324/haematol.2015.136671 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1833 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1833

L. Gilardin et al.

Introduction Thrombotic thrombocytopenic purpura (TTP) is a rare and severe autoimmune disease characterized by the occurrence of IgG autoantibodies against the metalloprotease "A Disintegrin And Metalloproteinase with Thrombospondin type 1 repeats, 13th member" (ADAMTS13).1–3 ADAMTS13 cleaves multimers of von Willebrand factor, a glycoprotein involved in hemostasis. Inhibition of ADAMTS13 by IgG leads to an accumulation of hyper-adhesive von Willebrand factor multimers causing microthrombi that occlude the lumen of the capillaries in the microcirculation, thus inducing red cell hemolysis and ischemia of downstream organs. TTP is thus characterized by a combination of microangiopathic hemolytic anemia, peripheral thrombocytopenia and organ failure of variable severity with typically neurological involvement.4 The physiopathological mechanisms underlying TTP and responsible for the synthesis of anti-ADAMTS13 antibodies, and particularly the mechanisms involved in the loss of tolerance of the immune system towards ADAMTS13, are poorly understood. Polyclonal antiADAMTS13 antibodies are directed against different domains of ADAMTS13.5 In most patients, antiADAMTS13 antibodies are of the IgG isotype with a predominance of the IgG4 subclass.6 IgG from all patients recognize immunodominant B-cell epitopes located in the spacer domain of ADAMTS13.7 The B-cell epitopes have been proposed to be located between the 660-661 and 665 amino-acids.8 The fact that anti-ADAMTS13 antibodies are of the IgG isotype, of high affinity and have undergone affinity maturation, strongly suggests the requirement of CD4+ T-cell help in the development of the disease.9 Besides, the HLA-DRB1*11 (DR11) haplotype was independently identified as a strong risk factor by three research groups.10–12 However, while CD4+ T cells are

thought to play a major role, the specificity and the properties of the CD4+ T lymphocytes involved in the pathogenesis of TTP have not been studied. Importantly, the HLA restriction hints at the existence of immunodominant peptides in ADAMTS13. Naïve CD4+ T-cell activation is initiated by the interaction of the T-cell receptor (TCR) with a peptide/MHC class II complex on professional antigen-presenting cells. Extracellular antigens are endocytosed, degraded into peptides in the early endosome and loaded onto MHC class II αβ heterodimer molecules. Sorvillo et al. identified, through mass spectrometry, ADAMTS13-derived peptides that are localized within the HLA-DR molecules expressed by dendritic cells derived from human monocytes (Mo-DC) and incubated with ADAMTS13 in vitro.13 The identified peptides belonged mainly to the CUB1 and CUB2 domains of ADAMTS13. The core sequence FINVAPHAR from the CUB2 domain was the only sequence detected for six donors with the DR11 haplotype among the 17 donors included in the study. In addition, the same group recently demonstrated that a peptide containing the FINVAPHAR amino-acid sequence is able to stimulate CD4+ T cells from an HLA-DR11 TTP patient.14 These studies did not, however, investigate the recognition of other peptides by CD4+ T cells. To address this issue, we focused on the HLA-DR1 haplotype for which a humanized transgenic HLA-DR1 mouse is available, and investigated the CD4+ T-cell epitopes for the HLA-DR1 allele. The identified epitopes were then studied in the context of the HLA-DR11 allele.

Methods Peptides and antigen ADAMTS13-derived peptides were selected based on their

Table 1. Affinity of ADAMTS13-derived peptides for HLA-DRB1*01:01 molecules.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16*

ADAMTS13 peptide

Domain

Sequence#

IC50 (nM)†

Predicted binding score‡

589-603 267-281 371-385 1239-1253 1392-1406 1355-1369 774-788 570-584 914-928 1224-1238 1281-1295 464-478 1325-1339 140-154 308-322 111-125

Spacer Metalloprotease Disintegrin CUB1 CUB2 CUB2 TSR 3 Spacer TSR 5 CUB1 CUB1 Cys-Rich CUB2 Metalloprotease Desintegrin Metalloprotease

RPLFTHLAVRIGGRY RRQLLSLLSAGRARC CRSLVELTPIAAVHG GDMLLLWGRLTWRKM EGFLKAQASLRGQYW ASYILIRDTHSLRTT GSLLKTLPPARCRAG YVTFLTVTPNLTSVY ELRFLCMDSALRVPV VVTLRVLESSLNCSA GVLLRYGSQLAPETF GASFYHLGAAVPHSQ CRLFINVAPHARIAI APNITANLTSSLLSV NEQCRVAFGPKAVAC GAELLRDPSLGAQFR

0.46 0.91 1.59 1.61 3.74 5.65 7.11 13.10 12.6 13.68 28.34 31.07 147.97 279.67 5000 >10000

7.01 1.36 4.22 7.08 1.62 6.65 5.65 0.79 3.24 6.21 6.90 8.42 1.99 9.74 8.11 20.93

ADAMTS13-derived peptides are sorted based on the measured IC50 values compared to the HA306-318 control peptide (IC50=4.61 nM). IC50 values are shown in nM. A lower IC50 value indicates stronger binding. IC50>10000 indicates no detectable binding in the assay. ‡Binding scores were generated by IEDB and reflect the expected binding affinities, with lower scores indicating stronger predicted affinities. #The nine-residue sequences (core peptides) predicted to fit into the HLA-DRB1*01:01 binding groove are underlined for each peptide. Anchor residues are highlighted in bold. *This ADAMTS13-derived peptide is used as a negative control, without detectable binding to HLA-DRB1*01:01. †

1834

haematologica | 2017; 102(11)

The ADAMTS131329-1253 CD4+ T-cell epitope in TTP

capacity to bind to HLA-DRB1*01:01 molecules. HLA class II– binding predictions of human ADAMTS13 (GenBank reference: AAL11095.1) were made on November 10, 2011, using the Immune Epitope Database analysis resource consensus tool, on the dedicated website www.iedb.org.15–17 Strong HLA-DRB1*01:01 binders were predicted in silico using overlapping 15-mer peptides that span the whole ADAMTS13 sequence. Altogether, 99 15-mer peptides were predicted to be strong binders to HLA-DRB1*01:01 with binding scores below 10 (i.e., with a probability of being good binders greater than 90%). Some of the predicted peptides shared common HLA-DRB1*01:01-binding core sequences (9-mer peptides). When considering only unique core sequences and after exclusion of two peptides located in the prodomain of ADAMTS13, the list came down to 15 9-mer core peptides (Table 1). The peptides were synthesized at greater than 80% purity (GL Biochem, Shanghai, China) and included the 9-mer core sequences with addition of the three residues from the N-terminal end and the three residues of the C-terminal end. Individual peptides were solubilized at 1 mg/mL in dimethylsulfoxide/water. Recombinant full-length human ADAMTS13 (rhADAMTS13) was a kind gift from Baxter (Vienna, Austria).18

HLA-peptide-binding assays HLA-DR molecules were purified from homozygous EpsteinBarr virus cell lines by affinity-chromatography using the monomorphic monoclonal antibody L243. The binding to HLADR molecules was assessed by competitive enzyme-linked immunosorbent assay (ELISA), using an automated workstation, as previously reported.19,20 Briefly, HLA heterodimers were incubated with a biotinylated indicator peptide and serial dilutions of competitor peptides. As reference, the unlabeled form of biotinylated reporter peptide was used as an internal control. After 24 h incubation at 37˚C, samples were neutralized with 450 mM Tris HCl (pH 7.5) (Sigma, St Quentin-Fallavier, France), 0.3% bovine serum albumin (Sigma), and 1 mM n-dodecyl β-D-maltoside buffer (Sigma) and applied to 96-well MaxiSorp ELISA plates (Nunc A/S, Roskilde, Denmark) coated with 10 mg/mL L243. Bound biotinylated peptide was detected by streptavidin-alkaline phosphatase conjugate (GE Healthcare, Saclay, France) after adding 4-methylumbelliferyl phosphate substrate (Sigma). Emitted fluorescence was measured at 450 nm upon excitation at 365 nm. The peptide concentration that prevented binding of 50% of the biotinylated peptide (IC50) was evaluated. The sequence of the biotinylated reporter hemagglutinin (HA) peptide was 306 PKYVKQNTLKLAT318; the mean IC50 values of the HA306-318 peptide for HLA DRB1*01:01 and DRB1*11:01 binding were 4.61 and 37.42 nM, respectively. For HLA DRB1*15:01, the sequence of the reporter biotinylated peptide A3152–166 was 152EAEQLRAYLDGTGVE166 with an IC50 value of 41.46 nM. Means and standard errors of the means were calculated from three independent experiments. Binding data are estimated as relative affinities defined as the ratio of the IC50 of each individual peptide to the IC50 of the reference peptide. Peptides with relative affinities of 20 or less were considered as strong binders to the HLA-DRB1 molecules and were retained for the rest of the study.

Human samples Heparinized blood from healthy donors (EFS, Ile-de-France, Créteil, France) or from TTP patients (French national reference center for TMA, Prof. P. Coppo, St Antoine Hospital, Paris, France and Dr L. Galicier, St Louis Hospital, Paris, France) was obtained after written informed consent. The use of samples for research purposes had been approved by the ethical committee from SaintAntoine Hospital (Authorization n. 04807/12/2005). The criteria for diagnosing acquired TTP were those previously reported,4 and haematologica | 2017; 102(11)

Figure 1. Identification of HLA-DR1-restricted T-cell epitopes using ADAMTS13-specific CD4+ T-cell hybridomas. (A) Anti-ADAMTS13 IgG titers in Sure-L1 mice immunized with rhADAMTS13. The serum from an HLA-DR1transgenic Sure-L1 mouse immunized with rhADAMTS13 was incubated in serial dilutions on ADAMTS13-coated ELISA wells. ADAMTS13-bound IgG were detected using horseradish peroxidase-coupled polyclonal goat anti-mouse IgG and substrate. Results are expressed in arbitrary units as optical density measured at 492 nm. (B) Activation of ADAMTS13-specific CD4+ T-cell hybridomas. A representative ADAMTS13-specific CD4+ T-cell hybridoma (clone 2G10δ) was incubated with AAPCDR1 (10:1 ratio) and ADAMTS13-derived peptides as listed in Table 1. Stimulation indices represent the ratio of IL-2 secreted by the T cells measured by ELISA upon incubation with ADAMTS13 peptides over IL-2 secreted in the absence of peptide. Means±SD are from two independent experiments. (C) Delineation of the core peptide for ADAMTS13-specific CD4+ T-cell hybridomas. A representative ADAMTS13-specific CD4+ T-cell hybridoma (clone 2G10δ) was incubated with AAPCDR1 and overlapping 15-mer peptides spanning the 1235-1256 peptide sequence. Stimulation indices were assessed as explained above.

1835

L. Gilardin et al. A included a severe deficiency (<5% of normal activity) of ADAMTS-13 activity along with a detectable inhibitor and/or detectable serum antibodies as assessed by ELISA (Technozym R ADAMTS13 INH, Technoclone, Vienna, Austria; threshold positivity >15 U/mL).

Blood processing All blood samples were shipped at room temperature within 4 h of being drawn. HLA-DR screening was performed by molecular typing using the Olerup SSP HLA*DR kit (Olerup SSP, Stockholm, Sweden) after extracting DNA from total blood (Qiagen, Vienna, Austria). Peripheral blood mononuclear cells were isolated by density gradient centrifugation on lymphocyte separation medium (PAA Laboratories GmbH, Les Mureaux, France) and immediately frozen in pooled human male AB serum containing 10% dimethylsulfoxide (Sigma).

Immunization of mice The animals used in this study were HLA-A2.1-/HLA-DR1transgenic H-2 class I-/class II (IA β-/- β2m-/-)-knockout mice (Sure-L1 mice), 8 to 12 weeks old.21 The mice were immunized subcutaneously with 150 mg of rhADAMTS13 in complete Freund’s adjuvant (Sigma), and 15 days later with 150 mg rhADAMTS13 in incomplete Freund’s adjuvant (Sigma). Blood was drawn by retro-orbital bleeding 4 days after the second administration of rhADAMTS13 and cells from the spleen and draining lymph nodes were collected. Animals were handled in agreement with local ethical authorities (Comité Régional d'Éthique p3/2005/002).

Titration of anti-ADAMTS13 IgG

Antibodies against human ADAMTS13 were measured in mice serum using ELISA. Briefly, ELISA plates were coated with rhADAMTS13 (1 mg/mL) for 1 h at 37°C and subsequently blocked for an additional hour with phosphate-buffered saline 1% bovine serum albumin (PBS-BSA). Serum serially diluted in PBS-BSA (dilution factor: 1/3) was then incubated for 1 h at room temperature. Bound IgG was revealed using a secondary horseradish peroxidase-coupled goat F(ab’)2 anti-mouse IgG antibody (Southern Biotech, Birmingham, AL, USA) and substrate. The mouse monoclonal anti-human ADAMTS13 IgG monoclonal antibody 20A5 (HYCULT, Uden, NL) was used as a positive control.

HLA-DRB1*01:01-positive antigen-presenting cells Three different sources of antigen-presenting cells were used for T-cell presentation assays: (i) purified splenocytes from Sure-L1 mice inactivated with mitomycin C (50 mg/mL, Sigma), (ii) immature Mo-DC from HLA-DRB1*01:01 healthy donors inactivated with mitomycin C (50 mg/mL), and (iii) artificial antigen-presenting cells derived from the NIH-3T3 cell line expressing HLADRB1*01:01 (AAPCDR1). To generate AAPCDR1, NIH-3T3 cells were transduced with the common HLA-DRα chain and the specific HLA-DRβ1*01:01 chain, and with the co-stimulatory and adhesion human molecules B7.1, ICAM-1 and LFA-3.22 AAPCDR1 does not produce murine interleukin-2 and does not, therefore, require inactivation with mitomycin C. Its ability to present and stimulate HLA-DRB1*01:01-restricted CD4+ T cells efficiently has been demonstrated.22 In the case of Mo-DC, human monocytes were isolated from peripheral blood mononuclear cells using anti-CD14 magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). Monocytes were cultured in RPMI-1640 medium supplemented with 10% human AB Serum (Sigma), L-glutamine and antibiotics (Life Technologies - Invitrogen, Carlsbad, CA, USA) in the presence of 500 IU human recombinant interleukin 4 (ImmunoTools, 1836

Figure 2. The ADAMTS131239-1253 peptide is an immunodominant T-cell epitope both for Sure-L1 mice as well as for a patient with TTP. (A) Stimulation by ADAMTS13derived peptides of spleen and lymph node cells from Sure-L1 mice immunized with rhADAMTS13. Cells (3x105/well) were incubated with each individual peptide (10 mg/mL), with the pool of peptides (Peptide mix) or with rhADAMTS13. Proliferation was assessed by incorporation of tritiated thymidine. Proliferation indices are defined as the ratio of cpm of stimulated cells versus cpm of control cells, and are expressed as means±SD for two Sure-L1 mice. (B) The ADAMTS1312391253 peptide is processed and presented to T cells by human antigen-presenting cells. A representative ADAMTS13-specific CD4+ T-cell hybridoma (clone 1F5γ) was incubated with immature dendritic cells from a healthy HLA-DR1+ blood donor and with the ADAMTS131239-1253 peptide, with rhADAMTS13 or with the ADAMTS13111-125 peptide. Stimulation indices represent the ratio of secretion of IL-2 by T cells stimulated in the presence of antigen versus the secretion of IL-2 by T cells incubated with Mo-DC alone. Means±SD are from two independent experiments. Similar results were obtained with other ADAMTS13-CD4+ T-cell hybridoma clones (not shown). The peptide incubated alone or with DR15+ artificial antigen-presenting cells failed to activate the T cells (not shown). (C) ADAMTS13 epitope specificity of CD4+ T-cell lines from a patient with acquired TTP. CD4+ T-cell lines were generated after stimulation with pooled ADAMTS13-derived peptides as defined in Table 1. T cells (3x105/well) were then incubated with AAPCDR1 (3x104 cells/well) alone or pulsed with individual peptide (10 mg/mL), or with the peptide pool (Peptide mix). The number of cells producing interferon-γ was then assessed by ELISPOT and is expressed as the mean number of spot-forming cells (SFC) per million cells calculated in the case of peptide-stimulated T cells minus the SFC obtained in the case of unstimulated cells. Means±SD are from two independent experiments. *The 111-125 ADAMTS13-derived peptide that did not bind to HLA-DRB1*01:01 was used as a negative control.

haematologica | 2017; 102(11)

The ADAMTS131329-1253 CD4+ T-cell epitope in TTP

Friesoythe, Germany) and 1000 IU human recombinant granulocyte-macrophage colony-stimulating factor (ImmunoTools) per 106 cells. After 5 days of culture, the non-adherent immature MoDC-enriched fraction was harvested and the immature status was confirmed by the expression of the following surface phenotypic markers: CD40, CD80, CD86, CD83, CD1a and HLA-DR (Becton Dickinson, Le Pont de Claix, France), by flow cytometry as explained previously.23

Proliferation response to ADAMTS13-derived peptide Cells from the spleen and draining lymph nodes from immunized Sure-L1 mice were pooled. Cells (105 cells/well) were incubated alone, with each individual synthetic ADAMTS13-derived peptide (10 mg/mL) or with a pool containing all the peptides (2.5 mg/mL of each peptide) in AIM-V medium (Life Technologies) for 4 days. Cell proliferation was determined by incorporation of 3Hthymidine (1 mCi per well, specific activity of 6.7 Ci/mmol) added during the final 16 h and measured as counts per minute (cpm) on a micro-β-counter (Perkin Elmer Applied Biosystems). Stimulation indices (SI) were calculated as the ratio between the cpm in the presence of a stimulus, and the average cpm of the same cells incubated alone.

Generation of mouse ADAMTS13-specific CD4+ T-cell hybridomas Cells from draining lymph nodes of immunized Sure-L1 mice were stimulated for 72 h with rhADAMTS13 (1 mg/mL) and fused with the BWZ.36 fusion (TCR−/−) partner cell line24 (a kind gift from Prof N Shastri, University of California, Berkeley, CA, USA) using polyethylene glycol. ADAMTS13-specific CD4+ T-cell hybridomas were obtained after culture in HAT and HT selective media (both from Sigma). Specificity for ADAMTS13 was evaluated using a T-cell stimulation assay.

T-cell hybridoma stimulation assay Mouse HLA-DR1-restricted CD4+ T-cell hybridomas (105 cells/well) were cultured in plates of 96 round-bottomed wells with 104 (ratio 1:10) DRB1*01:01 antigen-presenting cells (splenocytes or immature Mo-DC) in the presence of rhADAMTS13 (10 mg/mL) for 24 h at 37°C. Murine interleukin-2 secreted by activated T cells was measured in the supernatant using an ELISA kit according to the manufacturer’s protocol (eBioscience, Paris, France). The positive control was concanavalin A (Sigma) at a concentration of 1 mg/mL. For each hybridoma cell line growing in HAT media, the stimulation index was calculated. Hybridoma clones with an at least 2-fold increase in interleukin-2 release (ratio between ADAMTS13-stimulated cultures and control medium) were considered to be ADAMTS13-specific and were subsequently sub-cloned by limiting dilution (0.5 cells/well). Among ADAMTS13-specific T-cell hybridomas, screening for peptide specificity was realized with AAPCDR1 co-cultured for 24 h with peptides at 10 mg/mL, and interleukin-2 secretion by activated Tcell hybridomas was measured in the supernatant with the same ELISA kit.

Generation of ADAMTS13-specific human CD4+ T-cell lines Using the accelerated co-cultured dendritic cell assay,25 peripheral blood mononuclear cells from a TTP patient were stimulated for 10 days with the pool of 16 ADAMTS13-derived peptides (2.5 mg/mL for each peptide). Briefly, cells (5x106 cells/mL) were thawed and cultured in AIM-V medium supplemented with 1000 U/mL of granulocyte-macrophage colony-stimulating factor and 500 U/mL of interleukin-4 (ImmunoTools). After 24 h (day 1), the pool of ADAMTS13-derived peptides was added along with tumor necrosis factor-α (1000 U/mL), interleukin-1β (10 ng/mL), haematologica | 2017; 102(11)

interleukin-7 (0.5 ng/mL) (ImmunoTools) and prostaglandin E2 (1 mM; Merck Calbiochem, Saint-Quentin-en-Yvelines, France). On day 10, non-adherent cells were collected, washed, and CD4+ T lymphocytes were isolated by negative magnetic purification with a CD4+ T-cell isolation kit (Miltenyi). Cells were then incubated with irradiated DRB1*01 Mo-DC (45 Gy) from healthy donors (ratio 1:10), pulsed with the pool of ADAMTS13-derived peptides, in fresh AIM-V at 37°C for 6 h. Subsequently, CD4+ T cells reactive to ADAMTS13-derived peptides were isolated using the interferon (IFN)-γ secretion assay-cell detection kit according to the manufacturer’s protocol (Miltenyi). Cell lines were obtained after culture of the IFN-γ-secreting cells in AIM-V medium supplemented with 10% human AB serum, interleukin-2 (20 U/mL), interleukin-4 (5 ng/mL), anti-CD3 monoclonal antibody (OKT3, 30 ng/mL) (ImmunoTools) and irradiated (45 Gy) peripheral blood mononuclear cells from two unrelated donors. Cells were fed every 7 days with fresh cytokines. The stimulated CD4+ T cells were tested after ~3 weeks for peptide specificity by IFN-γ enzyme-linked immunospot (ELISPOT) assays after incubation with peptide-pulsed or non-pulsed antigen-presenting cells.

Enzyme-linked immunospot assays The human IFN-γ ELISPOT assay was performed according to the manufacturer’s recommendations. Ninety-six–well polyvinylidine fluoride plates (Millipore-Merck) were pre-wetted with ethanol. The individual ADAMTS13-derived peptides (10 mg/mL), the peptide pool, or medium alone were plated in triplicate wells, into the plates that had been pre-coated overnight with the antiIFN-γ capture antibody (U-CyTech, Utrecht, the Netherlands) and blocked with AIM-V plus 10% human serum. Irradiated AAPCDR1 (45 Gy) were then seeded and incubated (3x104/well) with the CD4+ T-cell line (3x105/well) for 24 h at 37°C in a CO2 incubator. Following removal of peripheral blood mononuclear cells, IFN-γ secretion was visualized with a biotin-conjugated anti–IFN-γ antibody (U-CyTech), alkaline phosphatase-conjugated ExtrAvidin, and Sigmafast 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NBT) tablets (Sigma). Plates were then air dried and scanned on an ImmunoSpot Analyzer (CTL, Shaker Heights, OH, USA). Spot-forming cells were automatically calculated by ImmunoSpot Software using the SmartCount™ and Autogate™ functions.26 ELISPOT readouts are expressed as spot-forming

Table 2. Affinity of ADAMTS131239-1253 peptide for HLA-DRB1 molecules.

HLA-DR molecule# DRB1*11:01 DRB1*15:01 DRB1*13:01 DRB1*08:01 DRB1*01:01 DRB1*03:01 DRB1*04:01 DRB1*12:01 DRB1*07:01 DRB1*09:01

Predicted binding score†

Relative affinity‡

1.2 1.40 1.72 4.47 7.08 7.92 12.10 25.89 31.19 33.67

16.04 3.48 ND ND 0.27 ND ND ND ND ND

# HLA-DRB1 molecules are classified based on their predicted binding score. †Binding scores were generated by IEDB using the consensus method prediction tool. ‡Binding data are estimated as relative affinities defined as the ratio of the IC50 of the ADAMTS131239-1253 peptide to the IC50 of each reference peptide. Relative affinities of 20 or less were considered high binders to the HLA-DR molecule. The sequences and mean IC50 values of the reference peptides were the following: HA306-318 (PKYVKQNTLKLAT) for DR1 (4.61 nM) and DR11 (37.42 nM); A3152-166 (EAEQLRAYLDGTGVE) for DR15 (41.46 nM). ND: not done.

1837

L. Gilardin et al.

cells/106 CD4+ T-cells. Peptides that generated at least twice the sport-forming cells over the control wells were defined as eliciting a positive response.

ADAMTS13-specific DRB1*01:01-restricted T-cell hybridomas is 1241MLLLWGRLTWR1251. Anchor residues for binding of peptide 1239-1353 to the HLA groove are highlighted in Table 1: L1, W4, R6 and W9.

Results

Few ADAMTS13-derived peptides are strong binders to HLA-DRB1*01:01 molecules We hypothesized that the T-cell epitopes of rhADAMTS13 correspond to ADAMTS13-derived peptides that bind stably to the groove of class II MHC molecules. The candidate peptides should thus bind with a high affinity to HLA-DR molecules. Using the IEDB consensus method prediction tool for MHC class II binding, we first predicted the array of ADAMTS13-derived 15mer peptides that bind to the HLA-DRB1*01:01 (DR1) allele. Altogether, 15 15-mer peptides with unique core sequences were predicted to be strong binders to HLADR1 with binding scores below 10 (Table 1). The corresponding synthetic 15 peptides were used in binding assays to determine in vitro affinities for purified HLADRB1*01:01 molecules by competition ELISA. The calculated IC50 values ranged from 0 to 5000 nM (Table 1). Five peptides bound to DR1 with an IC50 below 4 nM, an affinity greater than that of the influenza HA306-318 control peptide (IC50 = 4.61 nM). Three peptides revealed a poor binding capacity with IC50 values 20-fold greater than that of the HA control peptide, despite a good IEDB-predicted binding score.

ADAMTS13-derived peptides activate ADAMTS13-specific mouse CD4+ T-cell hybridomas restricted to HLADRB1*01:01 To assess whether the ADAMTS13-derived HLA-DR1 binding peptides are T-cell epitopes and are presented to T cells by antigen-presenting cells, we generated ADAMTS13-specific CD4+ T-cell hybridomas restricted to the HLA-DR1 allele. Induction of an anti-rhADAMTS13 immune response in a humanized transgenic HLA-DR1 mouse was confirmed by ELISA (Figure 1A). Cells from the draining lymph nodes were fused to a TCR-/- partner (BWZ36) and 95 CD4+ T-cell hybridomas were selected. Among them, 26 hybridomas produced interleukin-2 following incubation with splenocytes from Sure-L1 mice in the presence of rhADAMTS13. The 26 hybridomas were subsequently sub-cloned. We next evaluated the activation of the T-cell hybridomas by the 15-mer peptides. The ADAMTS131239-1253 peptide was the only peptide to induce a significant secretion of interleukin-2 in all the hybridoma cell lines tested (6 of the 26 rhADAMTS13-reactive T-cell hybridomas; a representative result is shown for 1 T cell-hybridoma in Figure 1B). To delineate the T-cell epitope recognized by the Tcell hybridomas more precisely, overlapping 15-mer peptides that span amino acids 1235 to 1256 were synthesized and assessed for their capacity to activate one of the T-cell hybridomas (Clone 2G10δ). Significant T-cell activation was observed for peptides 1237-1251, 1238-1252, 1239-1253, 1240-1254 and 1241-1255 (Figure 1C). Conversely, peptides 1235-1249, 1236-1250 and 12421256 failed to activate the T-cell hybridoma. Similar results were obtained with another T-cell hybridoma (clone 1F5γ, data not shown). These results demonstrate that the minimal core peptide sequence recognized by the 1838

Figure 3. The ADAMTS131239-1253 peptide is an immunodominant T-cell epitope for HLA-DRB1*11 individuals. CD4+ T-cell lines from HLA-DRB1*11 TTP patients at (A) acute phase or (B) in remission phase of the disease, or from (C) a HLADRB1*11 healthy donor, were stimulated with the ADAMTS13-derived peptides identified as immunodominant for HLA-DRB1*01. T cells (3x105 cells/well) were incubated with AAPCDR11 (3x104 cells/well) alone or pulsed with individual peptide (10 mg/mL), or with the peptide pool (Peptide mix). The number of cells producing interferon-γ was then assessed by ELISPOT and is expressed as mean number of spot-forming cells (SFC) per million cells calculated in the case of peptidestimulated T-cells minus the SFC obtained in the case of unstimulated cells. *The 111-125 ADAMTS13-derived peptide is used as a negative control, without detectable binding to HLA-DRB1*11:01 (data not shown).

haematologica | 2017; 102(11)

The ADAMTS131329-1253 CD4+ T-cell epitope in TTP

The ADAMTS131239-1253 peptide is recognized by CD4+ T cells from humanized Sure-L1 HLA-DRB1*01:01 mice We next asked whether other peptides than the one identified using the hybridoma technology may be recognized by polyclonal CD4+ T cells from Sure-L1 mice immunized with rhADAMTS13. Splenocytes and cells from the draining lymph nodes from three immunized Sure-L1 mice were stimulated with the individual ADAMTS13-derived peptides previously identified to bind to HLA-DRB1*01:01 with a high affinity. Significant cell proliferation was observed in the case of ADAMTS131239-1253 peptide (Figure 2A), confirming that this peptide is a dominant CD4+ T-cell epitope, at least in mice.

The ADAMTS13 T-cell epitope identified in humanized Sure-L1 HLA-DRB1*01:01 mice is presented by human antigen-presenting cells We then investigated whether the ADAMTS131239-1253 peptide is generated by professional human antigen-presenting cells and presented to CD4+ T cells in an HLADRB1*01:01 context. To this end, immature HLADRB1*01:01 Mo-DC were incubated with rhADAMTS13 and with T-cell hybridomas of known specificity for the 1239-1253 peptide. T cells incubated with Mo-DC in the presence of the ADAMTS131239-1253 peptide presented with a stimulation index of 22.2±1.9 (Figure 2B). In contrast, incubation of the cells with the ADAMTS13111-125 control peptide did not result in T-cell activation (stimulation index: 1.1±0.0). Interestingly, incubation of T cells and Mo-DC with rhADAMTS13 induced T-cell activation with a stimulation index of 3.3±0.6, indicating that rhADAMTS13 is efficiently processed by human antigenpresenting cells, and that the ADAMTS131239-1253 peptide is readily generated.

The ADAMTS131239-1253 peptide is a dominant T-cell epitope in a patient with thrombotic thrombocytopenic purpura with the HLA-DRB1*01:01 allele We then assessed whether the ADAMTS131239-1253 peptide is recognized by CD4+ T cells from patients with acquired TTP. CD4+ T-lymphocytes were isolated from the blood of a 43-year -old HLA-DRB1*01:01 male, 14 days after the onset of acute acquired TTP. ADAMTS13 activity in this patient was less than 5% of the normal value due to the presence of inhibitory antibodies directed to ADAMTS13. A CD4+ T-cell line specific for ADAMTS13-derived peptides was amplified from the patient’s peripheral blood mononuclear cells as explained. After 32 days of culture, the CD4+ T-cell line produced IFN-γ, as assessed by ELISPOT, following incubation with Mo-DC from HLA-DR1 healthy donors, pulsed with the pool of ADAMTS13-derived peptides with high binding affinity for HLA-DR1, but failed to produce IFN-γ when incubated with Mo-DC alone (data not shown). We then incubated the T-cell lines with AAPCDR1 and with each individual peptide. Together with the ADAMTS131392-1406 peptide, the ADAMTS131239-1253 was the only peptide able to activate human IFN-γ-producing CD4+ T cells (Figure 2C). Similar results were obtained when the T cells and ADAMTS131239-1253 peptide were incubated in the presence of HLA-DR1 Mo-DC (data not shown). Further evaluation of the ADAMTS131239-1253 peptide by IEDB predicted good binding scores to HLA-DR11, 15, 13, 8, 3 and 4, which haematologica | 2017; 102(11)

was confirmed in competitive ELISA at least in the case of HLA-DR11 and HLA-DR15 (Table 2). T-cell lines were also generated from one female TTP patient in acute phase (aged 19, Figure 3A), one female TTP patient in remission phase (aged 34, Figure 3B) and one 25-year old, healthy male donor (Figure 3C), all with the HLADRB1*11:01 allele. Reactivity towards the ADAMTS1312391253 peptide was detected in the case of the three individuals. Together, our data suggest that the ADAMTS131239-1253 peptide is a dominant and promiscuous T-cell epitope in TTP patients.

Discussion In the present study, we delineated the ADAMT13derived peptides that are potent T-cell epitopes in the context of the HLA-DR1 haplotype. An array of ADAMTS13derived peptides that are processed and presented by MoDC had been previously reported in the case of healthy donors with different HLA-DR haplotypes.13 As many as 11 core peptides had thus been retrieved from the HLADR-binding groove; the peptides spanned six different domains of ADAMTS13. In order to identify HLA-DRrestricted T-cell epitopes of ADAMTS13, we took advantage of the humanized Sure-L1 mouse strain. Sure-L1 mice are invalidated for murine MHC class I and II molecules and transgenic for HLA-DRB1*01:01. HLA transgenic mice have been proven to be an appropriate human counterpart for the characterization of T-cell epitopes derived from pathogens21 as well as therapeutic proteins.27 In agreement with this, our results demonstrate that the ADAMTS1312391253 peptide, identified as an immunodominant T-cell epitope using in silico prediction, in vitro binding assays and Tcell hybridoma generation, is processed by human MoDC in vitro and activates CD4+ T cells both in vivo in SureL1 mice, and ex vivo in the case of cells from patients with TTP. The in silico selection of the ADAMTS13-derived peptides was based on the hypothesis that the T-cell epitope of rhADAMTS13 binds with a high affinity to HLA-DR molecules. The prediction efficiency of IEDB for HLA class II–binding peptides has been validated for various antigens, including human plasma proteins, and correlates adequately with HLA binding, as measured by competitive ELISA.15-17 For instance, T-cell epitopes of the human pro-coagulant factor VIII were identified using the hybridoma technology in an HLA-DR15 transgenic mouse model upon screening a peptide library spanning the entire sequence of the protein.27 Our unpublished data demonstrate a strong correlation between the predicted IEDB binding score for HLA-DR15 and the CD4+ T-cell epitope found in this work through hybridoma technology (area under the receiver operating characteristic curve=0.90). In silico prediction may nevertheless be overpredictive and positively select peptides that do not activate T cells.28 In the case of ADAMTS13, HLA-DR11 has been identified as a major risk factor for TTP. Sorvillo et al. identified 11 hypothetical T-cell epitopes following purification from HLA-DR molecules expressed by ADAMTS13loaded Mo-DC from healthy donors with several HLA haplotypes.13 The ADAMTS13-derived core sequence 1328 FINVAPHAR1336 was shown to be preferentially presented by the HLA-DR11 molecule,13 and stimulated CD4+ T 1839

L. Gilardin et al.

cells from at least one TTP patient.14 In our hands, the ADAMTS131325-1339 peptide, that contains the same core sequence, was predicted to be a strong binder to HLADR1 peptide. However, it failed to bind HLA-DRB1*01:01 and to activate human CD4+ T cells from a DRB1*01:01 patient. Besides, the ADAMTS131325-1339 peptide did not activate CD4+ T cells from two TTP patients with the HLA-DRB11 allele. Whether this results from differences in peptide length and in the nature of the residues adjacent to the core peptide remains to be determined.29,30 Our rationale for focusing our study on the HLADRB1*01:01 allele is justified by the availability of an appropriate transgenic HLA-DRB1*01:01 mouse model, and by the fact that the HLA-DRB1*01:01 allele is found in 18% of the patients with acquired TTP10 and in 15 to 30% of the Caucasian population.31 Interestingly, 11 of the 15 peptides predicted to bind HLA-DR1 by IEDB were also predicted to have a high affinity for the HLADRB1*11:01 allele. In particular, the unique immunodominant HLA-DRB1*01:01-associated T-cell epitope characterized in our work, ADAMTS131239-1253, had an excellent binding score to HLA-DRB1*11:01 (score of 1.2) and HLA-DRB1*15:01, strong affinity for HLADRB1*11:01 and HLA-DRB1*15:01 in competitive ELISA (Table 2), and was a target epitope of CD4+ T cells from two TTP patients and one healthy donor with the HLADRB11 allele. Of note, the ADAMTS131240-1252 peptide, which shares the identical core peptide with the ADAMTS131239-1253 peptide, was identified by the study of Sorvillo et al. in the case of Mo-DC from a patient with the HLA-DR4/DR13 haplotype.13 Together with our IEDB prediction, the data suggest that the ADAMTS131239-1253 peptide exhibits promiscuity for at least three HLA-DR alleles: DR1, DR11 and DR15. Additional work including more HLA-DR alleles is required to determine whether pathogenic recognition of ADAMTS13 by CD4+ T cells is principally directed to promiscuous immunodominant epitopes such as the

References 1. Tsai HM, Lian EC. Antibodies to von Willebrand factor-cleaving protease in acute thrombotic thrombocytopenic purpura. N Engl J Med. 1998; 339(22):1585–1594. 2. Furlan M, Robles R, Solenthaler M, et al. Acquired deficiency of von Willebrand factor-cleaving protease in a patient with thrombotic thrombocytopenic purpura. Blood. 1998;91(8):2839–2846. 3. Levy GG, Nichols WC, Lian EC, et al. Mutations in a member of the ADAMTS gene family cause thrombotic thrombocytopenic purpura. Nature. 2001; 413(6855): 488–494. 4. Coppo P, Bengoufa D, Veyradier A, et al. Severe ADAMTS13 deficiency in adult idiopathic thrombotic microangiopathies defines a subset of patients characterized by various autoimmune manifestations, lower platelet count, and mild renal involvement. Medicine (Baltimore). 2004;83(4):233–244. 5. Zheng XL, Wu HM, Shang D, et al. Multiple domains of ADAMTS13 are targeted by autoantibodies against ADAMTS13 in patients with acquired idiopathic thrombotic thrombocytopenic purpura.

1840

ADAMTS131239-1253 peptide, or whether private T-cell epitopes dominate in the case of HLA-DR alleles that are predominantly associated with acquired TTP. We cannot exclude the existence of additional immunodominant T-cell epitopes potentially overlooked in our experimental approach. Besides, additional work is required to link our findings to the Th2 polarization of CD4+ T cells and production of anti-ADAMTS13 autoantibodies. Nevertheless, identification of the ADAMTS131239-1253 peptide as a promiscuous immunodominant T-cell epitope may have clinical implications at term. It should pave the way towards the characterization of the ADAMTS13-specific CD4+ T-cell response in TTP patients using ADAMTS131239-1253-loaded HLA-DR tetramers. Such a tool would provide new insights into the mechanisms involved in the breakdown of tolerance to ADAMTS13 and should be of clinical relevance for following, in a quantitative manner, populations of ADAMTS13-specific T cells in the course of TTP treatment and anticipate relapse. In particular, confirmation of our results would lead to identification of prognostic factors (persistence of specific Th17 T cells, generation of regulatory T cells) that would help in the decision process regarding pre-emptive use of rituximab to prevent relapse.32 Acknowledgments We thank Professor John Robinson (Newcastle University, Newcastle upon Tyne, United Kingdom) for his advice about generating the ADAMTS13-specific T-cell hybridomas. Funding This study was supported by Institut National de la Santé et de la Recherche Médicale, Centre National de la Recherche Scientifique, Université Pierre et Marie Curie – Paris 6 and the 2012 Prof. Heimburger Award from CSL-Behring (Marburg, Germany). LG was the recipient of a « Poste d'accueil INSERM» fellowship and Bourse Marcel Simon - Société Française de Médecine Interne (SNFMI).

Haematologica. 2010; 95(9):1555–1562. 6. Ferrari S, Palavra K, Gruber B, et al. Persistence of circulating ADAMTS13-specific immune complexes in patients with acquired thrombotic thrombocytopenic purpura. Haematologica. 2014; 99(4):779–787. 7. Thomas MR, de Groot R, Scully MA, et al. Pathogenicity of anti-ADAMTS13 autoantibodies in acquired thrombotic thrombocytopenic purpura. EBioMedicine 2015;2(8): 942-952. 8. Pos W, Crawley JTB, Fijnheer R, et al. An autoantibody epitope comprising residues R660, Y661, and Y665 in the ADAMTS13 spacer domain identifies a binding site for the A2 domain of VWF. Blood. 2010; 115(8):1640–1649. 9. Pos W, Luken BM, Sorvillo N, et al. Humoral immune response to ADAMTS13 in acquired thrombotic thrombocytopenic purpura. J Thromb Haemost. 2011;9(7):1285–1291. 10. Coppo P, Busson M, Veyradier A, et al. HLADRB1*11: a strong risk factor for acquired severe ADAMTS13 deficiency-related idiopathic thrombotic thrombocytopenic purpura in Caucasians. J Thromb Haemost JTH. 2010;8(4):856–859. 11. Scully M, Brown J, Patel R, et al. Human leukocyte antigen association in idiopathic

12.

13.

14.

15.

16.

17.

thrombotic thrombocytopenic purpura: evidence for an immunogenetic link. J Thromb Haemost. 2010;8(2):257–262. John M-L, Hitzler W, Scharrer I. The role of human leukocyte antigens as predisposing and/or protective factors in patients with idiopathic thrombotic thrombocytopenic purpura. Ann Hematol. 2012; 91(4):507–510. Sorvillo N, van Haren SD, Kaijen PH, et al. Preferential HLA-DRB1*11-dependent presentation of CUB2-derived peptides by ADAMTS13-pulsed dendritic cells. Blood. 2013;121(17):3502–3510. Verbij FC, Turksma AW, de Heij F, et al. CD4+ T cells from patients with acquired thrombotic thrombocytopenic purpura recognize CUB2 domain-derived peptides. Blood. 2016;127(12):1606–1609. Wang P, Sidney J, Dow C, et al. A systematic assessment of MHC class II peptide binding predictions and evaluation of a consensus approach. PLoS Comput Biol. 2008;4(4): e1000048. Wang P, Sidney J, Kim Y, et al. Peptide binding predictions for HLA DR, DP and DQ molecules. BMC Bioinformatics. 2010;11: 568. Vita R, Overton JA, Greenbaum JA, et al. The immune epitope database (IEDB) 3.0.

haematologica | 2017; 102(11)

The ADAMTS131329-1253 CD4+ T-cell epitope in TTP

18.

19.

20.

21.

22.

Nucleic Acids Res. 2015;43(Database issue):D405–412. Plaimauer B, Kremer Hovinga JA, Juno C, et al. Recombinant ADAMTS13 normalizes von Willebrand factor-cleaving activity in plasma of acquired TTP patients by overriding inhibitory antibodies. J Thromb Haemost. 2011;9(5):936–944. Texier C, Pouvelle-Moratille S, Busson M, et al. Complementarity and redundancy of the binding specificity of HLA-DRB1, -DRB3, DRB4 and -DRB5 molecules. Eur J Immunol. 2001;31(6):1837–1846. Texier C, Pouvelle S, Busson M, et al. HLADR restricted peptide candidates for bee venom immunotherapy. J Immunol. 2000; 164(6):3177–3184. Pajot A, Michel M-L, Fazilleau N, et al. A mouse model of human adaptive immune functions: HLA-A2.1-/HLA-DR1-transgenic H-2 class I-/class II-knockout mice. Eur J Immunol. 2004;34(11):3060–3069. Garnier A, Hamieh M, Drouet A, et al. Artificial antigen-presenting cells expressing HLA class II molecules as an effective tool

haematologica | 2017; 102(11)

23.

24. 25.

26. 27.

28.

for amplifying human specific memory CD4(+) T cells. Immunol Cell Biol. 2016;94(7):662–672. Bayry J, Lacroix-Desmazes S, Carbonneil C, et al. Inhibition of maturation and function of dendritic cells by intravenous immunoglobulin. Blood. 2003; 101(2):758–765. Sanderson S, Shastri N. LacZ inducible, antigen/MHC-specific T cell hybrids. Int Immunol. 1994;6(3):369–376. Martinuzzi E, Afonso G, Gagnerault M-C, et al. acDCs enhance human antigen-specific T-cell responses. Blood. 2011; 118(8):2128– 2137. Lehmann PV. Image analysis and data management of ELISPOT assay results. Methods Mol Biol. 2005;302:117-32. Steinitz KN, van Helden PM, Binder B, et al. CD4+ T-cell epitopes associated with antibody responses after intravenously and subcutaneously applied human FVIII in humanized hemophilic E17 HLA-DRB1*1501 mice. Blood. 2012 26;119(17):4073–4082. Lundegaard C, Lund O, Nielsen M. Predictions versus high-throughput experi-

29.

30.

31.

32.

ments in T-cell epitope discovery: competition or synergy? Expert Rev Vaccines. 2012;11(1):43–54. Carson RT, Vignali KM, Woodland DL, et al. T cell receptor recognition of MHC class IIbound peptide flanking residues enhances immunogenicity and results in altered TCR V region usage. Immunity. 1997;7(3):387– 399. Holland CJ, Cole DK, Godkin A. ReDirecting CD4(+) T cell responses with the flanking residues of MHC class II-bound peptides: the core is not enough. Front Immunol. 2013;4:172. González-Galarza FF, Takeshita LYC, Santos EJM, et al. Allele frequency net 2015 update: new features for HLA epitopes, KIR and disease and HLA adverse drug reaction associations. Nucleic Acids Res. 2015;43(Database issue):D784–788. Hie M, Gay J, Galicier L, et al. Preemptive rituximab infusions after remission efficiently prevent relapses in acquired thrombotic thrombocytopenic purpura. Blood. 2014;124 (2):204–210.

1841

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Myeloid Leukemia

Ferrata Storti Foundation

Treatment outcome in a population-based, ‘real-world’ cohort of patients with chronic myeloid leukemia

Inge G.P. Geelen,1 Noortje Thielen,2,3 Jeroen J.W.M. Janssen,2 Mels Hoogendoorn,4 Tanja J.A. Roosma,2 Sten P. Willemsen,5,6 Otto Visser,7 Jan J. Cornelissen8 and Peter E. Westerweel1

Haematologica 2017 Volume 102(11):1842-1849

Department of Hematology, Albert Schweitzer Hospital, Dordrecht; 2Department of Hematology, VU University Medical Center, Amsterdam; 3Department of Internal Medicine, Diakonessenhuis, Utrecht; 4Department of Hematology, Medical Center Leeuwarden; 5Department of Biostatistics, Albert Schweitzer Hospital, Dordrecht; 6 Department of Biostatistics, Erasmus University Medical Center, Rotterdam; 7 Netherlands Comprehensive Cancer Organisation (IKNL), Utrecht and 8Department of Hematology, Erasmus University Medical Center, Rotterdam, the Netherlands 1

ABSTRACT

Correspondence: p.e.westerweel@asz.nl

Received: June 21, 2017. Accepted: August 24, 2017. Pre-published: August 31, 2017. doi:10.3324/haematol.2017.174953 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1842

valuations of the ‘real-world’ efficacy and safety of tyrosine kinase inhibitors in patients with chronic myeloid leukemia are scarce. A nationwide, population-based, chronic myeloid leukemia registry was analyzed to evaluate (deep) response rates to first and subsequent treatment lines and eligibility for a treatment cessation attempt in adults diagnosed between January 2008 and April 2013 in the Netherlands. The registry covered 457 patients; 434 in chronic phase (95%) and 15 (3%) in advanced disease phase. Seventy-five percent of the patients in chronic phase were treated with imatinib and 25% with a second-generation tyrosine kinase inhibitor. At 3 years 44% of patients had discontinued their first-line treatment, mainly due to intolerance (21%) or treatment failure (19%). At 18 months 73% of patients had achieved a complete cytogenetic response and 63% a major molecular response. Deep molecular responses (MR4.0 and MR4.5) were achieved in 69% and 56% of patients, respectively, at 48 months. All response milestones were achieved faster in patients treated upfront with a second-generation tyrosine kinase inhibitor, but ultimately patients initially treated with imatinib also reached similar levels of responses. The 6-year cumulative incidence of eligibility for a tyrosine kinase cessation attempt, according to EURO-SKI criteria, was 31%. Our findings show that in a ‘real-world’ setting the long-term outcome of patients treated with tyrosine kinase inhibitors is excellent and the conditions for an attempt to stop tyrosine kinase inhibitor therapy are met by a third of the patients.

Introduction ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1842

Multiple, randomized controlled trials have provided solid evidence for the efficacy and safety of tyrosine kinase inhibitors (TKI) as treatment for chronic myeloid leukemia (CML), but analyses from observational studies, gathered in patients who did not participate in clinical trials (the 'real-world') are scarce. Clinical trials use tight inclusion criteria, strict rules for monitoring and treatment algorithms and may, therefore, not fully reflect results in the general treatment population.1-3 Moreover, randomized controlled trials mainly focus on the outcome of the core study treatment, while some patients will not be able to continue their initial study treatment and are subsequently switched to an alternative treatment outside the trial.4,5 To study treatment choices and patients’ outcome across different treatment lines, real-world data contain important information for clinical practice. Incidence and survival have been the main focuses of the published reports of nationwide haematologica | 2017; 102(11)

Real-world treatment outcome in CML

population-based registries. The large European population-based EUTOS registry was the first to provide insight into real-world first- and second-line treatment patterns in relation to cytogenetic and molecular response.6 However, this report did not cover deep molecular responses or the proportion of patients meeting the criteria to attempt cessation of TKI treatment. Discontinuing TKI therapy is a novel opportunity in CML for patients with a durable deep molecular remission, of whom approximately half may successfully stop their TKI treatment while retaining a 'treatment-free remission'.7,8 In the current article, we report findings from a nationwide population-based CML registry in the Netherlands capturing data from newly diagnosed CML patients in the majority of hospitals in our country. Detailed information was collected on the patients’ characteristics and their treatment, both at baseline and during follow-up. Importantly, all TKI are available in the Netherlands and the Dutch health care system includes mandatory health care insurance which covers all CML care making it accessible to all patients. The aim of the current study was to provide a detailed overview of all aspects of CML care including responses to first and subsequent treatment lines with a specific focus on the impact of first-line treatment with imatinib compared to that of the second-generation TKI, dasatinib and nilotinib. We also sought to evaluate what proportion of patients become eligible to attempt to stop their TKI treatment.

Definitions and end-points Disease phase according to the European LeukemiaNet criteria,11 Sokal risk score,12 EUTOS long term survival (ELTS) score13 and Charlson Comorbidity index14 were calculated as described in the original publications. Complete cytogenetic response was defined as the absence of Philadelphia chromosome-positive metaphases examined by chromosome banding. BCR-ABL1 levels of ≤ 0.1%, ≤0.01% and ≤0.0032% on the international scale or ≥3 log, ≥4 log and ≥4.5 log reduction of BCR-ABL1 mRNA transcripts from baseline (in molecular laboratories not able to report on the international scale at the time) were defined as the molecular response end-points major molecular response (MMR), MR4.0 and MR,4.5 respectively. Undetectable BCR-ABL1 levels were classified as a MMR when control gene numbers were not available to determine the sensitivity of the test. In the case of a switch in TKI therapy, the clinical chart was reviewed for the reason why the treating physician had changed the therapy (‘treatment failure’ or ‘TKI intolerance’). As a proxy for effective and tolerable first-line treatment 'effective first-line treatment' was reached when patients continued their first-line TKI for at least 1 year after achieving MMR. For the determination of eligibility to stop TKI, the inclusion criteria for the EURO-SKI trial (ClinicalTrials.gov Identifier: NCT01596114) were applied. In short, this required TKI treatment for a minimum of 3 years, MR4.0 for at least 1 year and no history of TKI switch for a less than optimal treatment response. Disease progression was defined according to European LeukemiaNet criteria.11 Death due to CML was defined as death preceded by disease progression.

Data analysis Methods Data sources Data from two complementary Dutch population-based registries on CML patients (PHAROS-CML and Hemobase) were combined to cover the nationwide population of adult (≥18 years) CML patients diagnosed between January 2008 and April 2013 in all 12 Dutch provinces. PHAROS-CML is an extension of the Dutch Cancer Registry and consists of real-world data collected by trained data managers from medical records of patients newly diagnosed with CML between January 2008 and April 2013, covering the Dutch population in 11 out of 12 provinces.9 Approval for this comprehensive data collection was obtained by the individual hospital boards. The PHAROS-CML registry is a joint initiative of the Dutch-Belgian Hemato-Oncology Group (HOVON), the Institute of Medical Technology Assessment at the Erasmus University Rotterdam and the Netherlands Comprehensive Cancer Organization. Hemobase is a multidisciplinary web-based electronic patients’ record in the north-eastern part of the Netherlands covering the one province that was not included in the PHAROS-CML registry, which is the province of Friesland. The data in Hemobase were registered by physicians and laboratory employees10 and extracted from Hemobase to be combined with the PHAROS-CML registry by the first author (IG) who verified each record to ensure comparability. Together, data on all new CML patients in 75 of approximately 90 hospitals in the Netherlands were available. Additional molecular data were retrieved from all 15 Dutch molecular laboratories performing BCR-ABL1 diagnostic testing. Data on vital status and causes of death were obtained from the Netherlands Cancer Registry with a follow-up until the 1st February, 2016. The Medical Ethics Committee of the Erasmus Medical Center in Rotterdam authorized this study and the exemption from informed consent. The study was conducted in accordance with the Declaration of Helsinki. haematologica | 2017; 102(11)

Descriptive statistics were used to compare baseline characteristics between treatment groups. Patients treated upfront with nilotinib and dasatinib were clustered in the second-generation TKI group for comparison with imatinib-treated patients. Overall survival was analyzed using the Kaplan-Meier method with logrank test. All other time-dependent analyses were performed using the cumulative incidence competing risks method. Death and progression were always considered a competing risk. Additional competing risks per specific analysis are shown in Online Supplementary Table S1. A P-value of less than 0.05 was considered statistically significant. All analyses were performed using SPSS version 24 and R-software15 version 3.2.4.

Results Population description The registries covered 457 patients newly diagnosed with CML between January 2008 and April 2013 in the Netherlands. The 75 out of 90 Dutch hospitals participating in the registry differed in size and type (academic versus non-academic) and therefore provided an accurate representation of the performance of CML treatment in the Netherlands. At diagnosis, 434 patients were in chronic phase ( 95%), eight patients (2%) in accelerated phase and seven patients (1%) in blast crisis. Disease phase was unknown for eight patients (2%). For all 457 patients, follow-up information on survival status and death due to CML was available until February 1st, 2016. Disease-specific follow-up (treatment and/or response) for more than 1 year was available for 413 patients. Disease-specific follow-up was not available for five patients (1%) and for 39 patients (9%) the disease-specific follow-up was less than 1 year (26 died, 13 were lost to follow-up). 1843

I.G.P. Geelen et al.

First-line treatment Of the patients with detailed treatment information available, 382 patients (97%) were treated with first-line TKI therapy (imatinib n=295, 75%; nilotinib n=65, 17%; dasatinib n=22, 6%) (Table 1), of whom 43% had received hydroxyurea prior to or simultaneous with the start of first-line TKI treatment. Hydroxyurea was the only reported treatment in six patients (2%) and two elderly patients with major comorbidities did not receive any treatment at all (0.5%). One patient was pregnant at diagnosis and was therefore treated first-line with interferon. One patient was in chronic phase at diagnosis, but was initially treated with leukapheresis and daunorubicin because of a hyperviscosity syndrome due to hyperleukocytosis at presentation (leukocytes 525x109/L). This patient died 1 day after diagnosis. Leukapheresis was performed in two other patients prior to imatinib treatment with white blood cell counts of 421x109/L and 606x109/L. The majority of patients received the first-line TKI at standard doses: imatinib 400 mg QD (82%), nilotinib 300 mg BID (89%) and dasatinib 100 mg QD (95%). First-line dose adjustments were most frequently observed during imatinib treatment, especially dose escalations and sequential dose adjustments and/or interruptions (Figure 1).

Response Cytogenetic response data were available for 246 out of 434 chronic-phase-CML patients (57%) and molecular response data for 326 out of 434 chronic-phase-CML patients (75%). Patients with no cytogenetic and/or no molecular response data available were significantly older, had a higher comorbidity index, were less frequently included in first-line clinical trials and more frequently treated in non-academic hospitals. The median time to first complete cytogenetic response was 10 months. The cumulative incidence of complete cytogenetic response was 55% (95% CI: 49-62%) at 12 months and 73% (95% CI: 67-79%) at 18 months (Figure 2). The median time to first MMR was 13 months. The cumulative incidence of MMR was 47% (95% CI: 42-52%) at 12 months and 63% (95% CI: 57-68%) at 18 months (Figure 2). Deeper molecular responses were achieved after a median treatment duration of 30.5 months (MR4.0) and 43 months (MR4.5) with cumulative incidence rates of 41% (95% CI: 3546%) and 69% (95% CI: 63-74%) for MR4.0 after 24 months and 48 months, respectively, and cumulative incidence rates of 30% (95% CI: 25-35%) and 56% (95% CI: 50-62%) for MR4.5 after 24 months and 48 months, respectively (Figure 2). Both cytogenetic and molecular response milestones were achieved faster in patients treated upfront with a second-generation TKI (Figure 3, Online Supplementary Figures S1-S3). On initial treatment cumulative incidence rates of the achievement of all response milestones were significantly lower in patients treated with first-line imatinib (Figure 3A, Online Supplementary Figures S1A, S2A and S3A), but eventually, after switching TKI treatment, the same response rates were achieved (Figure 3B, Online Supplementary Figures S1B, S2B and S3B).

Switching/discontinuation of tyrosine kinase inhibitor treatment and time to effective treatment Within the first 3 years after diagnosis, 44% (95% CI: 38-50%) of the 382 patients on first-line TKI treatment discontinued their first-line treatment; 21% (95% CI: 171844

26%) due to TKI intolerance, 19% (95% CI: 15-24%) due to treatment failure and 3% (95% CI: 1-5%) for other or unknown reasons (Figure 4A). The most frequently observed other reasons for first-line TKI discontinuation were trial inclusion within 4 months after starting treatment (n=5), inclusion in a TKI discontinuation trial (n=3) and pregnancy (n=3). The 3-year cumulative incidence of TKI discontinuation was not different between patients treated first-line with imatinib or second-generation TKI (46% versus 38%, P=0.104) (Figure 4B). TKI discontinuation due to TKI failure was significantly more common in patients treated with first-line imatinib (21% versus 13%, P=0.046) (Figure 4C). TKI discontinuation due to TKI intolerance did not differ according to whether the first-line treatment was with imatinib or a second-generation TKI (21% versus 24%, P=0.854) (Figure 4D). In total, up to five subsequent treatment switches were observed (Online Supplementary Figure S4). An overview of the reported types of intolerance on all treatment lines and the actions that followed as a result of the intolerance can be found in the Online Supplementary Results section and Online Supplementary Table S2. The median time to sustained effective and tolerable first-line treatment was 33 months and the cumulative incidence was 56% (95% CI: 49-64%) after 4 years of treatment: 51% (95% CI: 43-59%) with first-line imatinib

Table 1. Patients’ characteristics at diagnosis for the sub-analysis of patients with chronic-phase-CML treated with a TKI.

First line treatment

Imatinib, n=295

Male gender, n (%) 168 (57) Age, years median (IQR) 58 (42-69) Year of diagnosis, n (%) 2008 87 (30) 2009 77 (26) 2010 66 (22) 2011 40(14) 2012 – April 2013 25 (9) Age-adjusted Charlson Comorbidity index, n (%) 0 98 (33) 1-2 87 (30) 3-4 65 (22) ≥5 45 (15) Sokal risk group, n (%) Low 66 (26) Intermediate 112 (44) High 77 (30) Unknown 40 ELTS risk group, n (%) Low 124 (49) Intermediate 88 (35) High 43 (17) Unknown 40 Treating hospital, n (%) Non-academic 226 (77) Academic 69 (23) Unknown 0 Inclusion in 1st line clinical trial, n (%) No 247 (86) Yes 40 (14) Unknown 8

2GTKI, n=87* 51 (59) 57 (45-68) 3 (3) 0 (0) 14 (16) 32 (37) 38 (44) 27 (31) 33 (38) 18 (21) 9 (10) 14 (18) 36 (47) 26 (34) 11 33 (43) 30 (40) 13 (17) 11 54 (62) 33 (38) # 0 66 (76) 21 (24) # 0

*Nilotinib, n=65; dasatinib, n=22; #P<0.05 compared to imatinib-group. 2GTKI: secondgeneration tyrosine kinase inhibitor; ELTS: EUTOS long-term survival.

haematologica | 2017; 102(11)

Real-world treatment outcome in CML

treatment and 77% (95% CI: 57-97%) with first-line, second-generation TKI treatment (P<0.001).

patients treated with frontline imatinib and frontline second-generation TKI, respectively.

Progression during tyrosine kinase inhibitor therapy

Survival

The median observation period for progression of TKItreated chronic-phase-CML patients was 27 months (range, 0-82 months). A total of 17 patients progressed within 3 years, for a cumulative incidence of 3% after 1 year and 6% after 3 years. Progression occurred in 16 patients treated with first-line imatinib whereas it was observed in only one patient treated upfront with a second-generation TKI; the cumulative incidence rates for progression after 3 years were 7% versus 1% (P=0.193) in

The median observation time for survival of patients in chronic phase at diagnosis was 5 years and 7 months (range, 34-97 months). Eighty-two patients died during follow-up (19%). Overall survival rates after 1, 2 and 4 years were 96% (95% CI: 94-98%), 92% (95% CI: 9095%) and 85% (95% CI: 81-88%), respectively. No significant differences in overall survival were observed when overall survival was stratified by type of first-line TKI treatment (P=0.244). The cumulative incidence of death

Figure 1. Dose adjustments of firstline tyrosine kinase inhibitors. *P<0.05.

Figure 2. Achievement of cytogenetic and molecular response milestones. Cumulative incidence of first and unconfirmed achievement of CCyR, MMR, MR4.0 and MR4.5. CCyR: complete cytogenetic response; MMR: major molecular response; MR4.0: 0.01% on international scale; MR4.5: 0.0032% on international scale.

haematologica | 2017; 102(11)

1845

I.G.P. Geelen et al.

due to CML was 1% after 1 year, 2% after 2 years and 3% after 4 years. Again, death due to CML was not significantly different between patients treated first-line with imatinib or a second-generation TKI (P=0.208).

Tyrosine kinase inhibitor stop eligibiity Eligibility for a TKI stop attempt according to the EURO-SKI criteria was evaluated for all 382 patients who started first line TKI treatment. A total of 43 patients

Figure 3. Achievement of major molecular response on frontline imatinib and second-generation tyrosine kinase inhibitors. (A) Cumulative incidence of MMR on initial treatment. (B) Cumulative incidence of overall MMR achievement (including patients who switched TKI). MMR: major molecular response; 2GTKI: second-generation tyrosine kinase inhibitors.

Figure 4. Discontinuation of first-line tyrosine kinase inhibitors. (A) Cumulative incidence of all causes of TKI discontinuation. (B) Cumulative incidence of TKI discontinuation, a comparison of imatinib with 2GTKI. (C) Cumulative incidence of TKI discontinuation due to TKI failure. (D) Cumulative incidence of TKI discontinuation due to TKI intolerance. 2GTKI: second-generation tyrosine kinase inhibitor.

1846

haematologica | 2017; 102(11)

Real-world treatment outcome in CML

(11%) met the eligibility criteria. During follow up 131 patients (34%) experienced a competing risk making them ineligible for a (EURO-SKI) stop attempt: first-line TKI failure (43%), less than 3 years on first- and second-line TKI (29%), death (20%) and progression (8%). Censoring occurred for 208 patients (55%): lost to disease-specific follow-up within 3 years (80%), not (yet) achieved 3x MR4.0 in 1 year (19%), emigration (1%) and unknown reason for TKI switch (0.5%). The cumulative incidences of stop attempt eligibility after 4 and 6 years were 24% (95% CI, 17-30%) and 31% (95% CI, 23-38), respectively (Figure 5A). The cumulative incidence of TKI stop eligibility after first line 2GTKI treatment was numerically higher than after imatinib treatment, but this did not reach statistical significance (P=0.147) (Figure 5B).

Treatment and outcome of patients presenting in advanced-phase-chronic myeloid leukemia The 15 patients who presented in advanced disease phase at diagnosis (8 accelerated phase and 7 blast crisis) were mainly treated upfront with a TKI (6 accelerated phase and 4 blast crisis). One patient in blast crisis was started on chemotherapy + TKI treatment and treatment was unknown for two other patients in blast crisis. Allogeneic stem cell transplantation was performed in three patients with accelerated-phase-CML and three with blast crisis. The median survival after diagnosis was 1.4 years in patients with accelerated-phase-CML and not reached in the blast crisis patients. The 5-year overall survival rates were 38% and 57%, respectively.

Discussion Our nationwide, population-based study confirms the excellent results of TKI treatment in CML patients observed in randomized controlled trials outside clinical trials and adds insights into the patterns of TKI switching and patientsâ&#x20AC;&#x2122; outcome. Moreover, this study presents the first real-world data on the proportion of patients becoming eligible for an attempt to achieve treatment-free remission.

The baseline characteristics of the Dutch CML population are comparable to those of patients in other population-based CML registries.6,16,17 In contrast, patients included in the ENESTnd and DASISION randomized controlled trials18,19 were 10 years younger than those in our realworld cohort. Since nilotinib and dasatinib were both registered for first-line treatment in December 2010, the majority of patients treated with second-generation TKI were diagnosed in the second half of the inclusion period and therefore experienced a shorter follow-up period. Because of the relatively low number of individual patients treated upfront with nilotinib and dasatinib, it was decided to combine these two groups of patients for comparisons with the imatinib-treated patients. Our real-world observations show that TKI therapy is effective and tolerable in the majority of patients, but frequent dose adjustments and temporary treatment interruptions are required (45% of patients treated upfront with imatinib and 31-32% of patients treated with a second-generation TKI). Only half (54%) of all imatinib-treated patients with chronic-phase-CML were estimated to remain on first-line TKI 3 years after treatment initiation in our observational cohort. This rate of TKI treatment persistence is notably lower than the rates observed in the ENESTnd and DASISION randomized controlled trials in which 62% and 69% of patients were still on their firstline imatinib treatment after 3 years.20,21 For nilotinib and dasatinib these rates were 71-74% and 71%, respectively,20,21 while we observed a 3-year treatment persistence rate of only 62% in patients treated with second-generation TKI. These differences were mainly due to the higher TKI discontinuation rates due to TKI intolerance observed in our real-world setting. It can be hypothesized that the threshold for TKI switching due to intolerance for both physicians and patients is lower in subjects treated outside a clinical trial. In concordance with our study, Castagnetti et al. reported a treatment persistence rate of 59% in patients treated with imatinib after a median of 48 months follow-up in a real-world setting in Italy, also because of frequent TKI switches due to intolerance.22 The observations of earlier and higher rates of achieve-

Figure 5. Eligibility to attempt cessation of tyrosine kinase inhibitor therapy according to EURO-SKI criteria. (A) Cumulative incidence of overall eligibility of patients with chronic-phase-CML treated first-line with a TKI. (B) Comparison of cumulative incidences in patients treated upfront with imatinib or 2GTKI. 2GTKI: second-generation tyrosine kinase inhibitor.

haematologica | 2017; 102(11)

1847

I.G.P. Geelen et al.

ment of the important response milestones, complete cytogenetic response and MMR, in patients treated with first-line, second-generation TKI in our real-world cohort confirm the reproducibility of the superior efficacy results of first-line, second-generation TKI found in the two large randomized controlled trials.4,5 Of note, the 3-year MMR rate achieved in our real-world cohort with first-line, second-generation TKI (84%) is even higher than the rates recorded in the randomized controlled trials (69-73%). Differences in methodology and study design of our observational study and the randomized controlled trials may have contributed to the variation in findings. A limitation to population-based cohort studies is that patients are not monitored as strictly as during the closely supervised clinical follow-up in trials. Furthermore, in our registry, patients with more favorable baseline characteristics more often had cytogenetic and molecular response analyses performed. This may in part explain the relatively high remission rates we observed. Real-world data not only give insight into responses on first-line treatments, but also provide information on the overall response on subsequent treatment lines, in contrast to data from the currently published clinical trials. For example, the data from our nationwide registry demonstrate that 91% of all chronic-phase-CML patients treated with a TKI eventually achieved complete cytogenetic response and 88% reached a MMR after 4 years, whereas the results from randomized controlled trials on core treatment were significantly lower at this time point. The observational data on overall response do, therefore, reflect relevant patientsâ&#x20AC;&#x2122; outcomes much better than the clinical trial results do. Our overall response rates are comparable to those observed in other population-based registries.16,22,23 The observation that patients treated upfront with imatinib or a second-generation TKI eventually achieved comparable complete cytogenetic response and MMR rates, whether or not preceded by one or more treatment switches, has also been recognized before.23 An analysis of long-term molecular response on first-line treatment even showed that patients treated with imatinib 400 mg QD only can reach MMR rates nearly similar to those treated with imatinib 800 mg QD, nilotinib and dasatinib after 5 years.24 To our knowledge, this is the first observational study comparing deep molecular responses achieved on imatinib with deep molecular responses achieved on secondgeneration TKI. Of note, comparative analyses were hampered in our study by a relatively low number of patients receiving second-generation TKI and their shorter follow-

References 1. Pulte D, Gondos A, Redaniel MT, Brenner H. Survival of patients with chronic myelocytic leukemia: comparisons of estimates from clinical trial settings and populationbased cancer registries. Oncologist. 2011;16(5):663-671. 2. Latagliata R, Carmosino I, Vozella F, et al. Impact of exclusion criteria for the DASISION and ENESTnd trials in the front-line treatment of a 'real-life' patient population with chronic myeloid leukaemia. Hematol Oncol. 2017;35(2):232-236. 3. Mauro MJ, Davis C, Zyczynski T, Khoury

1848

up period. Despite this, we were still able to show that MR4.0 and MR4.5 were achieved significantly faster in the real-world when treatment was initiated with a secondgeneration TKI than when the initial treatment was imatinib, independently of subsequent treatment lines. In the randomized controlled trials both higher24 and lower4,5 cumulative response rates were observed than in our reallife cohort. These deep responses are especially interesting in the light of eligibility for an attempt to stop TKI therapy. Together with the duration of TKI treatment and switching history, a durable deep molecular remission is the main selection criterion currently used in trials investigating treatment-free remission. In a previous analysis of patients in first-line clinical trials, after 8 years of imatinib treatment, the cumulative incidence of stable (â&#x2030;Ľ 24 months) MR4.5, determined on an intent-to-treat basis, was 36.5%, suggesting that this proportion may be eligible for treatment-free remission.25 In our population-based study, the eligibility criteria of the largest treatment-free remission trial to date, EURO-SKI, were used to evaluate this endpoint and showed a cumulative incidence of 31% after 6 years. Although we observed a higher eligibility rate in patients treated upfront with a second-generation TKI than in those treated with imatinib, the difference did not reach statistical significance possibly because of the relatively low number of patients who had started on a second-generation TKI and had a follow-up beyond 3 years. In conclusion, this population-based analysis showed overall favorable treatment responses compared to those in randomized controlled trials, which could be attributed to dose adjustments and subsequent treatment lines in the real-world setting. It also showed that the long-term outcome of patients initially treated with imatinib is excellent when these patients are switched to second-generation TKI when needed. The cumulative incidence of patients eligible to attempt to stop their TKI to achieve treatmentfree remission was 31% after 6 years when the EURO-SKI criteria were applied. Acknowledgments The authors would like to thank Peter Huijgens, currently chairman of the Netherlands Comprehensive Cancer Organisation (IKNL) for initiating the Pharos registry. We thank Tom Wiggers, Wencke de Jager, Sanne Nijssen and Jolie Cheung for entering patientsâ&#x20AC;&#x2122; data into the Pharos database. We thank Marianne van der Mark from the IKNL for retrieving survival data from the Netherlands Cancer Registry and all the hospitals and molecular laboratories in the Netherlands that participated in the Pharos and Hemobase registries.

HJ. The role of observational studies in optimizing the clinical management of chronic myeloid leukemia. Ther Adv Hematol. 2015;6(1):3-14. 4. Cortes JE, Saglio G, Kantarjian HM, et al. Final 5-year study results of DASISION: The dasatinib versus imatinib study in treatment-naive chronic myeloid leukemia patients trial. J Clin Oncol. 2016;34(20): 2333-2340. 5. 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. 6. 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. 7. Hughes TP, Ross DM. Moving treatmentfree remission into mainstream clinical practice in CML. Blood. 2016;128(1):17-23. 8. Saussele S, Richter J, Hochhaus A, Mahon FX. The concept of treatment-free remission in chronic myeloid leukemia. Leukemia. 2016;30(8):1638-1647. 9. Huijgens P, Posthuma E, Coebergh J, van de Poll-Franse L, Uyl-de Groot C, Sonneveld P.

haematologica | 2017; 102(11)

Real-world treatment outcome in CML

10.

11.

12.

13.

14.

A 'population based registry' for hematooncology. Ned Tijdschr Hematol. 2010;7(8):321-325. [Dutch] Hoogendoorn M, Joosten P, Storm H, Kibbelaar R. Hemobase: an intelligent electronic patient file as aid for hemato-oncology care. Ned Tijdschr Hematol. 2009;6(3):104-110. [Dutch] Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia: 2013. Blood. 2013;122(6):872884. Sokal JE, Gomez GA, Baccarani M, et al. Prognostic significance of additional cytogenetic abnormalities at diagnosis of Philadelphia chromosome-positive chronic granulocytic leukemia. Blood. 1988;72 (1):294-298. Pfirrmann M, Baccarani M, Saussele S, et al. Prognosis of long-term survival considering disease-specific death in patients with chronic myeloid leukemia. Leukemia. 2016;30(1):48-56. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40(5):373-383.

haematologica | 2017; 102(11)

15. Team RC. R: A Language and Environment for Statistical Computing. 2016. 16. Hoglund M, Sandin F, Hellstrom K, et al. Tyrosine kinase inhibitor usage, treatment outcome, and prognostic scores in CML: report from the population-based Swedish CML registry. Blood. 2013;122(7):12841292. 17. Beinortas T, Tavoriene I, Zvirblis T, Gerbutavicius R, Jurgutis M, Griskevicius L. Chronic myeloid leukemia incidence, survival and accessibility of tyrosine kinase inhibitors: a report from population-based Lithuanian haematological disease registry 2000-2013. BMC Cancer. 2016;16:198. 18. Saglio G, Kim DW, Issaragrisil S, et al. Nilotinib versus imatinib for newly diagnosed chronic myeloid leukemia. N Engl J Med. 2010;362(24):2251-2259. 19. 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. 20. Larson RA, Hochhaus A, Hughes TP, et al. Nilotinib vs imatinib in patients with newly diagnosed Philadelphia chromosome-positive chronic myeloid leukemia in chronic phase: ENESTnd 3-year follow-up.

Leukemia. 2012;26(10):2197-2203. 21. 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. 22. Castagnetti F, Di Raimondo F, De Vivo A, et al. A population-based study of chronic myeloid leukemia patients treated with imatinib in first line. Am J Hematol. 2017;92(1):82-87. 23. Hoffmann VS, Baccarani M, Hasford J, et al. Treatment and outcome of 2904 CML patients from the EUTOS population-based registry. Leukemia. 2017;31(3):593-601. 24. Jain P, Kantarjian H, Alattar ML, et al. Longterm molecular and cytogenetic response and survival outcomes with imatinib 400 mg, imatinib 800 mg, dasatinib, and nilotinib in patients with chronic-phase chronic myeloid leukaemia: retrospective analysis of patient data from five clinical trials. Lancet Haematol. 2015;2(3):e118-128. 25. Branford S, Yeung DT, Ross DM, et al. Early molecular response and female sex strongly predict stable undetectable BCR-ABL1, the criteria for imatinib discontinuation in patients with CML. Blood. 2013;121(19): 3818-3824.

1849

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Myeloid Leukemia

Ferrata Storti Foundation

Senescence is a Spi1-induced anti-proliferative mechanism in primary hematopoietic cells Laure Delestré,*1,2 Hengxiang Cui,*1,2,3 Michela Esposito,1,2 Cyril Quiveron,1,2 Elena Mylonas,1,2 Virginie Penard-Lacronique,1,2 Oliver Bischof4,5,6 and Christel Guillouf,1,2,3,6

Institut Gustave Roussy, Université Paris-Saclay, Villejuif; 2INSERM U1170, Villejuif; Previous address: Institut Curie, Paris; 4Institut Pasteur, Unit of Nuclear Organization and Oncogenesis, Paris; 5INSERM U993, Paris and 6Centre national de la recherche scientifique (CNRS), Paris, France 1 3

Haematologica 2017 Volume 102(11):1850-1860

LD and HC contributed equally to this work

ABSTRACT

Correspondence: christel.guillouf@gustaveroussy.fr

Received: October 5, 2016. Accepted: September 6, 2017. Pre-published: September 14, 2017. doi:10.3324/haematol.2016.157636 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1850 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1850

ranscriptional deregulation caused by epigenetic or genetic alterations is a major cause of leukemic transformation. The Spi1/PU.1 transcription factor is a key regulator of many steps of hematopoiesis, and limits self-renewal of hematopoietic stem cells. The deregulation of its expression or activity contributes to leukemia, in which Spi1 can be either an oncogene or a tumor suppressor. Herein we explored whether cellular senescence, an anti-tumoral pathway that restrains cell proliferation, is a mechanism by which Spi1 limits hematopoietic cell expansion, and thus prevents the development of leukemia. We show that Spi1 overexpression triggers cellular senescence both in primary fibroblasts and hematopoietic cells. Erythroid and myeloid lineages are both prone to Spi1-induced senescence. In hematopoietic cells, Spi1-induced senescence requires its DNA-binding activity and a functional p38MAPK14 pathway but is independent of a DNA-damage response. In contrast, in fibroblasts, Spi1-induced senescence is triggered by a DNA-damage response. Importantly, using our well-established Spi1 transgenic leukemia mouse model, we demonstrate that Spi1 overexpression also induces senescence in erythroid progenitors of the bone marrow in vivo before the onset of the pre-leukemic phase of erythroleukemia. Remarkably, the senescence response is lost during the progression of the disease and erythroid blasts do not display a higher expression of Dec1 and CDKN1A, two of the induced senescence markers in young animals. These results bring indirect evidence that leukemia develops from cells which have bypassed Spi1-induced senescence. Overall, our results reveal senescence as a Spi1-induced antiproliferative mechanism that may be a safeguard against the development of acute myeloid leukemia.

Introduction Transcription factors (TFs) are major regulators of hematopoietic cell differentiation and are often deregulated in acute myeloid leukemia (AML). Spi1/PU.1 is a member of the ETS family, and accurate expression levels are critical for specifying cell fate and for proper hematopoietic differentiation.1 Spi1 plays a pivotal role in hematopoietic stem cell (HSC) self-renewal and in myeloid and B lymphoid differentiation.2-5 It acts by controlling the expression of a subset of lineage-specific genes involved in hematopoiesis6 and the expression of ubiquitous cell cycle regulators.5,7,8 Although the involvement of Spi1 alterations in tumor formation is well-established, the mechanisms by which Spi1 drives the development of AML are still not clear and seem to be complex. A reduction in Spi1 levels or an indirect inhibition of its activity by cooperating factors involved in leukemic transformation causes AML in humans.9-12 Rare cases of heterozygous inactivating mutations have also been haematologica | 2017; 102(11)

Spi1/PU.1 triggers cellular senescence

described in human AML.13,14 Studies using several mouse models of Spi1 reduction have corroborated the involvement of Spi1 in the development of AML.15-19 Consistent with the role of Spi1 in controlling growth arrest and promoting myeloid differentiation, its re-expression in knocked down or mutated Spi1 cells or in leukemic progenitors in which Spi1 expression is suppressed induces growth arrest and monocytic differentiation.10,15,20 Despite this tumor-suppressor function, Spi1 is required for the maintenance of leukemic cells in AMLs with specific fusion genes.21-23 Spi1 also displays oncogenic activity, promoting the proliferation of erythroid progenitors in mice.24,25 High Spi1 expression levels in mice cause a preleukemic syndrome characterized by an increase in the number of hyper-proliferative erythroid progenitors in which differentiation and apoptosis are blocked.25-27 In these cells, Spi1 induces replication stress and accelerates genetic mutability.28 haematologica | 2017; 102(11)

Figure 1. Ectopic expression of Spi1 and HRASV12 induces growth arrest and senescence in BJ and WI-38 fibroblast cells. (A) Western blot analysis of Spi1, HRASV12, and the senescence marker Dec1 in BJ and WI-38 cells subjected to the retroviralmediated expression of Spi1, HRASV12 or an empty vector as a control 10 days after puromycin selection. β-actin served as the loading control. (B) Population doublings (PDs) of BJ (left panel) and WI-38 (right panel) cells transduced as described in (A) over the indicated periods of time. Day 0 was the first day after puromycin selection. PDs for each time point are the mean of triplicate experiments. (C) Representative SA-βgal staining and percent of SA-βgal positive cells (histograms) in samples of BJ and WI-38 cells subjected to the retroviral-mediated expression of Spi1, HRASV12 or an empty vector as a control 10 days after puromycin selection. Magnification of images, 200X. The means ± SD of at least 3 independent experiments are shown. **P<0.005; ***P<0.0005 from two-tailed Student’s t-tests. (D) Flow cytometric detection of SA-βgal activity using C12FDG as a fluorogenic substrate in cells as described in (C).

Increasing evidence points to a critical role for cellular senescence as a barrier to malignant transformation. This tumor suppressive mechanism is activated when cells are exposed to exogenous or endogenous stresses such as supraphysiological oncogenic signaling. Oncogene-induced senescence (OIS) is a mechanism that limits cell hyperproliferation through a stable cell cycle arrest process,29 thus blocking the expansion of cells at the pre-cancerous stage in solid tumors.30,31 The expression of the hematopoietic oncogenes HRASV12, BCR-ABL, CBFBMYH11 or RUNX1-ETO in primary HSCs and committed progenitors (HSCPs) elicits a senescence response,32 and OIS acts as an antitumoral barrier in NRASV12-induced lymphomas and MLL-ENL-induced AML.33,34 Senescence can be triggered, at least in part, by DNA replication stress, mainly due to the over-activation of replication origin firing, and an associated DNA-damage response (DDR)33,35-37 or independently of DNA replication stress.32 1851

L. Delestré et al.

Although the role of OIS in limiting the proliferation of primary fibroblasts and epithelial cells and in protecting against the progression of solid tumorigenesis is now well characterized, the extent of the role of OIS in primary HSCPs and its protective effect against leukemic processes have yet to be fully explained. Because Spi1 is required to maintain murine HSCs in a quiescent state and to restrict HSC division,5 we examined whether cellular senescence is a mechanism by which Spi1 restricts cell proliferation and if it protects against the development of AML. Our results reveal that Spi1 restrains cell expansion by inducing senescence in primary HSCPs as well as in primary fibroblasts in vitro. The mechanistic underpinnings of this response are distinct for the two cell types. We took advantage of the well-characterized Spi1 transgenic (TgSpi1) mouse model, in which Spi1 overexpression results in oncogenic behavior,25 to analyze senescence in vivo. We established that Spi1 overexpression induces senescence of bone marrow cells in mice, and

1852

that this anti-proliferative process is lost during the leukemic progression observed in TgSpi1 mice. Overall, the study herein identifies a new pathway by which Spi1 controls hematopoietic cell growth.

Methods Mice and cell culture TgSpi1 mice have been described previously.25 Wild-type (WT) mice were obtained from crossing heterozygous TgSpi1 mice. Early passage foreskin human fibroblasts BJ cells and fetal lung fibroblasts WI-38 cells obtained from American Type Cell Culture collection (BJ batch number: 59899913; WI-38 batch number: 58483158) were cultured using Dulbecco's modified Eagle's medium (DMEM) with high glucose supplemented with 10% (v/v) fetal bovine serum (FBS) at 37°C in a humidified atmosphere containing 5% CO2 and 3% O2. HEK293EBNA and HeLa cells were maintained in DMEM medium with high glucose, 10% FBS at 37°C and 5% CO2.

Figure 2. Overexpression of Spi1 and HRASV12 in Lin-Kit+Sca1+ (LSK) cells leads to senescence. (A) Western blot analysis of Spi1, HRASV12 and the senescence marker Dec1 in hematopoietic cells subjected to the retroviral-mediated expression of Spi1, HRASV12 or an empty vector. Protein extracts of GFP-positive sorted cells were analyzed 7 days post-infection as described in Online Supplementary Figure S2. α-adaptin served as the loading control. (B) Number of total living cells retrovirally transduced with Spi1 and HRASV12 or an empty vector (control), at the indicated periods of time. The means ± SEM of at least 3 independent experiments are shown. (C) Representative SA-βgal staining and mean percentages of SA-βgal positive cells (histograms) in samples of hematopoietic cells subjected to the retroviral-mediated expression of Spi1, HRASV12 or an empty vector as a control. SA-βgal assays for sorted GFP-positive cells were performed 7 days post-infection as described in Online Supplementary Figure S2. The counting of GFP-positive SA-βgal cells was performed in 9 randomly selected fields with a total of more than 2000 cells from each group. Magnification of images, 200X. The means ± SD of at least 3 independent experiments are shown. **P<0.005; ***P<0.0005 from twotailed Student’s t-test. (D) Flow cytometric detection of SA-βgal activity using C12RG as a fluorogenic substrate in cells retrovirally transduced with Spi1, HRASV12 or an empty vector. The histograms represents the % of C12RG positive cells among GFP-positive cells. The means ± SD of at least 3 independent experiments are shown. **P<0.005 from two-tailed Student’s t-test.

haematologica | 2017; 102(11)

Spi1/PU.1 triggers cellular senescence

A detailed description of Methods is included in the Online Supplementary Appendix.

Results Ectopic expression of Spi1 induces senescence in primary fibroblasts and primary hematopoietic cells To understand the role of Spi1 in senescence, we first studied the consequences of its ectopic expression in the well-established senescence cell model based on normal

human primary fibroblasts (strains BJ and WI-38).38 We transduced cells with a retroviral construct expressing Spi1 or an activated RAS mutant (HRASV12) as a positive control for OIS (Figure 1A and Online Supplementary Figure S1). The ectopic expression of Spi1 or HRASV12 both led to senescence that was characterized by stable cell cycle arrest (Figure 1B), increased senescence-associated betagalactosidase (SA-βgal) activity, as measured by cytochemical staining and cytometric analyses (Figure 1C,D), and increased protein levels of the senescence biomarker Dec1 (Figure 1A).29,39

Figure 3. Spi1 triggers senescence in granulocytes, monocytes/macrophages and myeloid and erythroid progenitor cells. (A) Distribution of the cells according to CD11b, Gr1 and F4/80 myeloid markers and SA-βgal activity using C12RG as fluorogenic substrate by flow cytometry among total GFP+, GFP+C12RG– or GFP+C12RG+, 7 days after transduction of LSK cells with MSCV-Spi1 or MSCV control vectors. The means ± SEM of 3 independent experiments are shown. (B) Fold change (FC) of the % of C12RG positive cells between Spi1-overexpressing cells (hatched histograms) and control cells among GFP+ cells inside each indicated cells compartment, granulocytes (CD11b+F4/80-Gr1+), monocytes/macrophages (CD11b+F4/80+Gr1+), immature myeloid progenitors (CD11b+F4/80–Gr1–) or CD11b– cells. *P<0.05 from two-tailed Student’s t-test. (C and D) Representative SA-βgal staining using cytochemical staining and FC of the % of SA-βgal positive cells in MEP (C) and in GMP (D) progenitor cells subjected to the retroviral-mediated expression of Spi1 (MSCV-Spi1) relative to empty vector as a control (MSCV). SA-βgal assays for sorted GFP-positive cells were performed 4 or 6 days post-infection for MEP and GMP, respectively. The counting of GFP-positive SA-βgal cells was performed in 3 randomly selected fields with a total of more than 100 cells from each group. Magnification of images, 200X. Results are from 3 independent experiments. GFP: green fluorescent protein; MSCV: murine stem cell virus.

haematologica | 2017; 102(11)

1853

L. Delestré et al.

We next investigated the induction of senescence in primary hematopoietic stem/progenitor cells, defined as Lin–Kit+Sca1+ (LSK) cells (Online Supplementary Figure S2). LSK cells were transduced with retroviral vectors expressing green fluorescent protein (GFP) alone (vector) or coexpressing GFP with Spi1 or HRASV12 (Figure 2A). Total cell number was reduced over time by Spi1- and HRASV12 -overexpression compared with the numbers of cells infected with the control vector (Figure 2B). Moreover, Spi1 and HRASV12 expression caused an increase in the proportion of cells that stained positive for SA-βgal activity among GFP-positive cells compared with the control

cells (8.3% and 11.1% for Spi1- and HRASV12 -expressing cells, respectively, and 1.1% for control cells; Figure 2C). We also observed an increase in SA-βgal activity using C12RG as a fluorogenic substrate in Spi1- or H-RASV12expressing GFP-positive cells compared with activity in control cells (51% and 61 % for Spi1- and HRASV12expressing cells, respectively, and 24% for control cells; Figure 2D) and Spi1- or HRASV12-expressing GFP-negative cells (Online Supplementary Figure S3A). The proportion of GFP-positive cells was positively correlated with the proportion of senescent cells (Online Supplementary Figure S3B). In agreement with the increase of SA-βgal activity,

Figure 4. Spi1 induces senescence through distinct mechanisms in fibroblasts and in hematopoietic cells. (A-B) Western blot analysis of CDKN1A and CDKN2A in BJ and WI-38 cells subjected to the retroviral-mediated expression of Spi1, Dβ4-Spi1, HRASV12 or an empty vector (vector) 10 days after puromycin selection. α-adaptin served as the loading control. (C-D) Western blot analysis of S15-phosphorylated (p-p53) and total p53, S139-phosphorylated (p-H2AX) and total H2AX, and Thr180/182-phosphorylated and total p38MAPK14 (p-P38 and P38) in the same cells described in (A). (E, G) Western blot analysis of CDKN1A and CDKN2A (E) and S15-phosphorylated and total p53, S139-phosphorylated and total H2AX, and Thr180/182-phosphorylated and total p38MAPK14 (p-P38 and P38) (G) in hematopoietic cells subjected to the retroviral-mediated expression of Spi1, HRASV12 or an empty vector. Protein extracts of GFP-positive sorted cells were analyzed 7 days post-infection as described in Online Supplementary Figure S2. α-adaptin served as the loading control. (F). Thr180/182-phosphorylated and total p38MAPK14 (p-P38 and P38), and CDKN1A and CDKN2A expression were analyzed via Western blotting in hematopoietic cells subjected to the retroviral-mediated expression of Spi1, Dβ4-Spi1 or an empty vector. Protein extracts of GFP-positive sorted cells were analyzed 7 days post-infection as described in Online Supplementary Figure S2. α-adaptin served as the loading control.

1854

haematologica | 2017; 102(11)

Spi1/PU.1 triggers cellular senescence

are not different to the non-senescent GFP+ C12RG– cells (Figure 3A, middle histogram). Moreover, Spi1 induced 2.75-fold more senescent cells among the granulocytes compared to murine stem cell virus (MSCV)-transduced control cells (27% to 10%, respectively), and 1.5-fold more among the monocytes/macrophages (55% to 40%, respectively) (Figure 3B). These results indicate that granulocytes and monocytes/macrophages are prone to Spi1induced senescence even if Spi1 favors only differentiation of monocytes/macrophages. Spi1 serves as oncogene in erythroid but not in myeloid progenitors, wherein it is a tumor suppressor. Therefore, to investigate whether Spi1 overexpression is associated with senescence in both types of cells, we studied the senescence in Spi1-overexpressing megakaryocyte-erythroid (MEP) and granulocyte-monocyte (GMP) progenitor cells. Sorted MEP (Lin–Sca–Kit+CD34–CD16/32–) and GMP (Lin–Sca–Kit+CD34+CD16/32+) were transduced with retroviral vectors expressing GFP alone (vector) or coexpressing GFP with Spi1, and senescence was evaluated by measuring SA-βgal activity using cytochemical staining (Figure 3C). Spi1 overexpression caused an increase in the proportion of SA-βgal+ cells among GFP-positive GMP progenitor cells compared with the control cells (1.7-fold more for Spi1-overexpressing cells compared to the MSCV control cells; 80.7% and 48.8%, respectively), confirming that myeloid cells are prone to Spi1-induced senescence. As expected,26 we found that Spi1 blocked erythroid differentiation (Online Supplementary Figure S4C). Strikingly, Spi1 also triggered senescence when

the expression level of the senescence marker Dec1 was increased in cells transduced with Spi1 and HRASV12 (Figure 2A). Collectively, these results indicate that the ectopic expression of Spi1 in primary fibroblasts and hematopoietic cells induces cellular senescence. To characterize, in depth, which hematopoietic cells are prone to Spi1-induced senescence, we first combined measures of SA-βgal activity using the C12RG fluorogenic substrate and specific immunophenotypic markers for hematopoietic progenitors using flow cytometry. Mainly Kit–Sca1– cells were found at day 7, a time when senescence was measured (Online Supplementary Figure S4A), indicating that cells were engaged in differentiation. The distribution of cells in the different categories of Kit/Sca1 markers was similar in senescent (C12RG+) and in nonsenescent (C12RG–) cells (Online Supplementary Figure S4A). As the cytokines used for LSK culture are prone to push cells towards myeloid differentiation, and overexpression of Spi1 in primary hematopoietic cells causes differentiation primarily in macrophages,10,15,20 we looked at the immunophenotypic myeloid markers CD11b, Gr1 and F4/80 (Online Supplementary Figure S4B). Spi1 overexpression (GFP+ cells) reduced the proportion of granulocytes (CD11b+F4/80–Gr1+) and increased the proportion of monocytes/macrophages (CD11b+F4/80+Gr1+/-) (Figure 3A, upper histogram). All types of cells contributes to the Spi1-induced senescent cells as observed from the distribution of the CD11b, Gr1 and F4/80 population among the GFP+C12RG+ cells (Figure 3A, lower histogram), which

Figure 5. Spi1-induced senescence requires P38MAPK14 signaling in hematopoietic cells. (A) Mean percentage of SA-βgal positive cells (histograms) subjected to the retroviral-mediated expression of Spi1 or an empty vector, and maintained in cultures with or without 20 mM of SB203580 in samples of GFP-positive sorted hematopoietic cells 7 days post-infection. The means ± SD of at least 3 independent experiments are shown. *P<0.05; **P<0.005 from two-tailed Student’s t-tests. (B) Hematopoietic cells transduced with empty vectors (vector) or Spi1 expression vectors and maintained in cultures with or without 20 mM of SB203580 were sorted for GFP-positive cells 7 days post-infection and subjected to Western blot analyses of CDKN1A, CDKN2A, p-MAPKAPK2, MAPKAPK2, Dec1 and Spi1. α-adaptin served as the loading control. (C) Number of GFP-positive cells retrovirally transduced with empty vectors (vector) or Spi1 expression vectors and maintained in cultures with or without 20 μM of SB203580 from day 1 to day 6 at the indicated periods of time. The means ± SD of at least 3 independent experiments are shown.

haematologica | 2017; 102(11)

1855

L. Delestré et al.

overexpressed in MEP progenitors (3.5-fold more SA-βgal+ cells for Spi1-overexpressing cells compared to the MSCV control cells; 61.6% and 17.4%, respectively; Figure 3D). In conclusion, our results demonstrate that Spi1 is able to trigger senescence in myeloid and erythroid progenitors.

Spi1 triggers senescence through its DNA-binding activity To determine whether the DNA binding and transcriptional activities of Spi1 were required for inducing senescence, we used the Dβ4-Spi1 mutant with a deletion in the β4 region of the ETS domain, which we previously showed as being unable to bind DNA and transactivate40 (Online Supplementary Figure S5). The expression of Dβ4-Spi1 or an empty vector (Online Supplementary Figure 6A) did not impact the proliferation of BJ or WI-38 fibroblasts when compared with the proliferation of Spi1-WTexpressing cells (Online Supplementary Figure 6B). Moreover, cells overexpressing Dβ4-Spi1 were not found to be SA-βgal-positive (Online Supplementary Figure 6C,D) and did not increase Dec1 expression when compared with cells expressing the vector control (Online Supplementary Figure S6A). Similar to the results for fibroblasts, the expression of Dβ4-Spi1 in primary sorted LSK cells neither altered cell proliferation nor the percentage of cells staining positive for SA-βgal or Dec1 expression levels in hematopoietic cells (Online Supplementary Figure S6EG). Together, these results demonstrate that the DNA binding activity of Spi1 is required to induce senescence.

Spi1 induces senescence through distinct pathways in primary fibroblasts and hematopoietic cells To further understand the mechanisms underlying Spi1mediated senescence, we examined the expression of two cell cycle inhibitors, CDKN1A (alias p21/CIP) and CDKN2A (alias p16/INK4A), which are instrumental for inducing senescence. We detected a robust increase in CDKN1A and CDKN2A in Spi1- or HRASV12-expressing primary fibroblasts when compared with levels in cells transduced with the vector alone (Figure 4A). CDKN1A and CDKN2A expression levels were not increased in cells expressing the Dβ4-Spi1 protein (Figure 4B), which is consistent with the absence of senescence in those cells. It has been shown that OIS in primary fibroblasts is, at least in part, induced by DNA replication stress evoking a DDR, which is characterized by increased H2AX phosphorylation on Serine 139 (Ser139) and p53 phosphorylation on Serine 15 (Ser15). We found that Spi1, similar to the effects of HRASV12, induced the phosphorylation of p53 on Ser15 and of H2AX on Ser139 (Figure 4C), whereas this response was not detected in cells expressing the Dβ4-Spi1 mutant, as expected (Figure 4D). Next, we performed this analysis in HSCPs. Similar to the results observed in fibroblasts, the expression levels of CDKN1A and CDKN2A were increased in GFP-positive primary HSCPs 7 days post-infection with Spi1 or HRASV12 vectors when compared with levels in cells expressing control vectors (Figure 4E), whereas the overexpression of the Dβ4-Spi1 mutant had no effect (Figure 4F). Remarkably, in contrast to the results observed in fibroblasts and the effects of HRASV12, we did not detect any changes in the phosphorylation of p53 or H2AX in HSCPs overexpressing Spi1, indicating that Spi1 does not elicit a DDR in primary hematopoietic cells. These findings prompted us to test 1856

the role of the stress p38 mitogen-activated protein kinase (p38MAPK14) in Spi1-induced senescence, as OIS can be mediated by p38-dependent, p53-independent signaling pathways.41 We observed that HRASV12 induced the activation of p38MAPK14 based on an observed increase in the phosphorylated form (Thr180/182) of p38MAPK14 in both primary HSCPs and fibroblasts (Figure 4C,G). In contrast, Spi1 overexpression did not induce any increase in p38MAPK14 phosphorylation in BJ or WI-38 fibroblasts (Figure 4C,D), whereas in primary HSCPs, the overexpression of Spi1 produced an increase in the phosphorylation of p38MAPK14. Again, as shown by previous assays, Dβ4-Spi1 overexpression did not induce this response (Figure 4F). To corroborate the role of p38MAPK14 activity in Spi1induced senescence in hematopoietic cells, we treated cells undergoing Spi1-mediated senescence with the selective p38MAPK14 pharmacological inhibitor, SB203580. We found that the fraction of cells stained positive for SAβgal was strongly reduced in the presence of the inhibitor compared with the fraction in untreated cells (Figure 5A, compare Spi1-0 mM versus Spi1-20 mM). Spi1 did not increase Dec1, CDKN1A or CDKN2A levels of expression in the presence of the p38MAPK14 inhibitor (Figure 5B), further confirming the absence of senescence in SB203580-treated cells. The treatment of Spi1 overexpressing cells with SB203580 also increased, although partially, the number of GFP-positive cells (Figure 5C), indicating that p38MAPK14 controls senescence and additional mechanisms modulating cell number. The functional inhibition of p38MAPK14 activity by SB203580 was confirmed based on a decreased phosphorylation level of its downstream target mitogen-activated protein kinase-activated protein kinase 2 (MAPKAPK2) (Figure 5B). Together, these data support a model in which Spi1 elicits senescence in primary fibroblasts and hematopoietic cells through distinct molecular pathways. In fibroblasts, the induction of Spi1-mediated senescence involves a DDR, whereas in hematopoietic cells, it requires p38MAPK14 activation but is independent of a DDR.

Spi1-induced senescence in vivo is disrupted during leukemic progression To investigate the impact of dysregulated Spi1 expression on senescence in vivo, we utilized the TgSpi1 mouse model, which overexpresses Spi1 and develops AML from the erythroid lineage.25,42 Within 3.6 ± 1.5 months (14 weeks) after birth, TgSpi1 mice developed anemia with infiltration of the bone marrow and spleen associated with erythroid cells blocked at the colony forming uniterythroid (CFU-E) stage (CD71+Ter119–IL3Rα–Kit+). These mice are referred to as sick TgSpi1 mice (Figure 6A,B and Online Supplementary Figure S7A). These pre-leukemic cells, derived from sick TgSpi1 mice, are not tumorigenic when engrafted into nude mice.25 This is in contrast to the cells, referred to as leukemic cells, detected at a later stage that emerge due to the acquisition of mutations in the SCF receptor gene (Kit), and are characterized by fully malignant features.42 Here, we explored the effect of Spi1 up to the pre-leukemic stage.25,26 At this stage, the expansion of blast cells in the bone marrow and spleen of the sick TgSpi1 mice was mainly at the expense of the CD11bhigh Gr1high myeloid cells (Online Supplementary Figure S7B, compare 14-week-old WT and sick TgSpi1 mice). As expected, Spi1 was highly expressed at both the RNA and haematologica | 2017; 102(11)

Spi1/PU.1 triggers cellular senescence

F Figure 6. Spi1 induces senescence in vivo in the bone marrow of young TgSpi1 mice and is lost before the onset of the pre-leukemic syndrome. (A) Red blood cell numbers and hemoglobin concentrations of wild-type (WT) and TgSpi1 mice at the indicated ages. Bars indicate the mean values. *P<0.05; ****P<0.0001 from two-tailed Student’s t-tests. (B) Scatter plots represent the results of flow cytometry analyses of whole bone marrow cells for CFU-E markers (CD71+Ter119–Kit+IL3Rα–) in WT and TgSpi1 mice at the indicated ages. Bars indicate the mean values. *P<0.05; **P<0.001 from two-tailed Student’s t-tests. (C) Spi1 messenger ribonucleic acid (mRNA) levels in bone marrow cells from 7- and 14-week-old WT, healthy TgSpi1 and sick TgSpi1 mice were quantified via real-time quantitative polymerase chain reaction (qPCR) and normalized to the Polr2α mRNA level (DCt, Ctgene-CtPolr2α). Between 4 and 6 animals were analyzed for each category of mice. Bars represent the fold change relative to values for age-matched WT mice, as calculated from the 2−DDCt values. Statistical analysis of the 2−DCt values was carried out using Student’s t-test; *P<0.05. (D) Spi1 protein levels in bone marrow cells from 7- and 14-week-old WT, healthy TgSpi1 and sick TgSpi1 mice were analyzed by Western blotting. The histograms represent the quantified results, using ImageJ, relative to β-actin and to values for age-matched WT mice. (E) SA-βgal activity was examined in fresh bone marrow sections. Staining was performed on bone marrow from 7- and 14-week-old WT, healthy TgSpi1 and sick TgSpi1 mice. The number of mice displaying SA-βgal positive cells in their bone marrow is indicated for each category of mice. Bars represent 50μm (top) and 10μm (bottom) pictures. (F) MEP (Lin–Sca–Kit+CD34–CD16/32–) and GMP (Lin–Sca–Kit+CD34+CD16/32+) from bone marrow cells of 7-week-old WT and healthy TgSpi1 mice were analyzed for SA-βgal activity using C12RG as a fluorogenic substrate. The histograms represent the means of percentage of C12RG+ cells in TgSpi1 mice considering WT mice as negative control, as presented in Online Supplementary Figure S10A. N= 4 animals for WT and 5 animals for TgSpi1 mice. Ns: non-significant.

haematologica | 2017; 102(11)

1857

L. Delestré et al.

protein levels in the bone marrow cells of the preleukemic sick TgSpi1 mice (3.5- and 4-fold upregulation, respectively) compared with levels in age-matched WT mice (Figure 6C,D and Online Supplementary Figure S8). To explore the progression of the Spi1-mediated disease, we examined healthy TgSpi1 mice at 7 and 14 weeks of age. Mice were assumed to be healthy if their spleen weighed less than 0.2g and if there were no blast cells in the blood or bone marrow (Online Supplementary Figure 7C,D). The increase of Spi1 gene and protein expression levels was detectable in the whole bone marrow cells of TgSpi1 mice as early as 7 weeks after birth (Figure 6C,D). Moreover, using immunohistochemistry, we observed an increase in the number of cells that expressed Spi1 in the bone marrow of 7- and 14-week-old healthy TgSpi1 mice (Online Supplementary Figure S8). Seven-week-old TgSpi1 mice did not display any abnormalities compared with age-matched WT mice for any analyzed parameters, i.e., red blood cell and blast cell numbers, hemoglobin concentration, the fraction of CFU-E and the percentage of myeloid cells (Figure 6A,B, Online Supplementary Figure S7A,B and D). Early signs of pre-leukemic syndrome were detectable in 14-week-old TgSpi1 mice that presented a moderate but significant decrease in red cell number and hemoglobin concentration compared with WT mice concomitant with a moderate increase of CFU-E in the bone marrow (7% for 14-week-old healthy TgSpi1 mice compared with 3.5% for 14-week-old WT mice; Figure 6A,B).We then characterized the senescence program during the pre-leukemic progression phase using this wellcharacterized model. We detected an accumulation of SA-βgal in bone marrow sections from five out of seven TgSpi1 mice at 7 weeks of age, whereas this was not observed in age-matched WT mice (Figure 6E). A slightly increased expression of the senescence marker Dec1 was also observed in the total bone marrow of TgSpi1 mice compared to WT, consistent with the induction of senescence in the TgSpi1 mice (Online Supplementary Figure S9). Interestingly, more senescent cells were observed among the MEP cells of 7-week-old TgSpi1 mice compared with age-matched WT mice, while this was not the case for the GMP cells (Figure 6F and Online Supplementary Figure S10A). Consistently, Spi1 was significantly overexpressed in the MEP cells and weakly expressed in the GMP cells in TgSpi1 mice compared with age-matched WT mice (Online Supplementary Figure S10B). Overexpression of Spi1 in MEP cells was also found to be associated with an induction of Dec1 and IL1α gene expression, a cytokine of the senescence-associated secretory phenotype (SASP) (Online Supplementary Figure S10C), supporting senescence activation in those types of cells. These data indicate that only erythroid progenitors are prone to Spi1induced senescence in the TgSpi1 mice in vivo. Importantly, even though Spi1 expression was increased (Online Supplementary Figure S10B and S11), we did not detect any SA-βgal activity in the bone marrow sections from TgSpi1 mice at 14 weeks of age or from sick TgSpi1 mice (Figure 6E), indicating that the senescence program was not effective in either the 14-week-old TgSpi1 mice that displayed only early and minimal signs of a preleukemic syndrome or in the TgSpi1 mice with a preleukemic syndrome. Remarkably, while sorted erythroid blastic cells from sick TgSpi1 mice (CFU-E like, CD71+Ter119–IL3Rα–Kit+) still displayed a higher expression of CDKN2A compared with equivalent WT CFU-E 1858

cells, Dec1 and CDKN1A RNA expression was decreased (Online Supplementary Figure S11). These results are consistent with the lack of senescent phenotype in the pre-leukemic cells and bring molecular supports of the bypass of senescence in vivo. In conclusion, Spi1 overexpression is associated with senescence of erythroid progenitors in the bone marrow of asymptomatic young TgSpi1 mice. Erythroid cells give rise to blastic cells in sick TgSpi1 animals. The absence of a senescence program in older and sick TgSpi1 mice suggests that this process is lost during the completion of the pre-leukemic syndrome phase at a step that precedes the emergence of the pre-leukemic stage.

Discussion In the study herein, we demonstrated that Spi1 TF is a driver of senescence in several types of cells, ie., in primary fibroblasts, hematopoietic stem cells, MEP and GMP progenitors. It is well established that the OIS is a barrier that stands at pre-invasive stages of solid tumorigenesis. As Spi1 is not expressed in normal fibroblasts, and as its over-expression induces senescence through the activation of a DDR in primary fibroblasts, as shown for the HRASV12, our results suggest that Spi1 behaves as a “classical” oncogene for OIS which is associated with DNA replication stress.33,35-37 Of note, while HRASV12 -induced senescence is mediated by both DDR-dependent and DDR-independent processes, i.e., the activation of p38MAPK14, Spi1induced senescence is only associated with DDR-dependent signaling in primary fibroblasts. In contrast with the response observed in primary fibroblasts, DDR signaling is not involved in the Spi1induced senescence of primary hematopoietic cells, consistent with data from our previous study showing that Spi1 overexpression is not associated with the presence of DNA strand breaks in the K562 cell line and in the preleukemic cells of TgSpi1 mice.28 We found that p38MAPK14 signaling is required to elicit Spi1-induced senescence in hematopoietic cells. Contributions from either DDR or p38MAPK14 signaling in OIS have also been reported for leukemogenic fusion proteins in hematopoietic cells.32 The P38MAPK14 activation mediated by reactive oxygen species (ROS) plays a role in the exhaustion of hematopoietic stem cells.43 Interestingly, a p38MAPK14/p16 axis independent of the serine/threonine-protein kinase ataxia telangiectasia (ATR)- or Rad3related protein (ATM)-DDR has also been found to mediate senescence in epithelial cells.44 Therefore, p38MAPK14 seems to be an alternative and major signaling pathway by which cells others than fibroblasts can undergo senescence. In conclusion, we identified diverse molecular pathways for Spi1-induced senescence, either through DDR activation in primary fibroblasts or DDR-independent P38MAPK14 activation in primary hematopoietic cells. Interestingly, even if Spi1 triggers senescence at a slightly lower degree than H-RASV12 in the two cell types, the effect of Spi1 overexpression on growth arrest was stronger than that due to overexpression of H-RASV12. These results suggest that Spi1 also limits cell proliferation by additional mechanisms, other than that of senescence. As we have shown that Spi1 overexpression favors monohaematologica | 2017; 102(11)

Spi1/PU.1 triggers cellular senescence

cytes/macrophage differentiation and that growth arrest is concomitant to terminal differentiation, the reduced number of proliferating cells in hematopoietic cells may be the combined consequence of at least two mechanisms on a heterogeneous population, senescence and differentiationassociated growth arrest. Consistently, not all cells overexpressing Spi1 underwent senescence (up to 50% of LSK cells). Additionally, SA-βgal positive cells may not accumulate in culture and be eliminated as reported for apoptotic hematopoietic cells. We have shown that, in contrast to the opposite effects of Spi1 on erythroid and monocytic differentiation, monocytes/macrophages, granulocytes and erythroid cells undergo senescence when Spi1 was overexpressed in their respective immature progenitors, demonstrating that a wide range of cells are prone to Spi1-induced senescence. One query raised by the study herein is to determine the relationship between senescence and differentiation and, in particular, which program (senescence or differentiation) was the first to be initiated. However, the only parameter applicable to identify senescent cells and characterize their differentiation markers is the SA-βGal assay, which is a final parameter of senescence. Thus, the initial cell wherein senescence was instructed remains unknown. Spi1 has been shown to play an important role in normal HSC homeostasis.2-5 In particular, Spi1 prevents the self-renewal and exhaustion of HSCs by restraining the cell cycle through multiple downstream targets.45 Thus, our findings that Spi1 triggers senescence in vitro in hematopoietic cells and in vivo in the bone marrow of TgSpi1 mice are consistent with senescence as a new growth arrest mechanism by which Spi1 limits expansion. This result raises the question of whether the ability of Spi1 to induce senescence is only of pathological relevance, or if it also plays a role in normal hematopoiesis. As a gradient of Spi1 expression is observed during normal lineage commitment and in divergent hematopoietic lineages,1 we speculate that the senescence-inducing activity of Spi1 may participate in the regulation of cell number, depending on its level of expression and the hematopoietic context. While Spi1 is oncogenic in the erythroid lineage, considerable evidence points to a role for this TF as a tumor suppressor in myeloid malignancies.46 Thus, we propose that in addition to blocking myeloid differentiation, decreased Spi1 expression/activity may promote leukemogenesis through the loss of anti-proliferative activity. This may be particularly deleterious in the case of a cooperative proliferative signaling, such as replicative or oncogenic stress. In human acute promyelocytic leukemias (APLs) expressing the oncoprotein PML-RARα, a down-regulation of Spi1 occurs that is overridden by all-trans retinoic acid (ATRA) treatment.10 The restoration of Spi1 in APL cells triggers granulocytic differentiation.10 ATRA also rescues a senescence-like program, demonstrating that this treatment constitutes a senescence-driven cure for leukemia.47 As Spi1 expression and senescence are both increased by ATRA in APL cells, Spi1-induced senescence may contribute to the cure of APL. We used the TgSpi1 model to examine the process of

haematologica | 2017; 102(11)

senescence up to the appearance of the pre-leukemic syndrome. Notably, even though Spi1 induces senescence in erythroid progenitors of young asymptomatic TgSpi1 mice, senescence was not detectable in the bone marrow of sick TgSpi1 mice. Because senescence was likewise not detectable in the bone marrow of 14-week-old TgSpi1 mice that were asymptomatic, the senescence program is turned off prior to the emergence of pre-leukemic blast cells. Even if not proved, several results argue that senescent cells give rise to blastic cells. Senescent cells were detected among the MEP cells in the young TgSpi1 mice; this was not found for the GMP. Consistently, Spi1 was significantly overexpressed in the MEP cells and weakly in the GMP cells in those TgSpi1 mice. Remarkably, while the erythroid blastic cells of the sick pre-leukemic mice still displayed a higher expression of Spi1 and CDKN2A compared with equivalent WT CFU-E cells, Dec1 and CDKN1A RNA expression was decreased, indicating that the pre-leukemic cells lost some marks of senescence which they displayed in young animals. Knocking out p53 in the TgSpi1 mice accelerates the speed of the disease,48 in agreement with the fact that growth arrest pathway delays the leukemic progression. Altogether, our data suggest that, as is the case for solid tumors, senescence represents a mechanism for constraining cell number and is disrupted during the development of AML. High MCL1 expression found in the TgSpi1 erythroid blasts27 may be involved in the erasure of the Spi1induced senescence program, as MCL1 is a repressor of senescence.49 The anti-apoptotic function of Spi1 participates in the continuous expansion of pre-leukemic cells.26,27 Thus, the two main protective barriers to tumor development, apoptosis and senescence, are disrupted during erythroleukemic transformation in TgSpi1 mice. Further work will be needed to identify the pathways that are responsible for the bypass of senescence during leukemic progression and how targeting these pathways may rescue the senescence program in the pre-leukemic stage. Funding This work was supported by the Fondation de France, Inserm, the Institut National du Cancer (INCa-DGOS-INSERM 6043 and PL-BIO-06), ITMO Cancer de l’alliance nationale pour les sciences de la vie et de la santé (AVIESAN), Section régionale de la Ligue Nationale contre le Cancer. L. Delestré was supported by AVIESAN and Fondation de France; H. Cui by Cancéropole Ile-de-France; M. Esposito by the Institut National du Cancer (PL-BIO-06). Quiveron and E. Mylonas by a CDI-Mission (Institut Gustave Roussy). Acknowledgments The authors would like to thank Y. Lecluse, P. Rameau and Z. Maciorowski from the cytometry platform and A. Nicolas, P. Opolon, O. Bawa, S. Arrufat, R. Corre and E. Louvet from the IHC platform. We thank the animal facility housing at Curie and Gustave Roussy Institutes. We thank O. Bernard, M. David for comments on the manuscript and F. Rosselli and F. MoreauGachelin for helpful scientific discussions and critical reading of the manuscript.

1859

L. DelestrĂŠ et al.

References 1. DeKoter RP, Kamath MB, Houston IB. Analysis of concentration-dependent functions of PU.1 in hematopoiesis using mouse models. Blood Cells, Mol Dis. 2007;39(3): 316-320. 2. Iwasaki H, Somoza C, Shigematsu H, et al. Distinctive and indispensable roles of PU.1 in maintenance of hematopoietic stem cells and their differentiation. Blood. 2005;106(5): 1590-1600. 3. Kim HG, de Guzman CG, Swindle CS, et al. The ETS family transcription factor PU.1 is necessary for the maintenance of fetal liver hematopoietic stem cells. Blood. 2004;104 (13):3894-3900. 4. Nutt SL, Metcalf D, D'Amico A, Polli M, Wu L. Dynamic regulation of PU.1 expression in multipotent hematopoietic progenitors. J Exp Med. 2005;201(2):221-231. 5. Staber PB, Zhang P, Ye M, et al. Sustained PU.1 levels balance cell-cycle regulators to prevent exhaustion of adult hematopoietic stem cells. Mol Cell. 2013;49(5):934-946. 6. Turkistany SA, DeKoter RP. The transcription factor PU.1 is a critical regulator of cellular communication in the immune system. Arch Immunol Ther Exp (Warsz). 2011;59(6):431-440. 7. Kueh HY, Champhekar A, Nutt SL, Elowitz MB, Rothenberg EV. Positive feedback between PU.1 and the cell cycle controls myeloid differentiation. Science. 2013;341 (6146):670-673. 8. Wontakal SN, Guo X, Will B, et al. A large gene network in immature erythroid cells is controlled by the myeloid and B cell transcriptional regulator PU.1. PLoS Genet. 2011;7(6):e1001392. 9. Steidl U, Steidl C, Ebralidze A, et al. A distal single nucleotide polymorphism alters longrange regulation of the PU.1 gene in acute myeloid leukemia. J Clin Invest. 2007;117 (9):2611-2620. 10. Mueller BU, Pabst T, Fos J, et al. ATRA resolves the differentiation block in t(15;17) acute myeloid leukemia by restoring PU.1 expression. Blood. 2006;107(8):3330-3338. 11. Vangala RK, Heiss-Neumann MS, Rangatia JS, et al. The myeloid master regulator transcription factor PU.1 is inactivated by AML1-ETO in t(8;21) myeloid leukemia. Blood. 2003;101(1):270-277. 12. Mizuki M, Schwable J, Steur C, et al. Suppression of myeloid transcription factors and induction of STAT response genes by AML-specific Flt3 mutations. Blood. 2003;101(8):3164-3173. 13. Mueller BU, Pabst T, Osato M, et al. Heterozygous PU.1 mutations are associated with acute myeloid leukemia. Blood. 2002;100(3):998-1007. 14. Lavallee VP, Baccelli I, Krosl J, et al. The transcriptomic landscape and directed chemical interrogation of MLL-rearranged acute myeloid leukemias. Nat Genet. 2015;47(9): 1030-1037. 15. Cook WD, McCaw BJ, Herring C, et al. PU.1 is a suppressor of myeloid leukemia, inactivated in mice by gene deletion and mutation of its DNA binding domain. Blood. 2004;104(12):3437-3444. 16. Rosenbauer F, Wagner K, Kutok JL, et al. Acute myeloid leukemia induced by graded reduction of a lineage-specific transcription

1860

factor, PU.1. Nat Genet. 2004;36(6):624-630. 17. Metcalf D, Dakic A, Mifsud S, Di Rago L, Wu L, Nutt S. Inactivation of PU.1 in adult mice leads to the development of myeloid leukemia. Proc Natl Acad Sci USA. 2006;103(5):1486-1491. 18. Will B, Vogler TO, Narayanagari S, et al. Minimal PU.1 reduction induces a preleukemic state and promotes development of acute myeloid leukemia. Nat Med. 2015;21(10):1172-1181. 19. Walter MJ, Park JS, Ries RE, et al. Reduced PU.1 expression causes myeloid progenitor expansion and increased leukemia penetrance in mice expressing PML-RARalpha. Proc Natl Acad Sci USA. 2005; 102(35): 12513-12518. 20. Delgado MD, Gutierrez P, Richard C, Cuadrado MA, MoreauGachelin F, Leon J. Spi-1/PU.1 proto-oncogene induces opposite effects on monocytic and erythroid differentiation of K562 cells. Biochem Biophys Res Commun. 1998;252(2):383-391. 21. Staber PB, Zhang P, Ye M, et al. The RunxPU.1 pathway preserves normal and AML/ETO9a leukemic stem cells. Blood. 2014;124(15):2391-2399. 22. Aikawa Y, Katsumoto T, Zhang P, et al. PU.1-mediated upregulation of CSF1R is crucial for leukemia stem cell potential induced by MOZ-TIF2. Nat Med. 2010;16(5):580-585, 581p following 585. 23. Zhou J, Wu J, Li B, et al. PU.1 is essential for MLL leukemia partially via crosstalk with the MEIS/HOX pathway. Leukemia. 2014;28(7):1436-1448. 24. Moreau-Gachelin F, Tavitian A, Tambourin P. Spi-1 is a putative oncogene in virally induced murine erythroleukemia. Nature (London). 1988;331(6153):277-280. 25. Moreau-Gachelin F, Wendling F, Molina T, et al. Spi-1/PU.1 transgenic mice develop multistep erythroleukemias. Mol Cell Biol. 1996;16(5):2453-2463. 26. Rimmele P, Kosmider O, Mayeux P, Moreau-Gachelin F, Guillouf C. Spi-1/PU.1 participates in erythroleukemogenesis by inhibiting apoptosis in cooperation with Epo signaling and by blocking erythroid differentiation. Blood. 2007;109(7):3007-3014. 27. Ridinger-Saison M, Evanno E, Gallais I, et al. Epigenetic silencing of Bim transcription by Spi-1/PU.1 promotes apoptosis resistance in leukaemia. Cell Death Differ. 2013;20(9): 1268-1278. 28. Rimmele P, Komatsu J, Hupe P, et al. Spi1/PU.1 oncogene accelerates DNA replication fork elongation and promotes genetic instability in the absence of DNA breakage. Cancer Res. 2010;70(17):6757-6766. 29. Campisi J, d'Adda di Fagagna F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol. 2007;8(9):729-740. 30. Collado M, Gil J, Efeyan A, et al. Tumour biology: senescence in premalignant tumours. Nature. 2005;436(7051):642. 31. Michaloglou C, Vredeveld LC, Soengas MS, et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature. 2005;436(7051):720-724. 32. Wajapeyee N, Wang SZ, Serra RW, et al. Senescence induction in human fibroblasts and hematopoietic progenitors by leukemogenic fusion proteins. Blood. 2010;115(24):5057-5060. 33. Takacova S, Slany R, Bartkova J, et al. DNA

34.

35.

36.

37.

38.

39. 40.

41. 42.

43.

44.

45.

46.

47.

48.

49.

damage response and inflammatory signaling limit the MLL-ENL-induced leukemogenesis in vivo. Cancer Cell. 2012;21(4):517-531. Braig M, Lee S, Loddenkemper C, et al. Oncogene-induced senescence as an initial barrier in lymphoma development. Nature. 2005;436(7051):660-665. Di Micco R, Fumagalli M, Cicalese A, et al. Oncogene-induced senescence is a DNA damage response triggered by DNA hyperreplication. Nature. 2006;444(7119):638-642. Gorgoulis VG, Vassiliou LV, Karakaidos P, et al. Activation of the DNA damage checkpoint and genomic instability in human precancerous lesions. Nature. 2005;434(7035): 907-913. Bartkova J, Rezaei N, Liontos M, et al. Oncogene-induced senescence is part of the tumorigenesis barrier imposed by DNA damage checkpoints. Nature. 2006;444(7119):633-637. Cristofalo VJ, Volker C, Allen RG. Use of the fibroblast model in the study of cellular senescence. Methods Mol Med. 2000;38(2352. Collado M, Serrano M. The power and the promise of oncogene-induced senescence markers. Na Rev Cancer. 2006;6(6):472-476. Guillouf C, Gallais I, Moreau-Gachelin F. Spi-1/PU.1 oncoprotein affects splicing decisions in a promoter binding-dependent manner. J Biol Chem. 2006;281(28):1914519155. Han J, Sun P. The pathways to tumor suppression via route p38. Trends Biochem Sci. 2007;32(8):364-371. Kosmider O, Denis N, Lacout C, Vainchenker W, Dubreuil P, MoreauGachelin F. Kit-activating mutations cooperate with Spi-1/PU.1 overexpression to promote tumorigenic progression during erythroleukemia in mice. Cancer Cell. 2005;8(6):467-478. Ito K, Hirao A, Arai F, et al. Reactive oxygen species act through p38 MAPK to limit the lifespan of hematopoietic stem cells. Nat Med. 2006;12(4):446-451. Nassour J, Martien S, Martin N, et al. Defective DNA single-strand break repair is responsible for senescence and neoplastic escape of epithelial cells. Nat Commun. 2016;7:10399. Fukuchi Y, Ito M, Shibata F, Kitamura T, Nakajima H. Activation of CCAAT/enhancer-binding protein alpha or PU.1 in hematopoietic stem cells leads to their reduced self-renewal and proliferation. Stem Cells. 2008;26(12):3172-3181. Dakic A, Wu L, Nutt SL. Is PU.1 a dosagesensitive regulator of haemopoietic lineage commitment and leukaemogenesis? Trends Immunol. 2007;28(3):108-114. Ablain J, Rice K, Soilihi H, de Reynies A, Minucci S, de The H. Activation of a promyelocytic leukemia-tumor protein 53 axis underlies acute promyelocytic leukemia cure. Nat Med. 2014;20(2):167-174. Scolan EL, Wendling F, Barnache S, et al. Germ-line deletion of p53 reveals a multistage tumor progression in spi-1/PU.1 transgenic proerythroblasts. Oncogene. 2001;20(39):5484-5492. Bolesta E, Pfannenstiel LW, Demelash A, et al. Inhibition of Mcl-1 promotes senescence in cancer cells: implications for preventing tumor growth and chemotherapy resistance. Mol cell Biol. 2012;32(10):1879-1892.

haematologica | 2017; 102(11)

ARTICLE

Acute Myeloid Leukemia

Bortezomib as a new therapeutic approach for blastic plasmacytoid dendritic cell neoplasm

Laure Philippe,1,2 Adam Ceroi,2 Elodie Bôle-Richard,1,2 Alizée Jenvrin,2,3 Sabeha Biichle,2 Sophie Perrin,4 Samuel Limat,2,4 Francis Bonnefoy,2 Eric Deconinck,1,2 Philippe Saas,2,5 Francine Garnache-Ottou2,3 and Fanny Angelot-Delettre2,3 1 CHRU Besançon, Hematology; 2Univ. Bourgogne Franche-Comté, INSERM, EFS Bourgogne Franche-Comté, UMR1098, Interactions Hôte-Greffon-Tumeur – Ingénierie Cellulaire et Génique, LabEX LipSTIC; 3EFS Bourgogne Franche-Comté, Laboratoire d’Hématologie; 4 CHRU Besançon, Pharmacy and 5CHRU Besançon, INSERM, CIC-1431, Plateforme de BioMonitoring, Besançon, France

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1861-1868

ABSTRACT

lastic plasmacytoid dendritic cell neoplasm is an aggressive hematologic malignancy with a poor prognosis. No consensus regarding optimal treatment modalities is currently available. Targeting the nuclear factor-kappa B pathway is considered a promising approach since blastic plasmacytoid dendritic cell neoplasm has been reported to exhibit constitutive activation of this pathway. Moreover, nuclear factor-kappa B inhibition in blastic plasmacytoid dendritic cell neoplasm cell lines, achieved using either an experimental specific inhibitor JSH23 or the clinical drug bortezomib, interferes in vitro with leukemic cell proliferation and survival. Here we extended these data by showing that primary blastic plasmacytoid dendritic cell neoplasm cells from seven patients were sensitive to bortezomib-induced cell death. We confirmed that bortezomib efficiently inhibits the phosphorylation of the RelA nuclear factor-kappa B subunit in blastic plasmacytoid dendritic cell neoplasm cell lines and primary cells from patients in vitro and in vivo in a mouse model. We then demonstrated that bortezomib can be associated with other drugs used in different chemotherapy regimens to improve its impact on leukemic cell death. Indeed, when primary blastic plasmacytoid dendritic cell neoplasm cells from a patient were grafted into mice, bortezomib treatment significantly increased the animals’ survival, and was associated with a significant decrease of circulating leukemic cells and RelA nuclear factor-kappa B subunit expression. Overall, our results provide a rationale for the use of bortezomib in combination with other chemotherapy for the treatment of patients with blastic plasmacytoid dendritic cell neoplasm. Based on our data, a prospective clinical trial combining proteasome inhibitor with classical drugs could be envisaged.

Introduction Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare malignancy derived from plasmacytoid dendritic cells and is classified among acute myeloid leukemias by the 2008 World Health Organization (WHO). BPDCN is associated with a poor prognosis with a median overall survival of 8-12 months in the largest series of patients.1-3 The diagnosis is made from the typical cutaneous lesions that rapidly progress (90%) to bone marrow and extramedullary sites. The diagnosis is mainly based on histopathological and phenotypic characterization of blastic cells in the peripheral blood or bone marrow expressing the following markers CD123, BDCA2 (CD303), BDCA4 (CD304) and TCL1 as analyzed by flow cytometry.1-3 There is currently no consensus regarding optimal treatment modalities. Classical treatments such as CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) regimens show disappointing results.4 While intensive chemotherapy haematologica | 2017; 102(11)

Correspondence: fanny.delettre@efs.sante.fr

Received: March 23, 2017. Accepted: August 8, 2017. Pre-published: August 10, 2017. doi:10.3324/haematol.2017.169326 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1861 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1861

L. Philippe et al.

regimens (including those for acute myeloid leukemia and acute lymphoblastic leukemia) followed by allogeneic hematopoietic cell transplantation have been reported to improve the survival beyond 30 months in young patients,5-10 elderly patients are not eligible for this approach. Altogether, this makes it necessary to evaluate new therapeutic strategies. Recently, Sapienza et al. demonstrated a constitutive activation of the nuclear factor-kappa B (NF-κB) pathway in primary BPDCN cells which represents a potential therapeutic target.11 Using the proteasome inhibitor bortezomib, known to inhibit NF-κB activation,12 these authors demonstrated that treatment of the BPDCN cell line CAL1 inhibits cell proliferation and induces a significant cytotoxic effect. More recently, Ceroi et al. confirmed this constitutive activation of the NF-κB pathway in other primary BPDCN cells and demonstrated the induction of apoptosis of BPDCN cell lines (CAL-1 and GEN2.2) in vitro in response to the NF-κB p65 inhibitor, JSH23.13 Overall, targeting the NF-κB pathway by bortezomib would represent a promising, easily available therapeutic option for BPDCN patients if its efficacy were to be confirmed in vitro using primary BPDCN samples and in vivo in a preclinical BPDCN model. This was the goal of our work.

Methods

Cytotoxicity, proliferation and cell cycle assay by flow cytometry A panel of monoclonal antibodies against CD123, CD45, CD56 and BDCA4 was used to gate BPDCN cells (Online Supplement). BPDCN cells from seven patients and BPDCN cell lines (CAL-1, GEN2.2) were incubated at 106 cells/mL at 37°C in 5% CO2 with bortezomib at various concentrations (10 - 50 nM) for 24 or 48 h. The cytotoxic effects of drugs were evaluated in vitro using annexin-V and 7-amino actinomycin D (AV/7AAD, Beckman Coulter, Roissy, France) staining and flow cytometry.13,19 Cells were labeled by Dye eFluor® V450 (Ebioscience, San Diego, CA, USA) to assess cell proliferation.13 The percentage of cells in subG1, G1, S and G2 cell cycle phases was evaluated using CXP and MultiCycle software (Beckman Coulter).13

Nuclear factor-kappa B pathway activation CAL-1 cells or PDX (patient derived xenograft) cells obtained in vivo from blood of mice, after treatment with bortezomib for 6 h followed by stimulation with a TLR7 agonist (R848, 1 mg/mL, Invivogen, Toulouse, France) for 45 min were investigated by phospho-flow staining using phosphorylated-NF-κB subunit RelA (pRelA) staining, as described as described in the Online supplement.

Statistical analysis Statistical analyses were performed using the Student t-test or the Mann-Whitney test (GraphPad Prism software 5.0c, San Diego, CA, USA) (Online Supplement).

Patients’ cells, cell lines and culture Two human BPDCN cell lines (CAL-1, Dr. Maeda, Nagasaki University, Japan and GEN 2.2, patent #0215927, EFS, France)14,15 and samples from seven BPDCN patients (Online Supplementary Table 1) from our French national network (authorization #DC2016-2791) fully diagnosed as BPDCN by their phenotype (CD123+, CD56+, CD123+, CD303+, CD304+, TCL1+)1,16-18 were used. This study was approved by the Besançon local ethic committee (CPPEST II, Besançon, France).

Primary blastic plasmacytoid dendritic cell neoplasm cell xenograft model NOD/SCID/IL2Rγc-deficient (NSG) mice (6 to 8 weeks of age, The Jackson Laboratory, Sacramento, CA, USA) were irradiated (2.5 Gy) and inoculated intravenously with 2x106 primary BPDCN cells from patient #127. Mice were treated with bortezomib (0.25 mg/kg, intraperitoneally) once or twice a week for 2 or 4 weeks. Engraftment and quantification of the BPDCN cell line are described in the Online Supplement. These procedures were carried out in accordance with the guidelines for animal experimentation according to an approved protocol (protocol 11007R, Veterinary Services for Animal Health & Protection, issued by the Ministry for Agriculture, Paris, France).

Drugs Bortezomib was tested at different concentrations from 10 to 75 nM for in vitro evaluation, as previously described,11 and at 20 nM when associated with other drugs. BPDCN cells were cultured at 106 cell/mL in RPMI-1640 glutamax medium (Invitrogen, Cergy Pontoise, France) supplemented with 10% fetal calf serum (Invitrogen) and 1% penicillin/streptomycin (PAA Laboratoires, Vélizy-Villacoublay, France) at 37°C under 5% CO2 for 24 or 48 h (Online Supplement). Bortezomib was injected intraperitoneally into mice at a dose of 0.25 mg/kg. The NF-κB p65 inhibitor, JSH23 (Calbiochem-EMD Biosciences, Inc, San Diego, CA, USA) was used as a control at a dose of 40 mg/kg. Others drugs tested are described in the Online Supplement. 1862

Results Bortezomib is cytotoxic against blastic plasmacytoid dendritic cell neoplasm cell lines and primary cells Treatment of CAL-1 cells with bortezomib for 24 h (n=5, 50 nM) markedly decreased cell proliferation (from 51.7±7% to 16.8±7.9%, P<0.001) (Figure 1A) and cell survival (from 89.2±1.5% to 26.6±6.5%, P<0.001) (Figure 1B). After 48 h of treatment with bortezomib (50 nM), cell proliferation was also significantly reduced from 43.4±9.8% to 22.4±9.4% (n=4, P<0.001) (Figure 1A) and viability was significantly decreased (from 51.7±7% to 16.8±7.9%) (Figure 1B). Bortezomib treatment (30 nM) induced robust cytotoxicity in vitro, similarly to SL-401 (used as the positive control) (P<0.01) (Figure 1C). Similar data showing significant bortezomib-induced cytotoxity were also obtained for the GEN2.2 BPDCN cell line and primary BPDCN cells from seven different patients (Figure 1C). Moreover, exposure of CAL-1 cells to bortezomib induced a significant accumulation of BPDCN cells in the G2 phase of the cell cycle (from 5.8±1.4% to 24.1±3.02% at 24 h, n=4, P<0.05) (Figure 1D,E). These data were confirmed using primary cells isolated from patients and treated in vitro with bortezomib [patient #25 (n=2) and patient #127 (n=1)] (Figure 1D). Subsequently, CAL-1 cells underwent apoptosis, as attested by an increase of cell arrest in the subG1 phase (from 18.8±7.3% to 60.8±7.6% at 24 h, n=4, P=NS) (Figure 1D). Overall, this demonstrates that, in vitro, bortezomib inhibits BPDCN cell proliferation and induces cell death.

Bortezomib inhibits nuclear factor-kappa B pathway activation in blastic plasmacytoid dendritic cell neoplasm cells While bortezomib is a proteasome inhibitor with wellknown anti-NF-κB properties and was used to treat haematologica | 2017; 102(11)

Bortezomib for BPDCN

Figure 1. Bortezomib inhibits cell proliferation and survival of blastic plasmacytoid dendritic cell neoplasm cell lines and primary cells. Results are expressed as percentage ± SEM of (A) proliferation using the Dye eFluor® V450 dilution assay and (B) viable cells using AV–/7-AAD– staining of the CAL-1 cell line treated with bortezomib (10 - 50 nM) for 24 h (black) and 48 h (gray) (n=4). Untreated CAL-1 cells were arbitrarily assigned a value of 100%. (C) Percentage ± SEM of viable GEN 2.2 cells (n=3), CAL-1 (n=6) cells and primary BPDCN cells from seven patients was determined after incubation with bortezomib (30 nM), or SL-401 (365 pM) for 24 h. Untreated cells were considered as 100% viable. (D) One representative histogram showing the percentage of CAL-1 cells and primary cells from two patients (patient #25 and patient #127) in the different phases of the cell cycle: G1, S and G2 after treatment or not with bortezomib at 10 or 30 nM for 24 h. (E) Percentage of cells in the G2 phase in the CAL-1 cell line after treatment or not with bortezomib (10-50 nM) for 24 h (n=4). Histograms represent the mean ± SEM of four independent experiments, *P<0.05, **P<0.01, ***P<0.001 between bortezomib and untreated cells. (F) Percentage of viable CAL-1 (n=3) and GEN2.2 (n=3) cells after incubation with bortezomib (B, 20 nM) in association with: idarubicin (I) at 0.03 mM, dexamethasone (D) at 0.637 mM, vorinostat (S) at 1.25 μM, statins, such as pravastatin (P) and simvastatin (Sim) at 5 mM and 5-azacytidine (5-Aza) at 4 mM. Histograms represent the mean ± SEM of three independent experiments. *P<0.05, **P<0.01.

haematologica | 2017; 102(11)

1863

L. Philippe et al. BPDCN CAL-1 cells,11 inhibition of the NF-κB pathway in primary BPDCN cells was not demonstrated. Treatment of BPDCN cells with bortezomib (75 nM, 24 h) decreased R848-induced RelA phosphorylation in CAL-1 cells from 91.8±2.6% to 71 ±1.6% (n=4, P<0.05), in GEN 2.2 cells from 97.5±0.2% to 19.8±4.4% (n=3, P<0.05), and in five different primary BPDCN cells (#24, #25, #127, #66, #38) from 79.9±7.23% to 61.6±3.41%. The percentage of BPDCN cells (CAL-1, GEN 2.2 and BPDCN #66) exhibiting reduced pRelA expression increased after bortezomib treatment (from 6.9% to 29.5%, -11.4% to 63.4% and -8.2% to 27.1%, respectively, after treatment with bortezomib 75 nM) (Figure 2A). Moreover, pRelA analysis by confocal microscopy revealed a decrease of pRelA nuclear translocation in CAL-1 cells associated with a cytoplasmic retention of pRelA after bortezomib treatment (50 nM, 6 h, n=3) (Figure 2B). Similar results were also observed with the GEN 2.2 cell line (19±1% to 2±0.05%, n=3; data not shown).

Association of bortezomib with others drugs increases its cytotoxic effect Since limited treatment efficiency has been reported for BPDCN and no consensus exists on treatment modality, associations of bortezomib with other drugs were tested. Idarubicin was used since this drugs exerts potent in vitro cytoxicity on BPDCN,19 but idarubicin was used at a nontoxic concentration (0.03 mM). Since BPDCN has been shown to exhibit altered cholesterol metabolism,13 inhibitors of cholesterol synthesis (statins) were also tested. The viability of BPDCN cell lines (CAL-1 and GEN 2.2) treated with bortezomib (20 nM, a non-cytotoxic concentration) in association with other drugs was evaluated at 24 h (Figure 1F). The viability of CAL-1 cells was 51.2±4.8% (n=3) after treatment with suberoylanilide hydroxamic

acid (SAHA) alone and decreased to 26.1±2.6% when bortezomib and SAHA were associated together. The viability of CAL-1 cells (n=3) was 61.6±2% with idarubicin alone and decreased to 10.8±3.1% when bortezomib and idarubicin were associated together. In the same way the viability of CAL-1 cells (n=3) was 50.1±1.8% with simvastatin alone and 91.1±1.7% with pravastatin alone and decreased to 16.3±3.4% or to 13.9±1.1% when bortezomib and simvastatin or pravastatin were associated together. The viability of CAL-1 cells (n=3) was 87.03±0.73% with 5-azacytidine alone and decreased to 47.9±0.85% when bortezomib and 5-azacytidine were associated together. Only the association of bortezomib with dexamethasone (65.7±14.4% to 78.8±5.1%) did not induce a synergistic effect. The viability of GEN 2.2 cells (n=3) was 83.2±0.7% after treatment with SAHA alone and decreased to 5.9±0.6% when bortezomib and SAHA were associated together. The viability of GEN 2.2 cells (n=3) was 68.3±5.4% after treatment with idarubicin alone and decreased to 0.16±0.03% when bortezomib and idarubicin were associated together (P<0.01). The viability of GEN 2.2 cells (n=3) was 19.1±1.3% after treatment with simvastatin alone or 82±1.7% with pravastatin alone and decreased to 6±3% or to 8.5±1.5% when bortezomib and simvastatin or pravastatin were associated together. The viability of GEN 2.2 cells (n=3) was 68.6±4.3% after treatment with dexamethasone alone and decreased to 11.8±0.6% when bortezomib and dexamethasone were associated together. The viability of GEN 2.2 cells (n=3) was 37.66±1.38% after treatment with 5-azacytidine alone and decreased to 1.4±0.3% when bortezomib and 5-azacytidine were associated together. Thus, the association of bortezomib with idarubicin, SAHA, 5-azacytidine or statins increases the cytotoxic effect of the proteasome inhibitor on the two BPDCN cell lines.

Figure 2. Bortezomib inhibits the nuclear factor-kappa B signaling pathway in blastic plasmacytoid dendritic cell neoplasm cell lines and primary cells. (A-B) BPDCN cell lines (GEN 2.2 and CAL-1 cells, n=3) and primary BPDCN cells from a patient were incubated with bortezomib (50 nM and 75 nM) or vehicle for 6 h before TLR7 stimulation for 45 min (R848, 1 mg/mL). One representative example of intracellular expression of NF-κBp-65 evaluated in CAL-1, GEN 2.2 cell lines and in primary BPDCN cells from patient #66 were analyzed by (A) flow cytometry and by (B) confocal microscopy in the CAL-1 cell line.

1864

haematologica | 2017; 102(11)

Bortezomib for BPDCN

Luciferase-expressing CAL-1 cell xenograft in mice to assess antitumor efficacy In a xenograft model using the CAL-1 cell line, wildtype leukemic cells remained faintly detectable in mouse blood.20 We, therefore, developed a Luc+ CAL-1 cell line to assess leukemic cell proliferation by non-invasive imaging of the luminescent leukemic cells allowing the monitoring of disease progression by bioluminescence. Luciferaseexpressing CAL-1 cells obtained after retroviral transduction with a Luc-retroviral vector carrying luciferase (Luc+) and neomycin resistance (NeoR) genes were injected into NOG mice to develop a BPDCN xenograft mouse model. While disease progression was undetectable in the blood, injection of 0.5, 1 or 5x106 Luc+ CAL-1 cells into NOG mice provided a rapidly detectable total body bioluminescent imaging (BLI) signal (Figure 3A). Indeed, BLI signals were first detectable in the group injected with 5x106 at day 6, and at day 8 in the groups injected with 1x106 or 0.5x106 cells. By day 15, all mice that received 5x106 Luc+ CAL-1 cells were dead from leukemic progression. They developed paralysis of the lower limbs. The other six mice from groups given 0.5x106 and 1x106 cells were euthanized on day 15 and tissue infiltration was evaluated by BLI (Figure 3B). Luc+ CAL-1 cell infiltration was detectable

in the bone marrow, spleen, lungs and liver whereas it remained undetectable in the lymph nodes, kidneys, pancreas, ovary and spinal cord. After sacrificing the mice, immunostaining of the spleen and bone marrow confirmed the presence of cells with a BPDCN phenotype (human CD45+, murine CD45‒, CD56+, CD123+, BDCA4+) (Figure 2C) and cytological analysis showed, as previously described,21 large cells with blastic round or convoluted nuclei with slightly condensed chromatin, several nucleoli and a basophilic cytoplasm (Figure 3D). This model can be evaluated to assess the in vivo antitumor efficacy of bortezomib directly.

In vivo efficacy of bortezomib against primary blastic plasmacytoid dendritic cell neoplasm In order to assess the in vivo efficacy of bortezomib treatment in another way, a xenograft model was developed using primary BPDCN cells isolated from a patient. Weekly injections of bortezomib (4 weeks) significantly increased the overall survival of mice grafted with primary BPDCN cells compared to that of the same mice treated with phosphate-buffered saline (66±13 days versus 42±1 days, P<0.001). Twice-weekly injections of bortezomib further increased the overall survival of mice (77±11 days,

Figure 3. Development of a luciferase-expressing CAL-1 cell xenograft model. Luc+ CAL-1 cells (0.5, 1 or 5x106) were injected intravenously into NOG mice and animals were imaged at days 6, 8, 12, and 15 after the xenograft. Luciferin was administered and images were obtained by integrating the bioluminescent signal. (A) In vivo kinetics of tumor cell growth following the Luc+ CAL-1 cell xenograft. A pseudocolor luminescent image from blue (least intense) to red (most intense) is depicted. (B) Representative analysis of bioluminescent organs at sacrifice at day 15 after the xenograft. This mouse was injected with 0.5x106 Luc+ CAL-1 cells. (C) One representative example of the immunostaining of circulating peripheral blood mononuclear cells performed at day 6 after engraftment. Murine cells (green) and human BPDCN cells (red) are distinguishable based on specific human or murine CD45 antibody expression. Human BPDCN cells express CD56, CD123, and BDCA4. (D) Analysis of circulating cells from blood (left) and spleen cells (right) after May Grünwald Giemsa staining (standard MGG, magnification x1000). These cells were obtained at sacrifice from a mouse inoculated with Luc+ CAL-1 cells.

haematologica | 2017; 102(11)

1865

L. Philippe et al.

P<0.001, n=3-4 mice/group in 2 independent experiments) (Figure 4A). Circulating human BPDCN cells identified in murine blood as human CD45+, CD123+, BDCA4+, CD4+, CD56+ cells decreased from 2640±220 cells/mL (phosphate-buffered saline control mice) to 680±298 cells/mL or 622±158 cells/mL (weekly or twice-weekly bortezomibtreated mice, respectively) at 5 weeks (P<0.01) (Figure 4B,C). We also monitored hemoglobin and platelet counts in mice to assess leukemic cell bone marrow infiltration without observing any major cytopenia in any conditions (data not shown). Thus, in this in vivo model, bortezomib extended mouse survival and reduced circulating blast cells. Furthermore, using measurements of mean fluorescent intensity ratio (MFIR), we confirmed in vivo that PDX cells (patient #127) extracted from the blood of mice treated with bortezomib for 6 h exhibited significantly reduced pRelA expression after bortezomib treatment (mean of

MFIR: 3.2±1.1) compared to pRelA expression of PDX cells from untreated mice (mean of MFIR: 6.35±2.4). This in vivo inhibition of pRelA after bortezomib treatment was similar to that obtained with JSH23 treatment (mean of MFIR: 3.6±1.2) (Figure 2C).

Discussion BPDCN is an aggressive hematodermic neoplasia with a short-term survival.2,4 As there are no data supporting a particular regimen for this acute leukemia, treatments vary from chemotherapy based on a single agent used in B-cell lymphoma4,8 to poly-chemotherapy regimens similar to those given to patients with high-risk acute lymphocytic or acute myeloid leukemia,9,22 and allogeneic hematopoietic cell transplantation for consolidation.7,10,23-25

D E

Figure 4. Bortezomib treatment is efficient at controlling tumor growth in a xenograft model using primary blastic plasmacytoid dendritic cell neoplasm cells. NSG mice were irradiated (2 Gy) and then inoculated intravenously with 1x106 to 2x106 primary BPDCN cells from patient #127 on day 0. Treatment was started on day 100 (J1) after the graft with bortezomib (0.25 mg/kg/mouse intraperitoneally) given one or twice weekly for 4 weeks (n=7 and n=4 mice, respectively). Mice injected with phosphate-buffered saline (PBS) over the 4 weeks were used as the control (n=3). (A) Overall survival of BPDCN inoculated-mice treated with bortezomib (dotted line) or with PBS (solid line) is shown. (B) One example of the immunostaining of peripheral blood performed at day 89 after engraftment. Murine cells (blue) and primary BPDCN (red) cells are distinguishable based on specific human or murine CD45 antibody expression. Human BPDCN cells express CD123, BDCA4, and CD4. (C) Mean of BPDCN cell counts in the blood of mice following treatment with bortezomib (dotted line) or PBS (solid line). (*P<0.05 and ***P<0.001). Intracellular expression of pRelA (pS529 NF-κB p65) was evaluated in PDX cells (BPDCN patient #127) obtained in mouse blood at day 1 and day 15 after in vivo treatment with bortezomib (0.25 mg/kg/mouse intraperitoneally) for 6 h (n=3 mice). JSH23 was used as a positive control (40 mg/kg, n=3 mice) and PBS (control, n=3 mice) as a negative control. PDX cells were stimulated ex vivo with TLR7 for 45 min (R848, 1 mg/mL) before staining. (D) Representative examples of intracellular expression of pRelA and isotype control staining after ex vivo TLR7 stimulation in these different conditions. (E) This histogram represents the mean fluorescence intensity ratio (MFIR) ± SEM of intracellular NF-κBp-65 in PDX cells obtained after treatment with bortezomib on day 1 and day 15, *P<0.05, **P<0.01. NS: unstimulated; S: stimulated with R848. The MFIR was obtained by dividing the mean fluorescence intensity (MFI) obtained with the anti-NF-κBp-65 antibody by the MFI of the respective isotype control antibody.

1866

haematologica | 2017; 102(11)

Bortezomib for BPDCN

New approaches using more targeting therapies are needed for the majority of BPDCN patients unable to receive intensive chemotherapy regimens because of a median age of around 70 years at diagnosis.2,4 Bortezomib is a firstgeneration proteasome inhibitor approved by the Food and Drug Administration for the treatment of refractory multiple myeloma and mantle cell lymphoma.26 The efficacy of bortezomib is governed by its capacity to inhibit the NF-κB pathway, which plays an important role in the pathophysiology of BPDCN. Hirai et al. showed that bortezomib suppresses the survival and immunostimulatory functions of non-leukemic plasmacytoid dendritic cells by targeting intracellular trafficking of nucleic acidsensing Toll-like receptors and altering endoplasmic reticulum homeostasis.27 Using a genomics approach, Sapienza et al. showed that the NF-κB pathway is aberrantly activated in BPDCN, and they reported inhibition of the cell cycle progression and survival of CAL-1 cells after bortezomib treatment.11 The percentage of viable CAL-1 cells decreased significantly by more than 50% when the cells were treated with 30 nM bortezomib for 24 h.11 We recently confirmed constitutive NF-κB activation in BPDCN cells with upregulation of the NF-κB p105 precursor-coding gene (NFKB1) in 12 primary BPDCN samples and demonstrated that inhibition of NF-κB p65 subunit translocation by the specific inhibitor JSH-23 is sufficient to induce BPDCN cell death in vitro.13 Here, our study confirmed in vitro that two BPDCN cell lines (CAL-1 and GEN2.2) and seven primary samples from BPDCN patients are sensitive to bortezomib treatment in terms of cell cycle arrest, cell proliferation inhibition, and cell death induction. Bortezomib has been shown to induce G2/M cell cycle arrest in different tumor cell models.28,29 We confirmed that exposure of CAL-1 and GEN2.2 cells to bortezomib caused a significant accumulation in the G2 phase and in sub-G1 phase (related to apoptotic cells). Furthermore, we demonstrated for the first time in a mouse model that in vivo treatment using bortezomib decreases pRelA. These results reinforce published data from Sapienza et al. obtained in the CAL-1 cell line.11 Moreover, we were able to show that bortezomib is effective in vivo at extending the survival of a primary BPDCN xenograft model. In this model, bortezomib ‒ infused twice weekly for 4 weeks ‒ increased survival for at least 6 additional weeks. Bortezomib is not a “conventional” cytotoxic agent and is used, for example, in multiple myeloma.31,32 For instance, synergistic effects with dexamethasone30 and histone deacetylase inhibitors31,32 have been previously reported. In multiple myeloma, bortezomib is currently used in association with thalidomide and dexamethasone.33 In our hands, we observed an in vitro synergistic effect of bortezomib and a histone deacetylase inhibitor (SAHA), idarubicin, simvastatin and 5-azacytidine in CAL-1 and GEN2.2 cells lines. Recently, Ceroi et al. demonstrated that cholesterol homeostasis is modified in BPDCN cells: cholesterol accumulation within leukemic cells is responsible for these cells’ high proliferative properties and can be normalized by treatment with LXR agonists.13 LXR stimulation in BPDCN exerts an anti-leukemic effect that can be enhanced by increasing cholesterol efflux. Cholesterol dependency of BPDCN cells was confirmed, since inhibition of the mevalonate pathway (i.e., cholesterol synthesis) by atorvastatin was sufficient to induce significant BPDCN cell death. Here, we extend these data. Indeed, we observed an important effect of haematologica | 2017; 102(11)

statins against BPDCN cell lines and a synergistic effect with bortezomib mainly when associated with statins. Kim et al. recently observed a similar effect with simvastatin in combination with bergamottin ‒ an inhibitor of some cytochrome P450 isoforms ‒ that potentiates apoptosis through modulation of the NF-κB signaling pathway in human chronic myelogenous leukemia.34 These results suggest that statins could be a new approach for BPDCN treatment in combination with bortezomib. In the CAL-1 xenograft model, engraftment is not clearly detectable in blood early after infusion, and we, therefore, developed another model to track BPDCN cells easily using evaluation of luciferase-expressing BPDCN cells by measuring the BLI signal. Luc+ CAL-1 cells were preferentially detectable in the bone marrow, spleen, lungs and liver as described in BPDCN patients who exhibit BDPCN cell involvement in many tissues, including the spleen, liver, central nervous system, tonsils, mucous membranes, lungs, kidneys, and muscle.3 Nevertheless, the xenograft model using primary BPDCN cells revealed that bortezomib treatment induced a significant (up to 2-fold) increase of mouse survival with a significant reduction of circulating BPDCN cells. Although current treatment regimens for BPDCN can achieve complete responses, many patients relapse, even after allogeneic hematopoietic cell transplantation, underscoring the need for novel therapeutics. Bortezomib is effective at killing BPDCN cells in vitro and exerts an antileukemic effect in a xenograft mouse model of primary BPDCN. Given its low toxicity, it could be used in combination with other drugs, such as 5-azacytidine or simvastatin, in maintenance for several cycles of treatment to improve the response in elderly patients who cannot benefit from allogeneic hematopoietic cell transplantation. Several compounds of this proteasome inhibitor family are currently under development. Recently, carfilzomid, an irreversible and selective proteasome inhibitor, has shown superiority compared to bortezomib in a phase III myeloma clinical trial.35 Moreover, ixazomib is an orally bioavailable, reversible proteasome inhibitor, approved in combination with lenalidomide and dexamethasone for the treatment of patients with multiple myeloma.36 These molecules should be tested against BPDCN cells alone, or rather in combination with others drugs, such as hypomethylating agents like 5-azacytidine (tested here), which shows promising effects on patients with refractory acute myeloid leukemia,37 simvastatin (tested here), or lenalidomide, which has demonstrated efficacy in a xenograft mouse model of human BPDCN.20 The synergistic effect of these molecules can be evaluated in PDX mouse models. In conclusion, our preclinical results provide a rationale for the use of bortezomib in combination with classical chemotherapy for the treatment of BPDCN patients. A prospective clinical trial combining proteasome inhibitor with cytotoxic drugs should now be performed to prospectively validate these results. Funding This study was supported by grants from the Agence Nationale de la Recherche (LabEx LipSTIC, ANR-11-LABX-0021), the Conseil Régional de Bourgogne Franche-Comté (LabEx LipSTIC 2016 to PS), the DGOS and INCa (National PHRC #PHRCK 16-93 and PRT INCA 2015 #PRT-K15-175), as well as the Ligue Régionale Contre le Cancer (CCIRGE-BFC 2016). 1867

L. Philippe et al.

References 1. Garnache-Ottou F, Feuillard J, Ferrand C, et al. Extended diagnostic criteria for plasmacytoid dendritic cell leukaemia. Br J Haematol. 2009;145(5):624-636. 2. Feuillard J, Jacob MC, Valensi F, et al. Clinical and biologic features of CD4(+)CD56(+) malignancies. Blood. 2002;99(5):1556-1563. 3. Julia F, Dalle S, Duru G, et al. Blastic plasmacytoid dendritic cell neoplasms: clinicoimmunohistochemical correlations in a series of 91 patients. Am J Surg Pathol. 2014;38(5):673-680. 4. Petrella T, Bagot M, Willemze R, et al. Blastic NK-cell lymphomas (agranular CD4+CD56+ hematodermic neoplasms): a review. Am J Clin Pathol. 2005;123(5):662-675. 5. Dalle S, Beylot-Barry M, Bagot M, et al. Blastic plasmacytoid dendritic cell neoplasm: is transplantation the treatment of choice? Br J Dermatol. 2009;162(1):74-79. 6. Dietrich S, Andrulis M, Hegenbart U, et al. Blastic plasmacytoid dendritic cell neoplasia (BPDC) in elderly patients: results of a treatment algorithm employing allogeneic stem cell transplantation with moderately reduced conditioning intensity. Biol Blood Marrow Transplant. 2011;17(8):1250-1254. 7. Gilis L, Lebras L, Bouafia-Sauvy F, et al. Sequential combination of high dose methotrexate and L-asparaginase followed by allogeneic transplant: a first-line strategy for CD4+/CD56+ hematodermic neoplasm. Leuk Lymphoma. 2012;53(8):1633-1637. 8. Kharfan-Dabaja MA, Lazarus HM, Nishihori T, Mahfouz RA, Hamadani M. Diagnostic and therapeutic advances in blastic plasmacytoid dendritic cell neoplasm: a focus on hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2013;19(7): 1006-1012. 9. Pagano L, Valentini CG, Pulsoni A, et al. Blastic plasmacytoid dendritic cell neoplasm with leukemic presentation: an Italian multicenter study. Haematologica. 2013;98(2): 239-246. 10. Roos-Weil D, Dietrich S, Boumendil A, et al. Stem cell transplantation can provide durable disease control in blastic plasmacytoid dendritic cell neoplasm: a retrospective study from the European Group for Blood and Marrow Transplantation. Blood. 2013;121(3):440-446. 11. Sapienza MR, Fuligni F, Agostinelli C, et al. Molecular profiling of blastic plasmacytoid dendritic cell neoplasm reveals a unique pattern and suggests selective sensitivity to NFkB pathway inhibition. Leukemia. 2014;28(8):1606-1616. 12. Moreau P, Richardson PG, Cavo M, et al. Proteasome inhibitors in multiple myeloma: 10 years later. Blood. 2012;120(5):947-959. 13. Ceroi A, Masson D, Roggy A, et al. LXR agonist treatment of blastic plasmacytoid

1868

14.

15.

16.

17.

18.

19.

20.

21. 22.

23.

24.

25.

26.

dendritic cell neoplasm restores cholesterol efflux and triggers apoptosis. Blood. 2016;128(23):2694-2707. Chaperot L, Perrot I, Jacob MC, et al. Leukemic plasmacytoid dendritic cells share phenotypic and functional features with their normal counterparts. Eur J Immunol. 2004;34(2):418-426. Maeda T, Murata K, Fukushima T, et al. A novel plasmacytoid dendritic cell line, CAL1, established from a patient with blastic natural killer cell lymphoma. Int J Hematol. 2005;81(2):148-154. Angelot-Delettre F, Biichle S, Ferrand C, et al. Intracytoplasmic detection of TCL1--but not ILT7-by flow cytometry is useful for blastic plasmacytoid dendritic cell leukemia diagnosis. Cytometry. 2012;81(8):718-724. Garnache-Ottou F, Chaperot L, Biichle S, et al. Expression of the myeloid-associated marker CD33 is not an exclusive factor for leukemic plasmacytoid dendritic cells. Blood. 2005;105(3):1256-1264. Garnache-Ottou F, Feuillard J, Saas P. Plasmacytoid dendritic cell leukaemia/lymphoma: towards a well defined entity? Br J Haematol. 2007;136(4):539-548. Angelot-Delettre F, Roggy A, Frankel AE, et al. In vivo and in vitro sensitivity of blastic plasmacytoid dendritic cell neoplasm to SL401, an interleukin-3 receptor targeted biologic agent. Haematologica. 2015;100(2): 223-230. Agliano A, Martin-Padura I, Marighetti P, et al. Therapeutic effect of lenalidomide in a novel xenograft mouse model of human blastic NK cell lymphoma/blastic plasmacytoid dendritic cell neoplasm. Clin Cancer Res. 2011;17(19):6163-6173. Angelot-Delettre F, Garnache-Ottou F. Blastic plasmacytoid dendritic cell neoplasm. Blood. 2012;120(14):2784. Piccaluga PP, Paolini S, Sapienza MR, Pileri SA. Blastic plasmacytoid dendritic cell neoplasm: is it time to redefine the standard of care? Expert Rev Hematol. 2012;5(4):353355. Aoki T, Suzuki R, Kuwatsuka Y, et al. Longterm survival following autologous and allogeneic stem cell transplantation for blastic plasmacytoid dendritic cell neoplasm. Blood. 2015;125(23):3559-3562. Dalle S, Beylot-Barry M, Bagot M, et al. Blastic plasmacytoid dendritic cell neoplasm: is transplantation the treatment of choice? Br J Dermatol. 2010;162(1):74-79. Gruson B, Vaida I, Merlusca L, et al. Lasparaginase with methotrexate and dexamethasone is an effective treatment combination in blastic plasmacytoid dendritic cell neoplasm. Br J Haematol. 2013;163(4):543-545. Roy SS, Kirma NB, Santhamma B, Tekmal RR, Agyin JK. Effects of a novel proteasome inhibitor BU-32 on multiple myeloma cells. Cancer Chemother Pharmacol. 2014;73(6): 1263-1271.

27. Hirai M, Kadowaki N, Kitawaki T, et al. Bortezomib suppresses function and survival of plasmacytoid dendritic cells by targeting intracellular trafficking of Toll-like receptors and endoplasmic reticulum homeostasis. Blood. 2011;117(2):500-509. 28. Adams J, Palombella VJ, Sausville EA, et al. Proteasome inhibitors: a novel class of potent and effective antitumor agents. Cancer Res. 1999;59(11):2615-2622. 29. Piperdi B, Ling YH, Liebes L, Muggia F, Perez-Soler R. Bortezomib: understanding the mechanism of action. Mol Cancer Ther. 2011;10(11):2029-2030. 30. Koyama D, Kikuchi J, Hiraoka N, et al. Proteasome inhibitors exert cytotoxicity and increase chemosensitivity via transcriptional repression of Notch1 in T-cell acute lymphoblastic leukemia. Leukemia. 2014;28(6): 1216-1226. 31. Bastian L, Hof J, Pfau M, et al. Synergistic activity of bortezomib and HDACi in preclinical models of B-cell precursor acute lymphoblastic leukemia via modulation of p53, PI3K/AKT, and NF-kappaB. Clin Cancer Res. 2013;19(6):1445-1457. 32. Zhang QL, Wang L, Zhang YW, et al. The proteasome inhibitor bortezomib interacts synergistically with the histone deacetylase inhibitor suberoylanilide hydroxamic acid to induce T-leukemia/lymphoma cells apoptosis. Leukemia. 2009;23(8):1507-1514. 33. Cavo M, Tacchetti P, Patriarca F, et al. Bortezomib with thalidomide plus dexamethasone compared with thalidomide plus dexamethasone as induction therapy before, and consolidation therapy after, double autologous stem-cell transplantation in newly diagnosed multiple myeloma: a randomised phase 3 study. Lancet. 2010;376 (9758):2075-2085. 34. Kim SM, Lee EJ, Lee JH, et al. Simvastatin in combination with bergamottin potentiates TNF-induced apoptosis through modulation of NF-kappaB signalling pathway in human chronic myelogenous leukaemia. Pharm Biol. 2016;54(10):2050-2060. 35. Dimopoulos MA, Moreau P, Palumbo A, et al. Carfilzomib and dexamethasone versus bortezomib and dexamethasone for patients with relapsed or refractory multiple myeloma (ENDEAVOR): a randomised, phase 3, open-label, multicentre study. Lancet Oncol. 2016;17(1):27-38. 36. Kumar SK, Berdeja JG, Niesvizky R, et al. Safety and tolerability of ixazomib, an oral proteasome inhibitor, in combination with lenalidomide and dexamethasone in patients with previously untreated multiple myeloma: an open-label phase 1/2 study. Lancet Oncol. 2014;15(13):1503-1512. 37. Walker AR, Klisovic RB, Garzon R, et al. Phase I study of azacitidine and bortezomib in adults with relapsed or refractory acute myeloid leukemia. Leuk Lymphoma. 2014;55(6):1304-1308.

haematologica | 2017; 102(11)

ARTICLE

Acute Lymphoblastic Leukemia

Antigen receptor sequencing of paired bone marrow samples shows homogeneous distribution of acute lymphoblastic leukemia subclones

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Prisca M.J. Theunissen,1 David van Zessen,1,2 Andrew P. Stubbs,2 Malek Faham,3 Christian M. Zwaan,4 Jacques J.M. van Dongen1 and Vincent H.J. Van Der Velden1

Department of Immunology, Erasmus MC, University Medical Center Rotterdam, the Netherlands; 2Department of Bioinformatics, Erasmus MC, University Medical Center Rotterdam, the Netherlands; 3Adaptive Biotechnologies Corp., South San Francisco, CA, USA and 4Department of Pediatric Oncology, Sophia Childrenâ&#x20AC;&#x2122;s Hospital/Erasmus MC, University Medical Center Rotterdam, the Netherlands

Haematologica 2017 Volume 102(11):1869-1877

ABSTRACT

n B-cell precursor acute lymphoblastic leukemia, the initial leukemic cells share the same antigen receptor gene rearrangements. However, due to ongoing rearrangement processes, leukemic cells with different gene rearrangement patterns can develop, resulting in subclone formation. We studied leukemic subclones and their distribution in the bone marrow and peripheral blood at diagnosis. Antigen receptor gene rearrangements (IGH, IGK, TRG, TRD, TRB) were analyzed by nextgeneration sequencing in seven paired bone marrow samples and five paired bone marrow-peripheral blood samples. Background-thresholds were defined, which enabled identification of leukemic gene rearrangements down to very low levels. Paired bone marrow analysis showed oligoclonality in all 7 patients and up to 34 leukemic clones per patient. Additional analysis of evolutionary-related IGH gene rearrangements revealed up to 171 leukemic clones per patient. Interestingly, overall 86% of all leukemic gene rearrangements, including small subclones, were present in both bone marrow samples (range per patient: 72-100%). Paired bone marrow-peripheral blood analysis showed that 83% of all leukemic gene rearrangements in bone marrow were also found in peripheral blood (range per patient: 81-100%). Remarkably, in the paired bone marrow samples and paired bone marrow-peripheral blood samples the vast majority of leukemic gene rearrangements had a similar frequency (<5-fold frequency difference) (96% and 96% of leukemic rearrangements, respectively). Together, these results indicate that B-cell precursor acute lymphoblastic leukemia is generally highly oligoclonal. Nevertheless, the vast majority of leukemic clones, even the minor antigen receptor-defined subclones, are homogeneously distributed throughout the bone marrow and peripheral blood compartment.

v.h.j.vandervelden@erasmusmc.nl

Introduction

ÂŠ2017 Ferrata Storti Foundation

During B-cell and T-cell differentiation, genes that encode the B-cell receptor (Immunoglobulin; IG) and the T-cell receptor (TR) are rearranged, respectively. This process of IG and TR gene rearrangement includes the assembly of Variable (V), Diversity (D) and Joining (J) gene segments and the random insertion and deletion of nucleotides between these gene segments,1,2 resulting in unique V-(D)-J junctions. In normal B cells and T cells, productive V-(D)-J exons encode the antigen-binding domains of the IG and TR molecules. The large diversity of V-(D)-J junctions contributes significantly to the large diversity of IG and TR molecules, thereby providing the B and T lymphocytes with a wide variety of antigen specificities. B-cell precursor acute lymphoblastic leukemia (BCP-ALL) arises from a malignantly transformed B-cell precursor (BCP).3 Therefore, BCP-ALL cells contain IG gene haematologica | 2017; 102(11)

Correspondence:

Received: April 21, 2017. Accepted: August 24, 2017. Pre-published: August 31, 2017. doi:10.3324/haematol.2017.171454 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1869

Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1869

P. M.J. Theunissen et al. rearrangements in the vast majority of cases.4 In addition, most BCP-ALL cells also harbor certain TR gene rearrangements, which are uncommon in normal BCPs.5-7 Since BCP-ALL is thought to arise from a single cell, a BCP-ALL population is expected to be monoclonal with regard to IG/TR gene rearrangements, i.e. to consist of cells which all share the same IG/TR gene rearrangements. However, continuing rearrangement processes and V-segment replacements that occur in normal BCP cells can also occur during the development of BCP-ALL.8-10 As a result, a BCPALL population can be oligoclonal, i.e. can contain subclones with IG/TR gene rearrangements that deviate from those in the initial leukemic cell. Several previous studies had analyzed IG/TR gene rearrangements in BCP-ALL diagnosis samples by using PCR-based methods in combination with Southern blot (SB), which allowed identification of clonal IG/TR gene rearrangements down to the level of 1-5%.11,12 In these studies, oligoclonality was detected in 40% of patients when analyzing the IGH and TRD loci and in 10-20% of patients when analyzing the IGK and TRB loci.11-13 With regard to complete IGH gene rearrangements, a maximum of 9 clonal rearrangements could be found per patient.14 More recently, IGH gene rearrangements in BCP-ALL at diagnosis were analyzed with highly sensitive next-generation sequencing (NGS), which showed that many more clonal IGH gene rearrangements (up to 4024) can be present in a single patient.15 However, in NGS studies of IG/TR gene rearrangements in BCP-ALL patients, it is difficult to discriminate leukemic rearrangements at low-frequency levels from the background of normal lymphocytederived rearrangements and from technical artifacts. Furthermore, NGS-based analysis of gene rearrangements other than complete IGH gene rearrangements has, to the best of our knowledge, so far not been performed in patients with BCP-ALL. Therefore, the degree of oligoclonality in BCP-ALL patients, especially at low-frequency level, remains unclear. BCP-ALL subclones may develop at different locations in the body, after which they may disseminate throughout the bone marrow (BM) and peripheral blood (PB) compartment. Whereas minimal residual disease (MRD) levels

were shown to be comparable between paired BM samples during follow up,16 limited data are available on the distribution of BCP-ALL clones at the time of diagnosis. Our early SB study on IG/TR gene rearrangements in paired BM-PB samples at diagnosis reported that in 5 out of 10 oligoclonal patients, subclonal IGH gene rearrangements were found in the BM sample, but not in the corresponding PB sample.17 However, the relative frequency of these subclonal IGH gene rearrangements in PB might have been too low to be detected with SB. Therefore, it is still unknown to what extent BCP-ALL clones are distributed throughout the BM and PB compartments. Here, we studied the distribution of leukemic clones in BCP-ALL patients by performing NGS of IG/TR gene rearrangements on paired BM-BM samples as well as paired BM-PB samples.

Methods Patient and control samples Bone marrow and PB samples from a total of 12 children with newly diagnosed BCP-ALL were collected (Online Supplementary Table S1). These samples included seven left BM-right BM pairs, i.e. BM from the left and right pelvic bone of the same patient, and five BM-PB pairs, all taken at initial diagnosis. One BM sample from a healthy child (residual material from a donor for allogeneic BM transplantation), two BM samples from BCP-ALL patients at one year after end of therapy, and five regenerating BM samples from T-ALL patients at day 79 and month 5 during therapy were collected to be used as control samples. All control samples were MRD negative according to allele-specific oligonucleotide PCR assays, as described previously.4,18 All samples were obtained according to the guidelines of the local Medical Ethics Committees (MEC 2004-203 and MEC 2012-287) and in line with the Declaration of Helsinki Protocol.

Next-generation sequencing by the Adaptive Biotechnologies method Genomic DNA was isolated from BM and PB mononuclear cells (MNCs) by the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). A detailed description of the NGS-based immunose-

Table 1. Number and distribution of leukemic clones at B-cell precursor acute lymphoblastic leukemia (BCP-ALL) diagnosis in paired bone marrow (BM) samples.a

Vh-Jh Dh-Jh V κ-Jκ Vκ-KDE IntronRSS-KDE V γ-Jγ Vδ-Dδ Dδ-Dδ Vβ-Jβ Total

ID_15760 n clones paired/total

ID_15773 n clones paired/total

ID_15803 n clones paired/total

ID_16079 n clones paired/total

ID_16261 n clones paired/total

ID_16278 n clones paired/total

ID_16300 n clones paired/total

Total n clones paired/total

2/2 (100%) 1/1 (100%) 2/2 (100%) 1/1 (100%) 6/6 (100%)

4/4 (100%) 1/1 (100%) 3/3 (100%) 3/3 (100%) 11/11 (100%)

30/34 (88%) 4/5 (80%) 0/0 (NA) 1/1 (100%) 0/0 (NA) 12/26 (46%) 1/1 (100%) 48/67 (72%)

11/14 (79%) 1/1 (100%) 21/22 (96%) 9/9 (100%) 6/7 (86%) 4/4 (100%) 52/57 (91%)

6/7 (86%) 2/2 (100%) 3/3 (100%) 5/5 (100%) 1/1 (100%) 17/18 (94%)

19/25 (76%) 5/5 (100%) 2/2 (100%) 19/20 (95%) 8/9 (89%) 53/61 (87%)

2/2 (100%) 1/1 (100%) 2/2 (100%) 5/5 (100%) 2/2 (100%) 2/2 (100%) 14/14 (100%)

74/88 (84%) 4/5 (80%) 8/8 (100%) 8/8 (100%) 3/3 (100%) 55/57 (96%) 30/45(67%) 8/9 (89%) 11/11 (100%) 201/234 (86%)

The number of leukemic Vh-DJh, Dh-Jh,Vκ-Jκ,Vκ-KDE, intronRSS-KDE,Vγ-Jγ ,Vδ-Dδ, Dδ-Dδ and Vβ-Jβ rearrangements and the percentage of these leukemic rearrangements that was homogeneously distributed (i.e. found in both BM samples), in BCP-ALL patients at diagnosis. n: number; Intr-KDE: intronRSS-KDE. -: total RC <1000; NA: not applicable.

1870

haematologica | 2017; 102(11)

Homogeneous distribution of ALL subclones

quencing method has been published elsewhere.19 Briefly, a DNA quantity of 420-480 ng, corresponding to 70,000-80,000 MNCs, was used for the amplification of complete IGH gene rearrangements (Vh-DJh; three separate PCRs) and a DNA quantity of 120180 ng, corresponding to 20,000-30,000 MNCs, was used for the amplification of incomplete IGH gene rearrangements (Dh-Jh), IGK gene rearrangements (Vκ-Jκ, Vκ-KDE, intronRSS-KDE), TRG gene rearrangements (Vγ-Jγ), TRD gene rearrangements (Vδ-Dδ, Dδ-Dδ) and TRB gene rearrangements (Vβ-Jβ) (all single PCRs). These rearrangements were amplified in a first PCR reaction of 25 cycles, using locus-specific primer sets.20 Next, 1:100 of these PCR products was further amplified in a second PCR reaction of 14 cycles, using universal primers complementary to the adaptors that were linked to the locus-specific primers with sample-identifiers. The final PCR products were sequenced using the Illumina HiSEQ platform. Low-quality reads were filtered out and sequences with a single read were excluded.19 For each sequence, the absolute read count, i.e. the read count corrected for PCR amplification by the spike-in method, was calculated (Online Supplementary Methods). Also the frequency for each sequence was determined. Importantly, the frequency represents a fraction of the rearranged genes of the involved locus, rather than a fraction of all, including germline, IG and TR genes, as is the case in SB and

PCR-based IG/TR gene Supplementary Methods).

rearrangement

analysis

(Online

NGS clonal variant comparison pipeline and data analysis We developed a new analysis tool (PRISCA: PRecISe Clonal Analysis) in Galaxy21 to compare sequences from paired and triplicate samples (Online Supplementary Methods). The graphical output from this tool was used for further data analysis. The IG/TR gene rearrangements at diagnosis were subsequently analyzed according to the criteria described in the Online Supplementary Methods. In line with previous reports,19,22-24 an index clone was arbitrarily defined as a clonal IG/TR gene rearrangement with a frequency >5%.

Results Normal and regenerating BM samples define thresholds for the exclusion of normal lymphocyte-derived IG/TR gene rearrangements For in-depth analysis of IG/TR gene rearrangements in BCP-ALL patients, it is necessary to discriminate the IG/TR gene rearrangements derived from leukemic clones

Figure 1. Read counts (RC) of Vh-DJh rearrangements in bone marrow (BM) from a healthy child, a T- cell acute lymphoblastic leukemia (TALL) patient at a therapy interval (regenerating BM) and a B-cell precursor (BCP)-acute lymphoblastic leukemia (ALL) patient at one year after end of therapy (representative examples of all non-leukemic BM samples; total n=8). VhDJh with a read count of <1 (caused by amplification-correction) were displayed as rearrangements with a read count of 1. Black line indicates the established threshold for Vh-DJh rearrangements. Regenerating BM showed structurally lower numbers of unique Vh-DJh rearrangements compared to healthy control BM and BM one year after end of therapy. This was also true for Vκ-Jκ, Vκ-KDE and intronRSSKDE rearrangements, but not Dh-Jh rearrangements (Online Supplementary Figure S1). This was expected, since the B-cell population in regenerating BM mainly consists of pre-B-I stage cells.25,26

haematologica | 2017; 102(11)

1871

P. M.J. Theunissen et al.

from the background that also contains IG/TR gene rearrangements derived from normal B-cell and T-cell clones. We therefore aimed to define read count thresholds (RCTs) above which we can fairly assume normal lymphocyte-derived IG/TR gene rearrangements to be absent. To define these RCTs, we determined the maximum of the absolute read counts of normal lymphocytederived IG/TR gene rearrangements detected in eight nonleukemic BM samples, i.e. a normal BM sample from a healthy child, two normal BM samples from BCP-ALL patients one year after end of therapy, and five regenerating BM samples from T-ALL patients during therapy intervals. Importantly, regenerating BM samples did not contain residual leukemic IG/TR gene rearrangements (as determined by the absence of leukemic IG/TR gene rearrangements which were found by NGS in the corresponding diagnosis sample) (data not shown). We particularly included regenerating BM samples (and only limited healthy BM samples) as controls since we expected the highest background of IG/TR rearrangements in such samples with increased numbers of early B-cell precursors. Analysis of complete IGH gene rearrangements showed that the maximum read count of complete IGH gene rearrangements in the non-leukemic BM samples was 26 (Figure 1). The IGH Vh-DJh RCT was therefore set at 30 reads. The same strategy was applied to the other types of IG/TR gene rearrangements, resulting in an RCT of 30 reads for Dh-Jh rearrangements, 70 reads for Vκ-Jκ rearrangements, 40 reads for Vκ-KDE rearrangements, 170

reads for intronRSS-KDE rearrangements, 50 reads for Vγ-Jγ rearrangements, 20 reads for Vδ-Dδ rearrangements, 30 reads for Dδ-Dδ rearrangements, and 30 reads for Vγ-Jβ rearrangements (Online Supplementary Figure S1). The relatively high maximum read count of intronRSS-KDE rearrangements was probably due to the limited junctional variability of intronRSS-KDE rearrangements. Importantly, these RCTs are dependent on DNA input and are therefore not universal, but are specific for our data.

Samples run in triplicate show that our thresholds also exclude IG/TR gene rearrangements derived from technical artifacts We aimed to evaluate whether IG/TR gene rearrangements with a read count above the RCTs were truly clonal, i.e. were derived from leukemic clones and not from technical artifacts (non-clonal rearrangements whose read counts were insufficiently corrected after PCR amplification). We therefore prepared three independent libraries of the same sample and performed three independent NGSruns, based on the fair assumption that IG/TR gene rearrangements can be regarded as clonal if present in all three samples or if present in two of the three samples (since small clones can coincidently be absent in a sampling volume). Analysis of complete IGH gene rearrangements in a BM sample at diagnosis showed that all complete IGH gene rearrangements with a read count above the previously defined RCT, and thus considered to be derived from leukemic clones, were present in all three

Figure 2. Read counts of Vh-DJh rearrangements in bone marrow (BM) replicates from a B-cell precursor (BCP)-acute lymphoblastic leukemia (ALL) patient at diagnosis. This BM was sequenced in triplicate, resulting in IGH gene rearrangements that were found in one, in two or in all three replicates. Vh-DJh rearrangements with a read count of <1 (caused by amplification-correction) were displayed as rearrangements with a read count of 1. Black line indicates the previously defined threshold for Vh-DJh rearrangements (30 reads).

1872

haematologica | 2017; 102(11)

Homogeneous distribution of ALL subclones

replicates, indicating that these were not technical artefacts (Figure 2). Analysis of the other IG/TR gene rearrangement types in this diagnosis sample showed the same (Online Supplementary Figure S2). Thus, by using these RCTs, we ensured that the included rearrangements (above the RCT) were very likely to be derived from leukemic clones (as opposed to normal lymphocytes or technical artefacts), thereby accepting that leukemic rearrangements with a read count below the RCT will be ignored. Analysis of IG/TR gene rearrangements in a regenerating BM sample of a T-ALL patient showed, as expected, that none of the rearrangements had a read count above the RCT. Nevertheless, part of the IG/TR gene rearrangements with a read count below the RCT were present in two or three of the replicates, and could thus be derived from normal B-cell or T-cell clones (Online Supplementary Figure S3).

Paired BM samples show homogeneous distribution in BCP-ALL, also for small subclones Next, we aimed to analyze the number and particularly the tissue distribution of clones in BCP-ALL patients. Therefore, we performed NGS of IG/TR gene rearrangements in paired BM samples (from the left and the right pelvic bone) from newly diagnosed BCP-ALL patients (n=7). By applying the previously described thresholds (now expressed as frequency) and analysis criteria (Online Supplementary Methods) to complete IGH rearrangements, oligoclonality was detected in 6 out of 7 BCP-ALL patients (Figure 3). The number of index clones (frequency of >5%) ranged from 1 to 5 (median: 1), whereas the total number of (sub)clones ranged from 2 to 34. With regard to distribution of the clones, 84% (74 out of 88) of all leukemic Vh-DJh rearrangements were paired (i.e. present in both BM samples), ranging from 76% to 100% per patient (Figure 3). Importantly, 61 out of the 74 (82%) paired leukemic Vh-DJh rearrangements had a comparable frequency (<5-fold difference) in the two BM samples.

Within the 13 paired leukemic Vh-DJh index clones, all Vh-DJh rearrangements had a comparable frequency in the two BM samples. The leukemic Vh-DJh rearrangements which were present in only one of the two samples all had a very low frequency (below 0.2%) and often shared a common Dh-Jh stem with a paired leukemic VhDJh rearrangement (see below). Analysis of the other types of IG/TR gene rearrangements also showed that the vast majority of the leukemic rearrangements were paired: 4 out of 5 (80%) Dh-Jh rearrangements, 8 out of 8 (100%) Vκ-Jκ rearrangements, 8 out of 8 (100%) Vκ-KDE rearrangements, 3 out of 3 (100%) intronRSS-KDE rearrangements, 55 out of 57 (96%) Vγ-Jγ rearrangements, 30 out of 45 (67%) Vδ-Dδ rearrangements, 8 out of 9 (89%) Dδ-Dδ rearrangements, and 11 out of 11 (100%) Vβ-Jβ rearrangements (Online Supplementary Figure S4). Within all paired leukemic IG/TR gene rearrangements, 171 out of all 199 (86%) showed a comparable frequency in the two BM samples. The overall results per patient are summarized in Table 1.

Analysis of leukemic Dh-Jh stems at very low frequency levels suggests even higher numbers of homogeneously distributed subclones We subsequently aimed to estimate the number of leukemic IG/TR gene rearrangements that were present at very low frequencies and that were, based on the background-threshold, excluded from previous analyses. Therefore, we visualized all paired complete IGH gene rearrangements, irrespective of their frequency (i.e. also with a frequency below the threshold). Only paired IGH gene rearrangements (i.e. present in both BM samples) were visualized in order to exclude potential technical artifacts. The resulting graph showed that, besides the previously detected paired IGH gene rearrangements, many more paired IGH gene rearrangements were present below the threshold (Figure 4). However, a paired IGH gene rearrangement below the threshold can either be

Figure 3. Frequencies of leukemic Vh-DJh rearrangements in paired bone marrow (BM) samples [left (L) and right (R) pelvic bone] from B-cell precursor (BCP)acute lymphoblastic leukemia (ALL) patients at diagnosis. For each leukemic Vh-DJh rearrangement, the presence in both BM samples (L+R) or in only one of the two BM samples (L or R) is indicated. The background area, which also contains Vh-DJh rearrangements derived from normal B-cell clones, is indicated for each sample (gray). *Vh-DJh rearrangement sharing a common Dh-Jh stem with a paired leukemic Vh-DJh rearrangement.

haematologica | 2017; 102(11)

1873

P. M.J. Theunissen et al.

derived from a leukemic clone or derived from a normal mature B-cell clone which had been distributed throughout the BM compartment. To discriminate, at sub-threshold level, leukemia-derived IGH gene rearrangements from normal B-cell-derived IGH gene rearrangements, we searched for common Dh-Jh stems. An IGH gene rearrangement below the threshold that shared a common Dh-Jh stem with a leukemic IGH gene rearrangement above the threshold was also considered leukemic (since both were derived from the same ancestral clone). This analysis showed that, on top of the leukemic IGH gene rearrangements found in the previous analyses (above the threshold), even more leukemic IGH gene rearrangements are present at very low frequency levels (below the threshold) (summarized in Figure 5). The exact number of leukemic IGH gene rearrangement per patient remains unknown, since the origin (leukemia- or mature B-cellderived) of the IGH gene rearrangements with a frequency below the threshold and without a common Dh-Jh stem could not be determined. Of note, the IGH gene rearrangements with a common Dh-Jh stem also often had a common V-D junction, indicating that V to D-J joining as well as Vh-replacement had taken place.

Paired BM-PB samples confirm the homogeneous distribution of BCP-ALL clones via PB The above data indicate that BCP-ALL clones are distributed homogeneously over the BM compartment, implying that migration via PB is required. Therefore, we investigated whether leukemia-derived IG/TR gene rearrangements detected in BM can also be found in PB. To do so, we performed NGS of IG/TR gene rearrangements in paired BM-PB samples from 5 newly diagnosed BCP-ALL patients. Leukemic IG/TR gene rearrangements found in the BM sample (i.e. with a frequency above the previously established thresholds) were searched in the corresponding PB sample and classified as ‘paired’ when also present in the PB sample. Analysis of complete IGH gene rearrangements showed that 98% (43 out of 44) of all leukemic Vh-DJh rearrangements were paired (Figure 6). Forty-two of the 43 (98%) paired leukemic Vh-DJh

rearrangements had a comparable frequency (<5-fold difference) in the two corresponding samples. The only leukemic Vh-DJh rearrangement present in the BM sample but not in the PB sample, had a frequency of 0.04% and shared a common Dh-Jh stem with a paired leukemic VhDJh rearrangement. Analysis of the other IG and TR gene rearrangement types also showed that the majority of the leukemic rearrangements were paired: 30 out of 32 (94%) Dh-Jh rearrangements, 2 out of 2 (100%) Vκ-Jκ rearrangements, 1 out of 1 (100%) Vκ-KDE rearrangements, 9 out of 9 (100%) Vγ-Jγ rearrangements, 320 out of 403 (79%) Vδ-Dδ rearrangements, 17 out of 17 (100%) Dδ-Dδ rearrangements, and 4 out of 4 (100%) Vβ-Jβ rearrangements (no clonal intronRSS-KDE rearrangements were found) (Online Supplementary Figure S5). Within all paired leukemic IG/TR gene rearrangements, 411 out of all 426 (96%) showed a comparable frequency between the BM and the corresponding PB sample.

Discussion To study the number of clones and their distribution in patients with BCP-ALL, we performed NGS of IG/TR gene rearrangements in paired BM-BM and BM-PB samples at diagnosis. First, we aimed to carefully define background RCTs, which exclude normal lymphocyte-derived IG/TR gene rearrangements when using an identical amount of DNA per sample. Therefore, we determined the maximum read count of these normal lymphocytederived IG/TR gene rearrangements in non-leukemic BM, i.e. healthy control BM, regenerating BM, and BM one year after end of therapy. Second, by analyzing BCP-ALL samples at diagnosis in triplicate, we confirmed that all IG/TR gene rearrangements with a read count above the pre-set RCT represented clonal IG/TR gene rearrangements, and not technical artifacts. Of note, RCTs could not exclude IG/TR gene rearrangements derived from normal lymphocyte expansions directed against leukemic cells, since these expansions were not present in the studied follow-up samples (or in the healthy control BM sam-

Figure 4. Frequencies of Vh-DJh rearrangements in paired bone marrow (BM) samples [left (L) and right (R) pelvic bone] from B-cell precursor (BCP)-acute lymphoblastic leukemia (ALL) patients at diagnosis. All paired Vh-DJh rearrangements are shown, i.e. with a frequency above as well as below the threshold. Background area of each sample is indicated in gray.

1874

haematologica | 2017; 102(11)

Homogeneous distribution of ALL subclones

ple). However, since many Vh-DJh rearrangements shared a common stem with an index clone and since, besides clonal TR gene rearrangements, also clonal IG gene rearrangements were observed (whereas reactive expansions against leukemic cells would be mainly of T-cell origin), it is likely that most IG/TR gene rearrangements with read counts above the RCTs are of leukemic origin. Analysis of IG/TR gene rearrangements in paired BMBM diagnosis samples showed that all 7 patients were oligoclonal. Remarkably high numbers of leukemic VhDJh, Vγ-Jγ and Vδ-Dδ rearrangements per patient could be detected (e.g. up to 34 leukemic Vh-DJh rearrangements). Even higher numbers of leukemic rearrangements per patient were detected when common Dh-Jh stem (junction sequence) analysis was used to also include Vh-DJh rearrangements below the background threshold (up to 171 leukemic Vh-DJh rearrangements). Previous studies that had used SB to analyze IG/TR gene rearrangements reported that only 40% of the ALL-patients were oligoclonal, with maximally 9 leukemic rearrangements per patient.11,14 The discrepancies between these previous findings and our current results can be explained by the difference in sensitivity of the applied methods. A more recent study that used NGS to analyze IGH gene rearrangements, showed, in line with our results, that far more IGH gene rearrangements per patient can be found when analysis down to very low frequency levels is allowed.15 However, in contrast with our results, this

study also reported on a few patients (4 out of the 51 studied patients) in which more than 1000 leukemic IGH gene rearrangements were found. Patients with such exceptionally high numbers of IGH gene rearrangements were not present in our study, which is probably related to the lower number of patients analyzed. In any case, we showed that the number of leukemic IG/TR gene rearrangements per BCP-ALL patient is considerably higher than previously observed. IG/TR gene rearrangements are used as PCR targets for MRD detection.27-31 Insight into subclone formation and distribution of these subclones might improve selection of MRD targets. Our current findings indicate that MRD targets have so far been selected from only a small part of the total number of leukemic IG/TR gene rearrangements present at diagnosis, i.e. from those that were sufficiently large to be detected by PCR-based methods (generally index clones, i.e. with a frequency >5%). Still, more than 90% of the relapsed cases had maintained the IG and TR targets selected at diagnosis.30 Apparently, monitoring of only the index clones might be sufficient to predict relapse in most BCP-ALL patients. Furthermore, our observation that the vast majority of leukemic clones could be detected both in BM and PB implies that also PB samples may reliably be used for NGS-based selection of MRD targets. By analyzing paired BM samples, we showed that the vast majority of the leukemic rearrangements were present in both BM samples and that their frequency was gen-

Figure 5. Schematic representation of the relations between Vh-DJh rearrangements, based on common Dh-Jh stems, in bone marrow (BM) samples from B-cell precursor (BCP)-acute lymphoblastic leukemia (ALL) patients at diagnosis. The size of the circle represents the frequency of the Vh-DJh rearrangement. Blue or red indicates a leukemic origin (blue: based on frequency; red: based on a common Dh-Jh stem), whereas green indicates small clones of unknown origin (no leukemic Dh-Jh stem).

haematologica | 2017; 102(11)

1875

P. M.J. Theunissen et al.

Figure 6. Frequencies of leukemic Vh-DJh rearrangements in paired bone marrow (BM)-peripheral (PB) samples from B-cell precursor (BCP)-acute lymphoblastic leukemia (ALL) patients at diagnosis. For each leukemic Vh-DJh rearrangement in the BM sample (above the background threshold), the presence in both samples (BM+PB) or in only the BM sample (BM) is indicated. Diamonds represent Vh-DJh rearrangements found in BM, triangles represent Vh-DJh rearrangements found in PB. Background area of each sample is indicated in gray. *Vh-DJh rearrangement sharing a common Dh-Jh stem with a paired leukemic Vh-DJh rearrangement.

erally similar in both samples. The few leukemic rearrangements that were present in only one of the two samples generally had a low frequency (<2%). Probably, also these low-frequent clones were present in the whole BM compartment, but were absent in one of the two samples based on statistical chance. By analyzing paired BMPB samples (from other patients), we showed that almost all leukemic rearrangements that were present in BM, were also present in PB, generally at a similar frequency. As an exception to the above discussed results, relatively many leukemic Vδ-Dδ rearrangements in patient 15803 (child with a TCF3-PBX1 translocation) and in patient 15507 (infant with an MLL-ENL translocation) were present in only one of the paired BM samples (54% and 22%, respectively). Since infants with a MLL-R translocation are known to be highly oligoclonal,32,33 the specific genetic aberrancies might be associated with the deviating results on Vδ-Dδ rearrangements.34 It is important to note that we analyzed diagnosis samples in which tumor percentages were generally high, and that the situation during treatment (with lower numbers of ALL cells) might be different. Nevertheless, our previous study16 suggests that also in an MRD setting the ALL clones are homogeneously distributed over the BM compartment. Next-generation sequencing-based IG and TR gene rearrangement analysis in paired BM-BM and BM-PB samples from newly diagnosed BCP-ALL patients has, to our knowledge, not been performed previously. Our early SB study on IG gene rearrangements in paired BM-PB samples of newly diagnosed BCP-ALL patients showed that a relatively high number of leukemic IGH gene rearrangements were detected in the BM sample, but not or in a lower frequency in the corresponding PB sample,17 which seems to be in contrast with our current results. However, this discrepancy is probably due to technical differences between Southern blot and NGS analytical methods (see Methods section). Together, our paired BM-BM and paired BM-PB NGS analyses showed that almost all leukemic IG/TR gene 1876

rearrangements are present in comparable frequencies in the BM and PB compartment. We did not study leukemic IG/TR gene rearrangements in tissues other than BM or PB. However, it might be speculated that extramedullary dissemination is different from distribution within the BM and PB compartment, since only clones with specific characteristics (e.g. homing receptors) are able to migrate to extramedullary sites.35 Furthermore, it should be noted that we defined subclones based on IG/TR gene rearrangements. It may well be that subclones with different somatic mutations, related to leukemogenesis and prone to positive selection, are distributed in a less homogeneous way. Exome sequencing of paired BM samples may answer this question. In conclusion, by using NGS data in combination with carefully defined thresholds, we were able to study leukemic IG/TR gene rearrangements down to a low level (0.01-0.1%), without the interference of background rearrangements. By studying paired BM-BM and paired BM-PB samples, we showed that the relative quantity of IG/TR gene rearrangements within the total (rearranged) BCP-ALL population is generally similar throughout the BM compartment and between BM and PB. Apparently, after their development at different BM sites in the body, almost all IG/TR-defined clones within a BCP-ALL population, even the small subclones, disseminate homogeneously via PB throughout the BM compartment. Acknowledgments We gratefully thank Arjan Lankester and the involved technicians of the Laboratory for Medical Immunology, in particular Rianne Noordijk, Patricia Hoogeveen and Maaike de Bie. We acknowledge Auke Beishuizen for providing BM samples from the BCP-ALL patients. Funding The research for this manuscript was (in part) performed within the framework of the Erasmus Postgraduate School Molecular Medicine and was financially supported by PrioMedChild, project 40-41800-98-027. haematologica | 2017; 102(11)

Homogeneous distribution of ALL subclones

References 1. Tonegawa S. Somatic generation of antibody diversity. Nature. 1983; 302(5909):575-581. 2. Alt FW, Oltz EM, Young F, Gorman J, Taccioli G, Chen J. VDJ recombination. Immunol Today. 1992;13(8):306-314. 3. Pui CH, Campana D, Evans WE. Childhood acute lymphoblastic leukaemia-current status and future perspectives. Lancet Oncol. 2001;2(10):597-607. 4. van der Velden VH, van Dongen JJ. MRD detection in acute lymphoblastic leukemia patients using Ig/TCR gene rearrangements as targets for real-time quantitative PCR. Methods Mol Biol. 2009;538:115-150. 5. Chen Z, Le Paslier D, Dausset J, et al. Human T cell gamma genes are frequently rearranged in B-lineage acute lymphoblastic leukemias but not in chronic B cell proliferations. J Exp Med. 1987;165(4):10001015. 6. Szczepanski T, Beishuizen A, PongersWillemse MJ, et al. Cross-lineage T cell receptor gene rearrangements occur in more than ninety percent of childhood precursor-B acute lymphoblastic leukemias: alternative PCR targets for detection of minimal residual disease. Leukemia. 1999; 13(2):196-205. 7. Fronkova E, Krejci O, Kalina T, Horvath O, Trka J, Hrusak O. Lymphoid differentiation pathways can be traced by TCR delta rearrangements. J Immunol. 2005; 175(4):2495-2500. 8. Choi Y, Greenberg SJ, Du TL, et al. Clonal evolution in B-lineage acute lymphoblastic leukemia by contemporaneous VH-VH gene replacements and VH-DJH gene rearrangements. Blood. 1996;87(6):25062512. 9. Moreira I, Papaioannou M, Mortuza FY, et al. Heterogeneity of VH-JH gene rearrangement patterns: an insight into the biology of B cell precursor ALL. Leukemia. 2001; 15(10):1527-1536. 10. Stankovic T, Weston V, McConville CM, et al. Clonal diversity of Ig and T-cell receptor gene rearrangements in childhood B-precursor acute lymphoblastic leukaemia. Leuk Lymphoma. 2000;36(3-4):213-224. 11. Beishuizen A, Hahlen K, Hagemeijer A, et al. Multiple rearranged immunoglobulin genes in childhood acute lymphoblastic leukemia of precursor B-cell origin. Leukemia. 1991;5(8):657-667. 12. van der Velden VH, Bruggemann M, Hoogeveen PG, et al. TCRB gene rearrangements in childhood and adult precursor-BALL: frequency, applicability as MRD-PCR target, and stability between diagnosis and relapse. Leukemia. 2004;18(12):1971-1980. 13. van der Velden VH, Szczepanski T, Wijkhuijs JM, et al. Age-related patterns of immunoglobulin and T-cell receptor gene

haematologica | 2017; 102(11)

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

rearrangements in precursor-B-ALL: implications for detection of minimal residual disease. Leukemia. 2003;17(9):1834-1844. Beishuizen A, Verhoeven MA, van Wering ER, Hahlen K, Hooijkaas H, van Dongen JJ. Analysis of Ig and T-cell receptor genes in 40 childhood acute lymphoblastic leukemias at diagnosis and subsequent relapse: implications for the detection of minimal residual disease by polymerase chain reaction analysis. Blood. 1994; 83(8):2238-2247. Gawad C, Pepin F, Carlton VE, et al. Massive evolution of the immunoglobulin heavy chain locus in children with B precursor acute lymphoblastic leukemia. Blood. 2012;120(22):4407-4417. van der Velden VH, Hoogeveen PG, Pieters R, van Dongen JJ. Impact of two independent bone marrow samples on minimal residual disease monitoring in childhood acute lymphoblastic leukaemia. Br J Haematol. 2006;133(4):382-388. Beishuizen A, Verhoeven MA, Hahlen K, van Wering ER, van Dongen JJ. Differences in immunoglobulin heavy chain gene rearrangmeent patterns between bone marrow and blood samples in childhood precursor B-acute lymphoblastic leaukemia at diagnosis. Leukemia. 1993;7(6):60-63. van der Velden VH, Cazzaniga G, Schrauder A, et al. Analysis of minimal residual disease by Ig/TCR gene rearrangements: guidelines for interpretation of realtime quantitative PCR data. Leukemia. 2007;21(4):604-611. Faham M, Zheng J, Moorhead M, et al. Deep-sequencing approach for minimal residual disease detection in acute lymphoblastic leukemia. Blood. 2012; 120(26):5173-5180. Faham M. Methods of monitoring conditions by sequence analysis. 2011. Available from: http://www. google.sr/patents/ US9228232 Afgan E, Baker D, van den Beek M, et al. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2016 update. Nucleic Acids Res. 2016;44(W1):W3-W10. Sekiya Y, Xu Y, Muramatsu H, et al. Clinical utility of next-generation sequencing-based minimal residual disease in paediatric B-cell acute lymphoblastic leukaemia. Br J Haematol. 2017;176(2):248257. Pulsipher MA, Carlson C, Langholz B, et al. IgH-V(D)J NGS-MRD measurement preand early post-allotransplant defines very low- and very high-risk ALL patients. Blood. 2015;125(22):3501-3508. Ladetto M, Bruggemann M, Monitillo L, et al. Next-generation sequencing and realtime quantitative PCR for minimal residual disease detection in B-cell disorders. Leukemia. 2014;28(6):1299-1307. van Lochem EG, Wiegers YM, van den

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Beemd R, Hahlen K, van Dongen JJ, Hooijkaas H. Regeneration pattern of precursor-B-cells in bone marrow of acute lymphoblastic leukemia patients depends on the type of preceding chemotherapy. Leukemia. 2000;14(4):688-695. van Wering ER, van der Linden-Schrever BE, Szczepanski T, et al. Regenerating normal B-cell precursors during and after treatment of acute lymphoblastic leukaemia: implications for monitoring of minimal residual disease. Br J Haematol. 2000; 110(1):139-146. Szczepanski T, Orfao A, van der Velden VH, San Miguel JF, van Dongen JJ. Minimal residual disease in leukaemia patients. Lancet Oncol. 2001;2(7):409-417. van Dongen JJ, Seriu T, Panzer-Grumayer ER, et al. Prognostic value of minimal residual disease in acute lymphoblastic leukaemia in childhood. Lancet. 1998; 352(9142):1731-1738. van der Velden VH, Hochhaus A, Cazzaniga G, Szczepanski T, Gabert J, van Dongen JJ. Detection of minimal residual disease in hematologic malignancies by real-time quantitative PCR: principles, approaches, and laboratory aspects. Leukemia. 2003;17(6):1013-1034. Pieters R, de Groot-Kruseman H, Van der Velden V, et al. Successful Therapy Reduction and Intensification for Childhood Acute Lymphoblastic Leukemia Based on Minimal Residual Disease Monitoring: Study ALL10 From the Dutch Childhood Oncology Group. J Clin Oncol. 2016;34(22):2591-601. van Dongen JJ, van der Velden VH, Bruggemann M, Orfao A. Minimal residual disease diagnostics in acute lymphoblastic leukemia: need for sensitive, fast, and standardized technologies. Blood. 2015; 125(26):3996-4009. Bardini M, Woll PS, Corral L, et al. Clonal variegation and dynamic competition of leukemia-initiating cells in infant acute lymphoblastic leukemia with MLL rearrangement. Leukemia. 2015;29(1):3850. Jansen MW, Corral L, van der Velden VH, et al. Immunobiological diversity in infant acute lymphoblastic leukemia is related to the occurrence and type of MLL gene rearrangement. Leukemia. 2007;21(4):633641. Brumpt C, Delabesse E, Beldjord K, et al. The incidence of clonal T-cell receptor rearrangements in B-cell precursor acute lymphoblastic leukemia varies with age and genotype. Blood. 2000;96(6):22542261. van der Velden VH, de Launaij D, de Vries JF, et al. New cellular markers at diagnosis are associated with isolated central nervous system relapse in paediatric B-cell precursor acute lymphoblastic leukaemia. Br J Haematol. 2016;172(5):769-781.

1877

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

Targeting metabolism and survival in chronic lymphocytic leukemia and Richter syndrome cells by a novel NF-κB inhibitor

Tiziana Vaisitti,1,2 Federica Gaudino,1,2 Samedy Ouk,3 Maria Moscvin,2 Nicoletta Vitale,4 Sara Serra,1,2 Francesca Arruga,2 Johannes L. Zakrzewski,5 Hsiou-Chi Liou,3 John N. Allan,6 Richard R. Furman6 and Silvia Deaglio1,2

Haematologica 2017 Volume 102(11):1878-1889

1 Department of Medical Sciences, University of Turin, Italy; 2Italian Institute for Genomic Medicine, Turin, Italy; 3ImmuneTarget Inc., San Diego, CA, USA; 4Department of Molecular Biotechnologies and Health Sciences, University of Turin, Italy; 5Department of Pediatrics, Memorial Sloan Kettering Cancer Center, New York, NY, USA and 6 CLL Research Center, New York Presbyterian Hospital, Weill Cornell Medicine, New York, NY, USA

ABSTRACT

Correspondence: tiziana.vaisitti@unito.it or silvia.deaglio@unito.it.

Received: May 30, 2017. Accepted: August 28, 2017. Pre-published: August 31, 2017. doi:10.3324/haematol.2017.173419 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1878 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1878

T-901 is a novel and selective NF-κB inhibitor with promising activity in pre-clinical models. Here we show that treatment of chronic lymphocytic leukemia cells (CLL) with IT-901 effectively interrupts NF-κB transcriptional activity. CLL cells exposed to the drug display elevated mitochondrial reactive oxygen species, which damage mitochondria, limit oxidative phosphorylation and ATP production, and activate intrinsic apoptosis. Inhibition of NF-κB signaling in stromal and myeloid cells, both tumor-supportive elements, fails to induce apoptosis, but impairs NF-κB-driven expression of molecules involved in cell-cell contacts and immune responses, essential elements in creating a pro-leukemic niche. The consequence is that accessory cells do not protect CLL cells from IT901-induced apoptosis. In this context, IT-901 shows synergistic activity with ibrutinib, arguing in favor of combination strategies. IT-901 is also effective in primary cells from patients with Richter syndrome (RS). Its anti-tumor properties are confirmed in xenograft models of CLL and in RS patient-derived xenografts, with documented NF-κB inhibition and significant reduction of tumor burden. Together, these results provide pre-clinical proof of principle for IT-901 as a potential new drug in CLL and RS. Introduction Nuclear factor-kappa B (NF-κB) is a ubiquitous transcription factor, composed of a family of five structurally related proteins, including p50 (NF-κB1), p52 (NF-κB2), p65 (RelA), RelB and c-Rel, which can form homo- and hetero-dimers. While NF-κB is normally kept inactivated through binding to the inhibitory subunit (IκB), IκB phosphorylation and degradation releases the dimer that translocates to the nucleus and binds to target sequences on DNA.1-3 NF-κB signaling plays essential roles in inflammation, immune responses, proliferation, and cell survival.4-6 In cancer cells, NF-κB promotes tumor growth by contributing to maintenance/expansion of tumor-initiating cells and by shaping the tumor microenvironment.7 Deregulated NF-κB signaling is a common finding in most, if not all, B-lymphoid malignancies.8 Chronic lymphocytic leukemia cells (CLL) exhibit high constitutive NF-κB activation compared to normal B lymphocytes, with the p65 subunit being the most active and relevant for transcription.9-12 Moreover, p65 levels correlate with leukemic cell survival in vitro, as well as with tumor burden and shorter lymphocyte doubling time.13,14 NF-κB is activated by B-cell receptor (BCR) signaling, the driving force behind CLL pathogenesis and progression.15 However, additional signals from the micro-environment increase NF-κB activity and enhance CLL cell survival, with members of the Bcl-2 family among the most important transcriptional targets.16-19 Approximately 5-10% of CLL patients undergo transformation of the leukemia into an aggressive diffuse large B-cell lymphoma (DLBCL), a complication known as Richter syndrome (RS).20 The clinical outcome of RS is generally poor, with a median survival of a few months, mainly because of limited therapeutic options and responses.21 In a significant proportion of cases, RS patients carry activating haematologica | 2017; 102(11)

Targeting of NF-κB in CLL and RS cells

mutations in genes belonging to the NF-κB pathway.22 Given the central role of the NF-κB signals in CLL and in RS transformation, and specifically of the p65 subunit, this complex represents an important therapeutic target, even if so far none of the studied inhibitors have entered clinical trials. IT-901 was recently reported as an NF-κB inhibitor (acting through c-Rel and p65) and redox modulator, showing significant activity in mouse models of graft-versus-host disease (GvDH) and graft-versus-lymphoma (GvL), as well as in a xenograft model of humanB-cell lymphoma, mediating promising anti-lymphoma effects.23,24 The aim of this work is to determine the in vitro and in vivo effects of IT-901 in CLL and RS primary cells and derived line models.

Methods Cell lines and primary samples Leukemic cells were purified using Ficoll-Hypaque (SigmaAldrich, Milan, Italy) from peripheral blood (PB) of CLL patients or lymph node (LN) of RS patients presenting with typical morphology and immunophenotype.21 Samples were obtained at Weill Cornell Medicine after written informed consent in accordance with institutional guidelines and the Declaration of Helsinki. The referring physician provided molecular and genetic characterization of patients' samples. Normal circulating B cells were purified from healthy donors. Mec-1 and OSU-CLL CLL cell lines were obtained from the Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures and Ohio State University, respectively, and cultured in RPMI+10% fetal bovine serum (FBS). HS-5 stromal cells were obtained from ATCC and cultured in DMEM+10% FCS.

Metabolic assays Chronic lymphocytic leukemia cells were exposed to vehicle (0.02% DMSO in RPMI-1640, indicated as NT) or IT-901 (10 mM in the same solution as vehicle) for 6 hours (h), before dynamically measuring the metabolic profile using the XF96e Extracellular Flux Analyzer (Seahorse Bioscience, North Billerica, MA, USA). Cells (5x105 for primary cells and 105 for cell lines) were seeded in specialized tissue culture plates, coated with CellTak (BD Biosciences). An hour before measurement, cells were incubated at 37°C in a CO2-free atmosphere. Oxygen consumption rate (OCR), an indicator of mitochondrial respiration, was measured in basal conditions and following addition of specific drugs, oligomycin (1 mM), carbonilcyanide p-triflouromethoxyphenylhydrazone (FCCP, 1 mM) and Rotenone/Antimycin A (0.5 mM) able to interfere with different steps of the oxidative phosphorylation (OXPHOS) process (XF Cell Mito Stress test kit, Seahorse Bioscience). Maximal OCR and ATP production were measured. In all experiments, measurements were performed in quadruplicates.

In vivo experiments and treatments Mec-1 (5x105) cells were intravenously injected (i.v; tail vein) in 8-week old NOD/SCID/gamma chain–/– (NSG) mice and left to engraft for ten days before starting treatment. Mice received intra-peritoneal (i.p.) injection of IT-901 (15 mg/kg) or vehicle (Polyethene glycol-12 Glycerol-Dimyristate, GDM 4% in PBS). At the end of treatment, mice were euthanized, organs collected and partially dismantled to obtain single cell suspension or formalin-fixed for immunohistochemistry analyses. Mec-1 cell distribution in the different organs was analyzed by flow cytomehaematologica | 2017; 102(11)

try, after staining single cell suspensions with anti-humanCD19FITC and -CD45PerCP antibodies to identify leukemic cells. A different set of mice was monitored for survival.

Richter syndrome model Primary RS cells were obtained from PB or LN biopsies of clinically diagnosed RS patients. Purified cells (20x106) or LN fragments were injected sub-cutaneously (s.c., double flank) in 6-week old NSG mice and left to engraft. Tumor masses were then collected, partially dismantled and re-implanted in new animals for several passages to obtain a stable model of RS. Genetic stability and relationship to the original tumor was confirmed by exome sequencing (T Vaisitti and JN Allan, 2017, manuscript in preparation). After several passages in vivo, RS cells obtained from the tumor mass were put in culture over a stromal layer of HS-5 cells and were able to grow in vitro, maintaining the original genetic and phenotypic characteristics. RS-PDX models were established while TV and SD were Visiting Scientists at Weill Cornell Medicine. The Institutional Animal Care and Use Committee approved all the experiments involving mice. Additional details are provided in the Online Supplementary Appendix.

Results IT-901 blocks NF-κB activity in primary CLL cells and derived cell lines IT-901 is a recently described NF-κB inhibitor, with promising activity in mouse models of GvHD and lymphoma.23,24 In consideration of the high levels of constitutive NF-κB activation in CLL and in RS cells, the aim of this work is to provide pre-clinical evidence of the therapeutic effects of IT-901. In the experimental setting adopted, maximal NF-κB activation was obtained by culturing primary CLL cells for 6 h over a layer of HS-5, a stromal cell line with documented nurturing properties for CLL cells.25 By using an ELISA assay that measures NF-κB interactions with its DNA consensus sequences, we confirmed significant upregulation of p65 and p50 DNA-binding activities compared to the basal condition (Figure 1A). Conversely, addition of IT-901 during the co-culture period significantly decreased DNA binding of both subunits, with p65 being most sensitive (Figure 1A). In line with previous reports, IT-901 limited the DNA binding ability of c-Rel (Online Supplementary Figure S1A); however, in CLL cells this subunit is expressed and active at very low levels, suggesting that the main effects of the drug are through p65. Inhibition of NF-κB transcriptional activity was confirmed using nuclear extracts of Mec-1 and OSU-CLL, two CLL cell lines, where NF-κB activity is independent of microenvironmental conditions (Figure 1B and Online Supplementary Figure S1B). Diminished expression of p65 and p50 in nuclear and cytosolic fractions following IT-901 exposure was documented in primary cells and cell lines, suggesting that the drug induces degradation of the complex (Figure 1C, G and D, H, respectively). The same treatment also decreased the expression of the inhibitory subunit IκBα both in the phosphorylated and non-phosphorylated forms, in primary CLL cells (Figure 1E) and cell lines (Figure 1F), in line with the hypothesis that IT-901 interacts with the NF-κB-IκB cytoplasmic complex triggering its degradation. 1879

T. Vaisitti et al.

IT-901 induces mitochondrial reactive oxygen species accumulation Consistent with the documented cross-talk between NF-κB and ROS,26,27 a prominent and dose-dependent

increase in mitochondrial reactive oxygen species (mROS) was measured starting 6 h after treatment with IT-901, both in primary cells and in cell lines (Figure 2A and Online Supplementary Figure S2A, respectively).

Figure 1. IT-901 blocks nuclear factor-kappa B (NF-κB) activity in primary chronic lymphocytic leukemia (CLL) cells and derived cell line. (A) DNA binding activity of the p65 and p50 subunits in primary CLL cells (n=13) was analyzed using an ELISA kit, applying the same amount of nuclear extracts. Leukemic cells were co-cultured on a stromal layer (HS-5) to maximize the activation of the pathway, in the presence of increasing doses of drug or vehicle (NT) for 6 hours (h). (B) Cumulative DNA binding activity of NF-κB (p65 and p50) in 2 different cell line models of CLL, Mec-1 (n=5), and OSU-CLL (n=3). (C, D, G and H) Cytoplasmic (C) and nuclear (N) fractions obtained from primary CLL cells (C) or cell lines (D) cultured as indicated above were resolved by SDS-PAGE and blotted with specific antibodies to detect the expression of NF-κB1 subunits p105/p50 (arrow head) and RelA (p65). Lamin A/C and β-tubulin were used as nuclear and cytoplasmic markers, respectively. Nuclear p65 and p50 band intensities in primary CLL samples and CLL cell lines are reported in (G) and (H), respectively. (E and F) Total lysates from primary CLL cells (E) or CLL cell lines (F) cultured alone and exposed to different doses of IT-901 or vehicle (NT) for 6 h were resolved by SDS-PAGE and expression of the NF-κB complex analyzed using specific antibodies. Actin was used as a loading control. OD: optical density.

1880

haematologica | 2017; 102(11)

Targeting of NF-κB in CLL and RS cells

In principle, progressive mROS accumulation could be caused by increased production or impaired scavenging, due, at least in part, to the interruption of NF-κB-dependent transcription of genes coding for scavenging enzymes.26,27 The finding of a significantly decreased expression of the NF-κB-regulated catalase enzyme favored the hypothesis of a contribution of impaired scavenging to increased mROS (Figure 2B and Online Supplementary Figure S2B). Elevated mROS levels decreased mitochondrial membrane potential (Dfm), as shown by using the potentiometric probe JC-1, whose decrease in red fluorescence

depends solely on mitochondrial depolarization (Figure 2C). Similar results were obtained in CLL cell lines, with a dose-dependent loss of (Dfm), starting at the 2.5 mM dose (Online Supplementary Figure S2C). These results were confirmed using the cationic dye TMRM that binds to the inner membrane of intact mitochondria (Online Supplementary Figure S2D).

Mitochondrial respiration is severely compromised following IT-901 treatment Consistent with the finding of high levels of mROS in IT-901-treated CLL cells, mitochondrial energetic

Figure 2. IT-901 induces mitochondrial damage and compromises mitochondrial respiration. (A) Representative plots and cumulative data of mitochondrial reactive oxygen species (mROS) concentration in chronic lymphocytic leukemia (CLL) cells. Data are represented as fold change (FC) over the vehicle (n=10). (B) Box plot reporting the catalase (CAT) mRNA expression levels in vehicle (NT)- or IT-901-treated primary cells (n=10). (C) Representative plots and cumulative data of inner mitochondrial membrane potential (DΦm) in primary CLL cells (n=10) exposed to vehicle (NT) or increasing doses of IT-901 for 6 hours (h). CCCP was used as positive control. (D) Dynamic mitochondrial metabolic profile (OCR; pmoles/min) of a representative CLL patient treated with vehicle (red line) or IT-901 10 mM (blue line) for 6 h. Maximal respiration (calculated as: OCR after FCCP injection-late OCR measurement after RT/AA addition) and ATP production (calculated as last rate measurement of OCR before Oligo injection-minimum rate measurement after Oligo injection) in primary CLL patients (n=7). (E) Box plots reporting ATP-synthase (ATP5A1) and Cytochrome C Oxidase Assembly Protein (SCO2) mRNA expression levels in vehicle- or IT-901-treated primary cells (n=7). CCCP: carbonyl cyanide m-chlorophenyl hydrazone; Oligo: oligomycin; FCCP: carbonyl cyanide p-trifluoromethoxyphenylhydrazone; RT/AA: rotenone + antimycin A; OCR: oxygen consumption rate.

haematologica | 2017; 102(11)

1881

T. Vaisitti et al.

processes were severely compromised. By using the Seahorse metabolic analyzer, and in keeping with previous data,28 we observed that leukemic cells have a distinct preference for oxidative phosphorylation

(OXPHOS), with limited glycolytic capacity. After 6 h of culture with 10 mM IT-901, a dramatic drop in respiration was highlighted, leading to a marked decrease in ATP production (Figure 2D). In line with the hypothesis that

Figure 3. IT-901 rapidly induces apoptosis selectively in primary chronic lymphocytic leukemia (CLL) cells. (A) Representative plots and cumulative data of apoptotic staining with Annexin-V and propidium iodide (PI) of CLL patient cells (n=25) after 24-48 hours (h) of exposure to increasing doses of IT-901 or vehicle (NT). (B) Apoptotic data were plotted according to the mutational status of IgHV genes [mutated (M) vs. unmutated (UM)] and to the presence of specific genetic abnormalities (deletion 13/normal vs. deletion 17/mutation in TP53). (C) Total lysates from primary CLL cells cultured alone and exposed to different doses of IT-901 or vehicle (NT) for 6 h were resolved by SDS-PAGE and expression of pro- and anti-apoptotic proteins and activation of the caspase pathway analyzed using specific antibodies. Actin was used as a loading control. (D) Cumulative data of apoptosis obtained from Mec-1 (n=12) and OSU-CLL cell lines (n=9) at 24 h and both cell lines (n=21) at 48 h, exposed to increasing doses of IT-901. (E) Total lysates of CLL cell lines, cultured alone and exposed to different doses of IT-901 or vehicle (NT) for 6 h were resolved by SDS-PAGE and expression of pro- and anti-apoptotic proteins and activation of the caspase pathway analyzed using specific antibodies. Actin was used as a loading control. (F) Cumulative data of mROS and apoptosis of CLL cell lines (n=6) measured in the presence of N-acetyl-cysteine (NAC) and/or IT-901. (G) Cumulative data of apoptosis in normal B and T lymphocytes purified from 6 different healthy donors and exposed to increasing doses of IT-901 for the indicated time points. FL: full-length; CL: cleaved.

1882

haematologica | 2017; 102(11)

Targeting of NF-κB in CLL and RS cells

these metabolic alterations are due to the brusque interruption of NF-κB transcriptional circuits, as documented in other models,29 we found significant downmodulation of ATP5A1, the enzyme responsible for ATP synthesis, and of SCO2, a mitochondrial chaperone molecule that controls assembly of the cytochrome c oxidase subunit II of the electron transfer chain (Figure 2E). A similar behavior was measured in the cell lines, with a clear dose-dependent response to IT-901 (Online Supplementary Figure S2E and F).

IT-901 rapidly induces apoptosis in primary CLL cells, but not in normal B and T lymphocytes Induction of mROS with subsequent mitochondrial damage activated the apoptotic cascade in primary CLL cells from 25 different patients (Figure 3A). Apoptosis was time- and dose-dependent, starting at 5 mM after 24-h culture with IT-901 (Figure 3A). In the CLL cohort examined, the apoptotic response to IT-901 was apparently independent of the mutational status of the IgHV genes and of the presence of mutations or deletions in TP53, classically considered as markers of a more aggressive form of the disease (Figure 3B). Activation of the apoptotic cascade was confirmed by highlighting caspase 3 cleavage, downregulation of anti-apoptotic XIAP and upregulation of proapoptotic BIM proteins, evident at the 5 mM dose and peaking at 10 mM for primary cells (Figure 3C). These results were then confirmed in the two CLL cell line models, which were more sensitive to IT-901, in keeping with higher constitutive NF-κB activation (Figure 3D and E). Treatment of cell lines with the N-acetyl cysteine (NAC), a ROS scavenger, completely abrogated mROS accumulation and prevented apoptosis, clearly indicating that mROS are responsible for leukemic cell death (Figure 3F). In line with a markedly lower NF-κB expression and constitutive activation,10 PB, and B and T lymphocytes purified from healthy donors were less sensitive to apoptosis by IT-901 (Figure 3G). After 24h cultures at the 10 mM concentration, less than 10% (9.8±3.2) of normal B lymphocytes were dead, as opposed to 44% (44.7±17.7)

of primary CLL cells, indicating the presence of a therapeutic window.

NF-κB silencing recapitulates IT-901 functional effects To confirm that IT-901 effects on CLL cells were due to specific NF-κB inhibition, we silenced the p65 subunit of this complex. This choice was based on evidence of its high expression and critical role in CLL,3,13,14 also confirmed by our results. Lentiviral particles containing a control GFP-shRNA [Scramble (Scr)], as well as 4 different GFP-shRNA sequences targeting p65 were used to infect CLL cell lines. One of the shRNA p65 sequences (shp65A) significantly inhibited expression of p65 compared to the Scr sequence, as shown by western blot (Figure 4A). A second shRNA p65 sequence (shp65B) was used as an additional control because of its limited inhibition of expression of the target protein (Figure 4A). Infected cells analyzed for viability by flow cytometry at day 6 post infection, after staining for AnnexinV and PI, showed that silencing by shp65A significantly enhanced apoptosis of leukemic cells compared to cells infected with the Scr or shp65B sequences (Figure 4B). Consistently, expression of proteins related to the apoptotic pathway, such as XIAP and Caspase-3, was influenced by p65 silencing, confirming the results obtained in IT-901 treated primary CLL cells (Figure 4A).

IT-901 inhibits the nurturing properties of nurse-like cells (NLC) Given: i) the critical role played by the leukemic microenvironment in conditioning CLL cell survival and resistance to therapy; and ii) the central role of NF-κB in the interactions between CLL cells and stromal/myeloid components,30 we explored the effects of IT-901 on the stromal cell line HS-5 and on nurse-like cells (NLC), a population of alternatively activated macrophages typical of CLL patients.31 Exposure of both cell types to lipopolysaccharides (LPS) strongly activated NF-κB, as shown by nuclear translocation of the p65 subunit, which was almost completely abrogated by IT-901 treatment (Online Supplementary

Figure 4. Silencing of the p65 subunit confirms results obtained with IT-901. (A) chronic lymphocytic leukemia (CLL) cell lines were infected with lentiviral particles containing plasmids coding for an shRNA scramble (Scr) or shRNA p65 (A and B). Expression of the NF-κB complex was checked by western blot and p50 and p65 expression quantified over actin, showing a significant reduction for p65, while for p50 statistical significance was not reached. Silencing of p65 diminished also the expression of the anti-apoptotic protein XIAP and of the full-length (FL) Caspase3 protein. Actin was used as a loading control. (B) Apoptosis analysis by flow cytometry of cell lines infected with shRNA lentiviral particles (n=5).

haematologica | 2017; 102(11)

1883

T. Vaisitti et al.

Figure S3A), confirming that the drug is effective. We then co-cultured CLL cells on a layer of autologous NLC, documenting NF-κB activation in both cell types, which was prevented by IT-901 (Figure 5A). However, exposure of HS-5 or NLC to IT-901 did not significantly alter their viability (Figure 5B and Online Supplementary Figure S3B). Consistently, no IT-901-induced modulation of metabolic

genes, mROS and ROS scavengers could be highlighted, suggesting that CLL cells are uniquely sensitive to the drug, likely due to their basal metabolic conditions (Online Supplementary Figure S3C). NF-κB signaling in NLC induced transcription of genes encoding integrins, including ITGA4, ICAM1 and VCAM-1, and immune-modulatory molecules, such as

Figure 5. IT-901 inhibits the nurturing properties of nurse-like cells (NLC). (A) Confocal microscopy analysis of nuclear factor-kappa B (NF-κB) activation in NLC alone or co-cultured with autologous CLL cells in the presence of vehicle (+CLL) or IT-901 (+CLL+IT-901). Cumulative results of the nuclear p65 fluorescence intensity, in NLC, are reported in the graph. (B) Giemsa staining of vehicle- or IT-901-treated NLC. (C and D) mRNA expression level (C; n=8) and phenotypic analysis by flow cytometry (D) of molecules involved in cell-cell interactions (ITGA4, ICAM1 and VCAM1) and immune-modulatory pathway (CD274 and CD86) in NLC cultured as indicated above. (E) Cumulative data of apoptosis of primary leukemic cells co-cultured over HS-5 cells in the presence of increasing doses of IT901 for 24h. a.u.: arbitrary units.

1884

haematologica | 2017; 102(11)

Targeting of NF-κB in CLL and RS cells

CD274 and CD86.30 Consistent with the inhibitory effects of IT-901, expression of these target genes (Figure 5C) and derived proteins (Figure 5D) was inhibited at the 10 mM dose, indicating that IT-901 prevents the tumor-supportive phenotype of stromal or NLC cells. In line with this hypothesis, IT-901 was equally effective in inducing apoptosis in CLL cells cultured alone or over a layer of HS-5 or NLC, without measurable differences in degradation of the NF-κB complex, caspase 3 activation and modulation of apoptosis-related proteins (Figure 5E and Online Supplementary Figure S3D and E).

IT-901 co-operates with ibrutinib in inducing apoptosis of CLL cells We then asked whether IT-901 potentiated the effects of the btk inhibitor ibrutinib, selected because of its current use in the treatment of CLL patients, including the high-risk subgroups32 and considering that both drugs target the same signaling pathway.33 The combination of IT901 with ibrutinib markedly enhanced the effects of both drugs used alone, as determined both by treating CLL cells alone or after co-culturing leukemic cells on NLC with consequent protection from spontaneous apoptosis (Figure 6A and B). As an example, when cultured over NLC, treatment with ibrutinib 2.5 mM + IT-901 5 mM induced 48% (48±16.1) of the leukemic cells to undergo apoptosis after 48 h, compared to 36% (36±14.5) and 33% (33±17) when treating cells with ibrutinib or IT-901 alone, respectively (Figure 6B).34

IT-901 limits in vivo growth and spread of CLL cells We then tested whether treatment with IT-901 of mice xenografted with CLL cell lines induced an anti-tumor effect. To address this point, Mec-1 cells were injected in NSG mice,35 a model considered to be reproducible and instructive for therapeutic testing.36 Cells were injected in the tail vein of 8-week old mice, left to engraft for ten days, before beginning treatment with IT-901 (15 mg/kg, every day in PBS+GDM) or vehicle only. (A scheme of the

in vivo experiment is reported in Figure 7A). This experiment had two distinct end points: i) at the end of the treatment schedule, vehicle- and IT-901-treated animals were analyzed for tumor growth and leukemic cell infiltration of different target organs; ii) the remaining mice were analyzed for survival. Examination of the mice immediately after the last treatment revealed significant differences in tumor growth and localization. In agreement with previous data,35 vehicle-treated mice showed extensive Mec-1 colonization of the kidneys, significantly reduced in the IT-901-treated group (Figure 7B). Moreover, flow cytometry (Figure 7C) and immunohistochemistry (IHC) analyses, using an antihuman CD20 antibody (Figure 7D and Online Supplementary Figure S4A), indicated that IT-901 treatment significantly reduced the number of Mec-1 cells in liver, spleen and bone marrow (BM). In line with the specificity of action of IT-901, a significant decrease in total expression and nuclear localization of the p65 subunit was apparent in IT-901-treated mice (Figure 7D and Online Supplementary Figure S4B). Accordingly, Mec-1 cells obtained from this group of mice were characterized by diminished viability compared to cells from vehicle-treated mice (Figure 7E). Kaplan-Meier curves indicated that vehicle-treated mice were characterized by a significantly shorter survival (median: 31 days) when compared with those injected with IT-901 (median: 43 days; P<0.0001) (Figure 7F).

IT-901 is active on primary RS cells and in patient-derived xenograft We then focused on RS, typically a DLBCL developing in a minority of patients with a previous or concomitant diagnosis of CLL.37 Exposure of RS cells to increasing concentrations of IT901 induced apoptosis in a dose-dependent way. These cells were clearly more sensitive to the NF-κB inhibitor than CLL cells, independently of the presence of a protective stromal layer (Figure 8A). The mechanism behind increased apoptotic responses following IT-901 treatment

Figure 6. IT-901 co-operates with ibrutinib in inducing apoptosis of chronic lymphocytic leukemia (CLL) cells. Cumulative results of apoptosis of primary CLL cells (n=16) cultured alone (A) or over autologous nurse-like cells (NLC) (B) for 48 hours (h) in the presence of ibrutinib (2.5 mM) and IT-901 (5 and 10 mM) alone or in combination. Statistical significance of the combined effect of IT-901 and ibrutinib compared to single drugs alone was calculated according to the effect-based strategy, using the Highest Single Agent approach, as described by Foucquier and Guedj.34 *Over single box plot indicates statistical significance compared to untreated (NT).

haematologica | 2017; 102(11)

1885

T. Vaisitti et al.

was similar to the one highlighted for CLL cells, with decreased expression of the NF-ÎşB complex, including the inhibitory subunit. Concomitantly, there was the activation of the caspase pathway, with modulation of the expression of pro- and anti-apoptotic proteins (Figure 8B and Online Supplementary Figure S4C). To prove the efficacy of IT-901 in vivo, we then exploited 2 different patientderived xenograft (PDX) models, recently established in the lab. RS cells were s.c. injected in NSG mice and left to

engraft, until tumors became palpable. Mice were then randomized to receive vehicle or IT-901 treatment, with the same schedule adopted for the CLL xenografts. Results indicate that IT-901 significantly reduced tumor growth, as highlighted by size and weight of tumor masses (Figure 8C). Moreover, a significant reduction in p65 mRNA and protein levels were evident in IT-901 treated mice (Figure 8D and Online Supplementary Figure S4C). This diminished expression and activity was also confirmed by checking

1886

Figure 7. IT-901 limits in vivo growth and spread of chronic lymphocytic leukemia (CLL) cells. (A) Representative scheme of the in vivo model. Mec-1 cells were intravenously injected in tail vein of NSG mice, left to engraft for ten days before starting the treatment with IT-901 or vehicle. (B) Images of kidneys obtained from vehicleor IT-901-treated mice. (C) Mec-1 engraftment in different organs evaluated by flow cytometry after labeling of leukemic cells with anti-human-CD45 and -CD19 antibodies. Cumulative data of engraftment in kidneys, liver, spleen and bone marrow (BM) (n=8 different mice/group). (D) Immunohistochemical analyses and quantification of CD20 and nuclear p65 staining (reported as percentage of positive cells) in kidneys of vehicle- or IT-901-treated mice. (E) Viability of Mec-1 cells, purified from kidneys and liver of vehicle- or IT-901-treated mice, analyzed by flow cytometry (69% vs. 85% of viable cells in kidneys and 70% vs. 82% of viable cells in liver, respectively). (F) Kaplan-Meier curves showing survival of mice treated with IT-901 (n=8; red line; 43 days) compared to vehicle (n=8; blue line; 31 days). Engr: engraftment; Tx: treatment; Stop: drug holidays; Sac: euthanasia.

haematologica | 2017; 102(11)

Targeting of NF-κB in CLL and RS cells

localization of the NF-κB complex in cytoplasmic and nuclear extracts (Figure 8E). Together, these results provide a proof-of-principle that IT-901 is efficacious in CLL and RS cells.

Discussion This work studies the effects of IT-901, a novel inhibitor of the c-Rel/p65 NF-kB subunits in CLL and in RS. Our findings suggest that IT-901 induces cell death in both cell types, with tumor cells being significantly more sensitive to the drug than normal circulating B cells. Even if the mechanism of action of this novel inhibitor is still partially unclear, some conclusions can now be drawn. Firstly, our data indicate that treatment with IT901 reduces NF-κB binding to its consensus DNA sequences. This effect is particularly evident when considering p65, which is the most active subunit in CLL cells.

Biochemical analysis of the complex, however, reveals global degradation of all the subunits, including p50 and the inhibitory subunit IκB, suggesting that IT-901 binds to the complex in the cytosol, causing its degradation. Secondly, IT-901 exposure is followed by an increase of mROS, which become toxic after a few hours. In principle, increased mROS could be due to active production or impaired degradation. On the basis of RT-PCR data indicating that IT-901 decreases expression of genes favoring ROS scavenging, including catalase, impaired ROS degradation seems to be the main mechanism. According to our working model, IT-901 would bind the NF-κB complex in the cytosol, interrupting its cycle, which is constitutively active in neoplastic B cells, and hence blocking transcription of NF-κB-controlled genes. This would limit the cells' ability to scavenge mROS causing their accumulation over time, eventually leading over 6 h to mitochondrial damage, highlighted by the finding of altered membrane polarity and of severely impaired mitochondrial respiration

haematologica | 2017; 102(11)

Figure 8. IT-901 is active on primary Richter syndrome (RS) cells and in patient-derived xenograft (PDX). (A) Cumulative data of apoptosis of primary or PDX-tumor-derived RS cells. (B) Western blot analysis of the expression of the nuclear factor-kappa B (NF-κB) complex, pro- and anti-apoptotic proteins and caspase3 in RS cells exposed to the indicated doses of IT-901 for 6 hours (h). Actin was used as loading control. (C) Tumor masses from vehicle- or IT-901-treated RS-PDX mice compared for tumor volume (cm3) and weight (g) (6 mice/group; double-flank injected). (D) Immunohistochemistry analysis of p65 expression within the tumor mass. Staining was reported as percentage of positive cells. (E) Cytoplasmic (C) and nuclear (N) fractions obtained from RS cells purified from the tumor mass were resolved by SDS-PAGE and expression of the NF-κB complex analyzed. βtubulin and Lamin A/C were used as cytoplasmic and nuclear controls, respectively. FL: full-length; CL: cleaved.

1887

T. Vaisitti et al.

which in turn decrease ATP levels. The final result for the leukemic cell is the cleavage of caspase-3, resulting in induction of intrinsic apoptosis. The observation that the ROS scavenger N-acetyl cysteine fully prevents the apoptotic effects of IT-901 in cell lines confirms the central role of ROS in the process. This behavior in the presence of IT-901 seems to be specific for CLL cells as normal B lymphocytes need significantly higher doses to undergo apoptosis, while NLC, as well as stromal cells do not go into apoptosis with IT-901. The reason behind this different sensitivity to IT-901 remains unclear, even if it is tempting to speculate a connection to the basal metabolic state of the cell and to basal levels of mROS. According to this hypothesis, apoptotic responses to IT-901 would be specific to those cells that have constitutively high levels of mROS, with rapid toxicity following the interruption of genetic circuits regulating expression of the scavengers. On the contrary, NLC, which are essentially tumor-associated macrophages, possess a radically different basal metabolism, with low levels of mROS. In this context, the genes most affected by NFκB inhibition are those coding for surface molecules important for the crosstalk with tumor cells. These findings suggest that IT-901 treatment could be beneficial by acting through different mechanisms. On the one hand, IT-901 targets leukemic cells directly, leading to rapid toxicity and death by apoptosis. On the other hand, it perturbs molecular circuits driven by the leukemic cells and aimed at creating favorable growth conditions. The finding that PD-L1 expression is modulated when leukemic cells interact with NLC further suggests that IT901 helps restore an immunocompetent host. The third finding favoring further investigation of the therapeutic properties of IT-901 derives from the observation of a combined effect when administered together with the btk inhibitor ibrutinib. While this observation can be explained on the basis of the convergence of the two drugs on the same signaling pathway, it is interesting to note that addition of ibrutinib renders CLL cells sensitive to lower doses of IT-901, broadening its therapeutic window. Using an established xenograft model based on i.v.

References 1. Karin M, Lin A. NF-kappaB at the crossroads of life and death. Nat Immunol. 2002;3(3):221-227. 2. Perkins ND, Gilmore TD. Good cop, bad cop: the different faces of NF-kappaB. Cell Death Differ. 2006;13(5):759-772. 3. Oeckinghaus A, Ghosh S. The NF-kappaB family of transcription factors and its regulation. Cold Spring Harb Perspect Biol. 2009;1(4):a000034. 4. Hayden MS, Ghosh S. Shared principles in NF-kappaB signaling. Cell. 2008; 132(3):344-362. 5. Vallabhapurapu S, Karin M. Regulation and function of NF-kappaB transcription factors in the immune system. Annu Rev Immunol. 2009;27:693-733. 6. Perkins ND. NF-kappaB: tumor promoter or suppressor? Trends Cell Biol. 2004; 14(2):64-69. 7. Bradford JW, Baldwin AS. IKK/nuclear factor-kappaB and oncogenesis: roles in

1888

10.

11.

12.

injection of Mec-1 cells into NSG mice, we showed that IT-901 administered as a single agent for two weeks after documented disease engraftment significantly reduces tumor burden in all the organs examined. Importantly, a marked reduction in the nuclear expression of p65 was observed, confirming NF-κB targeting in vivo. Lastly, we investigated whether IT-901 is effective also in experimental models of RS. This is an important issue as RS remains a disease for which there is an urgent need for active drugs. One of the difficulties in studying RS is the lack of experimental models as there are no available cell lines. For this reason, we exploited two different patient-derived xenografts established from patients with RS (T Vaisitti and JN Allan, 2017, manuscript in preparation). Subsequent adoptive transfer of the lymphoma, which remained genetically similar to patients' cells, confirmed successful disease engraftment. By exploiting these models, we first demonstrated inhibition of NF-κB activation after IT-901 treatment, with rapid mROS accumulation, mitochondrial damage and death by apoptosis. Furthermore, the drug was used in monotherapy to treat xenografts obtained from RS patients. We found that IT901 markedly reduced tumor weight and volume and inhibited NF-kB expression, suggesting a similar mechanism of action. In conclusion, this work offers pre-clinical evidence supporting the use of IT-901 in the treatment of CLL and RS patients. Acknowledgments We thank K. Gizzi (Italian Institute for Genomic Medicine) for excellent technical support. Funding This work is supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC) (IG-17314 to SD), by the Italian Ministry of Health (Bando Giovani Ricercatori GR-201102349282 to TV and GR-2011-02346826 to SD), by the Leukemia and Lymphoma Society (6465-15 to JLZ), and by ImmuneTarget Inc.. IT-901 was provided by ImmuneTarget Inc.

tumor-initiating cells and in the tumor microenvironment. Adv Cancer Res. 2014; 121:125-145. Gasparini C, Celeghini C, Monasta L, Zauli G. NF-kappaB pathways in hematological malignancies. Cell Mol Life Sci. 2014; 71(11):2083-2102. Cuni S, Perez-Aciego P, Perez-Chacon G, et al. A sustained activation of PI3K/NFkappaB pathway is critical for the survival of chronic lymphocytic leukemia B cells. Leukemia. 2004;18(8):1391-1400. Furman RR, Asgary Z, Mascarenhas JO, Liou HC, Schattner EJ. Modulation of NFkappa B activity and apoptosis in chronic lymphocytic leukemia B cells. J Immunol. 2000;164(4):2200-2206. Lin Y, Bai L, Chen W, Xu S. The NF-kappaB activation pathways, emerging molecular targets for cancer prevention and therapy. Expert Opin Ther Targets. 2010;14(1):4555. Tracey L, Perez-Rosado A, Artiga MJ, et al. Expression of the NF-kappaB targets BCL2

13.

14.

15.

16.

17.

and BIRC5/Survivin characterizes small Bcell and aggressive B-cell lymphomas, respectively. J Pathol. 2005;206(2):123-134. Hewamana S, Alghazal S, Lin TT, et al. The NF-kappaB subunit Rel A is associated with in vitro survival and clinical disease progression in chronic lymphocytic leukemia and represents a promising therapeutic target. Blood. 2008;111(9):4681-4689. Hewamana S, Lin TT, Rowntree C, et al. Rel a is an independent biomarker of clinical outcome in chronic lymphocytic leukemia. J Clin Oncol. 2009;27(5):763-769. Sutton LA, Rosenquist R. The complex interplay between cell-intrinsic and cellextrinsic factors driving the evolution of chronic lymphocytic leukemia. Semin Cancer Biol. 2015;34:22-35. Bernal A, Pastore RD, Asgary Z, et al. Survival of leukemic B cells promoted by engagement of the antigen receptor. Blood. 2001;98(10):3050-3057. Barragan M, Bellosillo B, Campas C, Colomer D, Pons G, Gil J. Involvement of

haematologica | 2017; 102(11)

Targeting of NF-ÎşB in CLL and RS cells

18.

19.

20. 21. 22.

23.

24.

protein kinase C and phosphatidylinositol 3-kinase pathways in the survival of B-cell chronic lymphocytic leukemia cells. Blood. 2002;99(8):2969-2976. Zaninoni A, Imperiali FG, Pasquini C, Zanella A, Barcellini W. Cytokine modulation of nuclear factor-kappaB activity in Bchronic lymphocytic leukemia. Exp Hematol. 2003;31(3):185-190. Petlickovski A, Laurenti L, Li X, et al. Sustained signaling through the B-cell receptor induces Mcl-1 and promotes survival of chronic lymphocytic leukemia B cells. Blood. 2005;105(12):4820-4827. Tsimberidou AM, Keating MJ. Richter syndrome: biology, incidence, and therapeutic strategies. Cancer. 2005;103(2):216-228. Rossi D, Gaidano G. Richter syndrome: pathogenesis and management. Semin Oncol. 2016;43(2):311-319. Fabbri G, Khiabanian H, Holmes AB, et al. Genetic lesions associated with chronic lymphocytic leukemia transformation to Richter syndrome. J Exp Med. 2013; 210(11):2273-2288. Shono Y, Tuckett AZ, Ouk S, et al. A smallmolecule c-Rel inhibitor reduces alloactivation of T cells without compromising antitumor activity. Cancer Discov. 2014; 4(5):578-591. Shono Y, Tuckett AZ, Liou HC, et al. Characterization of a c-Rel Inhibitor That Mediates Anticancer Properties in

haematologica | 2017; 102(11)

25.

26. 27.

28.

29.

30.

Hematologic Malignancies by Blocking NFkappaB-Controlled Oxidative Stress Responses. Cancer Res. 2016;76(2):377-389. Seiffert M, Stilgenbauer S, Dohner H, Lichter P. Efficient nucleofection of primary human B cells and B-CLL cells induces apoptosis, which depends on the microenvironment and on the structure of transfected nucleic acids. Leukemia. 2007; 21(9):1977-1983. Morgan MJ, Liu ZG. Crosstalk of reactive oxygen species and NF-kappaB signaling. Cell Res. 2011;21(1):103-115. Nakano H, Nakajima A, Sakon-Komazawa S, Piao JH, Xue X, Okumura K. Reactive oxygen species mediate crosstalk between NF-kappaB and JNK. Cell Death Differ. 2006;13(5):730-737. Jitschin R, Hofmann AD, Bruns H, et al. Mitochondrial metabolism contributes to oxidative stress and reveals therapeutic targets in chronic lymphocytic leukemia. Blood. 2014;123(17):2663-2672. Mauro C, Leow SC, Anso E, et al. NFkappaB controls energy homeostasis and metabolic adaptation by upregulating mitochondrial respiration. Nat Cell Biol. 2011;13(10):1272-1279. Lutzny G, Kocher T, Schmidt-Supprian M, et al. Protein kinase c-beta-dependent activation of NF-kappaB in stromal cells is indispensable for the survival of chronic lymphocytic leukemia B cells in vivo.

Cancer Cell. 2013;23(1):77-92. 31. Burger JA, Tsukada N, Burger M, Zvaifler NJ, Dell'Aquila M, Kipps TJ. Blood-derived nurse-like cells protect chronic lymphocytic leukemia B cells from spontaneous apoptosis through stromal cell-derived factor-1. Blood. 2000;96(8):2655-2663. 32. Jain P, Keating MJ, Wierda W, et al. Long term follow up of treatment with ibrutinib and rituximab (IR) in patients with highrisk Chronic Lymphocytic Leukemia (CLL). Clin Cancer Res. 2017;23(9):2154-2158. 33. Wiestner A. BCR pathway inhibition as therapy for chronic lymphocytic leukemia and lymphoplasmacytic lymphoma. Hematology Am Soc Hematol Educ Program. 2014;2014(1):125-134. 34. Foucquier J, Guedj M. Analysis of drug combinations: current methodological landscape. Pharmacol Res Perspect. 2015;3(3):e00149. 35. Vaisitti T, Audrito V, Serra S, et al. The enzymatic activities of CD38 enhance CLL growth and trafficking: implications for therapeutic targeting. Leukemia. 2015;29 (2):356-368. 36. Bertilaccio MT, Scielzo C, Simonetti G, et al. Xenograft models of chronic lymphocytic leukemia: problems, pitfalls and future directions. Leukemia. 2013;27(3):534-540. 37. Rossi D. Richter's syndrome: Novel and promising therapeutic alternatives. Best Pract Res Clin Haematol. 2016;29(1):30-39.

1889

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

SYK inhibition thwarts the BAFF - B-cell receptor crosstalk and thereby antagonizes Mcl-1 in chronic lymphocytic leukemia

Cody Paiva,1 Taylor A. Rowland,1 Bhargava Sreekantham,1 Claire Godbersen,2 Scott R. Best,1 Prabhjot Kaur,2 Marc M. Loriaux,1 Stephen E.F. Spurgeon,1 Olga V. Danilova3 and Alexey V. Danilov1

Haematologica 2017 Volume 102(11):1890-1900

1 Knight Cancer Institute, Oregon Health and Science University, Portland, OR; 2Geisel School of Medicine at Dartmouth, Hanover, NH and 3Pathology, Portland VA Medical Center, Portland, OR, USA

ABSTRACT

Correspondence: danilov@ohsu.edu

Received: April 7, 2017. Accepted: August 18, 2017. Pre-published: August 24, 2017.

lthough small molecule inhibitors of B-cell receptor-associated kinases have revolutionized therapy in chronic lymphocytic leukemia (CLL), responses are incomplete. Pro-survival signaling emanating from the microenvironment may foster therapeutic resistance of the malignant B cells resident in the protective lymphoid niches. B-cell activating factor (BAFF) is critical to the survival of both healthy and neoplastic B cells. However, the pro-survival pathways triggered by BAFF have not been fully characterized. Here we show that BAFF elicited resistance to spontaneous and drug-induced apoptosis in stromal co-cultures, induced activation of both canonical and non-canonical NFκB signaling pathways, and triggered B-cell receptor signaling in CLL cells, independently of IGHV mutational status. SYK, a proximal kinase in the B-cell receptor signaling cascade, acted via STAT3 to bolster transcription of the anti-apoptotic protein Mcl-1, thereby contributing to apoptosis resistance in BAFF-stimulated cells. SYK inhibitor entospletinib downregulated Mcl-1, abrogating BAFF-mediated cell survival. BAFF-B-cell receptor crosstalk in neoplastic B cells was mediated by SYK interaction with TRAF2/TRAF3 complex. Thus, SYK inhibition is a promising therapeutic strategy uniquely poised to antagonize crosstalk between BAFF and B-cell receptor, thereby disrupting the pro-survival microenvironment signaling in chronic lymphocytic leukemia.

Introduction doi:10.3324/haematol.2017.170571 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1890 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1890

Soluble mediators derived from mesenchymal stromal cells, nurse-like cells, dendritic cells and T cells present in the protective niches (lymph nodes and bone marrow) prolong survival of neoplastic B cells in chronic lymphocytic leukemia (CLL).13 Lymph node-resident CLL cells exhibit gene signatures indicating activation of the B-cell receptor (BCR) and nuclear factor-κB (NFκB) pathways.4 Novel inhibitors of the BCR-associated kinases (BCRi) have made a significant clinical impact in CLL in part via induction of B-cell egress from niches wherein stromal support is lost. Ibrutinib and idelalisib, small molecule inhibitors of Bruton's tyrosine kinase (BTK) and phosphoinositide 3-kinase-δ (PI3K-δ), respectively, have improved outcomes in CLL.5 However, patients who progress on, or who are intolerant of BCRi therapy have poor outcomes.6,7 Improved understanding of microenvironment signaling will foster development of novel effective therapeutic approaches in CLL. Tumor necrosis factor receptor (TNFR) superfamily ligands, CD40L and BAFF/APRIL (B-cell activating factor/A proliferation-inducing ligand), are ubiquitously secreted in the stromal niches and promote fitness of the neoplastic clone.2 BAFF/APRIL ligands and their receptors are indispensable in B-cell survival.8-11 BAFF/APRIL share homology and are able to bind two TNFR - BCMA (B-cell maturation antigen) and TACI (transmembrane activator of the calcium modulator and cyclophilin ligand-interactor), whereas BAFF alone can bind BAFF receptor (BAFF-R, BR3).12 Like other TNFR ligands, BAFF/APRIL activate NFκB signaling, a major comhaematologica | 2017; 102(11)

SYK inhibition disrupts BAFF-BCR crosstalk in CLL

mon pathway which mediates anti-apoptotic responses in CLL cells through induction of Bcl-2 family proteins and chemokine networks.12-16 Both signal through BCMA/TACI to activate the canonical NFκB in CLL, where the IκB kinase complex phosphorylates IκB, triggering its ubiquitination and leading to nuclear translocation of the NFκB dimers, predominantly p50/RelA and p50/cRel.8,13 Meanwhile, BAFF-R/BR3 signals through an intermediary complex, which involves adaptor proteins TRAF2/TRAF3, NFκB-inducing kinase (NIK), and inhibitor of apoptosis (IAP) family proteins cIAP1/2.12 While the exact mechanism remains elusive, it is believed that, in unstimulated B cells, NIK is constitutively bound to TRAF3 and degraded. When BAFF engages BR3, the NIK/TRAF/cIAP complex is recruited to the receptor, followed by TRAF3 repression, thus allowing NIK to persist and activate IκB kinase-1 (IKK1). IKK1 catalyzes proteasome-assisted processing of NFκB2 (p100) precursor, thereby inducing the non-canonical (alternative) NFκB pathway.12 Despite significant progress in understanding the role of BAFF/APRIL signaling in healthy and neoplastic B cells, the role of BAFF-mediated NFκB activation in CLL has not been thoroughly studied. Furthermore, the mechanistic implications of targeting BCR signaling using novel BCRi have not been elucidated in this context. Here we explored the mechanistic underpinnings of CLL cell survival in response to BAFF signaling, uncovering the functional significance of the BCR-associated kinases and the pro-survival Bcl-2 family proteins in this setting.

ly described.16 Briefly, stromal cells were seeded to achieve 80100% confluence; on the following day, CLL cells were plated at a 50:1 ratio and incubated at 37°C in 5% CO2. Cultures were then treated with drugs as indicated. At harvest, CLL cells were gently washed off the stromal layer. When harvested for protein and mRNA analysis, CLL cells were transferred to a new plate and incubated for an additional 60 minutes (min) to allow re-attachment of stromal cells, thus minimizing contamination of CLL cells. Chronic lymphocytic leukemia cell apoptosis was quantified using the ApoScreen Annexin-V Apoptosis Staining Kit (Southern Biotech) in the CD19+ population, as previously described.16 Expression of BAFF ligand and receptors in paired peripheral blood-bone marrow samples and in BAFF-expressing stromal cells was quantified by flow cytometry using the following antibodies: CD257(BAFF)-PE (Clone 1D6, eBioScience, San Diego, CA, USA), CD256(APRIL)-PE, CD267(TACI)-PE, CD269(BCMA)-PE, CD268(BR3)-VioBlue (Miltenyi Biotech, San Diego, CA, USA). Chemotaxis assays across polycarbonate Transwell inserts were performed as described.18 Briefly, CLL cells (107/mL) were incubated at 37°C in 5% CO2 with 5 mg/mL of anti-IgM or 25 ng/mL BAFF with or without drugs. After 1 hour (h), CLL cells were washed and 100 mL of cell suspension (106 cells) was added to the top chamber of a Transwell culture insert (Corning) with a diameter of 6.5 mm and a pore size of 5 mm. Filters were then transferred to wells containing serum-free medium with or without 200 ng/mL CXCL12 (Cell Signaling). After a 3-h incubation, the cells in the lower chamber were aliquoted for counting by flow cytometry for 20 seconds in duplicates. A 1:20 dilution of input cells was counted under the same conditions.

Statistical analysis Methods Patients’ samples and cell culture Peripheral blood and bone marrow (where applicable) were obtained from patients with CLL at the Center for Hematologic Malignancies at the Oregon Health and Science University (Portland, OR, USA) after informed consent following approval by the Institutional Review Board (IRB#4422). Mononuclear cells were isolated using standard Ficoll-Hypaque techniques (Amersham, Piscataway, NJ, USA), rendering more than 90% CD5+/CD19+ cells, as determined by flow cytometry (FACSCanto). CLL cells were cultured in RPMI-1640 supplemented with 15% fetal bovine serum, 100 U/mL penicillin, 100 mg/mL streptomycin, 2 mM L-glutamine, 25 mM HEPES, 100 mM nonessential amino acids and 1 mM sodium pyruvate (Life Technologies, Grand Island, NY, USA). For stimulation with soluble factors, CLL cells were seeded at 1x106/mL in the presence of 5 mg/mL soluble goat F(ab’)2 anti-human IgM antibody (sol-IgM; Southern Biotech, Birmingham, AL, USA) or 25 ng/mL soluble human BAFF (sol-BAFF; Cell Signaling Technology, Danvers, MA, USA). CLL samples were analyzed for IGHV mutations using the IGH Somatic Hypermutation Assay v.2.0 (Invivoscribe, San Diego, CA, USA), as previously described.16 BAFF-expressing Chinese hamster ovary cells (BAFF-CHO) were obtained from Dr. Robert Woodland (University of Massachusetts, Worcester, MA, USA).17 Those cells were maintained in MEM-α supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 mg/mL streptomycin, and 1 mM nonessential amino acids. CHO-K1 cells not expressing BAFF were used as control [American Type Culture Collection (ATCC), Manassas, VA, USA]. Chronic lymphocytic leukemia cells were cultured on BAFFexpressing (or control) cells under the stromal conditions previoushaematologica | 2017; 102(11)

Paired or unpaired Student t-tests were performed in GraphPad Prism software (La Jolla, CA, USA). P<0.05 was considered statistically significant. Microarray data were analyzed for functional significance using Pathway Studio software (Ariadne Genomics/Elsevier, Rockville, MD, USA). Data are presented as mean±Standard Error (SE) throughout the manuscript.

Results BAFF predominantly activates non-canonical NFκB and up-regulates the pro-survival Bcl-2 proteins in CLL cells First, we analyzed expression of BAFF ligands and receptors in CLL cells sourced from peripheral blood and the bone marrow. BAFF receptors BR3, BCMA and TACI, and their ligands BAFF and APRIL were expressed in both compartments (Online Supplementary Figure S1). In our previous work, we had partially reconstituted the lymph node microenvironment by employing co-cultures of the primary CLL cells with CD40L-expressing fibroblasts. In this model, neoplastic B cells exhibited induction of NFκB pathways and Bcl-2 proteins, accompanied by protection from both spontaneous and drug-induced apoptosis, thereby partially replicating the resistant stromal niche.16 Here, to determine the effect of BAFF ligand on NFκB activation in primary CLL cells, we established an in vitro model where primary CLL cells were cultured in the presence of engineered BAFF-expressing CHO cells, as described in the Methods section (Online Supplementary Figure S2). In this model, cells were subjected to continuous stimulation by BAFF cytokine. BAFF-stimulated CLL cells were rescued from spontaneous apoptosis (12.3±3.2% cell apoptosis after 24-h incubation), com1891

C. Paiva et al.

pared to cells cultured off stroma (34.8±6.2%) or on control stroma (22.4±2.4%), and were resistant to chemotherapy-induced apoptosis (Figure 1A). We then employed gene expression profiling to determine the pathways induced by BAFF in CLL. Of the genes incorporated in the probe set, 7254 were expressed in CLL. Of these, 4844 were differentially expressed in

response to BAFF stimulation (P<0.01). Using a cut-off value of at least 1.5-fold change, we identified 2251 genes whose expression was significantly modulated by BAFF stimulation (P<0.01) (Figure 1B). We determined that receptor signaling and expression target pathways involving NFκB were most significantly associated with the upregulated genes (P<0.0001). Importantly, of the more than

Figure 1. B-cell activating factor (BAFF) promotes activation of pro-survival pathways in chronic lymphocytic leukemia (CLL) cells. (A) CLL cells (n=10) were cultured on BAFF-expressing or parental cells for 24 hours (h), followed by incubation with the indicated drugs, or vehicle control (ctrl), for an additional 24 h. As a reference, cells were treated off stroma. Apoptosis within CD19+ subset of cells was determined by Annexin V and 7-AAD staining. Data are presented as mean±Standard Error (SE). *P<0.05 compared to 'off stroma'. (B and C) CLL cells from 3 individual samples were co-cultured with BAFF-expressing stroma for 24 h. RNA was isolated from the purified CLL B cells and microarray analysis was performed as described in the Methods section. Heatmap shows hierarchical clustering of expression profiles of the 122 differentially expressed NFκB target genes (yellow: upregulation; blue: downregulation). (D) CLL cells from 3 individual patients were co-cultured with BAFFexpressing stroma for 4-24 h. Whole-cell protein lysates were subjected to immunoblotting. (E) CLL cells (n=4) were co-cultured with control, BAFF- or CD40L-expressing stroma for 24 h. p52 and p65/RelA activity was determined in whole-cell protein lysates using the TransAM NFκB activity assay. The dotted line represents activity measured in freshly isolated cells (at 0 h), which has been set at 1 (*P<0.05, **P<0.01 compared against that). (F) CLL cells (n=4) were co-cultured under the indicated conditions for 24 h. Whole-cell protein lysates were subjected to immunoblotting.

1892

haematologica | 2017; 102(11)

SYK inhibition disrupts BAFF-BCR crosstalk in CLL

400 known NFκB transcriptional targets, 181 genes were expressed in CLL and 122 were modulated by BAFF (P<0.01), among which 70 showed at least 50% induction, including anti-apoptotic genes and chemokines (Figure 1C). We confirmed up-regulated transcription of several NFκB target genes (Online Supplementary Figure S3A). We studied BAFF-mediated NFκB activation in additional detail. BAFF led to strong upregulation of the noncanonical NFκB pathway in CLL (Figure 1D-F). Processing of the non-canonical precursor protein NFκB2 (p100) occurred as early as 4 h after co-culture with BAFFexpressing stroma, and further increased by 24 h (Figure 1D). In agreement with this finding, BAFF led to a 4.5-fold increase in non-canonical (p52) activity in a DNA-binding ELISA assay, comparable with the effects of CD40L (Figure 1E). By contrast, the canonical NFκB pathway was less prominently induced by BAFF. Phosphorylation of IκBα, a negative pathway regulator, was detectable after 8 h (Figure 1D). RelA/p65 DNA binding was enhanced compared with cells cultured on control stroma, but was less pronounced compared with CD40L stimulation (Figure 1E). Deregulation of balance between the pro- and antiapoptotic Bcl-2 family members determines cell fate in response to TNFR signaling.15,16,19 We found that the antiapoptotic proteins Bcl-xL and Mcl-1 were induced in CLL cells in response to BAFF stimulation, while Bcl-2 expression was unchanged (Figure 1D). While CD40L predominantly up-regulated Bcl-xL (Figure 1F),20 BAFF mostly induced Mcl-1 (Figure 1D and Online Supplementary Figure S3B). It has been previously reported that focal CD40L expression by T cells may be restricted to lymphoid proliferation centers where it contributes to strong NFκB activation,21 while BAFF-R is ubiquitously expressed across many B-cell malignancies including CLL.22,23 Consistent with this, and together with our in vitro results, we found that, whereas Mcl-1 was diffusely expressed in CLL lymphatic tissue, Bcl-xL staining was scattered (Figure 2). Thus, in BAFF-expressing CLL-stromal cell co-cultures, non-canonical NFκB and Mcl-1 are induced to a greater extent than the canonical NFκB and Bcl-xL, a bona fide NFκB target.

that BAFF induced rapid activation of the proximal kinases SYK and LYN at the phosphorylation sites relevant to BCR signaling, which was sustained for 4 h in some samples (Figure 3A). IGHV mutational status is a strong determinant of response to BCR stimulation in CLL.27,29 We found that phosphorylation of the proximal BCR-associated kinases SYK, LYN, and BTK in response to BAFF ligation was variable between samples. However, their activation did not correlate with IGHV mutational status, with 3 of 4 unmutated and 4 of 6 mutated samples exhibiting SYK activation (Figure 3B and C). When present, ERK activation in response to BAFF occurred after 15 min, and waned by 120 min, mirroring IgM-mediated effects. By contrast, AKT phosphorylation occurred later (>2 h following stimulation), suggesting that BAFF activates AKT independent of BCR. BCR activation modulates cytokine synthesis, and CLL cell adhesion and migration.30 Consistent with earlier reports,18 we found that IgM crosslinking enhanced CLL migration toward CXCL12 2-fold, and this was partially inhibited in the presence of idelalisib, a PI3K-δ inhibitor (Figure 3D). Interestingly, we found that BAFF also induced CLL cell chemotaxis, suggesting that BAFF signaling may play a role in CLL cell homing. Finally, in our gene profiling experiments reported above, we found that of the 157 genes involved in the BCR pathway, 63 were significantly modulated by BAFF (Online Supplementary Figure S4).

BAFF induces activation of BCR-associated kinases independent of IGHV mutational status in CLL The mechanisms underlying canonical NFκB activation in BAFF-stimulated B cells are not well understood. In CLL and other B-cell neoplasia, BCR ligation has previously been implicated as a trigger of NFκB signaling.24 Therefore, and also because Mcl-1 is a recognized target of the BCR signaling cascade,25 we aimed to determine if BAFF coopts BCR to activate NFκB and induce Mcl-1 in CLL. In the short-term experiments which follow, we eliminated the effects of CLL-stromal contact and stromal-conditioned media by using soluble BAFF, and compared our findings against IgM crosslinking. Consistent with previous reports, we found that SYK was constitutively phosphorylated in a subset of CLL samples (data not shown).26 BCR engagement is known to lead to a rapid increase in phosphorylation of the BCRassociated kinases in CLL cells.26,27 While stimulation with soluble IgM leads to transient signaling, immobilized IgM leads to sustained activation of BCR kinases (>1 h), accompanied by enhanced CLL-cell survival.26,28 We found haematologica | 2017; 102(11)

Figure 2. Expression of Mcl-1 and Bcl-xL in chronic lymphocytic leukemia (CLL) lymph nodes. Lymphatic tissue from patients with CLL (n=10) were subjected to immunocytochemistry for Mcl-1 (A) and Bcl-xL (B), as described in the Methods section (40x).

1893

C. Paiva et al.

In summary, BAFF stimulation co-opts BCR signaling in CLL cells, independently of their IGHV mutational status, thereby contributing to CLL cell survival.

SYK inhibition thwarts BAFF-mediated survival in CLL by targeting Mcl-1 Given the co-operation between BAFF and BCR signaling, we explored the effects of targeting BCR signaling on BAFF-mediated events. CLL cells were cultured on BAFFexpressing stroma for 24 h and then exposed to BCRi. SYK inhibitors entospletinib (GS-9973) and R406 effectively antagonized survival of BAFF-stimulated CLL cells, while pharmacological inhibitors of BTK and PI3K were slightly less active (Figure 4A and Online Supplementary Figure S5). Entospletinib and idelalisib partially abrogated chemotaxis of BAFF-stimulated CLL cells (Figure 4B), echoing observations made in a setting of IgM crosslinking.31,32 As with IgM stimulation, the anti-chemotactic effect of BCRi was likely dependent on their ability to abrogate the autocrine secretion of chemokines, many of which are NFκB transcriptional targets.18,33 Next, we studied the mechanism of SYK inhibitioninduced apoptosis in this setting. It has been previously

shown that ibrutinib down-modulates NFκB signaling in CLL in vivo.34 Thus, we hypothesized that targeting kinases within the BCR signaling cascade would antagonize NFκB activation in BAFF-stimulated CLL cells. We found that all tested BCRi tested abrogated the canonical, but not the non-canonical NFκB (Figure 4C and Online Supplementary Figure S6A). By contrast, pevonedistat, an inhibitor of NEDD8-activating enzyme previously shown by us to disrupt NFκB activity in CD40-stimulated CLL cells,16 abrogated both NFκB pathways and induced apoptosis in BAFF-stimulated CLL cells (Online Supplementary Figure S6A-C). Since SYK is the key molecule in BCR activation, we further explored how SYK inhibition could counter BAFF-mediated CLL cell survival. SYK is involved in regulation of Mcl-1 in healthy and neoplastic B cells.25 Mcl-1 is a protein with a short half-life (approx. 30 min), whose expression is influenced by many regulatory networks.35 While CLL cells cultured off stroma quickly lose Mcl-1 (data not shown), BAFF-expressing stroma up-regulated Mcl-1 after 8-24 h (Figure 1D). The pro-survival effects of BAFF-expressing stroma were reversed by A-1210477, an Mcl-1 specific BH3-mimetic (Figure 5A).36 While less effi-

Figure 3. B-cell activating factor (BAFF) activates BCR signaling in chronic lymphocytic leukemia (CLL) cells. (A-C) CLL cells were stimulated with 5 mg/mL sol-IgM or 25 ng/mL sol-BAFF. Cells were lysed at the indicated time points and subjected to immunoblotting. A representative result of 10 independent experiments is shown (includes 4 unmut-IGHV and 6 mut-IGHV). Densitometry (C) was performed on immunoblots from 10 individual CLL samples after 15 minutes (min) stimulation with IgM or BAFF. (D) CLL cells were incubated or not with idelalisib (5 mM) for 1 hour (h) and stimulated with solIgM or sol-BAFF for 30 min. Cell migration using 200 ng/mL CXCL12 was evaluated as described in the Methods section. P<0.05 versus control.

1894

haematologica | 2017; 102(11)

SYK inhibition disrupts BAFF-BCR crosstalk in CLL

cacious than off stroma, BH3-mimetic venetoclax abrogated survival of BAFF-stimulated CLL cells, indicating their continued dependence on Bcl-2 (Figure 5A). While BAFF signaling resulted in partial rescue from low concentrations of venetoclax, CD40L-expressing stroma exhibited a high degree of protection from both BH3-mimetics (i.e. αMcl-1 and αBcl-2), suggesting that Bcl-xL may play a particularly important role in resistance to apoptosis in CLL. Treatment with pevonedistat did not down-regulate Mcl1 in CLL cells, confirming that BAFF-mediated induction of Mcl-1 is NFκB-independent. By contrast, inhibition of SYK reduced Mcl-1 protein levels more effectively than targeting other molecules within the BCR signaling cascade (Figure 5B and C). We further explored how SYK inhibition deregulated Mcl-1. Entospletinib did not enhance Mcl-1 degradation in BAFF-stimulated CLL cells (Figure 5D). Similarly, pharmacological targeting of MEK/ERK, which antagonize Mcl-1 degradation,37 did not modulate Mcl-1 levels in BAFF-stimulated CLL cells (Online Supplementary Figure S7A). By contrast, targeting SYK significantly repressed Mcl-1 mRNA (Figure 5E). SYK activation induces phosphorylation and nuclear translocation of signal transducer and activator of transcription 3 (STAT3),38 a transcriptional regulator of Mcl-1.39 STAT3 phosphorylation was lost in CLL cells cultured off stroma (Online Supplementary Figure S7B). By contrast, STAT3 remained activated in BAFF-expressing stromal co-cultures (Figure 5F). Inhibition of SYK, but not BTK or PI3K, abrogated STAT3 phosphorylation in this setting. Similarly, ruxolitinib, a JAK/STAT inhibitor, abrogated STAT3 activation, accompanied by downregulation of Mcl-1 transcript and protein (Figure 5E and F). Thus, BAFF-mediated induction of Mcl-1 contributes to

haematologica | 2017; 102(11)

CLL cell survival, and may be abrogated via disrupting the SYK-STAT3 axis.

BAFF induces SYK interaction with TRAF2/TRAF3 signaling complex Since BAFF promoted CLL cell survival via SYK-mediated upregulation of the canonical NFκB and Mcl-1, we asked how BAFF activates SYK. We supposed that SYK complexing with TRAF2/TRAF3 may be responsible for BAFF-induced SYK activation in neoplastic B cells. CLL cells and Raji B-cell lymphoma cells were used to address this question. Immunoprecipitation with SYK monoclonal antibodies showed association of SYK with TRAF3 and TRAF2 in both BAFF-stimulated CLL and Raji cells (Figure 6A and B). SYK binding was subsequently confirmed in the reverse experiments with TRAF2 monoclonal antibodies (Figure 6C). Since NIK participates in TRAF2/3 complex, it could be responsible for SYK phosphorylation. NIK expression was low in CLL and Raji lymphoma cells, complicating interpretation of experimental results involving NIK genetic knockdown (data not shown). While BAFF stimulation induced SYK phosphorylation in Raji cells, engineered expression of NIK40 did not modulate SYK activation either in the absence or in the presence of BAFF (Online Supplementary Figure S8A). At the same time, pharmacological targeting of IKK failed to prevent BAFF-mediated SYK activation and SYK did not complex with either NIK or IKK1 in immunoprecipitation experiments (Figure 6A and B, and Online Supplementary Figure S8B and C). Thus, BAFF-BCR crosstalk in neoplastic B cells is at least in part mediated by SYK interaction with TRAF2/TRAF3 signaling complex (Figure 7).

Figure 4. Inhibitors of BCR-associated kinases abrogate B-cell activating factor (BAFF)-mediated canonical NFκB activation in chronic lymphocytic leukemia (CLL). (A) CLL cells (n=6) were cultured on BAFF-expressing stroma for 24 hours (h), followed by incubation with the indicated drugs, or vehicle control (ctrl), for an additional 24 h. Apoptosis within the CD19+ subset of cells was determined by Annexin V and 7-AAD staining. Data are presented as mean±Standard Error (SE). *P<0.05 compared to vehicle control. (B) CLL cells (n=4) were incubated with the indicated drugs or vehicle control for 1 h, followed by stimulation with 25 ng/mL sol-BAFF for 30 minutes (min). Cell migration using 200 ng/mL CXCL12 was evaluated as described in the Methods section. (C) CLL cells (n=4) were co-cultured with BAFFexpressing stroma for 24 h, and incubated with the indicated drugs for an additional 24 h. p52/RelA activity was determined in nuclear protein lysates using the TransAM NFκB activity assay (ActivMotif). **P<0.01 compared to untreated control.

1895

C. Paiva et al.

Discussion Along with the BCR, many concurrently active pathways ensure survival of the neoplastic B cells in the protective niche. We and others previously demonstrated that

primary CLL cells co-cultured with CD40L-expressing stroma activate the canonical and non-canonical NFκB pathways, accompanied by upregulation of the pro-survival Bcl-2 family proteins (Bcl-xL), and acquire therapeutic resistance, including to BCRi.14-16 Here we demonstrate

F E

Figure 5. SYK inhibition down-regulates Mcl-1 via STAT3. (A) Chronic lymphocytic leukemia (CLL) cells were co-cultured with B-cell activating factor (BAFF)-expressing stroma for 24 hours (h), followed by incubation with the indicated drugs for 24 h. Cells were also treated off stroma for 24 h. Apoptosis within CD19+ subset of cells was determined by Annexin V and 7-AAD staining (n=6). Data are presented as mean±Standard Error (SE). *P<0.05, **P<0.01, compared to 'off stroma' control, or as shown. (B and C) CLL cells were co-cultured with BAFF-expressing stroma for 24 h, followed by incubation with the indicated drugs for 24 h in the presence of caspase inhibitor QVD-OPh (1 mM). Cells were lysed and subjected to immunoblotting. Densitometry chart (C) represents data from 6 individual CLL samples. Data are presented as mean±Standard Error (SE). *P<0.05 compared to control. (D) CLL cells (4 individual samples) were co-cultured with BAFF-expressing stroma for 24 h, followed by addition of 100 mg/mL cycloheximide and 1 mM entospletinib or vehicle control. Cells were lysed at the indicated time points and subjected to immunoblotting. (E and F) CLL cells (n=4) were co-cultured with BAFF-expressing stroma for 24 h, treated with SYK inhibitors (entospletinib, R406), JAK1/2 inhibitor (ruxolitinib), BTK inhibitor (ibrutinib) or PI-3Kδ inhibitor (idelalisib) for 24 h in the presence of caspase inhibitor QVD-OPh (1 mM), followed by collection of mRNA and protein.

1896

haematologica | 2017; 102(11)

SYK inhibition disrupts BAFF-BCR crosstalk in CLL

that BAFF signaling had a distinct effect on CLL cell survival. CLL cells co-cultured with BAFF-expressing stroma were resistant to both spontaneous and drug-induced apoptosis. BAFF triggered robust activation of the noncanonical and, to a lesser extent, canonical NFκB, leading to distinctive expression of Bcl-2 family proteins, predominantly Mcl-1. BAFF was previously shown to co-opt BCR signaling in mouse splenic B lymphocytes, manifested by phosphorylation of the BCR-associated CD79A subunit and SYK.41 Here, we present, for the first time, evidence that such crosstalk exists in primary human neoplastic B cells. We detected activation of the BCR-associated kinases, including SYK, BTK, and ERK in CLL cells stimulated with BAFF. Importantly, while BCR response to BAFF was transient in mouse B cells,41 it was sustained in CLL. Although CLL samples with unmutated IGHV have an enhanced capacity to respond to BCR engagement,28 both CLL subsets responded to BAFF. The ability of the mutated CLL to respond to BAFF suggests that surface IgM expression is an unlikely pre-requisite for successful transduction of BAFF signal to the BCR.24 In a mouse, engineered BCR loss prevented SYK activation by BAFF, suggesting that intracellular tyrosine activation motif (ITAM)-related sequences on CD79A/B may be involved in BAFF signal transmission.41 We observed direct interaction between TRAF2/TRAF3 adaptor protein complex and SYK in CLL cells, potentially implicating this interaction in BAFF-BCR crosstalk in neoplastic B cells (Figure 6D). It is possible that SYK induction in CLL by CD40L also occurs through TRAF2/3 interaction.42 When BAFF-R is engaged, the NIK/TRAF/cIAP complex is recruited to the receptor, thus allowing NIK to persist and activate IKK1. Therefore either NIK or IKK1 may be involved in BAFF-mediated SYK activation. Furthermore, the role of the BCR structures (CD79A/B) or LYN, a BCRassociated kinase constitutively active in CLL cells,43 as well as BCR-associated protein phosphatases, needs to be investigated in this setting. Additional experiments will be required to decipher the exact contributions of the individual kinases and BCR structural components in BAFFBCR crosstalk in CLL, and will be hampered by the technical challenges of eliminating those individual players in primary B cells. It is conceivable that many different conditions need to be fulfilled where co-operative action

involving an intact BCR structure and the SRC family protein kinases is required for BAFF-mediated activation of SYK and BCR signal propagation. It remains unclear whether activation of both canonical and non-canonical NFκB is necessary to ensure BAFFmediated CLL cell survival. Experiments in murine splenic B cells suggest that the two NFκB pathways may be redundant, since the double Nfkb1-/-Nfkb2-/- mouse, but not single mutations, recapitulated a BAFF/BAFF-R-deficiency phenotype.44 However, recent data suggest complementary roles for both pathways, which promote BAFFinduced B-cell survival and maturation via distinct gene expression programs.45 Earlier studies demonstrated that selectively blocking canonical NFκB abrogated survival advantage inferred by CLL cell exposure to soluble BAFF.8 Furthermore, although BCR signaling does not promote processing of p100, the latter is a positively regulated target of BCR-mediated gene transcription, where ultimately BCR signaling may enhance non-canonical NFκB signaling.46,47 We show that BCRi predominantly inactivated canonical NFκB and did not completely reverse protection in BAFF-expressing stromal co-cultures. Meanwhile, pevonedistat is an investigational small molecule that forms a covalent adduct with NEDD8, a ubiquitin-like modifier, thus disrupting the functionality of Cullin-RING ubiquitin ligases and leading to accumulation of their substrates, including inhibitor of NFκB (IκB) and NFκB2/p100, with a net outcome of inactivation of the canonical and non-canonical NFκB, respectively.16 While pevonedistat abrogated NFκB signaling in BAFF-stimulated CLL cells, apoptosis was not complete, suggesting that NFκB-independent signaling pathways, including BCR, contribute to BAFF-mediated survival in CLL. Additional studies are needed to elucidate the role of NFκB in BAFF-mediated CLL cell survival. Ultimately, a combination of strategies aimed at delivering a multi-pronged attack will be required to fully neutralize the pro-survival pathways induced by BAFF. Subsequent studies should help elucidate the exact ligand-receptor interactions leading to BCR activation in CLL. BAFF and APRIL bind the two TNFR superfamily members, BCMA and TACI, with high affinity. It would be important to confirm whether APRIL stimulation could replicate some of the BAFF ligand effects. While our data suggest that BAFF-BR3 interaction may be necessary for

Figure 6. SYK interacts with TRAF2/TRAF3 in neoplastic B cells. Cells were stimulated with 25 ng/mL sol-B-cell activating factor (BAFF) for 30 minutes (min). Proteins lysates were subjected to immunoprecipitation experiments using indicated antibodies as described in the Methods section. A representative blot of 3 independent experiments is shown. CLL: chronic lymphocytic leukemia.

haematologica | 2017; 102(11)

1897

C. Paiva et al.

Figure 7. B-cell activating factor (BAFF)-BCR crosstalk in chronic lymphocytic leukemia (CLL) cells. BAFF-R engagement stabilizes NIK within the NIK/TRAF2/TRAF3/cIAP1/2 complex, promoting the non-canonical NFκB pathway activity. SYK recruitment to TRAF2/TRAF3 assists BAFF-mediated activation of BCR signaling, which contributes to activation of the canonical NFκB. Concurrently, SYK induces STAT3 transcription factor, thereby up-regulating Mcl-1, a pro-survival Bcl2 family member.

BCR activation via TRAF2/TRAF3 complex, isolated and concurrent blockade of the individual signaling receptors9 will be necessary to elucidate their importance in regulation of NFκB and BCR signaling pathways, as well as response to BCRi in BAFF-simulated CLL cells. SYK, a key initiating kinase in the BCR signaling cascade, is an attractive therapeutic target in lymphoid malignancies. Entospletinib is a novel orally available kinase inhibitor which has greater selectivity for SYK compared with R406.48 Entospletinib showed a favorable toxicity profile and induced responses in 61% of patients with relapsed/refractory CLL.48 It has been previously demonstrated that SYK inhibitors, such as R406, block BCR engagement-mediated induction of Mcl-1 in CLL cells.25 While this manuscript was in preparation, Bojarczuk et al. reported that other BCRi (ibrutinib, idelalisib) were also able to inhibit Mcl-1; however, they were not as effective as SYK inhibitors.49 Interestingly, we found that neither BTK or PI3Kδ inhibition neutralized Mcl-1 to the same degree as entospletinib in BAFFexpressing stromal co-culture model. In a setting of BCR engagement, BCRi prevented inactivation of glycogen synthase kinase-3 (GSK-3), thereby presumably leading to Mcl-1 degrada1898

tion.35 By contrast, we did not observe enhanced Mcl-1 degradation following SYK inhibition in BAFF-stimulated CLL cells. It is possible that BAFF signaling incapacitates GSK-3 in a BCR-independent manner, e.g. through PI3K/AKT, assisted by the co-receptor CD19,10,50,51 such that concurrent inhibition of multiple pathways, resulting in abrogation of both AKT-dependent and AKT-independent (Protein kinase C-dependent) inactivation of GSK-3, may be necessary to enhance Mcl-1 turnover. Importantly, SYK inhibition resulted in decreased phosphorylation of STAT3, a known SYK target and a positive regulator of Mcl-1 transcription.38,39 Thus, transcriptional downmodulation rather than enhanced turnover underlies loss of Mcl-1 following SYK inhibition in BAFFstimulated CLL cells. This finding also suggests that BTK and PI3Kδ inhibitors may less effectively target Mcl-1 in the tumor microenvironment than SYK inhibitors, informing future therapeutic development of BCRi. High expression of Mcl-1 predicts adverse outcomes following chemo-immunotherapy in CLL.52 Given our findings, a relevant question would be whether responses in B-cell malignancies may be bolstered by concurrent inhibition of SYK and other BCR-associated kinases. While a clinical trial of a combination of entospletinib and idelalishaematologica | 2017; 102(11)

SYK inhibition disrupts BAFF-BCR crosstalk in CLL

ib, a PI3KÎ´ inhibitor, in patients with CLL was halted early due to treatment-emergent pneumonitis,53 clinical trials evaluating safety and efficacy of entospletinib and BTK inhibitor combinations are ongoing. Thus, our study throws light on the crosstalk between BAFF and BCR signaling pathways in neoplastic B cells,

References 1. Pascutti MF, Jak M, Tromp JM, et al. IL-21 and CD40L signals from autologous T cells can induce antigen-independent proliferation of CLL cells. Blood. 2013;122(17):30103019. 2. Caligaris-Cappio F, Bertilaccio MT, Scielzo C. How the microenvironment wires the natural history of chronic lymphocytic leukemia. Semin Cancer Biol. 2014;24:4348. 3. Burger JA, Tsukada N, Burger M, Zvaifler NJ, Dell'Aquila M, Kipps TJ. Blood-derived nurse-like cells protect chronic lymphocytic leukemia B cells from spontaneous apoptosis through stromal cell-derived factor-1. Blood. 2000;96(8):2655-2663. 4. Herishanu Y, Perez-Galan P, Liu D, et al. The lymph node microenvironment promotes B-cell receptor signaling, NF-kappaB activation, and tumor proliferation in chronic lymphocytic leukemia. Blood. 2011;117(2):563-574. 5. Danilov AV. Targeted therapy in chronic lymphocytic leukemia: past, present, and future. Clin Ther. 2013;35(9):1258-1270. 6. Jain P, Keating M, Wierda W, et al. Outcomes of patients with chronic lymphocytic leukemia after discontinuing ibrutinib. Blood. 2015;125(13):2062-2067. 7. Maddocks KJ, Ruppert AS, Lozanski G, et al. Etiology of Ibrutinib Therapy Discontinuation and Outcomes in Patients With Chronic Lymphocytic Leukemia. JAMA Oncol. 2015;1(1):80-87. 8. Endo T, Nishio M, Enzler T, et al. BAFF and APRIL support chronic lymphocytic leukemia B-cell survival through activation of the canonical NF-kappaB pathway. Blood. 2007;109(2):703-710. 9. Kern C, Cornuel JF, Billard C, et al. Involvement of BAFF and APRIL in the resistance to apoptosis of B-CLL through an autocrine pathway. Blood. 2004;103(2):679-688. 10. Mackay F, Figgett WA, Saulep D, Lepage M, Hibbs ML. B-cell stage and context-dependent requirements for survival signals from BAFF and the B-cell receptor. Immunol Rev. 2010;237(1):205-225. 11. Nishio M, Endo T, Tsukada N, et al. Nurselike cells express BAFF and APRIL, which can promote survival of chronic lymphocytic leukemia cells via a paracrine pathway distinct from that of SDF-1alpha. Blood. 2005;106(3):1012-1020. 12. Rickert RC, Jellusova J, Miletic AV. Signaling by the tumor necrosis factor receptor superfamily in B-cell biology and disease. Immunol Rev. 2011;244(1):115133. 13. Perkins ND. Integrating cell-signalling pathways with NF-kappaB and IKK function. Nat Rev Mol Cell Biol. 2007;8(1):49-62. 14. Tromp JM, Tonino SH, Elias JA, et al. Dichotomy in NF-kappaB signaling and chemoresistance in immunoglobulin variable heavy-chain-mutated versus unmutated CLL cells upon CD40/TLR9 triggering. Oncogene. 2010;29(36):5071-5082.

haematologica | 2017; 102(11)

and provides insights into the mechanistic effects of SYK inhibitors in CLL. Acknowledgments We would like to thank Allison Berger, Ph.D., for the helpful review of this manuscript.

15. Vogler M, Butterworth M, Majid A, et al. Concurrent up-regulation of BCL-XL and BCL2A1 induces approximately 1000-fold resistance to ABT-737 in chronic lymphocytic leukemia. Blood. 2009;113(18):44034413. 16. Godbersen JC, Humphries LA, Danilova OV, et al. The Nedd8-activating enzyme inhibitor MLN4924 thwarts microenvironment-driven NF-kappaB activation and induces apoptosis in chronic lymphocytic leukemia B cells. Clin Cancer Res. 2014;20(6):1576-1589. 17. Schmidt MR, Appel MC, Giassi LJ, Greiner DL, Shultz LD, Woodland RT. Human BLyS facilitates engraftment of human PBL derived B cells in immunodeficient mice. PLoS One. 2008;3(9):e3192. 18. Hoellenriegel J, Meadows SA, Sivina M, et al. The phosphoinositide 3'-kinase delta inhibitor, CAL-101, inhibits B-cell receptor signaling and chemokine networks in chronic lymphocytic leukemia. Blood. 2011;118(13):3603-3612. 19. Danilov AV, Soderquist RS, Bates DJ, Eastman A. Toward a cure for chronic lymphocytic leukemia: an attack on multiple fronts. Expert Rev Anticancer Ther. 2013;13(9):1009-1012. 20. Soderquist R, Bates DJ, Danilov AV, Eastman A. Gossypol overcomes stromamediated resistance to the BCL2 inhibitor ABT-737 in chronic lymphocytic leukemia cells ex vivo. Leukemia. 2013;27(11):22622264. 21. Herreros B, Rodriguez-Pinilla SM, Pajares R, et al. Proliferation centers in chronic lymphocytic leukemia: the niche where NF-kappaB activation takes place. Leukemia. 2010;24(4):872-876. 22. Rodig SJ, Shahsafaei A, Li B, Mackay CR, Dorfman DM. BAFF-R, the major B cellactivating factor receptor, is expressed on most mature B cells and B-cell lymphoproliferative disorders. Hum Pathol. 2005;36(10):1113-1119. 23. Nakamura N, Hase H, Sakurai D, et al. Expression of BAFF-R (BR 3) in normal and neoplastic lymphoid tissues characterized with a newly developed monoclonal antibody. Virchows Arch. 2005;447(1):53-60. 24. Stevenson FK, Krysov S, Davies AJ, Steele AJ, Packham G. B-cell receptor signaling in chronic lymphocytic leukemia. Blood. 2011;118(16):4313-4320. 25. Gobessi S, Laurenti L, Longo PG, et al. Inhibition of constitutive and BCR-induced Syk activation downregulates Mcl-1 and induces apoptosis in chronic lymphocytic leukemia B cells. Leukemia. 2009;23(4):686697. 26. Petlickovski A, Laurenti L, Li X, Marietti S, Chiusolo P, Sica S, Leone G, Efremov DG. Sustained signaling through the B-cell receptor induces Mcl-1 and promotes survival of chronic lymphocytic leukemia B cells. Blood. 2005;105(12):7. 27. Lanham S, Hamblin T, Oscier D, Ibbotson R, Stevenson F, Packham G. Differential signaling via surface IgM is associated with VH gene mutational status and CD38

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

expression in chronic lymphocytic leukemia. Blood. 2003;101(3):1087-1093. Deglesne PA, Chevallier N, Letestu R, et al. Survival response to B-cell receptor ligation is restricted to progressive chronic lymphocytic leukemia cells irrespective of Zap70 expression. Cancer Res. 2006;66(14):71587166. Mockridge CI, Potter KN, Wheatley I, Neville LA, Packham G, Stevenson FK. Reversible anergy of sIgM-mediated signaling in the two subsets of CLL defined by VH-gene mutational status. Blood. 2007;109(10):4424-4431. Packham G, Stevenson F. The role of the Bcell receptor in the pathogenesis of chronic lymphocytic leukaemia. Semin Cancer Biol. 2010;20(6):391-399. de Rooij MF, Kuil A, Geest CR, et al. The clinically active BTK inhibitor PCI-32765 targets B-cell receptor- and chemokine-controlled adhesion and migration in chronic lymphocytic leukemia. Blood. 2012; 119(11):2590-2594. Thijssen R, ter Burg, J, van Bochover GGW, et al. The pan phosphoinositide 3kinase/mammalian target of rapamycin inhibitor SAR245409 (voxtalisib/XL765) blocks survival, adhesion and proliferation of primary chronic lymphocytic leukemia cells. Leukemia. 2015;30(2):337-345. Bernard S, Danglade D, Gardano L, et al. Inhibitors of BCR signalling interrupt the survival signal mediated by the micro-environment in mantle cell lymphoma. Int J Cancer. 2015;136(12):2761-2774. Herman SE, Mustafa RZ, Gyamfi JA, et al. Ibrutinib inhibits BCR and NF-kB signaling and reduces tumor proliferation in tissueresident cells of patients with CLL. Blood. 2014;123(21):3286-3295. Ertel F, Nguyen M, Roulston A, Shore GC. Programming cancer cells for high expression levels of Mcl1. EMBO Rep. 2013;14(4):328-336. Leverson JD, Zhang H, Chen J, et al. Potent and selective small-molecule MCL-1 inhibitors demonstrate on-target cancer cell killing activity as single agents and in combination with ABT-263 (navitoclax). Cell Death Dis. 2015;6:e1590. Domina AM, Vrana JA, Gregory MA, Hann SR, Craig RW. MCL1 is phosphorylated in the PEST region and stabilized upon ERK activation in viable cells, and at additional sites with cytotoxic okadaic acid or taxol. Oncogene. 2004;23(31):5301-5315. Uckun FM, Qazi S, Ma H, Tuel-Ahlgren L, Ozer Z. STAT3 is a substrate of SYK tyrosine kinase in B-lineage leukemia/lymphoma cells exposed to oxidative stress. Proc Natl Acad Sci. 2010;107(7):2902-2907. Epling-Burnette PK, Liu JH, Catlett-Falcone R, et al. Inhibition of STAT3 signaling leads to apoptosis of leukemic large granular lymphocytes and decreased Mcl-1 expression. J Clin Invest. 2001;107(3):351-362. Xiao G, Sun SC. Negative regulation of the nuclear factor kappa B-inducing kinase by a cis-acting domain. J Biol Chem. 2000; 275(28):21081-21085.

1899

C. Paiva et al. 41. Schweighoffer E, Vanes L, Nys J, et al. The BAFF receptor transduces survival signals by co-opting the B cell receptor signaling pathway. Immunity. 2013;38(3):475-488. 42 Parente-Ribes A, Skanland SS, Burgler S, et al. Spleen tyrosine kinase inhibitors reduce CD40L-induced proliferation of chronic lymphocytic leukemia cells but not normal B cells. Haematologica. 2016;101(2):e59-62. 43. Contri A, Brunati AM, Trentin L, et al. Chronic lymphocytic leukemia B cells contain anomalous Lyn tyrosine kinase, a putative contribution to defective apoptosis. J Clin Invest. 2005;115(2):369-378. 44. Claudio E, Brown K, Park S, Wang H, Siebenlist U. BAFF-induced NEMO-independent processing of NF-kappa B2 in maturing B cells. Nat Immunol. 2002; 3(10):958-965. 45. Almaden JV, Liu YC, Yang E, et al. B-cell survival and development controlled by the

1900

46. 47.

48.

49.

coordination of NF-kappaB family members RelB and cRel. Blood. 2016; 127(10):1276-1286. Cancro MP. Signalling crosstalk in B cells: managing worth and need. Nat Rev Immunol. 2009;9(9):657-661. Castro I, Wright JA, Damdinsuren B, et al. B cell receptor-mediated sustained c-Rel activation facilitates late transitional B cell survival through control of B cell activating factor receptor and NF-kappaB2. J Immunol. 2009;182(12):7729-7737. Sharman J, Hawkins M, Kolibaba K, et al. An open-label phase 2 trial of entospletinib (GS-9973), a selective spleen tyrosine kinase inhibitor, in chronic lymphocytic leukemia. Blood. 2015;125(15):2336-2343. Bojarczuk K, Sasi BK, Gobessi S, et al. BCR signaling inhibitors differ in their ability to overcome Mcl-1-mediated resistance of CLL B cells to ABT-199. Blood.

2016;127(25):3192-3201. 50. Jellusova J, Miletic AV, Cato MH, et al. Context-specific BAFF-R signaling by the NF-kappaB and PI3K pathways. Cell Rep. 2013;5(4):1022-1035. 51. Henley T, Kovesdi D, Turner M. B-cell responses to B-cell activation factor of the TNF family (BAFF) are impaired in the absence of PI3K delta. Eur J Immunol. 2008; 38(12):3543-3548. 52. Awan FT, Kay NE, Davis ME, et al. Mcl-1 expression predicts progression-free survival in chronic lymphocytic leukemia patients treated with pentostatin, cyclophosphamide, and rituximab. Blood. 2009;113(3):535-537. 53. Barr PM, Saylors GB, Spurgeon SE, et al. Phase 2 study of idelalisib and entospletinib: pneumonitis limits combination therapy in relapsed refractory CLL and NHL. Blood. 2016;127(20):2411-2415.

haematologica | 2017; 102(11)

ARTICLE

Chronic Lymphocytic Leukemia

Toll-like receptor 9 stimulation can induce IκBζ expression and IgM secretion in chronic lymphocytic leukemia cells

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Eleonora Fonte,1 Maria Giovanna Vilia,1 Daniele Reverberi,2 Ilenia Sana,1 Lydia Scarfò,3,4 Pamela Ranghetti,3 Ugo Orfanelli,5 Simone Cenci,5 Giovanna Cutrona,2 Paolo Ghia3,4 and Marta Muzio1

Cell Signaling Unit, Division of Experimental Oncology, IRCCS San Raffaele Hospital, Milano; 2UOC Patologia Molecolare, IRCCS AOU S. Martino-IST, Genova; 3B-Cell Neoplasia Unit and Strategic Research Program on CLL, Division of Experimental Oncology, IRCCS San Raffaele Hospital, Milano; 4Università Vita-Salute San Raffaele, Milano and 5Age Related Diseases Unit, Division of Genetics and Cell Biology, IRCCS San Raffaele Hospital, Milano, Italy 1

Haematologica 2017 Volume 102(11):1901-1912

ABSTRACT

hronic lymphocytic leukemia cells strongly depend on external stimuli for their survival. Both antigen receptor and co-stimulatory receptors, including Toll-like receptors, can modulate viability and proliferation of leukemic cells. Toll-like receptor ligands, and particularly the TLR9 ligand CpG, mediate heterogeneous responses in patients’ samples reflecting the clinical course of the subjects. However, the molecular framework of the key signaling events underlying such heterogeneity is undefined. We focused our studies on a subset of chronic lymphocytic leukemia cases characterized by expression of CD38 and unmutated immunoglobulin genes, who respond to CpG with enhanced metabolic cell activity. We report that, while CpG induces NFKBIZ mRNA in all the samples analyzed, it induces the IκBζ protein in a selected group of cases, through an unanticipated post-transcriptional mechanism. Interestingly, IκBζ plays a causal role in sustaining CpG-induced cell viability and chemoresistance, and CpG stimulation can unleash immunoglobulin secretion by IκBζ-positive malignant cells. These results identify and characterize IκBζ as a marker and effector molecule of distinct key pathways in chronic lymphocytic leukemia.

Introduction Chronic lymphocytic leukemia (CLL) is the most frequent adult leukemia, characterized by the proliferation and accumulation of mature clonal B-lymphocytes in the peripheral blood and lymphoid tissues. Despite significant progress in treatment modalities, CLL remains incurable, with a sizeable fraction of patients experiencing progressive refractory disease. CLL is actually a heterogeneous disorder where aggressive cases are characterized by distinct clinical and biological markers including the presence of unmutated immunoglobulin heavy chain variable region (IGHV) genes, and CD38 expression.1 CLL cells strongly depend on external stimuli for survival and proliferation.2 Distinct cell types can support leukemia development and progression through different cytokine receptors. Moreover, both antigen receptor and co-stimulatory receptors, including Toll-like receptors (TLR), are involved in the pathobiology of CLL.3,4 Interestingly, novel B-cell receptor signaling inhibitors that have been recently approved for clinical use in CLL (i.e. BTK and PI3Kδ inhibitors) target pathways that regulate tumor-microenvironment interactions;5 a better understanding of all the molecules and pathways involved in these processes may, therefore, help to fully understand CLL biology and design more effective targeting strategies. TLR are transmembrane proteins devoted to the recognition of and binding to molecular patterns that can be derived either from microbes or from endogenous proteins (e.g. danger signals).6 TLR are mainly expressed by monocytes and macrophages where they trigger an innate immune response; nevertheless, both normal and malighaematologica | 2017; 102(11)

Correspondence: muzio.marta@hsr.it

Received: February 3, 2017. Accepted: August 1, 2017. Pre-published: August 3, 2017. doi:10.3324/haematol.2017.165878 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1901 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1901

E. Fonte et al.

nant B cells also express functional TLR and their stimulation influences leukemic cell survival and proliferation.3,4 Heterogeneity of response to different TLR ligands is observed in different patients, reflecting their clinical course. In detail, distinct TLR ligands binding to TLR1/2, TLR2/6 and TLR9 induce co-stimulatory molecules in virtually all cases, but they can induce proliferation and chemoresistance only in a select subset of cases characterized by adverse prognostic markers (e.g. unmutated IGHV genes and CD38).7-10 It should be noted that the levels of expression of TLR9 do not explain differential responsiveness of the cells to CpG.8 Cytokines (e.g. tumor necrosis factor-β), anti-apoptotic molecules (e.g. BclXL) and distinct miRNA are selectively induced in IGHV unmutated cases.11-13 However, the molecular framework of the key signaling molecules leading to such variability in response is poorly defined. We focused on the atypical IκBζ nuclear protein because it is specifically induced by TLR in different leukocyte populations but its role has never been characterized in CLL or in normal human B cells. IκBζ [also known as Interleukin-1 inducible nuclear ankyrin-repeat protein (INAP) or Molecule possessing ankyrin repeats induced by lipopolysaccharide (MAIL)] is selectively induced by interleukin-1 and TLR ligands but not by other inflammatory cytokines such as tumor necrosis factor.14-19 IκBζ expression is low to undetectable in unstimulated immune cells, and is rapidly induced after TLR stimulation by transcriptional, post-transcriptional and translational mechanisms.20-23 IκBζ is an atypical IκB family member; in contrast to the classic IκB family members (but similar to other atypical members), it is a nuclear protein and a direct transcriptional regulator. IκBζ can act as a negative or positive modulator in combination with distinct transcription factors including p50 and p65 nuclear factor-κB (NF-κB) family members. In particular IκBζ controls the induction of secondary response genes including interleukin-6 and interleukin10.17,18 Mouse models lacking IκBζ demonstrated that this protein is a key regulator of both innate and adaptive immune responses, such as Th17 development, natural killer-cellderived interferon-γ production, and interleukin-6 production in macrophages.19,24,25 In epithelial cells, a deficiency in IκBζ causes apoptosis, which induces Sjögren syndromelike inflammation.26 In the context of B-cell types, it has been recently shown that IκBζ controls the proliferation of mouse B-lymphocytes and triggers a TLR-dependent but Tindependent antibody response.27 We herein analyzed, for the first time, IκBζ expression, regulation and function in leukemic cells from CLL patients. Stimulation of TLR9 with the CpG ligand induced IκBζ in leukemic cells; this up-regulation was distinctively higher, as a result of a post-transcriptional mechanism, in a subgroup of CLL cases characterized by CD38 expression and unmutated IGHV genes. Moreover, we addressed the expression pattern and functional role of IκBζ in malignant cells. Our results provide novel insights into the pathobiology of CLL, and shed light onto the molecular pathways that mark and regulate distinct CLL cases.

Methods Chronic lymphocytic leukemia samples Leukemic lymphocytes were obtained from peripheral blood of CLL patients diagnosed according to the International Workshop 1902

on CLL/National Cancer Institute 2008 guidelines.28 All patients were either untreated or off therapy for at least 6 months before the study. The following parameters were analyzed for each patient: age, sex, disease stage at diagnosis, CD38 expression, and IGHV gene mutational status (Online Supplementary Table S1). All tissue samples were obtained with the approval of the institutional Ethics Committee of San Raffaele Scientific Institute (Milan, Italy), after informed consent.

Cell purification CLL cells were negatively selected and purified using a B-cell enrichment kit (RosetteSep; StemCell Technologies) following the manufacturer’s instructions. Normal B cells from buffy coat and B cells from tonsil were purified by negative selection (EasySep; StemCell Technologies). The purity of all leukemic CD19+CD5+ and normal CD19+ preparations was always >98% as checked by flow cytometry (FC500; Beckman Coulter). Buffy coats from anonymized healthy donors were obtained from ASST Rhodense Hospital (Rho, Italy), IRCCS AOU San Martino – IST (Genova, Italy) and “San Raffaele” Hospital (Milano, Italy). Tonsils from patients not affected by CLL were obtained from the “G. Gaslini” Hospital and “San Raffaele” Hospital.

Analysis of B-cell subpopulations The percentage of anti-ΙκΒζ-positive cells was determined in each B-cell subset, 4 h after CpG stimulation, by a multiparametric flow cytometry gating strategy described elsewhere.29,30 Briefly, cells were first stained for surface antigens with anti-IgD FITC, anti-CD38 PE-Cy7, anti-CD19 APC-H7, anti-CD5 Alexa-Fluor700, anti-IgM PerCP-Cy5.5, anti-CD27 PE-CF594 and anti-CD24 Alexa-Fluor647 monoclonal antibodies (BD Biosciences), then fixed, permeabilized and stained with anti-ΙκΒζ (see below for details). Flow cytometry analyses were performed using FACSAria II DIVA 6 software (BD Biosciences). Data were processed with Prism (GraphPad Software Inc.). Buffy-coat CD19+ B-cell subsets comprised naïve B cells (IgDbright CD27−), memory (MEM) B cells (IgD−/lowCD27−/+), CD5-positive (CD5+) B cells, CD38-positive (CD38+) B cells and CD38bright plasmablasts. MEM B cells comprised the following subsets: IgM memory (M-MEM) B cells (IgD−/low CD27+), switched memory B cells (S-MEM) (IgD−CD27+), and double-negative memory cells (IgD−CD27−) (DN-MEM). The tonsillar CD19+ B-cell subset comprised naïve B cells (IgDbrightIgMbrightCD38−CD27−), memory (MEM) B cells (IgD−/lowCD38−), germinal center (GC) B cells (IgD-CD38+CD24-), plasmablasts (CD38bright), transitional B cells (TRANS) (CD38+CD24+), activated B cells (IgD+CD38+). MEM B cells included three different memory B-cell subsets: IgM memory (M-MEM) (IgMbrightIgD−/lowCD27+), switched memory (S-MEM) (IgM-IgDCD27+), and double-negative memory (IgD-CD27-) (DN-MEM) subsets, which were analyzed separately. Online Supplementary Figures S1 and S2 describe the gating strategy for peripheral blood and tonsillar cells, respectively. Further details regarding the methods can be found in the Online Supplementary Files.

Results CpG induces IκBζ protein in a heterogeneous manner in patients with chronic lymphocytic leukemia IκBζ protein was undetectable in unstimulated CLL cells, but was markedly induced, in selected cases, after TLR9 stimulation with the CpG ligand; a representative case analyzed by flow cytometry before and after CpG haematologica | 2017; 102(11)

CpG induces ΙκΒζ and IgM in CLL

stimulation, and stained with anti-IκBζ or isotype control antibody is shown in Figure 1A. It should be noted that no significant cell death was induced at this time point, thus excluding apoptosis as influencing IκBζ staining (data not shown and Online Supplementary Figure S4).11

Western blot analysis demonstrated that a protein with a molecular weight of approximately 85 kDa, corresponding to the longer isoform of IκBζ, is detected in some but not all CLL samples after CpG stimulation (Figure 1B). Interestingly, analysis of a total of 75 CLL cases revealed

Figure 1. Heterogeneous IκBζ expression in chronic lymphocytic leukemia cells treated with CpG correlates with adverse prognostic markers. (A) IκBζ flow cytometry analysis of two representative patients; CLL cells were incubated with or without CpG for 24 h, and stained with anti-IκBζ-PE or IgG G1 K iso control-PE. (B) Western blot analysis of four representative patients; CLL cells were incubated with or without CpG for 24 h. (C-D) Higher percentage of IκBζ+ cells and higher IκBζ mean fluorescence Intensity (MFI) are observed in CD38-positive samples (% of CD38+ CLL cells ≥30); n=16 CD38+ and 56 CD38-. (E-F) Higher percentage of IκBζ+ cells and higher IκBζ MFI are observed in unmutated CLL cases); n=23 unmutated and 40 mutated. The Mann-Whitney test was used to analyze all the data. **P<0.01; ***P<0.001. (G) Receiver operating characteristic (ROC) curve analysis was performed with the percentage of IκBζ+ cells and IGHV mutational status; n=23 unmutated and 40 mutated. (H) Distribution of CLL samples based on the increase in percentage of IκBζ+ cells (n=75); the calculated cut-off value is indicated.

haematologica | 2017; 102(11)

1903

E. Fonte et al.

that IκBζ expression is different in patients characterized by different prognostic markers including CD38 expression and IGHV mutational status. Higher proportions of IκBζ positive cells and higher levels of expression of IκBζ protein (mean fluorescence intensity; MFI) were significantly associated with CD38 expression (Figure 1C,D) and the presence of unmutated IGHV genes (Figure 1E,F). No significant difference of IκBζ expression was found in CLL patients as a function of age, sex or in patients with progressive versus stable disease (Online Supplementary Table S1). We performed a receiver operating characteristic (ROC) analysis to discriminate IGHV-mutated versus -unmutated samples based on the levels of IκBζ; the ROC curve shown in Figure 1G demonstrated that the percentage of IκBζ -positive cells predicted IGHV mutated versus unmutated cases with an area under the curve (AUC) of 0.7804 (P=0.0002327). The Youden index (a measure of specificity and sensitivity) was calculated for each value of the curve, and the maximum index corresponded to a cut-off of 29.75% IκBζ -pos-

itive cells (Figure 1H). Using this cut-off value, we detected a higher frequency of IκBζ-positive samples among cases that were CD38-positive, IGHV unmutated, and higher Rai stage (Online Supplementary Figure S3). We thus applied 29.75% IκBζ-positive cells as the cut-off value for the subsequent biological analyses comparing IκBζ-positive to IκBζ-negative samples.

Metabolic activation driven by TLR9 stimulation is mediated by IκBζ In vitro TLR9 stimulation is known to upregulate costimulatory molecules in virtually all CLL cases after 24 h;11 in contrast, it induces metabolic cell activation, cell viability and proliferation only in those cases characterized by poor prognostic markers including disease progression, CD38 expression, unmutated IGHV genes and unfavorable cytogenetic aberrations.4 To study the specific signaling pathways that are involved in the TLR-mediated activation program, we cultured patient-derived leukemic CLL cells in the presence of the TLR9 ligand CpG

Figure 2. TLR9-induced ΙκΒζ controls chronic lymphocytic leukemia cell survival and chemoresistance. (A-B) Higher percentage of IκBζ+ cells (% in panel A) and higher IκBζ MFI (panel B) were observed in samples with increased metabolic cell activity induced by CpG (ATP+); n=33 ATP+ and 23 ATP-. (C-D) CLL cells stimulated with CpG for 24 h were analyzed for the percentage (C) and MFI (D) of IκBζ+ cells (n=21 “CpG-induced chemoresistant” samples and n=12 “CpG-reduced chemoresistant” samples analyzed). A Mann-Whitney test was used to analyze all the data. **P<0.01; ****P<0.0001. (E-G) CLL cells from eight CLL patients were electroporated with siRNA against the mRNA encoding for IκBζ protein (si-IκBζ) or a scrambled control siRNA (SCR) for 16 h; CLL cells were either stimulated with CpG or left untreated as indicated and subsequently treated with fludarabine. (E) Western blot analysis was performed to confirm the inhibition of IκBζ protein expression in one representative sample. CLL samples were incubated without or with 3 mM fludarabine, collected after 48 h of incubation and analyzed for cell viability. Mean and SEM of eight CLL patients are indicated. The Wilcoxon matched pairs test was performed to analyze all the data. **P value <0.01.

1904

haematologica | 2017; 102(11)

CpG induces ΙκΒζ and IgM in CLL

ODN2006 for 24 h, and analyzed the expression of IκBζ. In parallel, after an additional 24 h, we measured metabolic activation by a specific assay that quantifies the ATP present as an indirect indicator of cell viability.11 As expected, CpG induced metabolic activation in a group of patients, herein referred to as “ATP+” but had no effect or anti-metabolic influence in others (“ATP-“) (Online Supplementary Figure S4A). Strikingly, TLR9 stimulation induced higher levels of IκBζ in the group of “ATP+” samples (Figure 2A,B). Since we previously reported that CpG can induce chemoresistance to fludarabine in vitro in a proportion of cases,11 we investigated whether the observed rapid IκBζ induction also mediated this effect. We stimulated the cells with or without CpG for 24 h, and analyzed IκBζ expression (Figure 2C,D); subsequently, we treated the cells with

fludarabine for another 24 h, and measured metabolic activation at the end of the treatment (Online Supplementary Figure S4B). Both the percentage of IκBζ -positive cells and the mean expression of IκBζ were significantly higher in the group of samples in which CpG induced fludarabine resistance (Figure 2C,D). To test the functional role of IκBζ, CLL cells from “ATP+” cases were electroporated with IκBζ specific or control siRNA to inhibit IκBζ expression, as demonstrated by western blot analysis (Figure 2E); cellular metabolic activity was then analyzed 48 h after CpG treatment, in the absence or the presence of 3 mM fludarabine. Upon IκBζ silencing before CpG stimulation, a significant decrease in metabolic activation cells was evident in both fludarabine-treated and untreated cells (Figure 2F and G, respectively). These data implicate IκBζ in the mechanism whereby

Figure 3. ΙκΒζ is post-transcriptionally regulated in chronic lymphocytic leukemia. (A) Real time polymerase chain reaction analysis of NFKBIZ expression was performed at different time points (2 h, 4 h, 24 h as indicated) in the two groups of patients namely “ATP+” (n=15 at 2 h; n=18 at 4 h; n=4 at 24 h) or “ATP-” (n=9 at 2 h; n=9 at 4 h; n=4 at 24 h); Mean and SEM are indicated. (B) Western blot analysis of IκBζ expression was performed at different time points (4 h, 6 h, 24 h) in the two groups of patients, namely “IκBζ+” or “IκBζ_” as indicated (2 representative samples out of 6 analyzed: 3 ATP+ and 3 ATP-. See Online Supplementary Figure S6 for quantifications). (C) CLL cells were cultured for 4 h with CpG, fixed and stained with anti-IκBζ antibody (green), and DAPI (blue). One IκBζ+ and one IκBζ- samples are shown (representative of 3 each). Confocal analysis was performed at 63x. (D) Unselected CLL cells were treated with IRAK inhibitor and stimulated with CpG; 24 h later cells were analyzed for IκBζ expression by flow cytometry (n=7: mean and SEM are indicated) The Mann-Whitney test was used to analyze all the data. *P<0.05.

haematologica | 2017; 102(11)

1905

E. Fonte et al.

CpG induces metabolic cell activation, and modulates the sensitivity of CLL cells to fludarabine.

TLR9 stimulation differentially induces ΙκΒζ protein through a post-transcriptional mechanism.

To better characterize the molecular mechanism underlying differential induction of IκBζ among different patients’ samples, we measured NFKBIZ mRNA levels at different time points in cells from two subgroups of patients (the previously characterized “ATP+” and “ATP-” groups). NFKBIZ

mRNA was up-regulated by CpG in all patients’ samples at all time points (Figure 3A; “ATP+” cases: n=15 at 2 h; n=18 at 4 h; n=4 at 24 h. “ATP-” cases: n=9 at 2 h; n=9 at 4 h; n=4 at 24 h); the amount of mRNA detected peaked 2 h after CpG stimulation, remained high at 4 h, and declined in a similar manner in both groups after 24 h (Figure 3A). In order to understand whether different splicing isoforms are differentially induced by TLR stimulation in CLL samples, we performed a real-time polymerase chain reaction analysis with primers specifically recognizing the long

Figure 4. TLR9 stimulation unleashes IgM secretion from ΙκΒζ-positive cells. (A-B) IgM expression (% IgM-positive CLL cells in panel A and MFI levels in panel B) was analyzed by flow cytometry in CLL cells 72 h after CpG stimulation (20 patients’ samples analyzed; mean and SEM are indicated). Both IκBζ+ and IκBζ– samples are shown. Wilcoxon matched pairs test was performed and *indicates a P value <0.05. (C) IgM expression (MFI level) was analyzed by flow cytometry in CLL cells before and after CpG stimulation for 72 h. Both IκBζ+ and IκBζ– samples are shown (11 and 9 patients’ samples, respectively; mean and SEM are indicated). (D) CLL cells were stimulated with CpG for 96 h; 11 CLL samples were IκBζ+ and five were IκBζ–. IgM secretion was measured by ELISA (mean and SEM are indicated). A Mann-Whitney test was performed and *indicates P value <0.05. (E) Real-time analysis was performed in CLL cells stimulated with CpG for 48 h; eight CLL samples were IκBζ+ and six were IκBζ– (mean and SEM are indicated). A Mann-Whitney test was performed and * indicates a P value <0.05. (F) Two representative samples (out of 8; 6 IκBζ+ cases and 2 IκBζ– cases) were analyzed for Blimp-1, and β-actin (as an internal control) expression by western blot; cell lysates were prepared after 48 h of CpG treatment. Results of densitometric analysis of BLIMP-1/β-actin expression are reported above the blot (additional quantifications are reported in Online Supplementary Figure S6).

1906

haematologica | 2017; 102(11)

CpG induces ΙκΒζ and IgM in CLL

isoform A, in “ATP+” and “ATP-“ patients, 4 h after TLR stimulation. Both groups of patients expressed this IκBζ isoform; counter-intuitively, “ATP-” patients’ samples (n=4) showed even higher mRNA levels of the long isoform A, as compared to “ATP+” samples (n=4) (Online Supplementary Figure S5). Altogether, these data rule out differential mRNA expression of the long isoform as a probable cause of the differential induction of IκBζ protein by CpG in CLL cells. Strikingly, time course analyses of protein expression by western blot revealed that IκBζ was expressed soon after 4 h of CpG stimulation, with a peak at 6 h in IκBζ-positive samples; IκBζ protein remained detectable after 24 h of TLR stimulation in IκBζ-positive samples (Figure 3B for 1 representative case). In contrast, IκBζ protein was barely detectable if present in IκBζ-negative samples (Figure 3B for 1 representative sample). Densitometric analysis of the western blots for all the six samples studied is shown in Online Supplementary Figure S6. We further analyzed IκBζ expression by immunofluorescence microscopy, which revealed IκBζ mainly in the nucleus, in punctate structures of IκBζ-positive cases (Figure 3C); low to undedectable immunofluorescence for IκBζ was observed in IκBζ-negative cases before and after CpG treatment (Figure 3C). To dissect the signaling pathways involved in this specific IκBζ upregulation, we treated CpG-stimulated CLL cells with an IRAK1/4 inhibitor that blocks the activity of the

TLR-signaling complex including IRAK kinases; drug treatment significantly blunted IκBζ protein expression (Figure 3D), while no significant cell death was induced at this time point (data not shown). In conclusion, IκBζ protein is specifically upregulated after TLR9 ligation by an IRAK-dependent post-transcriptional mechanism in a distinct group of CLL samples.

TLR9 stimulation promotes IgM secretion in IκBζ-positive chronic lymphocytic leukemia It was recently shown that stimulation of CLL cells with a mix of CpG and cytokines can induce immunoglobulin secretion, at least in a proportion of cases.31-33 We hypothesized that IκBζ could mark and/or be involved in CpGinduced immunoglobulin production. To test this hypothesis, we first analyzed the expression of surface and intracellular IgM by flow cytometry, upon incubation of CLL cells in the presence or absence of CpG; we calculated both the percentage of IgM-positive cells and MFI of IgM expression. Surface IgM was similar in the two conditions (Figure 4A,B). Conversely, we found significant induction of intracellular IgM levels (MFI but not percentage of positive cells) after 72 h of CpG stimulation (Figure 4B). There was no difference between IκBζ-positive and IκBζ-negative patients in terms of intracellular IgM induction (Figure 4C). In contrast, when we quantified IgM secreted into the culture medium, we detected

Figure 5. CpG induced autophagy is required for IgM secretion. (A) Immunoblot analysis of converted LC3 (LC3-II) following treatment with CpG for the indicated times. (B) Autophagic fluxes were assessed as the rate of lysosomal digestion of LC3-II (quantiﬁed as the difference of LC3-II band intensity in the presence or absence of NH4Cl normalized by β-actin) by lysosomal inhibition with 50 mM NH4Cl for 1 h. One out of four experiments from one representative IκBζ+ CLL patient is shown. (C-D) IgM expression analysis was performed by flow cytometry (C) and its secretion was quantified by ELISA (D) on eight IκBζ+ CLL samples (mean and SEM are indicated). CLL cells were stimulated with CpG for 24 h and treated with 50 nM of bafilomycin for an additional 48 h. A Wilcoxon matched pairs test was performed and ** indicates a P<0.001.

haematologica | 2017; 102(11)

1907

E. Fonte et al.

significantly higher levels in the group of IκBζ-positive patients (Figure 4D). Next, we analyzed the expression of Activation-induced (cytidine) deaminase (AICDA or AID) and Blimp-1, both representing key regulators of antibody response. AID mRNA transcripts were significantly induced in the group of IκBζ-positive samples (Figure 4E). Moreover, a trend of increased Blimp-1 protein expression was observed 48 h after CpG stimulation (Figure 4F shows 2 representative IκBζ-positive samples); a densitometric analysis of different cases is reported in Online Supplementary Figure S6. As expected, at the same time point low but detectable levels of IκBζ protein were revealed (Figure 4F).

Autophagy flux induced by CpG is required for IgM secretion We analyzed autophagy, a selective lysosomal recycling strategy recently implicated in the differentiation of antibody-secreting cells,34 and in immunoglobulin secretion induced by TLR stimulation.35,36 We treated CLL cells with or without CpG, and analyzed the expression of LC3 protein, whose short-lived faster-migrating lipidated form (LC3-II, 14 kDa) is an established biochemical marker of autophagosomes. CpG induced autophagy in CLL cells, as indicated by increased steady-state abundance of LC3-II, which further accumulated under short treatment with the lysosomal inhibitor NH4Cl, attesting to increased autophag-

ic flux (Figure 5A,B). Moreover, when bafilomycin-A1, an inhibitor of autophagosome-lysosome fusion, was coadministered to CLL cells stimulated with CpG, IgM accumulated intracellularly (Figure 5C) at the expense of secretion (Figure 5D). Overall, these data reveal that TLR9 stimulation unleashes IgM secretion selectively in IκBζ-positive cases through an autophagy-dependent mechanism.

IκBζ expression and regulation in normal and leukemic B cells Having defined a novel molecular mechanism sustaining IκBζ overexpression in selected CLL cases, we compared the levels of this protein among different CLL cases and normal B-cell populations. We thus analyzed the expression of IκBζ in human purified B lymphocytes freshly isolated from buffy coats and tonsils by flow cytometry upon 24 h stimulation with CpG. As shown in Figure 6A,B, IκBζ was induced by TLR9 in both types of samples. Interestingly, CLL cells from “ATP+” cases showed higher proportions of IκBζ-positive cells and higher IκBζ MFI compared to circulating and tonsillar B cells, respectively (Figure 6C,D). Of note, 4 h after TLR9 stimulation, both circulating and tonsillar B cells expressed high levels of IκBζ protein which decreased rapidly over time, especially in the buffy-coat cells (Online Supplementary Figure S7). This further suggests that IκBζ-

Figure 6. IκBζ expression in normal and leukemic B cells. (A-B) Flow cytometry analysis of IκBζ expression in circulating B-lymphocytes (n=6 samples) and tonsillar B cells (n=4 samples). Cells freshly purified from buffy coat or tonsils were stimulated with CpG for 24 h and subsequently intracellularly stained with an anti- IκBζ antibody and analyzed as percentage of positive cells (A) or MFI (B). (C-D) Percentage of IκBζ+ cells (C) or MFI values (D) in “ATP+” CLL cells (n=33) were compared to those in “ATP-“ CLL (n=23), normal purified B cells from buffy coats (n=6) and tonsillar B cells (n=4). For each sample the data were normalized to the unstimulated condition. A Mann-Whitney test was used for the statistical analysis. * indicates a P<0.05.

1908

haematologica | 2017; 102(11)

CpG induces ΙκΒζ and IgM in CLL

negative CLL resemble normal B lymphocytes while IκBζpositive CLL show overexpression of this protein. To address whether IκBζ-positive cells resembled a specific sub-population of normal B cells, we analyzed IκBζ expression by flow cytometry in different B-cell populations in peripheral blood and tonsil. Naïve, memory, germinal center, transitional, activated, and plasmablast B-cell subsets were detected by specific surface markers (see Methods section and Online Supplementary Figures S1 and S2). Concomitant IκBζ immunolabeling demonstrated high percentages of IκBζ-positive cells among naïve and memory B lymphocytes in buffy coats (Figure 7A; naïve=58.4±9%; memory=31.2±8.7% mean and SD). In particular, a high percentage of peripheral blood B cells expressing CD38, excluding plasmablasts (CD38bright), were IκBζ-positive (58.8±14.5% mean and SD) whereas very few if any CD5+ B cells expressed IκBζ (Figure 7A). It is of note that CD38bright B cells from both buffy coats and tonsils stimulated by CpG were mainly plasmablasts (CD138 negative, not shown). Among memory B cells, switch memory and IgM memory expressed the highest levels of IκBζ (Figure 7B). Finally, when we analyzed tonsil samples we observed an enrichment of IκBζ-positive cells among naïve and memory subsets (naïve=50.3±9.2%, memory=36.3±11.7%) (Figure 7C). Among memory B cells, in the tonsils, IκBζ was expressed more by switch memory than by IgM memory and double-negative memory cells (Figure 7D).

Discussion The aim of our project was to identify the mechanism that accounts for the differential response of CLL cells to TLR9 in regulating metabolic cell activation and immunoglobulin secretion. We focused on the atypical IκBζ nuclear protein because it is specifically induced by TLR in different leukocyte populations; however, its role has never been characterized in CLL. We observed that CpG stimulation strongly induced IκBζ protein in a selected group of CLL samples characterized by CD38 expression and unmutated IGHV genes. Moreover, we observed that IκBζ was specifically induced in cells that showed metabolic cell activation. By knocking down IκBζ protein before TLR9 stimulation, we proved that it was essential for CpG-induced cell survival, spontaneous or following treatment with fludarabine. Interestingly, recent findings suggest that IκBζ is either mutated or upregulated in an aggressive type of lymphoma in which it controls cell survival.37,38 Along this line, our results suggest that IκBζ controls an oncogenic pathway relevant in mature B-cell neoplasia, and its dysregulation can occur via genetic mechanisms (mutations) or following microenvironmental stimulation (e.g. through TLR). Because IκBζ is also an essential transcriptional regulator during T-independent antibody responses,27 we focused on the control of immunoglobulin production in CLL. Leukemic cells are mature clonal B-lymphocytes characterized by the expression of intracellular and surface IgM and/or IgD, although other Ig classes can rarely be expressed. These cells are somehow blocked throughout terminal differentiation. However, different cytokines together with the TLR9 ligand CpG were recently shown to induce immunoglobulin secretion, at least in a group of haematologica | 2017; 102(11)

CLL cases.31-33,39 Nevertheless, the molecular mechanisms accounting for this regulation are poorly understood. Here we studied whether the observed upregulation of IκBζ in select CLL cases could also explain this dichotomy of response. Indeed, we observed that IκBζ is a hallmark of IgM-secreting CLL cells. Wagner et al. recently described a signaling framework that links TLR9 activation, in ZAP70positive CLL, with protection from apoptosis mediated by IgM secretion.40 Our results support and extend this observation, and suggest that TLR9-induced IgM secretion and metabolic cell activation in IκBζ-positive cases may play a key role in tuning CLL cell survival. We also observed that CpG stimulation induces IgM secretion through an autophagy-dependent pathway, and that IκBζ is overexpressed in CLL as compared to normal B cells. TLR9 signaling is known to induce autophagy in different cell types.41 Hence, while the observed induction of autophagy by CpG stimulation was expected, it was surprising to observe that IgM secretion – but not production – following TLR9 engagement was abolished by lysosomal inhibition in CLL cells. The underlying selective autophagic control on TLR-dependent immunoglobulin secretion warrants investigation. Being operative specifically in IκBζ-positive CLL cells, such a mechanism may unveil relevant prognostic markers and therapeutic targets. A previous publication reported a differential response of CLL cells to CpG (as well as interleukin-21) in terms of Blimp-1 induction and immunoglobulin secretion, while similar levels of IκBα and IκBζ mRNA expression were observed (no protein analysis of IκBζ was performed). Such a differential response was attributed to specific epigenetic regulation marked by H3K4 trimethylation and to the presence or absence of “poised” Blimp-1 promoter.33 It should be noted that IκBζ is a key regulator of the transcriptional activity of inflammatory cytokine genes during the socalled “secondary response”. Indeed, it can bind to “poised” gene promoters to facilitate the recruitment of the transcription pre-initiation complex and H3K4 trimethylation. Moreover, in mouse models, IκBζ mediates LPS-induced Blimp-1 induction in splenic B cells.27 Based on these observations and on our results, it is tempting to speculate that differentially induced IκBζ protein may directly regulate Blimp-1 promoter activity in CLL cells. The overexpression of IκBζ in a selected group of CLL cases may represent an abnormal characteristic of the malignant clone, or may reflect the phenotype of a putative cell of origin. The cellular origin of CLL is still debated;1,42 one study demonstrated that CLL cells are strongly related to memory B cells in terms of gene expression profile43 while another study showed a significant similarity between CLL cells and naïve and mature CD5+ B cells, suggesting a heterogeneous origin of CLL cells from CD5+ Bcell subpopulations (CD27- or CD27+).44 MicroRNAome analysis of CLL in comparison to normal B-cell subsets highlighted similarities of CLL cells to antigen-experienced and marginal zone–like B cells but also naïve B cells.30 When we analyzed IκBζ induction in different B-cell populations from buffy coat and tonsil we observed an enrichment of IκBζ positivity in both naïve and memory B-cells, which is in agreement with some aforementioned results supporting the hypothesis that CLL could resemble a subset of normal B cells intermediate between a naive B cell and antibodysecreting cells. Analyzing in detail memory B-cell subsets we observed 1909

E. Fonte et al.

that IκBζ was preferentially expressed in conventional memory B-cell subsets, both switched and not switched (CD27+), but was absent in double-negative memory B cells (IgD-CD27-) which are a special subpopulation particularly increased in systemic lupus erythematosus autoimmune disease.45,46 Interestingly, the CLL cases positive for CD38 were those that responded to stimulation with CpG with IκBζ induction; in a similar way, a high percentage of CD38+ normal B cells expressed IκBζ following TLR9 stimulation. Looking for the mechanisms regulating IκBζ protein expression, we demonstrated that while NFKBIZ mRNA was induced by CpG in all the samples analyzed, a spe-

cific post-transcriptional mechanism was operating in a selected group of CLL cases only. Of note, both transcriptional and post-transcriptional mechanisms of regulation of IκBζ have been documented.20-23 Nevertheless, this is the first example of selective regulation of IκBζ protein in CLL cases expressing CD38 and unmutated IGHV genes; this may underscore a novel biochemical framework that is specifically activated in a group of patients only, and that is highlighted by higher IκBζ induction. Overall, our results support a non-redundant role for TLR-induced IκBζ in regulating leukemic metabolic cell activation and suggest that a novel biochemical pathway regulating its protein expression levels may be implicated in immunoglobulin secretion toward the first step of dif-

Figure 7. ΙκΒζ expression in normal B-cell subpopulations. (A, B) Flow cytometry analysis of IκBζ expression in normal B cells enriched from buffy coat (peripheral blood; panel A) and in normal tonsillar B cells (panel C), after 4 h of culture in the presence of CpG. The box plots represent the mean values and the whiskers indicate minimum and maximum values of five samples analyzed. The subpopulations analyzed in the peripheral blood were: naïve, memory (MEM), plasmablasts and plasma cells (PBs+PC, CD38bright), CD5-positive B cells (CD5+) and CD38-positive B cells (CD38+). The B-cell subpopulations investigated in the tonsil, were: naïve, memory (MEM), germinal center (GC), plasmablasts and plasma cells (PBs+PC), transitional (TRANS) and activated (ACT). (B, D) Flow cytometry analysis of IκBζ expression in memory B cell subsets in the peripheral blood (panel B) or tonsil (panel D): IgM memory (M-MEM), switched memory (S-MEM) and double negative memory (DN-MEM) cells. Phenotypic features of the B-cell subsets and the gating strategy applied for the analysis are shown in Online Supplementary Figures S1 and S2.

1910

haematologica | 2017; 102(11)

CpG induces ΙκΒζ and IgM in CLL

ferentiation into antibody-secreting cells. Differentiation therapy has been proposed for different types of leukemia and lymphoma with the general aim of converting malignant clones into terminally differentiated cells which may be specifically targeted by other drugs. In this context, the TLR-induced - IκBζ regulated - CLL differentiation observed in this study may open new therapeutic perspectives. Finally, the molecular mechanisms that we have identified in CLL may shed light onto the regulation of these processes during physiological humoral immune responses as well as in autoimmune diseases.

References 1. Gaidano G, Foà R, Dalla-Favera R. Molecular pathogenesis of chronic lymphocytic leukemia. J Clin Invest. 2012;122(10):3432–3438. 2. Fowler NH, Cheah CY, Gascoyne RD, et al. Role of the tumor microenvironment in mature B-cell lymphoid malignancies. Haematologica. 2016;101(5):531–540. 3. Isaza-Correa JM, Liang Z, van den Berg A, Diepstra A, Visser L. Toll-like receptors in the pathogenesis of human B cell malignancies. J Hematol Oncol. 2014;7(1):499–510. 4. Ntoufa S, Vilia MG, Stamatopoulos K, Ghia P, Muzio M. Toll-like receptors signaling: a complex network for NF-κB activation in Bcell lymphoid malignancies. Semin Cancer Biol. 2016;39:15–25. 5. Wiestner A. BCR pathway inhibition as therapy for chronic lymphocytic leukemia and lymphoplasmacytic lymphoma. Hematology Am Soc Hematol Educ Program. 2014;2014(1):125–134. 6. O'Neill LAJ, Golenbock D, Bowie AG. The history of Toll-like receptors — redefining innate immunity. Nat Rev Immunol. 2013;13(6):453–460. 7. Ntoufa S, Vardi A, Papakonstantinou N, et al. Distinct innate immunity pathways to activation and tolerance in subgroups of chronic lymphocytic leukemia with distinct immunoglobulin receptors. Mol Med. 2012;18(9):1281–1291. 8. Longo PG, Laurenti L, Gobessi S, et al. The Akt signaling pathway determines the different proliferative capacity of chronic lymphocytic leukemia B-cells from patients with progressive and stable disease. Leukemia. 2007;21(1):110–120. 9. Tarnani M, Laurenti L, Longo PG, et al. The proliferative response to CpG-ODN stimulation predicts PFS, TTT and OS in patients with chronic lymphocytic leukemia. Leuk Res. 2010;34(9):1189–1194. 10. Efremov DG, Bomben R, Gobessi S, Gattei V. TLR9 signaling defines distinct prognostic subsets in CLL. Front Biosci (Landmark Ed). 2013;18:371–386. 11. Fonte E, Apollonio B, Scarfò L, et al. In vitro sensitivity of CLL cells to fludarabine may be modulated by the stimulation of Toll-like receptors. Clin Cancer Res. 2013;19(2):367– 379. 12. Tromp JM, Tonino SH, Elias JA, et al. Dichotomy in NF-kappaB signaling and chemoresistance in immunoglobulin vari-

haematologica | 2017; 102(11)

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Acknowledgments We thank Prof. Federico Caligaris-Cappio for his helpful suggestions and support. Funding This work was supported by Associazione Italiana per la Ricerca sul Cancro Milano, Italy (AIRC IG 16777 to MM, AIRC IG 15189 to PG, AIRC IG 14691 to SC and AIRC IG 14326 to Manlio Ferrarini; FIRC fellowship awarded to EF; AIRC Special Program Molecular Clinical Oncology – 5 per mille #9965 to MM, PG, and SC), and Ricerca Finalizzata 2010 (RF-20102318823 to PG) – Ministero della Salute, Roma, Italy.

able heavy-chain-mutated versus unmutated CLL cells upon CD40/TLR9 triggering. Oncogene. 2010;29(36):5071–5082. Bomben R, Gobessi S, Dal Bo M, et al. The miR-17 92 family regulates the response to Toll-like receptor 9 triggering of CLL cells with unmutated IGHV genes. Leukemia. 2012;26(7):1584–1593. Kitamura H, Kanehira K, Okita K, Morimatsu M, Saito M. MAIL, a novel nuclear I kappa B protein that potentiates LPS-induced IL-6 production. FEBS Lett. 2000;485(1):53–56. Haruta H, Kato A, Todokoro K. Isolation of a novel interleukin-1-inducible nuclear protein bearing ankyrin-repeat motifs. J Biol Chem. 2001;276(16):12485–12488. Yamazaki S, Muta T, Takeshige K. A novel IkappaB protein, IkappaB-zeta, induced by proinflammatory stimuli, negatively regulates nuclear factor-kappaB in the nuclei. J Biol Chem. 2001;276(29):27657–27662. Annemann M, Plaza-Sirvent C, Schuster M, et al. Atypical IκB proteins in immune cell differentiation and function. Immunol Lett. 2016;171:26–35. Schuster M, Annemann M, Plaza-Sirvent C, Schmitz I. Atypical IκB proteins - nuclear modulators of NF-κB signaling. Cell Commun Signal. 2013;11(1):23. Yamamoto M, Yamazaki S, Uematsu S, et al. Regulation of Toll/IL-1-receptor-mediated gene expression by the inducible nuclear protein IkappaBzeta. Nature. 2004;430 (6996):218–222. Dhamija S, Doerrie A, Winzen R, et al. IL-1induced post-transcriptional mechanisms target overlapping translational silencing and destabilizing elements in IκB mRNA. J Biol Chem. 2010;285(38):29165–29178. Yamazaki S, Muta T, Matsuo S, Takeshige K. Stimulus-specific induction of a novel nuclear factor-kappaB regulator, IkappaBzeta, via Toll/interleukin-1 receptor is mediated by mRNA stabilization. J Biol Chem. 2005;280(2):1678–1687. Ohba T, Ariga Y, MaruYama T, et al. Identification of interleukin-1 receptor-associated kinase 1 as a critical component that induces post-transcriptional activation of IκB-ζ. FEBS J. 2012;279(2):211–222. Rossato M, Curtale G, Tamassia N, et al. IL10-induced microRNA-187 negatively regulates TNF-α, IL-6, and IL-12p40 production in TLR4-stimulated monocytes. Proc Natl Acad Sci USA. 2012;109(45):E3101–3110. Okamoto K, Iwai Y, Oh-Hora M, et al. IkappaBzeta regulates T(H)17 development

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

by cooperating with ROR nuclear receptors. Nature. 2010;464(7293):1381–1385. Miyake T, Satoh T, Kato H, et al. IκB is essential for natural killer cell activation in response to IL-12 and IL-18. Proc Natl Acad Sci USA. 2010;107(41):17680–17685. Okuma A, Hoshino K, Ohba T, et al. Enhanced apoptosis by disruption of the STAT3-IκB-ζ signaling pathway in epithelial cells induces Sjögren's syndrome-like autoimmune disease. Immunity. 2013;38(3): 450–460. Hanihara-Tatsuzawa F, Miura H, Kobayashi S, et al. Control of Toll-like receptor-mediated T cell-independent type 1 antibody responses by the inducible nuclear protein IκB-ζ. J Biol Chem. 2014;289(45):30925– 30936. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008;111(12): 5446–5456. Colombo M, Cutrona G, Reverberi D, et al. Expression of immunoglobulin receptors with distinctive features indicating antigen selection by marginal zone B cells from human spleen. Mol Med. 2013;19:294–302. Negrini M, Cutrona G, Bassi C, et al. microRNAome expression in chronic lymphocytic leukemia: comparison with normal B-cell subsets and correlations with prognostic and clinical parameters. Clin Cancer Res. 2014;20(15):4141–4153. Gutierrez A, Arendt BK, Tschumper RC, et al. Differentiation of chronic lymphocytic leukemia B cells into immunoglobulin secreting cells decreases LEF-1 expression. PLoS ONE. 2011;6(10):e26056. Hoogeboom R, Reinten RJA, Schot J-J, et al. In vitro induction of antibody secretion of primary B-cell chronic lymphocytic leukaemia cells. Leukemia. 2014;29(1):244– 247. Duckworth A, Glenn M, Slupsky JR, Packham G, Kalakonda N. Variable induction of PRDM1 and differentiation in chronic lymphocytic leukemia is associated with anergy. Blood. 2014;123(21):3277–3285. Pengo N, Scolari M, Oliva L, et al. Plasma cells require autophagy for sustainable immunoglobulin production. Nat Immunol. 2013;14(3):298–305. Eriksen AB, Torgersen ML, Holm KL, et al.

1911

E. Fonte et al. Retinoic acid-induced IgG production in TLR-activated human primary B cells involves ULK1-mediated autophagy. Autophagy. 2015;11(3):460–471. 36. Arnold J, Murera D, Arbogast F, et al. Autophagy is dispensable for B-cell development but essential for humoral autoimmune responses. Cell Death Differ. 2015;23 (5):853–864. 37. Nogai H, Wenzel S-S, Hailfinger S, et al. IκBζ controls the constitutive NF-κB target gene network and survival of ABC DLBCL. Blood. 2013;122(13):2242–2250. 38. Zhao S, Dong X, Shen W, Ye Z, Xiang R. Machine learning-based classification of diffuse large B-cell lymphoma patients by eight gene expression profiles. Cancer Med. 2016;5(5):837–852.

1912

39. Ghamlouch H, Ouled-Haddou H, Guyart A, et al. TLR9 ligand (CpG oligodeoxynucleotide) induces CLL B-cells to differentiate into CD20(+) antibody-secreting cells. Front Immunol. 2014;5(1):292. 40. Wagner M, Oelsner M, Moore A, et al. Integration of innate into adaptive immune responses in ZAP-70-positive chronic lymphocytic leukemia. Blood. 2016;127(4):436– 448. 41. Shibutani ST, Saitoh T, Nowag H, Münz C, Yoshimori T. Autophagy and autophagyrelated proteins in the immune system. Nat Immunol. 2015;16(10):1014–1024. 42. Chiorazzi N, Ferrarini M. Cellular origin(s) of chronic lymphocytic leukemia: cautionary notes and additional considerations and possibilities. Blood. 2011;117(6):1781–1791.

43. Klein U, Tu Y, Stolovitzky GA, et al. Gene expression profiling of B cell chronic lymphocytic leukemia reveals a homogeneous phenotype related to memory B cells. J Exp Med. 2001;194(11):1625–1638. 44. Seifert M, Sellmann L, Bloehdorn J, et al. Cellular origin and pathophysiology of chronic lymphocytic leukemia. J Exp Med. 2012;209(12):2183–2198. 45. Bagnara D, Squillario M, Kipling D, et al. A reassessment of IgM memory subsets in humans. J Immunol. 2015;195(8):3716–3724. 46. Wei C, Anolik J, Cappione A, et al. A new population of cells lacking expression of CD27 represents a notable component of the B cell memory compartment in systemic lupus erythematosus. J Immunol. 2007;178 (10):6624–6633.

haematologica | 2017; 102(11)

ARTICLE

Non-Hodgkin Lymphoma

Efficacy and safety of subcutaneous and intravenous rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone in first-line diffuse large B-cell lymphoma: the randomized MabEase study

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Pieternella Lugtenburg,1 Irit Avivi,2 Henriette Berenschot,3 Osman Ilhan,4 Jean Pierre Marolleau,5 Arnon Nagler,6 Antonio Rueda,7 Monica Tani,8 Mehmet Turgut,9 Stuart Osborne,10 Rodney Smith11 and Michael Pfreundschuh12

Department of Hematology, Erasmus MC Cancer Institute, Rotterdam, the Netherlands; The Sackler Faculty of Medicine, Tel Aviv University, Israel; 3Department of Hematology, Alberts Schweitzer Hospital, Dordrecht, the Netherlands; 4Department of Hematology, Ankara University School of Medicine, Turkey; 5Unit of Hematology, University Hospital of Amiens, France; 6Division of Hematology, Chaim Sheba Medical Center, Tel Hashomer, Israel; 7Área de Oncología. Unidad de Oncología Médica, E.P. Hospital Costa del Sol, Marbella, Spain; 8Hematology Unit, Santa Maria Delle Croci Hospital, Ravenna, Italy; 9 Hematology Department, Ondokuz Mayis University, Samsun, Turkey; 10PDMA Operations (Biometrics), F. Hoffmann-La Roche Ltd., Basel, Switzerland; 11Pharma Development Oncology, F. Hoffmann-La Roche Ltd., Basel, Switzerland and 12 Department of Internal Medicine I, University Hospital of Saarland, Homburg, Germany 1 2

Haematologica 2017 Volume 102(11):1913-1922

ABSTRACT

Intravenous rituximab plus chemotherapy is standard treatment for diffuse large B-cell lymphoma. A subcutaneous formulation of rituximab is expected to simplify and shorten drug preparation and administration, and to reduce treatment burden. MabEase (clinicaltrials.gov Identifier: 01649856) examined efficacy, safety and patient satisfaction with subcutaneous rituximab plus chemotherapy in treatment-naïve patients with diffuse large B-cell lymphoma. Patients were randomized 2:1 to subcutaneous rituximab (intravenous 375 mg/m2 cycle 1; subcutaneous 1,400 mg cycles 2-8) or intravenous rituximab (375 mg/m2 cycles 1-8) plus cyclophosphamide, doxorubicin, vincristine, and prednisone every 14 or 21 days. The primary endpoint was investigatorassessed complete response/unconfirmed complete response. Secondary endpoints included safety, treatment satisfaction (Cancer Treatment Satisfaction Questionnaire and Rituximab Administration Satisfaction Questionnaire), time savings, and survival. Of 576 randomized patients, 572 (378 subcutaneous; 194 intravenous) received treatment. End of induction complete response/unconfirmed complete response rates were 50.6% (subcutaneous) and 42.4% (intravenous). After a median 35 months, median overall, event-free and progression-free survivals were not reached. Grade ≥3 adverse events (subcutaneous 58.3%; intravenous 54.3%) and administration-related adverse events (both groups 21%) were similar between arms. Injection-site reactions were more common with subcutaneous injections (5.7% versus 0%, respectively). Rituximab Administration Satisfaction Questionnaire scores for ‘impact on activities of daily living’, ‘convenience’, and ‘satisfaction’ were improved with subcutaneous versus intravenous injections; Cancer Therapy Satisfaction Questionnaire scores were similar between arms. Median administration time (6 minutes vs. 2.6 to 3.0 hours), chair/bed and overall hospital times were shorter with subcutaneous versus intravenous rituximab. Overall, subcutaneous and intravenous rituximab had similar efficacy and safety, with improved patient satisfaction and time savings. haematologica | 2017; 102(11)

Correspondence: p.lugtenburg@erasmusmc.nl

Received: June 2, 2017. Accepted: September 14, 2017. Pre-published: September 21, 2017. doi:10.3324/haematol.2017.173583 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1913 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1913

P. Lugtenburg et al.

Introduction Intravenous (IV) rituximab plus chemotherapy is standard treatment for diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL)1,2 on the basis of improved efficacy versus chemotherapy alone.3-8 In addition to the primary treatment goal of optimizing patient outcomes, such as response rate, progression-free and overall survival, simplifying treatment and reducing treatment burden are also important aims for patients and healthcare providers. Subcutaneous (SC) dosing has the potential to simplify administration, reduce the treatment burden for patients, and reduce resource utilization at the treatment facility.9-13 To address these needs in patients with non-Hodgkin lymphoma (NHL), a SC formulation of rituximab has been developed. Two studies using a pharmacokinetic and clinical bridging approach, SABRINA and SparkThera, have demonstrated pharmacokinetic noninferiority for rituximab SC compared with the IV formulation in patients with FL.14-16 In SABRINA, when given in a three-weekly dosing schedule as first-line treatment, geometric mean rituximab trough concentrations at cycle 7 were 83.13 mg/mL for rituximab IV and 134.58 mg/mL for rituximab SC, with comparable efficacy and safety between formulations.14,15 In addition, the phase Ib SAWYER study demonstrated pharmacokinetic non-inferiority and similar safety profiles for rituximab SC and IV (both with fludarabine and cyclophosphamide) in firstline chronic lymphocytic leukemia patients.17,18 The accumulating clinical data, along with the established efficacy and safety of rituximab IV,19,20 supported the authorization of rituximab SC for patients with NHL, and rituximab SC has subsequently been approved in Europe and elsewhere for the treatment of DLBCL and FL.21-23 The SC formulation takes approximately 5 minutes to administer versus 1.5 to 6 hours for rituximab IV,19,21 and studies have confirmed that this new presentation offers improved patient convenience and healthcare resource savings over the IV form.13,24 Herein we report the final analysis of the MabEase study in patients with previously untreated CD20+ DLBCL, an aggressive form of NHL that is treated with curative intent. The objectives of MabEase were to examine the efficacy and safety of rituximab SC versus the IV formulation as part of a rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (RCHOP) regimen. Patient satisfaction with treatment was also assessed.

Methods Study design MabEase (clinicaltrials.gov Identifier: 01649856) is a phase IIIb, multicenter, randomized, open-label study. The study was conducted in line with International Conference on Harmonisation E6 guidelines for Good Clinical Practice and the Declaration of Helsinki. The study protocol was approved by independent ethics committees at each center. The first patient was enrolled on 22 August 2012, and the data cut-off for the current analysis was 18 September 2016. All patients provided written informed consent. Additional methodological details are provided in the Online Supplementary Appendix.

1914

Patients Eligible patients were aged 18–80 years with untreated histologically confirmed CD20+ DLBCL, International Prognostic Index (IPI) 1–5 or IPI 0 with bulky disease (one lesion ≥7.5 cm), had at least one bidimensionally measurable lesion ≥1.5 cm at its largest dimension by computed tomography (CT), positron emission tomography-CT (PET-CT), or magnetic resonance imaging (MRI), had adequate hematologic function, and Eastern Cooperative Oncology Group (ECOG) performance status ≤2 (detailed inclusion/exclusion criteria are supplied in the Online Supplementary Appendix).

Randomization Patients were randomized 2:1 via a centralized interactive voice/web response system to receive rituximab SC or IV, and were stratified according to age (<60 or ≥60 years), IPI risk category (low, low-intermediate, high-intermediate, high), and chemotherapy regimen (CHOP-14, CHOP-21).

Procedures All patients were scheduled to receive eight cycles of rituximab in accordance with the prescribing information for rituximab in DLBCL.19-21 In addition, patients received six to eight cycles of CHOP chemotherapy every 14 (CHOP-14) or 21 (CHOP-21) days (see the Online Supplementary Appendix). The planned CHOP regimen for each patient was chosen by the center prior to randomization. However, patients scheduled to receive eight cycles of CHOP who achieved a complete response (CR)/unconfirmed complete response (CRu) after cycle 4 could be reduced to two additional rituximab plus CHOP cycles (for a total of six CHOP cycles), followed by two rituximab monotherapy cycles (for a total of eight rituximab cycles). Patients randomized to rituximab SC received rituximab IV 375 mg/m2 on day one of cycle 1, then rituximab SC 1,400 mg on day one of the subsequent seven cycles (eight rituximab doses in total). Patients randomized to the IV arm received rituximab 375 mg/m2 on day one of each cycle. All patients received rituximab IV during cycle 1 in order to allow appropriate intervention in the event of an administration-related reaction (ARR).

Outcomes The primary endpoint was investigator-assessed CR/CRu rate according to Cheson 1999 criteria25 at the end of induction (EOI) in the intent-to-treat (ITT) population. Secondary endpoints included patient satisfaction measured by the Cancer Therapy Satisfaction Questionnaire (CTSQ) and Rituximab Administration Satisfaction Questionnaire (RASQ) and time savings; namely rituximab administration time, chair/bed time, and hospital time. Other secondary endpoints were progression-free survival (PFS), event-free survival (EFS), disease-free survival (DFS), and overall survival (OS). The end of the study was defined as the last patient visit in the follow-up period when all patients had been followed for at least 24 months after their last dose of study treatment.

Statistical analyses This was a descriptive study designed to exclude major differences in efficacy between treatment arms as measured by CR/CRu rates. There was no formal statistical hypothesis for treatment comparison. The calculated sample size (600 patients) was based on the primary endpoint, and no power calculation was performed for PFS. The final analysis was performed in the ITT population when the last patient had completed at least 24 months of follow up

haematologica | 2017; 102(11)

SC vs. IV rituximab plus CHOP in 1st line DLBCL

after the EOI, or when one of the following had been documented for all randomized patients: disease recurrence, withdrawal from study, loss to follow up or death, or whichever occurred first. An exploratory analysis was conducted to investigate potential correlations between the incidence of adverse events (AEs) and serious adverse advents (SAEs) in different body surface area (BSA) categories and age, sex, or type of AE (MedDRA v17.1 System Organ Class preferred terms).

Results

Figure 1). In the randomized patient population, 311 of 381 SC patients and 159 of 195 IV patients completed the treatment period. There were 217 (69.8%) SC patients and 117 (73.6%) IV patients still in follow up at the end of the study. The median duration of rituximab exposure was 4.9 months in both groups (safety population). Overall, 82.2% of patients received eight rituximab cycles. Median durations of CHOP exposure were 4.6 and 4.4 months in the rituximab SC and IV arms, respectively. Overall, 85.7%, 47.4% and 46.7% of patients received six, seven, and eight CHOP cycles, respectively.

Study population Overall, 576 patients were enrolled and randomized (SC, 381; IV, 195) from 151 in- or outpatient treatment centers in 25 countries. Of these, 572 received at least one dose of rituximab; however, nine patients randomized to receive rituximab SC only received the first rituximab IV infusion. The safety population therefore comprised 369 patients in the SC arm and 203 in the IV arm (369 and 188, respectively, from cycle 2 onwards). The ITT population comprised 519 patients (SC, 342; IV, 177; Figure 1). Baseline demographics and disease characteristics were balanced between treatment groups (Table 1). Of 576 randomized patients, 102 (SC, 67 [17.6%]; IV, 35 [17.9%]) discontinued study treatment before the end of cycle 8, predominantly because of AEs (SC, 36 [9.4%]; IV, 15 [7.7%]), and 217 patients (SC, 146 [38.3%]; IV, 71 [36.4%]) withdrew from the study altogether, predominantly because of death (SC, 73 [19.2%]; IV, 30 [15.4%];

Efficacy In the ITT population at EOI, rates of investigatorassessed CR/CRu (95% CI) were 50.6% (45.3%–55.9%) and 42.4% (35.1%–49.7%), P=0.076, in the SC and IV groups, respectively. Partial response (PR) and progressive disease (PD) rates were similar between treatment arms. CR/CRu rates (95% CI) for all randomized patients were 45.7% (40.7%–50.7%) for rituximab SC and 38.5% (31.6%–45.3%) for IV, P=0.099. When stratified by age, sex, BSA, CHOP regimen, and IPI score, statistically significantly higher CR/CRu rates with SC treatment versus IV were seen in patients with low-intermediate IPI scores along with a trend towards higher rates in patients aged ≥60 years (Table 2). Overall, CR/CRu rates were higher in patients receiving CHOP-21 than CHOP-14 (Table 2), although the latter regimen was used in only small numbers of patients.

Figure 1. Patient disposition. *Nine patients randomized to SC only received the first IV infusion and were analyzed as IV in the safety population. AE: adverse event; ITT: intent-to-treat; IV: intravenous: SC: subcutaneous; PD: progressive disease; SD: stable disease.

haematologica | 2017; 102(11)

1915

P. Lugtenburg et al.

After a median follow-up of 35 months, median survival in the ITT population was not reached for PFS, EFS, DFS or OS. Statistical analyses showed no significant differences between treatment groups (Figure 2). At the time of the final data lock, 56 of 342 rituximab SC patients (16.4%) had died, while 22 of 177 (12.4%) rituximab IV patients had died. Analysis of PFS and EFS showed that 72.2% of SC and 78.5% of IV patients had not progressed, relapsed or died, while 66.1% of SC and 71.2% of IV patients had not experienced an event (Figure 2). PFS was also generally similar between treatments for the subgroups with high SC CR/CRu rates (aged ≥60 years and with low-intermediate IPI scores; Table 2; Online Supplementary Appendix), although a higher proportion of SC patients receiving CHOP-14 had progressed, relapsed or died (14/36; 38.9% vs. 2/22; 9.1%; P=0.041). In addition, significantly higher proportions of SC patients with low BSA had progressed, relapsed or died (43/115; 37.4% vs. 9/56; 16.1%; P=0.01) or experienced an event (51/115; 44.3% vs. 13/56; 23.2%; P=0.02). At 24 months of follow up, PFS (95% CI) was 75.0% (69.9%–79.4%) in the SC group and 81.5% (74.7%– 86.6%) in the IV group (P=0.175), and EFS (95% CI) was 68.6% (63.3%–73.4%) and 73.4% (66.0%– 79.4%), respectively (P=0.456).

Safety Safety profiles were similar between arms, with no unexpected safety signals (Online Supplementary Appendix). Most AEs were grade 2 or 3 (339 [60.9%] of 557 patients in the safety population with cycle 2 dosing or beyond completed). In cycle 2 or later (all patients received rituximab IV in cycle 1), 58.3% of SC and 54.3% of IV patients experienced at least one AE of grade ≥3. ARRs were reported in 20.9% of SC patients and 21.3% of IV patients on or after cycle 2. Ten patients (2.7%) receiving rituximab SC experienced an ARR of grade ≥3 on or after cycle 2, compared with nine patients (4.8%) receiving IV. Injection site reactions were reported by 5.7% of patients receiving SC therapy; there were no such reactions with IV administration (P=0.0002). One injection site reaction (an episode of injection site pain in the SC group) was grade ≥3; the remainder were grade <3. In cycle 2 or later, 141 SC (38.2%) and 62 IV patients (33.0%) reported at least one SAE, most commonly febrile neutropenia (FN), neutropenia, and pneumonia. A higher proportion of patients experienced FN as an SAE in the SC versus the IV arm (11.7% vs. 6.4%, P=0.0515), consistent with the higher incidence of grade 3/4 FN in the SC arm (12.5% vs. 6.9%, P=0.0575). A similar proportion of patients in each group in the safety population discontinued rituximab treatment because of AEs (SC, 30 [8.1%]; IV, 19 [9.4%]), the most common of which were infections and infestations (2.4% and 2.5% of patients in the SC and IV groups, respectively; all treatment cycles). When cycle 1 was excluded, discontinuation rates were 7.9% for SC and 5.3% for IV rituximab; infection and infestation rates were 2.4% (SC) and 1.1% (IV). More SC patients (138 [37.4%]) had an interruption/delay in their rituximab treatment due to AEs when compared with the IV arm (56 [27.6%]; safety population, all cycles). The most common reasons (>2% of patients) were neutropenia (SC, 34 [9.2%]; IV, 14 [6.9%]), FN (SC, 13 [3.5%]; IV, 1 [0.5%]), pneumonia (SC, 8 [2.2%]; IV, 6 [3.0%]), neutrophil count decreased (SC, 20 [5.4%]; 1916

IV, 8 [3.9%]), and white blood cell count decreased (SC, 10 [2.7%]; IV, 3 [1.5%]). Death due to treatment-emergent AEs (i.e., grade 5 AE) was reported in 32 patients in the SC group (8.7%) and 14 in the IV group (6.9%). The main causes were infections and infestations (SC, 14 [3.8%]; IV, 4 [2.0%]), cardiac disorders (SC, 4 [1.4%]; IV, 4 [2.0%]), and respiratory, thoracic, and mediastinal disorders (SC, 2 [0.5%]; IV, 3 [1.5%]). In the exploratory analysis, treatment effect on grade ≥3 AEs and SAEs (i.e., rituximab SC vs. IV) was not modified by BSA, age group or sex (Online Supplementary Appendix).

Treatment satisfaction Mean RASQ scores for 'impact on activities of daily living', 'convenience', and 'satisfaction' were improved with SC versus IV rituximab. Overall, 428 patients (SC, 284; IV, 144) completed the RASQ at cycle 7 and were included in the RASQ analysis. The mean RASQ scores were higher across all domains for rituximab SC versus IV (Table 3), with mean satisfaction scores of 89.6 and 77.4 for the SC and IV groups, respectively (Figure 3). More patients in the rituximab SC group versus the IV group thought that

Table 1. Patient baseline demographics and disease characteristics (safety population).

Characteristic

Median age, years (range) <60 ≥60 Sex Male Female Median BSA, m2 (range) BSA Low (≤1.7) Medium (>1.7-≤1.9) High (>1.9) ECOG PS 0 1 2 Ann Arbor stage I–II III–IV IPI score Low/low-intermediate High/high-intermediate Chemotherapy regimen CHOP-14 CHOP-21

Rituximab SC plus CHOP (n=369)

Rituximab IV plus CHOP (n=203)

64.0 (18-80) 142 (38.5) 227 (61.5)

64.0 (24-80) 79 (38.9) 124 (61.1)

204 (55.3) 165 (44.7) 1.83 (1.35-2.62)

103 (50.7) 100 (49.3) 1.84 (1.27-2.79)

122 (30.4) 119 (32.2) 138 (37.4)

59 (29.1) 68 (33.5) 76 (37.4)

180 (48.8) 148 (40.1) 41 (11.1)

105 (51.7) 73 (36.0) 25 (12.3)

122 (33.1) 247 (66.9)

61 (30.0) 142 (70.0)

224 (60.7) 145 (39.3)

126 (62.1) 77 (37.9)

36 (9.8) 333 (90.2)

22 (10.8) 181 (89.2)

Data are n (%), unless otherwise indicated. BSA: body surface area; CHOP: cyclophosphamide, doxorubicin, vincristine, and prednisone; CHOP-14: CHOP 14-day cycle; CHOP-21: CHOP 21-day cycle; ECOG PS: Eastern Cooperative Oncology Group performance status; IPI: International Prognostic Index; IV: intravenous; SC: subcutaneous. .

haematologica | 2017; 102(11)

SC vs. IV rituximab plus CHOP in 1st line DLBCL

the length of time taken for the SC injection/IV infusion was 'just right' (78.9% SC vs. 57.6% IV). When patients in the SC group were asked which treatment they would prefer, if given the option, 90.8% stated a preference for SC over IV. CTSQ scores were similar between arms (Figure 3 and Table 3). A total of 421 patients were included in the CTSQ analysis, with 421 (280 SC and 141 IV) completing the questionnaire at cycle 7. The mean CTSQ satisfaction score and scores for individual domains were similar

between the treatment arms (Table 3). RASQ and CTSQ results for cycle 3 were similar to those for cycle 7.

Time savings The median administration time (cycles 2â&#x20AC;&#x201C;8) was substantially shorter for SC (6 minutes) than IV rituximab (range: 2.6 to 3.0 hours). Chair/bed and overall hospital times were also shorter with SC treatment. In cycle 2, 82.9% of patients in the SC arm had a chair/bed time â&#x2030;¤4 hours, whereas 61.2% in the IV arm had a chair/bed

D Figure 2. Secondary time-to-event endpoints for rituximab SC and rituximab IV (intent-to-treat population). Analyses presented are (A) progression-free survival, (B) event-free survival, (C) disease-free survival, and (D) overall survival. CI: confidence interval; DFS: disease-free survival; EFS: event-free survival; HR: hazard ratio; IV: intravenous; OS: overall survival; PFS, progression-free survival; SC: subcutaneous.

haematologica | 2017; 102(11)

1917

P. Lugtenburg et al. time ≥4 hours. For each cycle from cycle 2 onwards, a higher proportion of SC than IV patients spent <2 hours in a chair/bed receiving rituximab (ranges: 27%–56% SC vs. <1%–5% IV). In cycle 2, 64.8% of SC patients required ≤6 hours of hospital time overall, whereas 51.6% of those receiving rituximab IV required ≥6 hours.

Discussion MabEase was a randomized, phase IIIb study designed to exclude major differences in efficacy and safety between rituximab SC and IV treatment arms in newly diagnosed DLBCL patients treated with R-CHOP. The primary endpoint results indicated similar efficacy of the rituximab SC and IV formulations in our overall study population. We note that patient demographics in MabEase differed from those in many key trials of R-CHOP in DLBCL patients. For example, the patient population in the MabThera International Trial group (MInT) study was

younger (18–60 years), with low IPI risk scores,4 whereas the populations in the LNH98.53 LNH03-6B,26 and rituximab with CHOP over age 60 years (RICOVER-60)8 studies were older (60–80 years). However, we suggest that comparisons of outcomes with these studies, although indirect, are valid and informative once demographic differences are taken into consideration. Overall, previous trials of R-CHOP regimens including rituximab IV reported CR/CRu rates ranging from 58% to 86% in patients with previously untreated DLBCL.3,4,8,26,27 The CR/CRu rates in our study (SC, 50.6%; IV, 42.4%) were lower, which may relate to the number of patients in the ITT population who did not complete the planned course of treatment. The MabEase ITT population included all patients who completed a baseline assessment and at least one on-treatment efficacy assessment, with the first efficacy assessment conducted at interim staging at the end of cycle 4. However, approximately 18% of patients discontinued study treatment before the end of cycle 8, predominantly because of AEs. For SC and IV patients who completed all eight cycles of induction,

Table 2. Efficacy endpoints at EOI treatment.

Efficacy endpoint, % (95% CI)

Rituximab SC plus CHOP*

Rituximab IV plus CHOP

177 177 177 177

42.4 (35.1-49.7) 35.6 (28.6-43.1) 6.2 (3.1-10.8) 78.0 (71.1-83.8)

0.076 -

195

38.5 (31.6-45.3)

0.099

CR/CRu PR PD ORR

342 342 342 342

CR/CRu Subgroups (all-randomized population) Age (years) <60 ≥60 Sex Male Female CHOP regimen CHOP-14 CHOP-21 IPI score Low Low-intermediate High-intermediate High BSA (m2) Low (≤1.7) Medium (1.7-1.9) High (>1.9)

381

ITT population 50.6 (45.3-55.9) 31.6 (26.7-36.8) 3.8 (2.0-6.4) 82.2 (77.7-86.1) All-randomized population 45.7 (40.7-50.7)

147 234

42.9 (34.9-50.9) 47.4 (41.0-53.8)

76 119

40.8 (29.7-51.8) 37.0 (28.3-45.6)

0.767 0.062

209 172

46.9 (40.1-53.7) 44.2 (36.8-51.6)

100 95

37.0 (27.5-46.5) 40.0 (30.1-49.9)

0.102 0.508

36 345

41.7 (25.6-57.8) 46.1 (40.8-51.3)

22 173

22.7 (5.2-40.2) 40.5 (33.1-47.8)

0.146 0.224

118 114 94 55

50.0 (41.0-59.0) 52.6 (43.5-61.8) 41.5 (31.5-51.4) 29.1 (17.1-41.1)

61 57 47 30

54.1 (41.6-66.6) 35.1 (22.7-47.5) 31.9 (18.6-45.2) 23.3 (8.2-38.5)

0.603 0.031 0.272 0.569

115 123 143

44.3 (35.3-53.4) 45.5 (36.7-54.3) 46.9 (38.7-55.0)

56 65 74

42.9 (29.9-55.8) 38.5 (26.6-50.3) 35.1 (24.3-46.0)

0.854 0.353 0.099

*Three patients initially recorded as CR/CRu in the SC arm subsequently had their response downgraded to PR due to bone marrow data analysis. BSA: body surface area; CHOP: cyclophosphamide, doxorubicin, vincristine, and prednisone; CHOP-14: CHOP 14-day cycle; CHOP-21: CHOP 21-day cycle; CI: confidence interval; CR: complete response; CRu: complete response unconfirmed; EOI: end of induction; IPI: International Prognostic Index; ITT: intent-to-treat; IV: intravenous; ORR: overall response rate; PD: progressive disease; PR: partial response; SC: subcutaneous.

1918

haematologica | 2017; 102(11)

SC vs. IV rituximab plus CHOP in 1st line DLBCL

CR/CRu rates were 57% and 47%, respectively. In addition, distinguishing PR from CRu using a CT scan alone is difficult. Current criteria recommend using PET scans where possible in order to better delineate disease extent and treatment response in DLBCL.28 However, due to limited PET availability, the MabEase study protocol prespecified the use of CT scans only for all tumor assessments. This may also have contributed to the apparently low CR rate. Another limitation was the lack of centralized radiologic review. For these aforementioned reasons the CR rates in MabEase should be interpreted with caution.

Despite the lower CR rate, the overall response rate in our study (CR/CRu plus PR; approximately 80%) was similar to observations in previous studies.3,4,8,26 Some trends towards higher CR/CRu rates were seen with SC treatment in some subgroups, but these did not translate into improvements in PFS or EFS, with a small number of subgroups showing increased rates of progression, relapse or death, or increased event rates. We note that these analyses were exploratory only, however, and that patient numbers in the subgroups were too small to permit any conclusions to be drawn. In particular, very

Figure 3. Patient satisfaction and preference. (A) Patient satisfaction assessed by (i) RASQ and (ii) CTSQ at cycles 3 and 7, (B) Time taken to receive SC injection/IV infusion (RASQ individual question) at cycle 7, (C) Treatment preferences (RASQ individual question) at cycles 3 and 7. CTSQ: Cancer Therapy Satisfaction Questionnaire; IV: intravenous; RASQ: Rituximab Administration Satisfaction Questionnaire; SC: subcutaneous.

haematologica | 2017; 102(11)

1919

P. Lugtenburg et al. Table 3. Mean (SD) RASQ and CTSQ scores at cycles 3 and 7 (ITT RASQ and CTSQ populations).

Domain RASQ Physical impact Psychological impact Impact on activities of daily living Convenience Satisfaction CTSQ Expectations of therapy Feelings about side effects Satisfaction with therapy

Visit

Rituximab SC

Rituximab IV

Cycle 3 Cycle 7 Cycle 3 Cycle 7 Cycle 3 Cycle 7 Cycle 3 Cycle 7 Cycle 3 Cycle 7

87.1 (12.9) 86.2 (14.0) 84.2 (14.2) 85.7 (13.9) 83.0 (16.8) 83.8 (16.1) 82.3 (13.5) 82.3 (13.4) 87.6 (12.9) 89.6 (12.1)

84.5 (15.1) 81.5 (16.8) 77.3 (17.4) 78.7 (18.2) 60.4 (20.0) 57.4 (19.2) 62.4 (19.8) 60.1 (17.5) 78.3 (16.9) 77.4 (18.2)

Cycle 3 Cycle 7 Cycle 3 Cycle 7 Cycle 3 Cycle 7

79.7 (17.8) 79.4 (17.4) 63.4 (18.7) 60.7 (21.6) 86.0 (11.1) 85.9 (11.4)

82.1 (18.1) 82.9 (16.5) 62.7 (21.2) 57.6 (23.3) 83.3 (12.6) 83.6 (13.5)

CTSQ: Cancer Therapy Satisfaction Questionnaire; ITT: intent-to-treat; IV: intravenous; RASQ: Rituximab Administration Satisfaction Questionnaire; SC: subcutaneous; SD: standard deviation.

few patients (approximately 10%) received CHOP-14 in MabEase, and it is therefore not possible to draw meaningful comparisons with response rates in the CHOP-21 or overall populations. Survival results overall were similar for both formulations in the ITT population. In MabEase, similar safety profiles were observed in the SC and IV arms. There were no new safety signals, and the rate of treatment-related deaths was comparable with rates reported in other studies.3,8,26 ARRs with rituximab have been well characterized in previous studies, particularly during cycle 1.29 To minimize ARR risk, rituximab was infused at a low initial rate, which was then increased incrementally. All patients received rituximab IV during cycle 1. Consistent with the reported SparkThera (phase Ib) and SABRINA (phase III) studies,14-16 we observed a higher rate of ARRs with rituximab SC versus IV. As expected, injection site reactions were more common with SC than with IV treatment, but these were mostly mild/moderate (< grade 3) and manageable. The most frequent SAE in this study was FN (SC, 11.7%; IV 6.4%), although there was no difference between groups in rates of treatment discontinuation due to AEs or infections. FN and grade 3 neutropenia were more frequently reported in the rituximab SC arm, and we do not have a comprehensive explanation for this observation. However, in DLBCL studies with rituximab dose intensification, and thus higher rituximab serum levels, more neutropenia and/or FN were also reported.30,31 Compared with other studies in DLBCL, the overall incidence of FN in our study (9.9%) was similar to that reported by Cunningham et al. among patients receiving the CHOP-21 regimen (11%),27 and to specific analysis of FN among the DLBCL cohort of the PrefMab crossover phase IIIb study in which patients with DLBCL or FL received rituximab SC and IV in different sequences (9.4%; data 1920

not reported by Rummel et al.).24 Analysis of patient subgroups showed a trend towards higher incidence of AEs and SAEs in patients with low BSA. These effects were not significant, however, and no significant interaction effect was found for AEs of grade â&#x2030;Ľ3 or SAEs for any of the covariates (BSA, age, or sex). CTSQ results showed that patients had similar levels of satisfaction with the R-CHOP treatment when used with either SC or IV rituximab, which is consistent with CTSQ data obtained in the PrefMab study.24 However, RASQ data suggested that most rituximab SC patients, given the option, preferred to have the SC injection over the IV infusion; again this finding reiterates the preference for rituximab SC expressed by 80.7% of patients in PrefMab.24 Consistent with our findings, in PrefMab, rituximab SC scored more highly for satisfaction with therapy (87.5% vs. 75.0% for IV), impact on activities of daily living and convenience of therapy (both 83.3% vs. 58.3%).24 Of note, the CTSQ was designed for use in a wide range of cancer types and stages.32 In contrast, the RASQ was developed specifically for the assessment of patientsâ&#x20AC;&#x2122; perceptions of the impact of treatment administration route.33 The use of rituximab SC resulted in substantial savings in clinic time. These findings concur with a time and motion analysis based on data collected within the MabCute study34 in patients with indolent NHL, which evaluated aspects of SC and IV administration of rituximab in real-world clinical practice.13 Reductions in chair time could potentially reduce waiting lists, increase the efficiency of oncology units, and increase the availability of appointments. In addition, the healthcare practitioner time gained could be deployed in other activities. Cost minimization data from The Netherlands also indicate the potential for cost savings with rituximab SC when comhaematologica | 2017; 102(11)

SC vs. IV rituximab plus CHOP in 1st line DLBCL

pared with IV dosing.35 Although reductions in clinic time could also reduce the opportunity for healthcare professionals to provide patient support, the RASQ findings indicated that patients felt that this factor did not compromise their treatment, and that they had sufficient time to discuss their treatment with their healthcare providers. In conclusion, the MabEase study showed no major differences between the efficacy of rituximab SC and IV therapy in treatment-naĂŻve patients with DLBCL. Overall, safety was similar between arms but with a higher incidence of FN and injection site reactions in the rituximab SC arm. The higher RASQ scores in the rituximab SC arm suggest that patient satisfaction, convenience, and effect on daily living were improved with rituximab SC compared with IV. Attempts to improve care for patients with

References 1. Tilly H, Gomes da Silva M, Vitolo U, et al. Diffuse large B-cell lymphoma (DLBCL): ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2015;26(Suppl 5):v116-125. 2. Dreyling M, Ghielmini M, Marcus R, Salles G, Vitolo U, Ladetto M. Newly diagnosed and relapsed follicular lymphoma: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2014; 25(Suppl 3):iii76-82. 3. Coiffier B, Thieblemont C, Van Den Neste E, et al. Long-term outcome of patients in the LNH-98.5 trial, the first randomized study comparing rituximab-CHOP to standard CHOP chemotherapy in DLBCL patients: a study by the Groupe d'Etudes des Lymphomes de l'Adulte. Blood. 2010;116(12):2040-2045. 4. Pfreundschuh M, Trumper L, Osterborg A, et al. CHOP-like chemotherapy plus rituximab versus CHOP-like chemotherapy alone in young patients with good-prognosis diffuse large-B-cell lymphoma: a randomised controlled trial by the MabThera International Trial (MInT) Group. Lancet Oncol. 2006;7(5):379-391. 5. Marcus R, Imrie K, Belch A, et al. CVP chemotherapy plus rituximab compared with CVP as first-line treatment for advanced follicular lymphoma. Blood. 2005;105(4):1417-1423. 6. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximab compared with CHOP alone in elderly patients with diffuse large-B-cell lymphoma. N Engl J Med. 2002;346(4):235-242. 7. Habermann TM, Weller EA, Morrison VA, et al. Rituximab-CHOP versus CHOP alone or with maintenance rituximab in older patients with diffuse large B-cell lymphoma. J Clin Oncol. 2006;24(19):31213127. 8. Pfreundschuh M, Schubert J, Ziepert M, et al. Six versus eight cycles of bi-weekly CHOP-14 with or without rituximab in elderly patients with aggressive CD20+ Bcell lymphomas: a randomised controlled trial (RICOVER-60). Lancet Oncol. 2008;9(2):105-116. 9. Lundin J, Kimby E, Bjorkholm M, et al. Phase II trial of subcutaneous anti-CD52 monoclonal antibody alemtuzumab

haematologica | 2017; 102(11)

10.

11.

12.

13.

14.

15.

16.

17.

18.

DLBCL, including short treatment intervals, consolidation with high-dose chemotherapy and autologous stem cell transplantation and replacement of rituximab with the type II CD20 antibody obinutuzumab, have so far been unsuccessful.29,36-38 Combined with previous evidence, the results of this study provide support for the use of rituximab SC in this setting. Funding This study was sponsored by F. Hoffmann-La Roche Ltd. Editorial support under the direction of the lead author was provided by Cheryl Wright, PhD and Susan Browne, PhD, of Gardiner-Caldwell Communications (Macclesfield, UK) and funded by F. Hoffmann-La Roche Ltd.

(Campath-1H) as first-line treatment for patients with B-cell chronic lymphocytic leukemia (B-CLL). Blood. 2002;100(3):768773. Pivot X, Gligorov J, Muller V, et al. Preference for subcutaneous or intravenous administration of trastuzumab in patients with HER2-positive early breast cancer (PrefHer): an open-label randomised study. Lancet Oncol. 2013;14(10):962-970. Shpilberg O, Jackisch C. Subcutaneous administration of rituximab (MabThera) and trastuzumab (Herceptin) using hyaluronidase. Br J Cancer. 2013;109(6): 1556-1561. Stilgenbauer S, Zenz T, Winkler D, et al. Subcutaneous alemtuzumab in fludarabine-refractory chronic lymphocytic leukemia: clinical results and prognostic marker analyses from the CLL2H study of the German Chronic Lymphocytic Leukemia Study Group. J Clin Oncol. 2009;27(24):3994-4001. De Cock E, Kritikou P, Sandoval M, et al. Time savings with rituximab subcutaneous injection versus rituximab intravenous infusion: a time and motion study in eight countries. PLoS One. 2016;11(6):e0157957. Davies A, Merli F, Mihaljevic B, et al. Pharmacokinetics and safety of subcutaneous rituximab in follicular lymphoma (SABRINA): stage 1 analysis of a randomised phase 3 study. Lancet Oncol. 2014;15(3):343-352. Davies AJ, Merli F, Mihaljevic B, et al. Efficacy and safety of subcutaneous rituximab versus intravenous rituximab for firstline treatment of follicular lymphoma (SABRINA): a randomised, open-label, phase 3 trial. Lancet Haematol. 2017; 4(6):e272â&#x20AC;&#x201C;e282. Salar A, Avivi I, Bittner B, et al. Comparison of subcutaneous versus intravenous administration of rituximab as maintenance treatment for follicular lymphoma: results from a two-stage, phase IB study. J Clin Oncol. 2014;32(17):1782-1791. Assouline S, Buccheri V, Delmer A, et al. Pharmacokinetics and safety of subcutaneous rituximab plus fludarabine and cyclophosphamide for patients with chronic lymphocytic leukaemia. Br J Clin Pharmacol. 2015;80(5):1001-1009. Assouline S, Buccheri V, Delmer A, et al. Pharmacokinetics, safety, and efficacy of

19.

20.

21.

22.

23.

24.

subcutaneous versus intravenous rituximab plus chemotherapy as treatment for chronic lymphocytic leukaemia (SAWYER): a phase 1b, open-label, randomised controlled non-inferiority trial. Lancet Haematol. 2016;3(3):e128-e138. MabThera Summary of Product Characteristics. European Medicines Agency; 2017 [updated 6 July 2017; cited 25 September 2017]. Available from: http://www.ema.europa.eu/docs/en_GB/d o c u m e n t _ l i b r a r y / E PA R _ - _ P r o d u c t _ Information/human/000165/WC50002582 1.pdf. Last accessed 25 September 2017. Rituxan Highlights of Prescribing Information. US Food and Drug Administration; 2012 [updated October 2012; cited 25 September 2017]. Available from: http: // www. accessdata. fda. gov/ drugsatfda_docs/label/2012/103705s5367s 5388lbl.pdf. Last accessed 25 September 2017. MabThera 1400 mg Solution for Subcutaneous Injection. European Medicines Agency; 2017 [updated 15 August 2017; cited 25 September 2017]. Available from: <http://www.medicines.org.uk/emc/medicine/28732/SPC/MabThera+1400+mg+Sol ution+for+Subcutaneous+Injection>. Last accessed 25 September 2017. Rituximab (MabThera SC) Australian Public Assessment Report. Therapeutic Goods Administration; 2014 [updated 13 October 2014; cited 25 September 2017]. Available from: https://www.tga.gov.au/auspar/auspar-rituximab-3. Last accessed 25 September 2017. Rituxan SC Product Monograph. Hoffmann-La Roche Ltd.; 2016 [updated 9 September 2016; cited 25 September 2017]. Available from: http:// www. rochecanada.com/content/dam/roche_canada/en_C A/documents/Research/ClinicalTrialsForm s/Products/ConsumerInformation/Monogr aphsandPublicAdvisories/RituxanSC/Ritux anSC_PM_E.pdf. Last accessed 25 September 2017. Rummel M, Kim TM, Aversa F, et al. Preference for subcutaneous or intravenous administration of rituximab among patients with untreated CD20+ diffuse large B-cell lymphoma or follicular lymphoma: results from a prospective,

1921

P. Lugtenburg et al.

25.

26.

27.

28.

29.

1922

randomized, open-label, crossover study (PrefMab). Ann Oncol. 2017;28(4):836842. Cheson BD, Horning SJ, Coiffier B, et al. Report of an international workshop to standardize response criteria for nonHodgkin's lymphomas. NCI Sponsored International Working Group. J Clin Oncol. 1999;17(4):1244. Delarue R, Tilly H, Mounier N, et al. Dosedense rituximab-CHOP compared with standard rituximab-CHOP in elderly patients with diffuse large B-cell lymphoma (the LNH03-6B study): a randomised phase 3 trial. Lancet Oncol. 2013;14(6):525-533. Cunningham D, Hawkes EA, Jack A, et al. Rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisolone in patients with newly diagnosed diffuse large B-cell non-Hodgkin lymphoma: a phase 3 comparison of dose intensification with 14day versus 21-day cycles. Lancet. 2013;381(9880):1817-1826. 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. Vogel WH. Infusion reactions: diagnosis,

30.

31.

32.

33.

34.

assessment, and management. Clin J Oncol Nurs. 2010;14(2):E10-21. Murawski N, Pfreundschuh M, Zeynalova S, et al. Optimization of rituximab for the treatment of DLBCL (I): dose-dense rituximab in the DENSE-R-CHOP-14 trial of the DSHNHL. Ann Oncol. 2014;25(9):18001806. Lugtenburg PJ, de Nully Brown P, van der Holt B, et al. Randomized phase III study on the effect of early intensification of rituximab in combination with 2-weekly CHOP chemotherapy followed by rituximab or no maintenance in patients with diffuse large B-cell lymphoma: Results from a HOVON-Nordic Lymphoma Group study. J Clin Oncol. 2016;34(suppl; abstr 7504). Trask PC, Tellefsen C, Espindle D, Getter C, Hsu MA. Psychometric validation of the cancer therapy satisfaction questionnaire. Value Health. 2008;11(4):669-679. Rule S, Briones J, Smith R, et al. Preference for rituximab subcutaneous (SC) and intravenous (IV) among patients with CD20+ Non-Hodgkin's Lymphoma (NHL) completing the RASQ measure in randomized phase III studies Prefmab and Mabcute. Value Health. 2014;17(7):A537. Rule S, Briones J, Carella AM, et al. A ran-

35.

36.

37.

38.

domized comparison of maintenance therapy with subcutaneous rituximab for 2 years versus until progression in patients with indolent non-Hodgkin’s lymphoma: interim safety data from the Mabcute Study. Blood. 2013;122(21):3052. Bax P, Postma MJ. Cost-minimization of MabThera intravenous versus subcutaneous administration. Value Health. 2013;16:A390-A391. Stiff PJ, Unger JM, Cook JR, et al. Autologous transplantation as consolidation for aggressive non-Hodgkin's lymphoma. N Engl J Med. 2013;369(18):1681– 1690. Schmitz N, Nickelsen M, Ziepert M, et al. Conventional chemotherapy (CHOEP-14) with rituximab or high-dose chemotherapy (MegaCHOEP) with rituximab for young, high-risk patients with aggressive B-cell lymphoma: an open-label, randomised, phase 3 trial (DSHNHL 2002-1). Lancet Oncol. 2012;13(12):1250–1259. Vitolo U, Trnêný M, Belada D, et al. Obinutuzumab or rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone in previously untreated diffuse large B-cell lymphoma. J Clin Oncol. 2017 Aug 10:JCO.2017733402. [Epub ahead of print]

haematologica | 2017; 102(11)

ARTICLE

Non-Hodgkin Lymphoma

CUDC-907 in relapsed/refractory diffuse large B-cell lymphoma, including patients with MYC-alterations: results from an expanded phase I trial

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Yasuhiro Oki,1 Kevin R Kelly,2 Ian Flinn,3 Manish R. Patel,3,4 Robert Gharavi,5 Anna Ma,5 Jefferson Parker,5 Amir Hafeez,5 David Tuck5 and Anas Younes6

Lymphoma/Myeloma, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX; 2Jane Anne Nohl Division of Hematology and Center for the Study of Blood Diseases, University of Southern California, Los Angeles, CA; 3Sarah Cannon Research Institute, Nashville, TN; 4Florida Cancer Specialists, Sarasota, FL; 5Curis Inc., Lexington, MA and 6Memorial Sloan Kettering Cancer Center, New York, NY, USA 1

Haematologica 2017 Volume 102(11):1923-1930

ABSTRACT

UDC-907 is a first-in-class, oral small molecule inhibitor of both HDAC (class I and II) and PI3K (class Iα, β, and δ) enzymes, with demonstrated anti-tumor activity in multiple pre-clinical models, including MYC-driven ones. In this report, we present the safety and preliminary activity results of CUDC-907, with and without rituximab, in patients with relapsed/refractory diffuse large B-cell lymphoma (DLBCL), with a particular focus on those with MYC-altered disease. Thirty-seven DLBCL patients were enrolled, 14 with confirmed MYCaltered disease. Twenty-five patients received monotherapy treatment, and 12 received the combination of CUDC-907 with rituximab. CUDC907 monotherapy and combination demonstrated similar safety profiles consisting primarily of Grade 1/2 hematologic and gastrointestinal events. The most frequently reported Grade ≥3 treatment-related events were thrombocytopenia, neutropenia, diarrhea, fatigue, and anemia. Eleven responses (5 complete responses and 6 partial responses) were reported, for a response rate of 37% (11 out of 30) in evaluable patients [30% (11 out of 37) including all patients]. The objective response rate in evaluable MYC-altered DLBCL patients was 64% (7 out of 11; 4 complete responses and 3 partial responses), while it was 29% (2 out of 7) in MYC unaltered, and 17% (2 out of 12) in those with unknown MYC status. Median duration of response was 11.2 months overall; 13.6 months in MYC-altered patients, 6.0 months in MYC unaltered, and 7.8 months in those with MYC status unknown. The tolerable safety profile and encouraging evidence of durable anti-tumor activity, particularly in MYC-altered patients, support the continued development of CUDC907 in these populations of high unmet need. (clinicaltrials.gov identifier: 01742988). Introduction Diffuse large B-cell lymphoma (DLBCL) represents an aggressive and heterogeneous group accounting for approximately 35% of all malignant lymphomas.1-4 Since the introduction of rituximab in 2004, response rates and survival in DLBCL patients have greatly improved.5 However, approximately 30-40% of patients still develop relapsed or refractory disease.1 Treatment options in this setting are limited to salvage therapies with the intent to bridge to hematopoietic stem cell transplantation, or clinical trials.5,6 Multiple studies of relapsed/refractory DLBCL patients have demonstrated that the progression-free survival (PFS) and overall survival (OS) times are less than one year. Furthermore, over half the patients with haematologica | 2017; 102(11)

Correspondence: rgharavi@curis.com

Received: May 30, 2017. Accepted: August 29, 2017. Pre-published: August 31, 2017. doi:10.3324/haematol.2017.172882 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1923 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1923

Y. Oki et al.

relapsed/refractory disease are ineligible for stem cell transplantation.1,6 These patients thus represent a population of urgent unmet clinical need.7-9 Given the heterogeneity of DLBCL and the need for new therapeutic options, there is increasing interest in determining treatments based on the genetic/molecular features of the disease.10 Numerous studies in the past decade have demonstrated that alterations of MYC in DLBCL patients, defined as rearrangements or amplification of the MYC gene and/or MYC protein overexpression, confer dismal outcomes and poorer prognoses.2,11-17 MYC is a transcription factor responsible for many cellular functions including cell proliferation and growth, and the upregulation of MYC is a common driver event in multiple human cancers.2,3,18,19 Certain rearrangements of MYC [determined by fluorescent in situ hybridization (FISH)] lead to activation of the gene, increased protein expression [determined by immunohistochemistry (IHC)], uncontrolled cell growth, and increasingly aggressive disease.2 MYC gene rearrangements and high MYC protein overexpression (defined as ≥ 40% of lymphoma cells) are estimated to be present in approximately 10% and 30% of all newly diagnosed DLBCL patients, respectively,2,16,20-22 with similar rates also reported in relapsed patients.13 The emergence of MYC alterations as defining features of DLBCL is highlighted in the current National Comprehensive Cancer Network (NCCN) guidelines on non-Hodgkin lymphoma and the most recent World Health Organization (WHO) revisions on lymphoma classifications.5 The 2016 revision of the 2008 WHO classification includes an up-dated subpopulation classified as “high grade B-cell lymphoma (HGBL) with rearrangements in MYC and BCL2 and/or BCL6”.23 The histone deactylase (HDAC) and phosphatidylinositol 3-kinase (PI3K) enzymes and their associated signaling pathways are established therapeutic targets in multiple hematologic cancers. Responses have been reported from the use of HDAC inhibitors in a limited number of relapsed/refractory DLBCL patients.8,24,25 Individually, HDAC inhibitors have been reported to reduce the expression of MYC and associated tumorigeneic oncogenes such as BCL2, while PI3K inhibitors have been reported to decrease MYC family protein stability by disrupting their regulation at the post-transcriptional level.2634 When combined, reports from multiple DLBCL cell lines and mouse xenograft models have demonstrated that dual HDAC and PI3K inhibition results in synergistic anti-cancer effects, including suppression of MYC-induced oncogenic transcriptional programs.8,29,35-40 CUDC-907 is a rationally designed, first-in-class, oral small molecule that dually inhibits HDAC (class I and II) and PI3K (class Iα, β, and δ) enzymes.8,19,35 Pre-clinical data in multiple solid and hematologic cell lines and animal models have shown that the pro-apoptotic and tumor growth inhibition activities of CUDC-907 are more potent than single-targeting HDAC or PI3K inhibitors. In addition, CUDC-907 treatment has demonstrated decreased MYC gene and protein expression and anti-tumor activity in multiple MYC-driven tumor models, including DLBCL.19 Despite increased awareness, the optimal treatment strategies for relapsed/refractory DLBCL and MYC-altered DLBCL patients remain poorly defined, highlighting the need for development of novel therapies targeting MYC.3,13,41 The mechanistic rationale and pre-clinical 1924

Table 1. Screening characteristics of diffuse large B-cell lymphoma (DLBCL) patients.

Male, n (%)

27 (73)

Age, median years (range)

60.6 (20-85)

Ethnicity, n (%) White Other Histology, n (%) MYC-altered t-FL/DLBCL Both Non-MYC-altered MYC status unknown

30 (81) 7 (19) 14 (38) 13 (16) 5 (14) 8 (22) 15 (40)

Cell of origin, n (%) GCB Non-GCB ABC Unknown/unclassifiable Years since diagnosis, median (range) MYC-altered patients t-FL/DLBCL patients

6 (16) 3 (8) 1 (3) 27 (76) 2.4 (0.6-20.9) 2.5 (0.6-15.1) 3.6 (0.6-11.4)

ECOG performance status, n (%) 0 1 2 Number of previous treatments, median (range) MYC-altered patients t-FL/DLBCL patients

13 (35) 22 (60) 2 (5) 4 (2-10) 3 (2-10) 4 (2-10)

Previous treatments, n (%) Prior HDAC or PI3K inhibitor Stem cell transplants Autologous Allogeneic

2 (5) 12 (32) 11 (30) 1 (3)

n: number; ECOG: Eastern Co-operative Oncology Group; t-FL/DLBCL: transformed follicular lymphoma to DLBCL; ABC: activated B-cell like, GCB: germinal B-center

observations support the investigation of CUDC-907 in DLBCL and MYC-altered DLBCL patients. Preliminary safety and activity results from the dose escalation of CUDC-907 across multiple lymphoma types and multiple myeloma have been presented previously.8 Here, we present results from the dose escalation and expansion of CUDC-907 in relapsed/refractory DLBCL, with a particular focus on those with MYC-altered disease. The trial is registered at clinicaltrials.gov identifier: 01742988.

Methods Results presented here are from dose escalation and expansion of CUDC-907 in relapsed/refractory DLBCL patients. In dose expansion, the CUDC-907 recommended phase II dose (RP2D) of 60 mg for 5 days on/2 days off (5/2) was further explored as a monotherapy and in combination with rituximab at 375 mg/m2 on day (d)1 of every cycle for 6 cycles (R-907). In total, 37 adult patients with relapsed/refractory DLBCL were enrolled across 6 US cancer centers. Major inclusion criteria included confirmed relapsed/refractory disease, at least 2 prior anti-canhaematologica | 2017; 102(11)

CUDC-907 in relapsed/refractory DLBCL

cer regimens, measurable or evaluable disease, adequate hematologic, renal, and hepatic functions, and the ability to provide written informed consent and follow protocol requirements. Patients with DLBCL that had transformed from follicular lymphoma (tFL/DLBCL) were eligible to enroll. Major exclusion criteria included active central nervous system (CNS) involvement of disease, gastrointestinal disorders interfering with study drug absorption, uncontrolled or severe cardiovascular disease, HIV infection, and hepatitis B or C infection. The trial was approved by the institutional review boards of all participating centers, and was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonisation guidelines for Good Clinical Practice. Patients received CUDC-907 capsules (Pharmatek Laboratories Inc., San Diego, CA, USA) orally, within 30 minutes of a meal, in 21-day cycles until disease progression or other criteria for treatment discontinuation were met. Dose delays and modifications were permitted to ameliorate toxicities. Safety was assessed at screening, throughout the trial, and at follow up 30 days after the last dose of CUDC-907. The incidence and severity of adverse events were assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events (v.4.03). Standard, single 12-lead electrocardiograms (ECGs) were performed and monitored locally during screening, and before and after dosing on d1, d2, d8 and d15 of the first cycle of treatment, and at the end of treatment visit. Parameters measured included heart rate, PR, QRS, QT, and QTc (Bazett's formula) intervals. The safety population includes all DLBCL patients of the trial who received at least one dose of CUDC-907. The responseevaluable population includes all DLBCL patients who received at least one dose of study drug and completed at least one post-baseline disease assessment. DLBCL patients were re-staged according to the Revised Response Criteria for Malignant Lymphoma, during the last week of cycles 2, 4, and 6, then every 4 cycles thereafter, and at the end of treatment.42 MYC-altered disease was defined per central testing of patient tumor tissue as MYC gene translocation or amplification (≥3 copies in >20% of cells) as determined by FISH, or MYC protein expression in ≥40% of lymphoma cells as determined by IHC. Where central testing could not be conducted, local results from pathology reports were used when available. Patients without abnormality by FISH and without protein overexpression were designated as MYC negative. Patients without tumor tissue available for testing, or without prior testing results of their MYC status were designated as MYC unknown. Cell-of-origin (COO), and translocation and expression status of BCL2 and BCL6 were also assessed or collected where available.

Table 2. Disposition of diffuse large B-cell lymphoma (DLBCL) patients.

Discontinued study treatment

34 (92)

Progressive disease Physician’s decision Withdrawal of consent Toxicity Lost to follow up Other*

20 (54) 5 (14) 4 (11) 3 (8) 1 (3) 1 (3)

Duration on treatment, median months (range) 1.3 (0.1-35.4) Responders 15.5 (1.0-20.8+) MYC-altered patients 2.8 (0.2-25.4) Monotherapy 1.4 (0.2-35.5) MYC-altered patients 2.8 (0.2-25.4) R-907 1.2 (0.5-21.9+) MYC-altered patients 1.3 (0.7-21.9+) *Patient discontinued study treatment to proceed to autologous stem cell transplantation after achieving a complete response at cycle 2 of study treatment. R-907: CUDC907 on 60 mg 5/2 schedule with rituximab at 375 mg/m2 on day 1 of every cycle for 6 cycles.

Among responding patients, the median duration of treatment was 15.5 (range: 1.0-20.8+) months (Table 3). Through central testing of available tissue and/or information from pathology reports, 14 (38%) patients were determined to have MYC-altered disease, 8 (22%) had non-MYC altered disease, and 15 (40%) had disease of unknown MYC status. Among 13 t-FL/DLBCL patients, 5 (38%) were determined to have MYC-altered disease, 4 (31%) were non-MYC-altered, and 4 (31%) had disease of unknown MYC status. BCL2 and BCL6 protein expression and/or gene alteration (translocation and/or amplification) status was available for 11 (30%) patients. Two patients had confirmed MYC and BCL2 protein overexpression (double-expressors) while no patients had both MYC and BCL2 translocations present (double-hit). Three (8%) patients had some form of MYC, BCL2, and BCL6 protein expression and/or gene alteration status available. There were no reported double-hit, triple-hit, or triple-expressor lymphoma patients on the study. Cell-of-origin subtypes were determined for 11 (32%) patients, of which 7 were germinal center B-cell (GCB), 3 were non-GCB, and one was activated B-cell (ABC).

Safety Results Patients’ characteristics In total, 37 relapsed/refractory DLBCL patients were enrolled between 23 January 2013 and 12 May 2016. Twenty-five patients received CUDC-907 monotherapy and 12 received R-907. Demographic and study disposition information are summarized in Tables 1 and 2, respectively. As of the 7th July 2017, 3 DLBCL patients (all MYC-altered) remained on active treatment; one patient was receiving the RP2D, and 2 patients were receiving R-907. Median duration of treatment for all patients was 1.3 (range: 0.1-35.4) months, 1.4 (range: 0.2-35.4) months for monotherapy patients, and 1.2 (range: 0.5-21.9+) for R-907 patients. Fifteen (41%) patients stayed on treatment beyond cycle 2 (42 days). haematologica | 2017; 102(11)

Adverse events were generally mild to moderate in severity (Grade 1/2) in both monotherapy and R-907, and reversible with standard medications, or with dose holds or reductions (Table 3). The most frequently reported adverse events reported were diarrhea [21 (57%)], thrombocytopenia [20 (54%)], fatigue [15 (41%)], nausea [14 (38%)], constipation [9 (24%)], vomiting [9 (24%)], and neutropenia [8 (22%)]. Grade ≥3 adverse events were reported in 16 (43%) patients, and Grade ≥3 treatmentrelated adverse events were reported in 15 (40%). The most frequently reported Grade ≥3 treatment-related events were thrombocytopenia [12 (32%)], neutropenia [6 (16%)], anemia [2 (5%)], diarrhea [2 (5%)], and fatigue [2 (5%)] (Table 4). Serious adverse events were reported in 4 (28%) patients, none of which were considered treatment-related. They consisted of Grade 2 atrial fibrillation, Grade 3 1925

Y. Oki et al.

abdominal pain, Grade 5 worsening of lymphoma, and Grade 3 pleural effusion that eventually progressed to Grade 5 respiratory failure. In this case, the patient entered the study with an Eastern Co-operative Oncology Group (ECOG) performance status of 2 and extra-nodal lung involvement of the disease. Symptoms of pleural effusion were reported within 11 days of starting R-907; treatment was then discontinued and the patient was reported deceased due to respiratory failure 13 days later. Two other adverse events resulted in treatment discontinuations: Grade 4 hypercalcemia and Grade 5 sepsis, neither of which were considered treatment-related. Adverse events resulting in dose holds were reported in 16 (43%) patients, consisting primarily of treatment-related Grade 1-4 diarrhea, thrombocytopenia, and neutropenia. Adverse events leading to dose reductions were reported on 10 occasions in a total of 5 (13.5%) patients. The events were all considered treatment-related and consisted of Grade 2-4 thrombocytopenia and Grade 3 neutropenia, fatigue, and diarrhea. The onset of diarrhea varied across the study and was generally well controlled by the use of anti-diarrheal medication and/or dose modifications. Patients were recommended to continue prophylactic use of anti-diarrheals from the first onset throughout study treatment. The majority of diarrhea events were Grade 1/2, though 2 patients reported Grade 3 diarrhea (both related to study treatment); both events were resolved within one week of onset. The onset in these 2 cases varied from during the first week of treatment to after one year of treatment. There were no events of colitis or use of colonoscopies noted. Grade â&#x2030;Ľ3 hematologic toxicities such as thrombocytopenia, neutropenia, and anemia were reported from within the first few days of treatment to more than one year after starting treatment. The majority of these events resolved within one week through dose holds and/or reductions, though 2 events persisted for over a month and did not result in any dose modifications. There were no reports of any major pulmonary toxicities such as pneumonitis, though 2 patients reported unrelated Grade 1-2 pneumonia events. One patient also reported Grade 1 metapneumovirus infection that was considered unrelated to study treatment and resolved within two weeks without any dose modification. The study protocol did not mandate the use of prophylactic measures for potential pulmonary toxicities or infections. One event of atrial fibrillation and tachycardia each were reported (both unrelated to study treatment) and no events of QTc prolongation were noted. Minimal evidence of hepatotoxicity was reported, with 2 patients reporting unrelated Grade 1/2 events of alanine transaminase increase and no reports of aspartate transaminase increase.

Efficacy Of the 37 patients enrolled, 30 were evaluable for response. Seven patients discontinued treatment prior to a post-baseline disease assessment. Two monotherapy patients discontinued due to adverse events (Grade 4 unrelated hypercalcemia and Grade 5 unrelated sepsis), and 4 patients withdrew consent. The single non-evaluable R-907 patient was lost to follow up during the second week of treatment. Eleven DLBCL patients achieved objective responses; 9 receiving monotherapy and 2 receiving R-907, for an over1926

Figure 1. Extended progression-free survival (PFS) in MYC-altered patients. Kaplan-Meier PFS curve for all patients (n=37) on the trial (solid black line) along with subsets based on MYC status. MYC-altered patients (n=14, blue dotted line), MYC negative patients (n=8, red dashed line) and MYC unknown (n=15, green dash-dotted line). Median PFS was 2.9 months for all diffuse large B-cell lymphoma (DLBCL) patients and 21.8 months for the MYC-altered patients, respectively. x-axis in months; y-axis is the proportion of patients.

all response rate of 37% among evaluable patients (30% for all 37 enrolled patients). Objective responses were reported in 9 of 19 (47%) with monotherapy and 2 of 11 (18%) with combination therapy. Three responses with monotherapy and both responses in R-907 were complete responses (CR). Median duration of response for all 11 responding patients was 11.1 (range: 1.0-20.8+) months; 6.0 (range: 1.0-16.4) months for monotherapy-treated patients. Both R-907 complete responses were ongoing at durations of 10.2 and 20.8 months, respectively (Table 5 and Figure 2). Median time to response was 2.4 (range: 1.2-15.3) months. Eight of 11 responses were reported within the first 2-3 disease assessments (by end of cycle 6), while 3 were reported at cycle 10 or beyond (cycles 10, 17 and 22). Of the 3 monotherapy patients achieving CR, one was reported at the first disease assessment (and the subject ultimately discontinued treatment to pursue a stem cell transplant), one PR at cycle 10 converted to CR at cycle 14, and one PR at cycle 2 converted to a CR at cycle 10. Median PFS for all DLBCL patients on the study was 2.9 (range: 0.2-35.5) months. Among patients receiving monotherapy, the median PFS was 5.7 (range: 0.5-35.5) months and for patients treated with R-907 1.3 (range: 0.521.9+) months (Table 5 and Figure 1).

Efficacy in MYC-altered and other subgroups Of the 19 evaluable monotherapy patients, 7 were MYC-altered, and, of the 11 evaluable R-907 patients, 4 were MYC-altered. Objective responses in evaluable MYC-altered patients were reported in 5 of 7 (71%) monotherapy patients, and 2 of 4 (50%) patients receiving R-907, for an overall response rate of 64% (7 of 11) in this group. Two CRs in the monotherapy group and both CRs in the R-907 group occurred in MYC-altered patients. In MYC non-altered patients, the evaluable response rate was 29% (2 of 7) and 17% (2 of 12) in those with unknown MYC status. Median duration of response was 13.6 (range: 1.0-20.8+) months for MYC-altered patients. Among monotherapy haematologica | 2017; 102(11)

CUDC-907 in relapsed/refractory DLBCL

Table 3. Treatment emergent adverse events (>10% of patients).

Event

Diarrhea Thrombocytopenia Fatigue Nausea Neutropenia Constipation Vomiting Fever Anemia Cough Hypokalemia Abdominal pain Edema Hyperglycemia Hypomagnesemia

Grades 1-2

Monotherapy (n = 25) Grade Grade Total 3 4

Grades 1-2

R-907 (n = 12) Grade Grade Total 3 4

Grades 1-2

Grade 3

12 (48) 4 (16) 10 (40) 7 (28) 3 (12) 5 (20) 2 (8) 2 (8) 1 (4) 3 (12) 2 (8) 2 (8) 3 (12) 0 0

2 (8) 8 (32) 1 (4) 0 3 (12) 0 0 0 0 0 0 0 0 1 (4) 0

7 (28) 4 (16) 3 (12) 7 (28) 1 (4) 4 (16) 6 (24) 4 (16) 3 (12) 3 (12) 3 (12) 2 (8) 1 (4) 3 (12) 4 (16)

0 2 (8) 1 (4) 0 2 (8) 0 1 (4) 1 (4) 2 (8) 0 1 (4) 1 (4) 0 0 0

19 (51) 8 (22) 13 (35) 14 (38) 4 (11) 9 (24) 8 (22) 6 (16) 4 (11) 6 (16) 5 (14) 4 (11) 4 (11) 3 (8) 4 (11)

2 (5) 10 (27) 2 (5) 0 5 (14) 0 1 (3) 1 (3) 2 (5) 0 1 (3) 1 (3) 0 1 (3) 0

0 1 (4) 0 0 1 (4) 0 0 0 0 0 0 0 0 0 0

14 (56) 13 (52) 11 (44) 7 (28) 7 (20) 5 (20) 2 (8) 2 (8) 1 (4) 3 (12) 2 (8) 2 (8) 3 (12) 1 (4) 0

0 1 (4) 0 0 0 0 0 0 0 0 0 0 0 0 0

7 (28) 7 (28) 4 (16) 7 (28) 3 (12) 4 (16) 7 (28) 5 (20) 5 (20) 3 (12) 4 (16) 3 (12) 1 (4) 3 (12) 4 (16)

Overall (n = 37) Grade Total 4 0 2 (5) 0 0 1 (3) 0 0 0 0 0 0 0 0 0 0

21 (57) 20 (54) 15 (41) 14 (38) 10 (22) 9 (24) 9 (24) 7 (19) 6 (16) 6 (16) 6 (16) 5 (14) 4 (11) 4 (11) 4 (11)

Number (n) of diffuse large B-cell lymphoma patients (%) experiencing treatment-emergent adverse events. Only events reported in >10% of patients are presented. Six Grade 5 events were reported, none of which were treatment related; sepsis (n = 2) and worsening of lymphoma/progression of disease (n = 4). Note: subjects with multiple intensities under the same preferred term were counted only once in the highest severity.

MYC-altered responses, median duration was 7.5 (range: 1.0-16.4) months. Both R-907 CRs in MYC-altered patients were ongoing with durations of 10.2 and 20.8 months each, and the median duration of response was not reached. Median duration of response in MYC negative and MYC status unknown patients was 6.0 (range: 3.4-8.7) months and 7.7 (range: 1.4-14 months) months, respectively, none of which were ongoing (Table 5 and Figure 2). Median PFS was 21.8 months (range: 1.0+ - 25.4+ months) for MYC-altered patients, with a median PFS of 21.8 (range: 1.0-16.4) months for patients treated with monotherapy, and not reached for patients treated with R907. Median PFS in MYC negative and MYC status unknown patients was 1.3 (range: 0.4-15.5) months and 1.3 (range: 0.2 -35.3) months, respectively (Table 5 and Figure 1). Other subgroups included patients with t-FL/DLBCL and those tested for COO, BCL2, and BCL6 status. Among the 10 evaluable t-FL/DLBCL patients, 6 responses were noted (4 PR, 2 CR), 4 of which occurred in MYCaltered patients. When considering COO, 3 of 7 GCB patients reported responses while none of the 3 non-GCB patients did. Of the 2 double-expressor patients, one discontinued after eight days of treatment due to unrelated Grade 4 hypercalcemia, while the other achieved a CR at cycle 2 before discontinuing to pursue a stem cell transplant. Patients who did not report objective responses almost all discontinued treatment within the first 4 cycles. One exception was an MYC-altered patient who maintained stable disease for over two years before leaving the study due to physicianâ&#x20AC;&#x2122;s decision.

Pharmacokinetics and pharmacodynamics Plasma pharmacokinetics (PK) of CUDC-907 and major metabolites M1 and M2 were determined for each dose haematologica | 2017; 102(11)

Figure 2. Extended duration of response in MYC-altered patients. Kaplan-Meier duration of response curve for all responding patients (n=11) on the trial (solid black line) along with subsets based on MYC status. MYC-altered patients (n=7, blue dotted line), MYC negative patients (n=2, red dashed line) and MYC unknown (n= 2, green dash-dotted line). x-axis in months; y-axis is the proportion of patients.

level and schedule on the basis of samples obtained on d1 and d15 of cycle 1. In vitro studies have demonstrated that M1 and M2 both have reduced PI3K inhibitory activity (31% and 76%, respectively) and no HDAC inhibitory activity. CUDC-907 demonstrated rapid absorption with maximum plasma concentrations occurring approximately two hours post administration (Tmax) at the RP2D schedule. The metabolites reached their highest concentrations at later times; the Tmax of M1 was approximately 16-20 hours (h), and the Tmax of M2 was approximately 20-24 h. There was no accumulation of CUDC-907 in plasma on d15 after two weeks of dosing at the RP2D with an area 1927

Y. Oki et al. Table 4. Grade ≥3 treatment-related adverse events.

Event Grade 3 Thrombocytopenia Neutropenia Anemia Diarrhea Fatigue Hyperglycemia Hypokalemia Vomiting

Monotherapy (n = 25) Grade Total 4

8 (32) 3 (12) 0 2 (8) 1 (4) 1 (4) 0 0

14 14 0 0 0 0 0 0

Grade 3

9 (36) 4 (16) 0 2 (8) 1 (4) 1 (4) 0 0

2 (17) 2 (17) 2 (17) 0 1 (8) 0 1 (8) 1 (8)

R-907 (n = 12) Grade Total 4 1 (8) 0 0 0 0 0 0 0

Grade 3

3 (25) 2 (17) 2 (17) 0 1 (8) 0 1 (8) 1 (8)

Overall (n = 37) Grade Total 4

10 (27) 5 (14) 2 (5) 2 (5) 2 (5) 1 (3) 1 (3) 1 (3)

2 (5) 1 (3) 0 0 0 0 0 0

12 (32) 6 (16) 2 (5) 2 (5) 2 (5) 1 (3) 1 (3) 1 (3)

Number (n) of diffuse large B-cell lymphoma patients (%) reporting Grade ≥3 treatment-related events. No related Grade 5 events were reported. Note: subjects with multiple intensities under the same preferred term were counted only once in the highest severity

Table 5. Summary of disease assessments in diffuse large B-cell lymphoma (DLBCL) patients.

Monotherapy MYC+ R-907 MYC+ Total MYC+

Evaluable, n

CR, n

PR, n

Evaluable ORR, %

DoR, median months (range)

SD, n

PD, n

PFS, median months (range)

25 10 12 4 37 14

19 7 11 4 30 11

3 2 2 2 5 4

6 3 0 0 6 3

47 71 18 50 37 64

6.0 (1.0-16.4) 7.5 (1.0-16.4) NR (10.2-20.8+) NR (10.2-20.8+) 11.1 (1.0-20.8+) 13.6 (1.0-20.8+)

4 1 1 0 5 1

6 1 8 2 14 3

5.7 (0.5-35.5) 21.8 (0.8-25.4) 1.3 (0.5-21.9+) NR (0.7-21.9+) 2.9 (0.5-35.5) 21.8 (0.7-25.4)

n: number; MYC+: MYC translocation or amplification by FISH, or MYC protein expression in ≥40% lymphoma cells per IHC; CR: complete response; PR: partial response; ORR: objective response rate; SD: stable disease; PD: progressive disease; DoR: duration of response; PFS: progression-free survival; NR: not reached; + : ongoing as of 7th July 2017 data cut-off date.

under the concentration-time curve from time 0 to 24 h (AUC0-24h) of 37.8 ng·h/mL on d1 and 24.5 ng·h/mL on d15 (Online Supplementary Table S1 and Online Supplementary Figure S1). As reported previously, peripheral blood mononuclear cell western blot analysis from patients in the RP2D dose schedule demonstrated post-treatment accumulation of acetylated histone H3 and decreased AKT phosphorylation.8 These pharmacodynamic effects further support the mechanism of action of CUDC-907 and the use of 60 mg 5/2 as the RP2D.

Discussion CUDC-907 monotherapy and R-907 demonstrated similar and tolerable safety profiles at doses able to achieve responses in heavily pre-treated relapsed/refractory DLBCL patients. Adverse events were generally mild to moderate in severity, reversible, and manageable with standard medications or by dose modifications. Diarrhea, fatigue, nausea, thrombocytopenia, and neutropenia were the most frequently associated adverse events of CUDC907. These events are all also associated class toxicities of HDAC inhibitors.43,44 As described above, diarrhea events were largely Grade 1/2 and managed with over-the-counter medications (i.e. loperamide) and/or dose modifications, and there were no cases of colitis reported. Grade ≥3 cytopenias were generally considered treatment related, varied in their time of onset, and were most often resolved within one week 1928

through dose holds or reductions. Minimal adverse events reported were attributed solely to rituximab. Overall, the safety profile of CUDC-907 shares characteristics with currently FDA approved HDAC and PI3K inhibitors, consisting primarily of gastrointestinal and hematologic events, while not associated with other serious associated risks such as colitis, pneumonitis, hepatotoxicity, and cardiac toxicities.8,43,45-47 CUDC-907 demonstrated an encouraging ORR with durable responses in relapsed/refractory DLBCL, particularly in patients with MYC-altered disease. The monotherapy response rate was greater than that of R907, but the small sample size of R-907 (11 evaluable patients) limits comparisons. A similar finding regarding response rates with and without rituximab was reported in a phase II study of the HDAC inhibitor panobinostat in relapsed/refractory DLBCL patients. Interestingly, the same study also reported no responses and an inferior PFS in patients with MYC-altered disease.48 Median duration of treatment among responders and the median duration of response times suggest durable anti-tumor activity for all responders. The durability of responses may be due to the novel dual inhibitory mechanism of CUDC-907, mitigating the development of resistance through simultaneous inhibition of multiple signaling pathways, as supported by pharmacodynamic data. MYC-altered patients demonstrated a notable median PFS time of 21.8 months, with all 3 active patients remaining being MYC-altered, further supporting the durable benefits of CUDC-907 treatment in this population. haematologica | 2017; 102(11)

CUDC-907 in relapsed/refractory DLBCL

MYC protein overexpression alone and with BCL2 have been reported to be among the worst prognostic factors in DLBCL, independent of MYC gene alterations or other prognostic parameters.17 Five of the 7 MYC-altered responders were positive by IHC and one CR was reported in a double-expressor. This patient left the study early to pursue a stem cell transplant, and although this resulted in censored duration of response and PFS times, CUDC907 has demonstrated activity in these particularly poor prognosis populations. CUDC-907 also demonstrated promising activity in t-FL/DLBCL, with 6 responses occurring in this therapeutically challenging patient population.9 Patients with relapsed/refractory DLBCL and MYCaltered DLBCL represent populations of unmet medical needs requiring the incorporation of novel agents into their treatment paradigms.8,32,41,48 In this phase I study, CUDC-907 demonstrated a moderate safety profile and durable anti-tumor activity in these populations, particularly in those with MYC-altered disease. However, the relatively small sample sizes and incomplete capture of MYC information and other disease characteristics (BCL2, BCL6, and COO status) highlight the need for additional investigations. A phase II study is currently ongoing to further explore

References 1. Colosia A, Njue A, Trask PC, et al. Clinical efficacy and safety in relapsed/refractory diffuse large B-cell lymphoma: a systematic literature review. Clin Lymphoma Myeloma Leuk. 2014;14(5):343-355.e6. 2. de Jonge AV, Roosma TJA, Houtenbos I, et al. Diffuse large B-cell lymphoma with MYC gene rearrangements: Current perspective on treatment of diffuse large B-cell lymphoma with MYC gene rearrangements; case series and review of the literature. Eur J Cancer. 2016;55:140-146. 3. Rosenthal A, Younes A. High grade B-cell lymphoma with rearrangements of MYC and BCL2 and/or BCL6: Double hit and triple hit lymphomas and double expressing lymphoma. Blood Rev. 2017;31(2):3742. 4. Miyazaki K. Treatment of Diffuse Large BCell Lymphoma. J Clin Exp Hematop. 2016;56(2):79-88. 5. Gisselbrecht C, Schmitz N, Mounier N, et al. Rituximab maintenance therapy after autologous stem-cell transplantation in patients with relapsed CD20(+) diffuse large B-cell lymphoma: final analysis of the collaborative trial in relapsed aggressive lymphoma. J Clin Oncol. 2012; 30(36):4462-4469. 6. NCCN Clinical Practice Guidelines. NonHodgkinâ&#x20AC;&#x2122;s Lymphomas. V 4.2014. 7. Van Den Neste E, Schmitz N, Mounier N, et al. Outcome of patients with relapsed diffuse large B-cell lymphoma who fail second-line salvage regimens in the International CORAL study. Bone Marrow Transplant. 2016;51(1):51-57. 8. Younes A, Berdeja JG, Patel MR, et al. Safety, tolerability, and preliminary activity

haematologica | 2017; 102(11)

10.

11.

12.

13.

14.

15.

the activity of CUDC-907 at the RP2D in relapsed/refractory DLBCL patients, with a primary analysis population consisting of patients with centrally confirmed MYCaltered disease by IHC analysis (clinicaltrials.gov identifier: 02674750). In addition, the study aims to better understand the potential relationship of other disease characteristics with CUDC-907 activity by implementing central testing of MYC, BCL2, and BCL6 protein expression and gene status, as well as collecting the COO status for all patients. CUDC-907 has also demonstrated synergistic anti-tumor activity when combined with the BCL2 inhibitor venetoclax in multiple DLBCL cell lines.49 Although only preliminary results are available, this may highlight a promising potential for CUDC-907 combinations with other therapies targeting MYC-related pathways or signals in DLBCL. Acknowledgments The authors would like to thank all the patients and their families and clinical sites participating in this trial. Funding This trial was funded by Curis Inc., with financial support from the Leukemia and Lymphoma Society.

of CUDC-907, a first-in-class, oral, dual inhibitor of HDAC and PI3K, in patients with relapsed or refractory lymphoma or multiple myeloma: an open-label, doseescalation, phase 1 trial. Lancet Oncol. 2016;17(5):622-631. Elstrom RL, Martin P, Ostrow K, et al. Response to second-line therapy defines the potential for cure in patients with recurrent diffuse large B-cell lymphoma: implications for the development of novel therapeutic strategies. Clin Lymphoma Myeloma Leuk. 2010;10(3):192-196. Puvvada S, Kendrick S, Rimsza L. Molecular classification, pathway addiction, and therapeutic targeting in diffuse large B cell lymphoma. Cancer Genet. 2013;206(7-8):257-265. Barrans S, Crouch S, Smith A, et al. Rearrangement of MYC is associated with poor prognosis in patients with diffuse large B-cell lymphoma treated in the era of rituximab. J Clin Oncol. 2010;28(20):33603365. Zhou K, Xu D, Cao Y, Wang J, Yang Y, Huang M. C-MYC aberrations as prognostic factors in diffuse large B-cell lymphoma: a meta-analysis of epidemiological studies. PloS One. 2014;9(4):e95020. Cuccuini W, Briere J, Mounier N, et al. MYC+ diffuse large B-cell lymphoma is not salvaged by classical R-ICE or R-DHAP followed by BEAM plus autologous stem cell transplantation. Blood. 2012;119(20):46194624. Nitsu N, Okamoto M, Miura I, Hirano M. Clinical significance of 8q24/c-MYC translocation in diffuse large B-cell lymphoma. Cancer Sci. 2009;100(2):233-237. Savage KJ, Johnson NA, Ben-Neriah S, et al. MYC gene rearrangements are associated with a poor prognosis in diffuse large B-cell

16.

17.

18.

19.

20. 21.

22.

lymphoma patients treated with R-CHOP chemotherapy. Blood. 2009;114(17):35333537. Horn H, Ziepert M, Becher C, et al. MYC status in concert with BCL2 and BCL6 expression predicts outcome in diffuse large B-cell lymphoma. Blood. 2013; 121(12):2253-2263. Valera A, LĂłpez-Guillermo A, CardesaSalzmann T, et al. MYC protein expression and genetic alterations have prognostic impact in patients with diffuse large B-cell lymphoma treated with immunochemotherapy. Haematologica. 2013;98(10):1554-1562. Landsburg D, Falkiewicz M, Petrich A, et al. Sole rearrangement but not amplification of MYC is associated with a poor prognosis in patients with diffuse large B cell lymphoma and B cell lymphoma unclassifiable. Br J Haematol. 2016; 175(4):631-640. Sun K, Atoyan R, Borek MA, et al. Dual HDAC and PI3K Inhibitor CUDC-907 Downregulates MYC and Suppresses Growth of MYC-dependent Cancers. Mol Cancer Ther. 2017;16(2):285-299. Aukema SM, Siebert R, Schuuring E, et al. Double-hit B-cell lymphomas. Blood. 2011;117(8):2319-2331. Johnson NA, Slack GW, Savage KJ, et al. Concurrent expression of MYC and BCL2 in diffuse large B-cell lymphoma treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone. J Clin Oncol. 2012;30(28):3452-3459. Green TM, Young KH, Visco C, et al. Immunohistochemical double-hit score is a strong predictor of outcome in patients with diffuse large B-cell lymphoma treated with rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone. J

1929

Y. Oki et al. Clin Oncol. 2012;30(28):3460-3467. 23. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood. 2016;127(20):2375-2390. 24. West AC, Johnstone RW. New and emerging HDAC inhibitors for cancer treatment. J Clin Invest. 2014;124(1):30-39. 25. Crump M, Coiffier B, Jacobsen ED, et al. Phase II trial of oral vorinostat (suberoylanilide hydroxamic acid) in relapsed diffuse large-B-cell lymphoma. Ann Oncol. 2008;19(5):964-969. 26. Kurland JF, Tansey WP. Myc-mediated transcriptional repression by recruitment of histone deacetylase. Cancer Res. 2008; 68(10):3624-3629. 27. Zhang X, Zhao X, Fiskus W, et al. Coordinated silencing of MYC-mediated miR-29 by HDAC3 and EZH2 as a therapeutic target of histone modification in aggressive B-Cell lymphomas. Cancer Cell. 2012;22(4):506-523. 28. Chambers AE, Banerjee S, Chaplin T, et al. Histone acetylation-mediated regulation of genes in leukaemic cells. Eur J Cancer. 1990 2003;39(8):1165-1175. 29. Gui C-Y, Ngo L, Xu WS, Richon VM, Marks PA. Histone deacetylase (HDAC) inhibitor activation of p21WAF1 involves changes in promoter-associated proteins, including HDAC1. Proc Natl Acad Sci USA. 2004;101(5):1241-1246. 30. Duan H, Heckman CA, Boxer LM. Histone deacetylase inhibitors down-regulate bcl-2 expression and induce apoptosis in t(14;18) lymphomas. Mol Cell Biol. 2005; 25(5):1608-1619. 31. Kenney AM, Widlund HR, Rowitch DH. Hedgehog and PI-3 kinase signaling converge on Nmyc1 to promote cell cycle progression in cerebellar neuronal precursors. Development. 2004;131(1):217-228. 32. Asano T, Yao Y, Zhu J, Li D, Abbruzzese JL,

1930

33.

34.

35.

36.

37.

38.

39.

40.

Reddy SAG. The PI 3-kinase/Akt signaling pathway is activated due to aberrant Pten expression and targets transcription factors NF-kappaB and c-Myc in pancreatic cancer cells. Oncogene. 2004;23 (53):8571-8580. Kumar A, MarquĂŠs M, Carrera AC. Phosphoinositide 3-kinase activation in late G1 is required for c-Myc stabilization and S phase entry. Mol Cell Biol. 2006; 26(23):9116-9125. Cross DA, Alessi DR, Cohen P, Andjelkovich M, Hemmings BA. Inhibition of glycogen synthase kinase-3 by insulin mediated by protein kinase B. Nature. 1995;378(6559):785-789. Qian C, Lai C-J, Bao R, et al. Cancer network disruption by a single molecule inhibitor targeting both histone deacetylase activity and phosphatidylinositol 3-kinase signaling. Clin Cancer Res. 2012; 18(15):4104-4113. Rahmani M, Aust MM, Benson EC, Wallace L, Friedberg J, Grant S. PI3K/mTOR inhibition markedly potentiates HDAC inhibitor activity in NHL cells through BIM- and MCL-1-dependent mechanisms in vitro and in vivo. Clin Cancer Res. 2014;20(18):4849-4860. Mondello P, Derenzini E, Asgari Z, et al. Dual inhibition of histone deacetylases and phosphoinositide 3-kinase enhances therapeutic activity against B cell lymphoma. Oncotarget. 2017;8(8):14017-14028. Wendel H-G, De Stanchina E, Fridman JS, et al. Survival signalling by Akt and eIF4E in oncogenesis and cancer therapy. Nature. 2004;428(6980):332-337. Sander S, Calado DP, Srinivasan L, et al. Synergy between PI3K signaling and MYC in Burkitt lymphomagenesis. Cancer Cell. 2012;22(2):167-179. Chapuy B, McKeown MR, Lin CY, et al. Discovery and characterization of superenhancer-associated dependencies in dif-

41.

42.

43.

44.

45. 46.

47.

48.

49.

fuse large B cell lymphoma. Cancer Cell. 2013;24(6):777-790. Petrich AM, Gandhi M, Jovanovic B, et al. Impact of induction regimen and stem cell transplantation on outcomes in double-hit lymphoma: a multicenter retrospective analysis. Blood. 2014;124(15):2354-2361. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007; 25(5):579586. HDAC Inhibitors in Cancer Care | Cancer Network. Available from: http:// www. cancernetwork.com/cancer-andgenetics/hdac-inhibitors-cancer-care [Last accessed 27 March 2017] Aggarwal R, Thomas S, Pawlowska N, et al. Inhibiting Histone Deacetylase as a Means to Reverse Resistance to Angiogenesis Inhibitors: Phase I Study of Abexinostat Plus Pazopanib in Advanced Solid Tumor Malignancies. J Clin Oncol. 2017;35(11):1231-1239. Farydak (panobinostat) Product Characteristics leaflet. Chia S, Gandhi S, Joy AA, et al. Novel agents and associated toxicities of inhibitors of the pi3k/Akt/mtor pathway for the treatment of breast cancer. Curr Oncol. 2015;22(1):33-48. Barr PM, Saylors GB, Spurgeon SE, et al. Phase 2 study of idelalisib and entospletinib: pneumonitis limits combination therapy in relapsed refractory CLL and NHL. Blood. 2016;127(20):2411-2415. Assouline SE, Nielsen TH, Yu S, et al. Phase 2 study of panobinostat with or without rituximab in relapsed diffuse large B-cell lymphoma. Blood. 2016;128(2):185-194. Sun K, Atoyan R, Borek MA, et al. The Combination of Venetoclax and CUDC907 Exhibits Synergistic Activity in Venetoclax-Refractory DLBCL. Blood. 2016;128(22):4184-4184.

haematologica | 2017; 102(11)

ARTICLE

Non-Hodgkin Lymphoma

Italian real-life experience with brentuximab vedotin: results of a large observational study of 40 cases of relapsed/refractory systemic anaplastic large cell lymphoma

Alessandro Broccoli,1* Cinzia Pellegrini,1* Alice Di Rocco,2 Benedetta Puccini,3 Caterina Patti,4 Guido Gini,5 Donato Mannina,6 Monica Tani,7 Chiara Rusconi,8 Alessandra Romano,9 Anna Vanazzi,10 Barbara Botto,11 Carmelo Carlo-Stella,12 Stefan Hohaus,13 Pellegrino Musto,14 Patrizio Mazza,15 Stefano Molica,16 Paolo Corradini,17 Angelo Fama,18 Francesco Gaudio,19 Michele Merli,20 Angela Gravetti,21 Giuseppe Gritti,22 Annalisa Arcari,23 Patrizia Tosi,24 Anna Marina Liberati,25 Antonello Pinto,26 Vincenzo Pavone,27 Filippo Gherlinzoni,28 Virginia Naso,29 Stefano Volpetti,30 Livio Trentin,31 Maria Cecilia Goldaniga,32 Maurizio Bonfichi,33 Amalia De Renzo,34 Corrado Schiavotto,35 Michele Spina,36 Sergio Storti,37 Angelo Michele Carella,38 Vittorio Stefoni,1 Lisa Argnani1 and Pier Luigi Zinzani1

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1931-1935

*AB and CP contributed equally to this work.

Institute of Hematology "L. e A. Seràgnoli", University of Bologna; 2Hematology, Department of Cellular Biotechnologies and Hematology, Sapienza University, Rome; 3 Hematology Department, University and Hospital Careggi, Firenze; 4Department of Hematology, Azienda Ospedali Riuniti Villa Sofia Cervello, Palermo; 5Department of Hematology, Ospedali Riuniti di Ancona; 6Hematology Unit, Ospedale Papardo, Messina; 7 Hematology Unit, Santa Maria delle Croci Hospital, Ravenna; 8Division of Hematology, Niguarda Hospital, Milan; 9Division of Hematology, AOU, Catania; 10European Institute of Oncology, Milan; 11SC Ematologia, Azienda Ospedaliera Universitaria Città della Salute e della Scienza, Torino; 12Oncology and Hematology, Humanitas Cancer Center, Humanitas Clinical and Research Center, Rozzano; 13Institute of Hematology, Catholic University, Rome; 14Scientific Direction, IRCCS-CROB, Referral Cancer Center of Basilicata, Rionero in Vulture (Pz); 15Ospedale Moscati Department of Hematology-Oncology, Taranto; 16 Department of Hematology, Ciaccio-Pugliese Hospital, Catanzaro; 17Department of Hematology, Fondazione IRCCS Istituto Nazionale dei Tumori, University of Milan; 18 Hematology Unit, Arcispedale Santa Maria Nuova, Istituto di Ricovero e Cura a Carattere Scientifico, Reggio Emilia; 19Hematology Unit, Policlinico di Bari; 20Hematology, Ospedale di Circolo, Fondazione Macchi, Varese; 21Division of Hematology and Stem Cell Transplantation Unit, Cardarelli Hospital, Napoli; 22Department of Hematology, Hospital Papa Giovanni XXIII, Bergamo; 23Division of Hematology, Guglielmo da Saliceto Hospital, Piacenza; 24Hematology Unit, Infermi Hospital Rimini; 25Hematology, Ospedale Perugia; 26 Hematology-Oncology and Stem Cell Transplantation Unit, National Cancer Institute, Fondazione Pascale, IRCCS, Napoli; 27Division of Hematology, Ospedale G. Panico, Tricase, Lecce; 28Hematology Unit, Ca' Foncello Hospital, Treviso; 29Sant’Andrea Hospital – Sapienza, Rome; 30Department of Hematology, DISM, Azienda Sanitaria Universitaria Integrata, Udine; 31Hematology and Clinical Immunology Unit, Department of Medicine, University of Padua; 32OncoHematology Unit, Fondazione Ca' Granda IRCCS Ospedale Maggiore Policlinico, Milan; 33Hematology, IRCCS Policlinico San Matteo, Pavia; 34 Hematology, AOU Federico II Napoli; 35Hematology, San Bortolo Hospital, Vicenza; 36 Division of Medical Oncology A, National Cancer Institute, Aviano; 37Hematology, Università Cattolica Sacro Cuore Campobasso and 38Division of Hematology 1, IRCCS A.O.U. San Martino IST, Genova, Italy 1

ABSTRACT

etween November 2012 and July 2014, in accordance with national law 648/96, brentuximab vedotin was available in Italy for patients with relapsed systemic anaplastic large cell lymphoma outside a clinical trial context. A large Italian observational retrospective study was conducted on the use of brentuximab vedotin in everyday clinical practice to check whether clinical trial results are confirmed in a real-life context. The primary endpoint of this study was best response; secondary endpoints were the overall response rate at the end of the treatment, duration of response, survival and safety profile. A total of 40 heavily pretreated patients were enrolled. Best response was observed after a median of four cycles in 77.5%: globally, 47.5% patients obtained a complete response, 64.2% in the elderly subset. The overall response rate was 62.5%. At the latest follow up, 15/18 patients are still in complete remission (3 with consolidation). The progression-free survival rate at 24 months was 39.1% and the disease-free survival rate at the same time haematologica | 2017; 102(11)

Correspondence: pierluigi.zinzani@unibo.it

Received: April 26, 2017. Accepted: July 27, 2017. Pre-published: August 3, 2017. doi:10.3324/haematol.2017.171355 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1931 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1931

A. Broccoli et al.

was 54% (median not reached). All the long-term responders were aged <30 years at first infusion. The treatment was well tolerated even in this real-life context and no deaths were linked to drug toxicity. Brentuximab vedotin induces clinical responses quite rapidly, i.e. within the first four cycles of treatment in most responders, thus enabling timely use of transplantation. For patients ineligible for transplant or for those in whom a transplant procedure failed, brentuximab vedotin may represent a feasible effective therapeutic option in everyday clinical practice.

Introduction Approximately 40% to 65% of patients with systemic anaplastic large-cell lymphoma (ALCL) develop recurrent disease after front-line therapy.1 Historically, at relapse the disease is resistant to conventional multiagent chemotherapy regimens and there is no established standard of care. High-dose therapy and autologous stem cell transplantation (SCT) may result in long-term remission in 30% to 40% of patients, but the benefit is limited to patients with chemotherapy-sensitive disease.2-6 Given that most patients with relapsed or refractory (R/R) systemic ALCL are scheduled to undergo a highly toxic high-dose chemotherapy regimen, any strategy aimed at achieving minimal residual disease, specifically a positron emission tomography-negative status before autologous SCT, without severe toxicity would represent a major advance in the overall management of these patients. Furthermore, despite the role of autologous SCT, the outcomes remain poor in patients with primary chemorefractory disease, in whom long-term survival rarely exceeds 1517%.1 In fact, disease recurrence still remains the principal cause of failure of autologous SCT, and early disease progression after transplantation, i.e. within 6 months of highdose conditioning, emerges as the most important predictor of an unfavorable outcome. No standard treatment options exist for patients whose disease relapses after autologous SCT or for patients not eligible for autologous SCT. In fact, while allogeneic SCT may induce long-term progressionfree survival in a fraction of patients, only a few are candidates for this procedure, mainly because of unsatisfactory pre-transplant cytoreduction and the substantial risk of morbidity due to the heavy load of previous therapies. In this light, optimization of the outcomes obtained with high-dose regimens and autologous SCT still remains a strategic priority, in order to offer the best chance of cure for the largest fraction of patients with R/R disease. Brentuximab vedotin (BV) is an antibody-drug conjugate targeting CD30 which may be an excellent candidate among the newly developed agents for the treatment of R/R systemic ALCL.7 In fact, systemic ALCL is characterized by the expression of CD30. In the initial phase 1 study of BV in patients with CD30+ lymphoid diseases, the two patients with systemic ALCL both achieved a complete response.7 The favorable activity of this agent in R/R systemic ALCL was clearly documented by Pro et al. in a phase 2 study involving 58 patients: 86% obtained a response, which was a complete response in 57% of cases.8 The median progression-free survival of these patients was 13.3 months, and the median overall survival was not reached (estimated 64% at 4 years). The same relevant proportion of complete responses in this subset of patients also emerged from the data collected by Zinzani et al. regarding the BV Named Patient Program experiences across Europe.9,11,15-17 1932

A high response rate is important not only in pretreated patients with a poor prognosis, but also in first-line R/R patients because a complete response obtained before transplantation is one of the stronger predictors of longterm survival.10 BV could represent an optimal therapeutic option as a bridge to either autologous or allogeneic SCT in patients achieving a suboptimal response after salvage treatment.11,12 Recent updates on the pivotal study have shown that BV can induce long-lasting complete responses in pretreated cases of systemic ALCL even without additional consolidation therapies, suggesting that BV may be curative for some patients.13,14 The pooled overall response rate and complete response rate reported for patients with R/R systemic ALCL (globally 46) in the Named Patient Program cohorts were both 69.5%.9,11,15-17 After accelerated approval by the US Food and Drug Administration, eligible patients in Italy were granted early access to BV through a Named Patient Program. After closure of this program in 2012, BV was available in Italy for patients with R/R systemic ALCL, based on a local disposition of the Italian Drug Agency (AIFA) issued in accordance with a national law (Law 648/96: â&#x20AC;&#x153;medicinal products that are provided free of charge on the national health serviceâ&#x20AC;?): a boundary zone in the passage from clinical trials to marketing and free use phases in which patients can be treated in any case. On the basis of our previous exploratory study,18 a large observational retrospective study was conducted on the use of BV in R/R systemic ALCL patients in everyday clinical practice in Italy to check whether clinical trial results are confirmed in a real-life context.

Methods An observational retrospective study was conducted among patients with systemic ALCL treated from November 2012 to July 2014 with BV in 38 Italian centers outside of clinical trials, in accordance with national law n. 648/96.19 The study was approved by the institutional board of the Policlinico S.Orsola-Malpighi Hospital in Bologna, the coordinating center of the study, and by all the ethical committees involved and registered in the Italian Registry of Observational Studies. All participants gave written informed consent in accordance with the Declaration of Helsinki. A shared database was used after the approval of all the co-investigators and variables were strictly defined to avoid bias in reporting data.19 We obtained special permission (for scientific purposes) from our ethical committee to collect data regarding patients who could not consent because they had died or been lost to follow up. BV was administered as a 30-min infusion at the dose of 1.8 mg/kg of body weight every 3 weeks for a maximum of 16 cycles. A dose reduction to 1.2 mg/kg was recommended in the case of grade 3 toxicity and the treatment had to be interrupted in the case of grade 4 toxicity. The primary endpoint of the study was the best response haematologica | 2017; 102(11)

Brentuximab vedotin in ALCL: real-life experience

achieved during BV therapy; secondary endpoints were the overall response rate at the end of the treatment, duration of response, overall survival, progression-free survival, disease-free survival, and the drug’s safety and tolerability. Duration of therapy was defined as the number of cycles of treatment administered. Effectiveness was also evaluated through the occurrence of longterm responders, defined as patients who had a response (complete or partial) lasting ≥12 months. Response was assessed by positron emission tomography or computed tomography scanning after cycles 4, 8, 12 and at drug discontinuation by each investigator using the International Working Group revised response criteria for malignant lymphoma.20 Safety and tolerability were evaluated by recording the incidence, severity, and type of any adverse events according to the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0. Overall survival was defined as the time from initiation of therapy to death from any cause and was censored at the date of the last available follow up. Progression-free survival was measured from initiation of therapy to progression, relapse, or death from any cause and was censored at the date of the last available follow up. Disease-free survival was calculated for patients who achieved a complete response from the first documentation of response to the date of relapse or death due to lymphoma or acute toxicity of treatment. Duration of response was calculated from the first objective tumor response (complete or partial) to first documentation of progression or death.20 Patients who were lost to follow up (n=2) were censored at the date of their last available information. Demographics and patients’ characteristics as well as adverse events were summarized by descriptive statistics. Survival functions were estimated using the Kaplan-Meier method and compared using the log-rank test. Statistical analyses were performed with Stata 11 (StataCorp LP, TX, USA) and P values <0.05 were considered statistically significant.

Results Of the estimated 40 patients who received BV under Law 648/96, all participated in this observational study. All had histologically confirmed CD30+ disease. Their median age at the time of being treated with BV was 47 years (range, 17-80 years) with 14 (35.0%) being considered elderly (age >60 years). There were 28 males and 12 females. Eleven (27.5%) had systemic symptoms at baseline (Table 1). The median number of prior lymphoma-related systemic regimens was two (range, 2-10) and included high-dose

Figure 1. Overall survival.

haematologica | 2017; 102(11)

chemotherapy and autologous SCT (in 13, 32.5% of the patients). Eight patients (20%) had received prior radiation therapy. Eighteen were negative for anaplastic lymphoma kinase (ALK-negative), while 22 were positive (ALK-positive). Each patient’s status after both frontline therapy and most recent therapy was recorded: 24 (60%) patients had disease that was refractory to frontline therapy and 25 patients (62.5%) had disease that was refractory to the last therapy before BV.

Response to treatment Best response was observed after a median of four cycles of treatment in 31 (77.5%) patients: 19 (47.5%) obtained a complete response and 12 (30%) achieved a partial response. The overall response rate at the end of the treatment was 62.5% (25 patients), represented by 18 (45%) complete responses and seven (17.5%) partial responses; of the remaining patients, one had stable disease, and 14 patients showed disease progression. The best response rate was higher in the elderly subset (>60 years): nine (64.3%) complete responses and three (21.4%) partial responses for an overall response rate of 85.7%. Four patients who were in complete remission at first restaging relapsed during further BV courses; two patients who had a partial response at first restaging converted to complete response status after the four subsequent infusions. None of the patients who had stable or progressive disease at first restaging had an improvement in their status at the end of therapy. The median number of treatment cycles administered was eight (range, 1-16). With a median follow up of 18 months, the global overall survival rate was 56.9% at 24 months (Figure 1) and the median had not been reached. The progression-free survival rate at 24 months was 39.1%, with the median achieved at 12.5 months (Figure 2). The disease-free survival rate at 24 months was 54% (Figure 3). Of the 19 patients who had a complete response, 4 (21%) relapsed and 15 were in continuous complete response at the last follow up with a median duration of response of 12 months (range, 9-24 months). After controlling for confounding variables, no differences were observed between ALK-negative and ALK-positive patients for any times to endpoints. Among the patients who achieved complete responses, three were given consolidation with transplantation (1 autologous and 2 allogeneic SCT). Currently, 15 patients are

Figure 2. Progression free-survival.

1933

A. Broccoli et al.

still in complete response, including the three who underwent a consolidation procedure. Of the nine elderly patients, six (66.7%) are still in continuous complete response without any consolidation procedure after a median of 14 months. There were five long-term responders, all of whom were still in continuous complete response at the last available follow up. Of note, they were all aged <35 years at the time of starting BV therapy and only one of them had a subsequent consolidative transplant. At the latest follow-up, 27 (67.5%) patients were alive and 13 had died. Of the 13 deaths, 11 were due to lymphoma and two were caused by complications after allogeneic SCT (1 case of respiratory failure related to graft-versus-host disease and 1 case of pneumonia).

Safety All patients who received at least one BV infusion were included in the safety analysis. In general, the treatment was well tolerated and the toxicity profile was very similar to that previously published. Twelve patients had at least one toxicity. All hematologic toxicities were grade 1-2, except one case of grade 3 neutropenia. The extra-hematologic side effects were mostly represented by peripheral sensory neurological toxicity (15/20), including three cases of grade 3. The other adverse events were nausea grade 12 (2 patients), erythema grade 2 (2 patients), and hyposthenia grade 2 (1 patient). Neurological toxicity always reversed completely after the end of treatment. No longterm toxicity related to BV was observed during the followup period, even in patients later subjected to consolidation with transplantation.

term remission, lasting for more than 5 years, in response to single agent BV without any additional anticancer therapy, other than transplantation. In our study the estimated disease-free survival rate at 2 years was 54% and 15 patients (37.5%) were in continuous complete response with a median duration of response of 12 months (range, 9-24 months). As only three of those 15 patients had had transplant consolidation, a comparison between these patients and those who did not undergo a SCT procedure was not possible. Thus, the duration of response and disease-free survival in the real-life experience, confirming the findings of the pivotal study, indicate that a substantial subset of patients with R/R systemic ALCL who have a complete response with single agent BV obtain long-term disease control and may potentially be cured. One important question remains unanswered: which of the patients in complete response may benefit from transplant consolidation? In our series there were five long-term responders, all of whom were still in continuous complete response at the latest follow-up, but of whom only one had undergone a consolidative allogeneic SCT procedure. Updates from the piv-

Discussion This retrospective, large, multicenter Italian study on 40 patients with R/R systemic ALCL treated with BV outside a clinical trial represents the largest ever reported in a realworld context. Our results are in accordance with the pivotal phase II study and its updates and the other national experience studies with an overall response rate of 77.5% and a complete response rate of 47.5% in terms of best response.8,9,11,14-18 In addition, we gained some interesting insights into the role of BV in everyday clinical practice. First, both the best response rate and overall response rate were higher among elderly patients: 85.7% versus 77.5% and 64.3% versus 62.5%, respectively. It was confirmed that having a complete response after four cycles of treatment is very important for classifying a patient as a real good responder; however, the best number of cycles to give with a view to evaluating potential consolidation with transplantation (in most cases with allogeneic transplantation) or continuation with BV until cycle 16 remains an open issue, mainly because in the case of a complete response the choice between the two options is at the physicianâ&#x20AC;&#x2122;s discretion. According to the recent update from Pro et al. on the pivotal phase II study, the 5-year progression-free survival rate was 68% in patients who achieved a complete response and underwent allogeneic SCT versus 47% in patients who continued BV treatment even though they had obtained a complete response after the first four cycles of treatment.8,14 In this update the authors reported that 27.6% of the whole study population achieved long1934

Figure 3. Disease-free survival.

Table 1. Patientsâ&#x20AC;&#x2122; demographics and characteristics at baseline.

Total population, N

ALK-positive ALK-negative Median age, years (range) Median time from diagnosis-BV, years (range) Male, n (%) Stage, n (%) I/II III IV Systemic symptoms, n (%) Refractory to most recent therapy, n (%) Refractory to first-line therapy, n (%) Median number of previous therapies (range) Prior autologous stem cell transplant, n (%) Prior radiotherapy, n (%)

22 18 51.4 (22.6-80.7) 2 (1-16) 28 (70.0) 9 (22.5) 5 (12.5) 26 (65.0) 11 (27.5) 24 (60.0) 25 (62.5) 2 (2-10) 13 (32.5) 8 (20)

ALK: anaplastic lymphoma kinase; BV: brentuximab vedotin.

haematologica | 2017; 102(11)

Brentuximab vedotin in ALCL: real-life experience

otal study and our data could indicate that long-term disease control can be obtained even without transplant consolidation, with a real chance of curing a subset of patients with R/R systemic ALCL with BV alone.14 Physicians are still divided on whether or not to offer a consolidative transplant to patients in complete response because solid clinical trial data are lacking on this issue. A large, welldesigned randomized control study is needed, but ALCL is so rare that we are unlikely ever to have a definitive answer. Differences in survival outcomes between ALK-positive and ALK-negative patients have often been reported: no statistically significant difference was observed between the two subgroups in our sample.21 Our study indicated that for patients who had stable or progressive disease after four cycles of BV the potential conversion rate to partial or complete response with further cycles is close to zero. The final message is that when patients have stable or progressive disease at first restaging, they should be changed rapidly to another treatment. On the other hand, for patients who achieve a partial response by the first restaging, it could be important to continue the treatment: in our series 2/12 (16.7%) patients showed a conversion from a partial to a complete response. In conclusion, the results of this large retrospective study on 40 cases of R/R systemic ALCL in daily practice support the efficacy of single-agent BV, which appears to be a treatment with manageable toxicity without evidence of cumulative toxic effects with previous regimens. We acknowledge that this kind of report has a potential bias given the lack of a predictable and calculated sample

References 8. 1. Savage KJ, Harris NL, Vose JM, et al. ALKanaplastic large-cell lymphoma is clinically and immunophenotypically different from both ALK+ ALCL and peripheral T-cell lymphoma, not otherwise specified: report from the International Peripheral T-Cell Lymphoma Project. Blood. 2008;111(12): 5496-5504. 2. Kewalramani T, Zelenetz AD, TeruyaFeldstein J, et al. Autologous transplantation for relapsed or primary refractory peripheral T-cell lymphoma. Br J Haematol. 2006;134 (2):202-207. 3. Vose JM, Peterson C, Bierman PJ, et al. Comparison of high-dose therapy and autologous bone marrow transplantation for T-cell and B-cell non-Hodgkin’s lymphomas. Blood. 1990(2);76:424-431. 4. Rodriguez J, Munsell M, Yazji S, et al. Impact of high-dose chemotherapy on peripheral T-cell lymphomas. J Clin Oncol. 2001(17);19:3766-3770. 5. Moskowitz CH, Nimer SD, Glassman JR, et al. The International Prognostic Index predicts for outcome following autologous stem cell transplantation in patients with relapsed and primary refractory intermediate-grade lymphoma. Bone Marrow Transplant. 1999;23(6):561-567. 6. Shipp MA, Abeloff MD, Antman KH, et al. International Consensus Conference on high- dose therapy with hematopoietic stem cell transplantation in aggressive nonhodgkin’s lymphomas: report of the jury. J Clin Oncol 1999;17(1):423-429. 7. Younes A, Bartlett NL, Leonard JP, et al. Brentuximab vedotin (SGN-35) for relapsed

haematologica | 2017; 102(11)

10.

11.

12.

13.

14.

size and the risk of underreporting toxicity. However, ALCL is a very rare disease, accounting for approximately 2% to 3% of all lymphoid neoplasms. The phase II study that led to Food and Drug Administration-accelerated approval of BV enrolled 58 patients globally, thus 40 ALCL patients from a single nation is a substantial sample related to this pathology. Nevertheless, we could not analyze prognostic features due to the small sample and we have reported the raw observed data. Compared with randomized controlled trials, observational studies may better identify clinically important adverse events for several reasons. These include longer follow-up times, the inclusion of patients with concomitant illnesses who may be more likely to experience drug interactions or other side effects and the probability of detecting infrequent or rare complications. With regards to the retrospective nature of this specific study, AIFA demands strict monitoring of drugs prescribed under law 648/96 and physicians must report any adverse event occurring during treatment: thus, all the safety data were already in the patients’ chart at the time our retrospective study started. Our report confirms the activity of BV in elderly patients, the duration of the clinical response independently of transplant consolidation, and the relevance to the final response of achieving a complete response after four cycles of treatment. BV is the first drug which has led to a drastic change in the management of ALCL, with an overall response rate of 80%. The next research efforts could be aimed at developing combination regimens with BV to reach a 100% rate of responses in patients with R/R ALCL

CD30-positive lymphomas. N Engl J Med. 2010;363(19):1812-1821. Pro B, Advani R, Brice P, et al. Brentuximab vedotin (SGN-35) in patients with relapsed or refractory systemic anaplastic large-cell lymphoma: results of a phase II study. J Clin Oncol. 2012;30(18):2190-2196. Zinzani PL, Sasse S, Radford J, Shonukan O, Bonthapally V. Experience of brentuximab vedotin in relapsed/refractory Hodgkin lymphoma and relapsed/refractory systemic anaplastic large-cell lymphoma in the Named Patient Program: review of the literature. Crit Rev Oncol Hematol. 2015;95(3): 359-369. Sureda A, Constans M, Iriondo A, et al. Prognostic factors affecting long-term outcome after stem cell transplantation in Hodgkin's lymphoma autografted after a first relapse. Ann Oncol. 2005;16(4):625-633. Gibb A, Jones C, Bloor A, et al. Brentuximab vedotin in refractory CD30+ lymphomas: a bridge to allogeneic transplantation in approximately one quarter of patients treated on a Named Patient Programme at a single UK center. Haematologica. 2013;98(4): 611-614. Illidge T, Bouabdallah R, Chen R, et al. Allogeneic transplant following brentuximab vedotin in patients with relapsed or refractory Hodgkin lymphoma and systemic anaplastic large cell lymphoma. Leuk Lymphoma. 2015;56(3):703-710. Pro B, Advani R, Brice P, et al. Four-year survival data from an ongoing pivotal phase 2 study of brentuximab vedotin in patients with relapsed or refractory systemic anaplastic large cell lymphoma. Blood. 2014;124(21 suppl): Abstract 3095. Pro B, Advani R, Brice P, et al. Five-year sur-

15.

16.

17.

18.

19.

20. 21.

vival data from a pivotal phase 2 study of brentuximab vedotin in patients with relapsed or refractory systemic anaplastic large cell lymphoma. Blood. 2016; 128(22):4144. Broccoli A, Derenzini E, Pellegrini C, et al. Complete response of relapsed systemic and cutaneous anaplastic large cell lymphoma using brentuximab vedotin: 2 case reports. Clin Lymphoma Myeloma Leuk. 2013;134 (4):493–495. Monjanel H, Malphettes M, Deville L, et al. Brentuximab vedotin in heavily treated Hodgkin and anaplastic lymphoma, a single center study on 45 patients. Hematol Oncol. 2013;31(s1):268. Lamarque M, Bossard C, Contejean A, et al. Brentuximab vedotin in refractory or relapsed peripheral T-cell lymphomas: the French named patient program experience in 56 patients. Haematologica. 2016; 101(3):e103-106. Gandolfi L, Pellegrini C, Casadei B, et al. Long-term responders after brentuximab vedotin: single-center experience on relapsed and refractory Hodgkin lymphoma and anaplastic large cell lymphoma patients. Oncologist. 2016; 21(12):14361441. Argnani L, Zinzani PL. Reporting real-life experience with drugs in lymphoma patients. Hematol Oncol. 2016 Mar 21. doi: 10.1002/hon.2284 [Epub ahead of print] Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5): 579-586. Beaven AW, Diehl LF. Peripheral T-cell lymphoma, NOS, and anaplastic large cell lymphoma. Hematology Am Soc Hematol Educ Program. 2015;2015:550-558.

1935

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Cell Therapy & Immunotherapy

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1936-1946

Amina Kariminia,1 Sabine Ivison,1 Bernard Ng,2 Jacob Rozmus,1 Susanna Sung,1 Avani Varshney,1 Mahmoud Aljurf,3 Sylvie Lachance,4 Irwin Walker,5 Cindy Toze,6 Jeff Lipton,7 Stephanie J. Lee,8 Jeff Szer,9 Richard Doocey,10 Ian Lewis,11 Clayton Smith,12 Naeem Chaudhri,3 Megan K. Levings,13 Raewyn Broady,6 Gerald Devins,7 David Szwajcer,14 Ronan Foley,5 Sara Mostafavi,2 Steven Pavletic,15 Donna A. Wall,16 Stephan Couban,17 Tony Panzarella7 and Kirk R. Schultz1

Michael Cuccione Childhood Cancer research Program, BC Children’s Hospital, Department of Pediatrics, University of British Columbia, Vancouver, BC, Canada; 2 Department of Statistics, University of British Columbia, Centre for Molecular Medicine and Therapeutics, Vancouver, BC, Canada; 3King Faisal Specialist Hospital and Research Center, Riyadh, Kingdom of Saudi Arabia; 4Hôpital Maisonneuve-Rosemont, Université de Montréal, QC, Canada; 5Hamilton Health Sciences Centre and McMaster University, Hamilton, ON, Canada; 6Leukemia/Bone Marrow Transplant Program of BC, Vancouver General Hospital, British Columbia Cancer Agency and the University of British Columbia, Vancouver, BC, Canada; 7Princess Margaret Cancer Centre University of Toronto, ON, Canada; 8Fred Hutchinson Cancer Research Centre, Seattle, WA, USA; 9 Royal Melbourne Hospital and University of Melbourne, Australia; 10Auckland City and Starship Children’s Hospital, Auckland, New Zealand; 11Institute of Medical and Veterinary Sciences, Adelaide, Australia; 12General Hematology, Blood Cancers and Bone Marrow Transplant Program, University of Colorado Hospital, Aurora, CO, USA; 13 BC Children’s Hospital Research Institute and Department of Surgery, University of British Columbia, Vancouver, BC, Canada; 14CancerCare Manitoba, Winnipeg, MB, Canada; 15Center for Cancer Research, National Cancer Institute, Bethesda, MD, USA; 16 The Hospital for Sick Children and University of Toronto, ON, Canada and 17Nova Scotia Health Authority and Dalhousie University, Halifax, NS, Canada 1

Correspondence: kschultz@mail.ubc.ca

ABSTRACT Received: April 15, 2017. Accepted: September 15, 2017. Pre-published: September 21, 2017. doi:10.3324/haematol.2017.170928 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1936 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1936

andomized trials have conclusively shown higher rates of chronic graft-versus-host disease with filgrastim-stimulated apheresis peripheral blood as a donor source than unstimulated bone marrow. The Canadian Blood and Marrow Transplant Group conducted a phase 3 study of adults who received either filgrastim-stimulated apheresis peripheral blood or filgrastim-stimulated bone marrow from human leukocyte antigen-identical sibling donors. Because all donors received the identical filgrastim dosing schedule, this study allowed for a controlled evaluation of the impact of stem cell source on development of chronic graft-versus-host disease. One hundred and twenty-one evaluable filgrastim-stimulated apheresis peripheral blood and filgrastim-stimulated bone marrow patient donor products were immunologically characterized by flow cytometry and tested for their association with acute and chronic graft-versus-host disease within 2 years of transplantation. The immune populations evaluated included, regulatory T cells, central memory and effector T cells, interferon γ positive producing T cells, invariate natural killer T cells, regulatory natural killer cells, dendritic cell populations, macrophages, and activated B cells and memory B cells. When both filgrastim-stimulated apheresis peripheral blood and filgrastimstimulated bone marrow were grouped together, a higher chronic graftversus-host disease frequency was associated with lower proportions of CD56bright natural killer regulatory cells and interferon γ-producing T helper cells in the donor product. Lower CD56bright natural killer regulatory cells displayed differential impacts on the development of extensive chronic graft-versus-host disease between filgrastim-stimulated apheresis haematologica | 2017; 102(11)

NKreg cells Decrease Chronic GvHD in G-PB Transplants

peripheral blood and filgrastim-stimulated bone marrow. In summary, while controlling for the potential impact of filgrastim on marrow, our studies demonstrated that CD56bright natural killer regulatory cells had a much stronger impact on filgrastim-stimulated apheresis peripheral blood than on filgrastim-stimulated bone marrow. This supports the conclusion that a lower proportion of CD56bright natural killer regulatory cells results in the high rate of chronic graft-versus-host disease seen in filgrastim-stimulated apheresis peripheral blood. clinicaltrials.gov Identifier: 00438958.

Introduction Filgrastim granulocyte-colony stimulating factor (GCSF)-stimulated apheresis peripheral blood (G-PB) as a donor source is clinically well established due to rapid engraftment, ease of collection, and similar survival to marrow as a donor source. G-PB is limited by a significantly higher rate of chronic graft-versus-host disease (cGvHD),1-8 purported to be due to the infusion of increased donor product T cell numbers.9 Other studies have suggested that the CD34+ cell donor load,10,11 activated HLA-DR+ T cells,12 and possibly the total nucleated cell dose13 impact on the development of cGvHD after G-PB transplantation, however, all such studies are limited by the lack of comparison to a control marrow transplanted population. Cell populations are associated with acute GvHD (aGvHD) after G-PB which includes dendritic cells14 and Treg cells15 but do not have any association with cGvHD. Other donor cell populations have found no association for either aGvHD or cGvHD.16 To date no study has definitely established which immune cell populations are most responsible for the higher rate of cGvHD associated with the G-PB donor source compared to marrow. Until the specific, unique components in G-PB versus marrow as a source are identified, it remains difficult to develop graft manipulation strategies to modulate cGvHD. The Canadian Blood and Marrow Transplant Group (CBMTG) undertook a definitive phase 3 trial comparing G-bone marrow (BM) with G-PB in sibling allografts for adults with hematologic malignancies. In that study, the CBMTG showed that cGvHD was lower with G-BM (HR=0.66; 95% CI 0.46 – 0.95; P=0.007).17 This study presented an unprecedented opportunity to evaluate the impact of graft source on the development of cGvHD with both donor sources receiving G-CSF treatment using an identical regimen. The population was relatively homogenous as only human leukocyte antigen (HLA)identical sibling donor (8/8 or 7/8 HLA-match) was used for predominantly myeloid malignancies with the only variable being the method of collection (i.e., marrow harvest versus apheresis). We hypothesized that one of the immune cell populations previously identified by correlative cGvHD biology and biomarkers studies would correlate with the induction of cGvHD by G-PB donor product. To test this hypothesis, we evaluated both G-BM and GPB donor grafts combined for donor product immune cells for any specific cell types correlation with the development of cGvHD. Once identified, we evaluated the relative impact of each immune cell population on cGvHD for the relative impact of the two donor sources, G-PB versus G-BM, on the development of cGvHD. The immune populations evaluated included: regulatory T cells, central memory and effector T cells, interferon (IFN)γ+ producing T cells, regulatory natural killer (NK) cells, invariant natural killer T (iNKT) cells, plasmacytoid haematologica | 2017; 102(11)

and myeloid dendritic cells, macrophages, activated B cells, and memory B cells.

Methods Clinical Study Design Samples for the current study were obtained as part of a larger clinical study (CBMTG 0601), a randomized phase 3, parallel group trial conducted by the CBMTG at 13 centers in Canada, Saudi Arabia, Australia, New Zealand, and the USA. The institutional research ethics board at each center approved the trial and recipients and donors both gave informed consent before randomization. Recipients were between 16 and 65 years of age and with a hematologic malignancy. Donors were 7/8 or 8/8 HLA-matched siblings medically fit to receive G-CSF and undergo a marrow harvest or apheresis. This study has been described previously.17

Patient and Donor Characteristics CBMTG 0601 comprised 223 donor-recipient pairs randomized between April 2007 and January 2012 with 223 evaluable pairs. Of the entire 223 evaluable patients from the clinical trial, 121 had evaluable samples for the current correlative studies. The primary analysis was performed on patients who had survived up to 2 years after BM transplantation (BMT) (> 95% of patients developed overall cGvHD by 2 years), with the omission of patients due to death and leukemia relapses that occurred before the onset of cGvHD (Table 1; n = 89). We found no significant difference between the 121 evaluated and the 102 not included in the analysis for cGvHD (65% vs. 59%), death (34% vs. 44%), relapses (29% vs. 24%), or time to cGvHD (day 180±112 vs. day 185±124), respectively. We defined overall cGvHD as including both limited and extensive cGvHD, and will from now on refer to overall cGvHD as cGvHD, unless specified as extensive cGvHD. Confirmatory analysis was performed on all evaluable patients including those with a death or relapse before 2 years. Both subgroups had similar patient characteristics to the entire population in the study (Table 1). A comprehensive immune evaluation of T cells, B cells, NK cells, iNKT cells, macrophages, and dendritic cell populations was performed on the donor product for a number of immune populations (Online Supplementary Table S1) and tested for association with aGvHD and cGvHD within 2 years of transplantation. Additional evaluations examined the association of identified immune cell populations and the development of aGvHD and cGvHD for the graft source (G-PB or G-BM) and other clinical factors, including transplant related mortality (TRM), relapse, previous aGvHD before the onset of cGvHD, donor-recipient sex differences, acute myeloid leukemia (AML) versus no AML, total body irradiation (TBI) versus no TBI, recipient age and donor age.

Sample processing for biological studies using immunophenotyping and functional assays Samples from allografts were couriered overnight at room temperature to a central laboratory located at BC Children’s Hospital 1937

A. Kariminia et al.

Research Institute in Vancouver, Canada; peripheral blood mononuclear cells (PBMCs) were isolated and frozen on arrival. Batched samples were thawed using 1 x 106 viable cells per assay. Immunophenotyping and functional assays evaluated T cell, B cell, dendritic cell, monocyte, and NK cell populations (Online Supplementary Table S1). Data were acquired using LSR II flow cytometer (BD Biosciences) and analyzed by FlowJo v10 (TreeStar, Ashland, OR, USA). Details of the cell immunophenotyping strategy can be found in Table 1. Graft composition was evaluated as the percentage of cell population per donor lymphocytes. On smaller cell count samples, there was a prioritization for assays with immunophenotyping to be carried out first, followed by functional stimulations assays for cytokine production if sufficient samples were available.

Study endpoints The primary biologic endpoint of the analysis was an association of cGvHD with a number of subpopulations of T cell, B cell, NK cell, iNKT cell, macrophages, and dendritic cell populations. aGvHD and cGvHD were characterized according to Przepiorka et al.18 and Sullivan et al.,19 respectively. Any GvHD was defined as chronic GvHD and/or acute GvHD (grade 1 – IV aGvHD). Chronic GvHD was defined as an initial diagnosis of cGvHD within 2 years of transplantation.

Statistical analysis The effect of candidate immune cell populations on the development of cGvHD (within a 2 year period from transplantation) was tested using a univariate logistic regression model. Patients

Table 1. Baseline characteristics: overall vs. studied populations.

Overall Population in clinical trial (N = 223)

Variable Disease Stage Early Late Disease CML AML MYELO Other Conditioning Regimen BU+CY CY+TBI Other: Fludarabine + Melphalan VP-16 + TBI Fludarabine+Busulfan Donor Sex Male Female Donor median age (range) Donor CMV Status Positive Negative Recipient Sex Male Female Recipient median age (range) Recipient CMV Status Positive Negative

Confirmatory population including early deaths, relapse1 (N = 121)

Primary studied population excluding early deaths, relapses1 (N = 89)

G-BM (%) N=113

G-PB (%) N=110

G-BM (%) N=52

G-PB (%) N=69

G-BM (%) N=39

G-PB (%) N=50

71 (63) 42 (37)

77 (70) 33 (30)

34 (65) 18 (35)

49 (71) 20 (29)

26 (67) 13 (33)

33 (66) 17 (34)

7 (6) 53 (47) 15 (13) 38 (34)

6 (5) 54 (49) 9 (8) 41 (37)

5 (10) 26 (50) 5 (10) 16 (30)

5 (7) 35 (58) 7 (10) 22 (32)

5 (13) 18 (46) 4 (11) 12 (31)

4 (8) 27 (54) 6 (12) 13 (26)

65 (58) 2 (2)

57 (51) 2 (2)

40 (35) 5 (4) 1 (1)

41 (37) 7 (6) 3 (3)

35 (67) 0 0 16 (30) 1 (3) 0

36 (52) 0 0 24 (35) 6 (9) 2 (3)

29 (74) 0 0 9 (23) 1 (3) 0

29 (58) 0 0 13 (26) 5 (10) 2 (4)

66 (58) 47 (42)

66 (60) 44 (44)

30 (58) 22 (42)

44 (64) 25 (36)

22 (57) 17 (43)

32 (64) 18 (36)

43 (18 – 66)

41 (18 – 64)

45.5 (18-62)

44 (18-66)

49 (20-62)

46 (19-64)

69 (61) 44 (39)

57 (52) 53 (48)

26 (50) 26 (50)

26 (38) 43 (62)

19 (49) 20 (51)

20 (40) 30 (60)

69 (61) 44 (39) 43 (16 – 63)

57 (52) 53 (48) 46 (16 – 64)

34 (65) 18 (35) 48 (20-63)

39 (57) 30 (43) 46 (16-64)

25 (64) 14 (36) 49 (22-63)

29 (58) 21 (42) 47 (17-59)

62 (55) 51 (45)

59 (54) 51 (46)

23 (44) 29 (56)

32 (46) 37 (54)

17 (44) 22 (66)

22 (44) 28 (56)

1 Early deaths and relapses were defined as occurring before 2 years post BMT or before the onset of cGvHD. BU: busulfan; CY: cyclophosphamide; VP=16 = etoposide; CMV: cytomegalovirus; MYELO: myelodysplastic disease; CML: Chronic Myeloid Leukemia; AML; Acute Myeloid Leukemia; G-BM: G bone marrow; G-PB: G peripheral blood; TBI: total body irradiation.

1938

haematologica | 2017; 102(11)

NKreg cells Decrease Chronic GvHD in G-PB Transplants

who relapsed or those who died before occurrence of cGvHD were excluded from the logistic regression analysis, since it could not be established whether or not they would have developed cGvHD. Of the initial 121 patients, 89 met the inclusion criteria. As these were exploratory analyses no statistical adjustments were made for multiple comparisons, thus a P-value threshold of 0.01 was used in the primary analysis and 0.05 in all other secondary analyses. All analyses were performed using MATLAB. For the two cell populations that were found to be significant, due to smaller cell numbers in some donor graft samples, 7 of the 89 patients did not have immunophenotyping for CD56bright NKreg cells and 11 of the 89 patients did not have functional stimulation and immunophenotyping for the CD4+ T cell IFNγ+ (Figure 1). To confirm that excluding patients who displayed relapse or died before cGvHD did not introduce biases, we also employed a univariate Cox proportional hazards model20 with those patients included, and used the time to cGvHD onset as the response. The time to cGvHD for said patients was considered as censored under the Cox model. As to visualization, the identified cell populations were plotted with values split by GvHD status. Multivariate analysis on all significant cell populations identified in the univariate analysis was also performed to test for the unique effect of each cell population. This analysis was applied only on patients that had all these cell populations, resulting in 75 and 94 patients being evaluated for the logistic regression model and Cox model, respectively (Figure 1). Furthermore, logistic regression was applied to examine the effect of aGvHD, sex, TBI, recipient age, donor age, AML, death, relapse as well as donor source on the identified cell populations. Moreover, with patients split by cGvHD status, optimal cut points for the identified cell populations were determined by plotting their receiver operating characteristic (ROC) curve and by finding the point on the ROC curve that is closest to the point of perfect sensitivity and specificity. Lastly, we examined the interaction effect between donor source and the identified significant markers on cGvHD status using logistic regression and the Cox model.

Role of the funding source The National Cancer Institute of the National Institutes of Health (NIH) funded the study herein following peer review, but had no direct influence on the study design, the collection, analysis, and interpretation of data, in the writing of the report or in the decision to submit the paper for publication. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results Evaluation of donor graft composition for immune populations associated with the development of cGvHD Immunophenotypic and functional evaluations were performed for a large number of CD4+ and CD8+ T cell, NK cell, B cell, macrophages, plasmacytoid and myeloid dendritic cell, iNKT cell, and regulatory T cell populations as outlined in Online Supplementary Table S1 and correlated with the presence of cGvHD. We also evaluated the activation status of CD4+and CD8+ T cells by CD25 and human leukocyte antigen – antigen D related (HLA-DR) expression and found no difference. Initial analyses evaluated candidate immune cell populations for correlation with cGvHD followed by analysis for extensive cGvHD. The two donor sources, G-PB and G-BM, were grouped together for these analyses. We found no significant associations (at P<0.01) with cGvHD except for two populations, CD56bright NKreg cells (Table 2; P=0.003) and IFNγ+ CD4+ T cells (P=0.002). A confirmatory analysis was further performed that included patients who either died or developed a leukemia relapse before the onset of cGvHD and before 2 years following transplantation. These included 107 patients for CD56bright NKreg cells and 100 patients for IFNγ+ CD4+ T cells as a marker (Figure 1). We confirmed the results of

Table 2. Association between donor cell population numbers with development of cGvHD (excludes all patients with a death or relapse before 2 years or overall cGvHD).

Variable CD56bright NKreg cells Overall cGvHD (N = 821) Extensive cGvHD (N = 78) CD4+ T cells IFNγ Overall cGvHD (N = 78) Extensive cGvHD (N = 75)

Percentage cells per total lymphocytes – Independent variable1 P (Wald test) Odds Ratio (95% CI) 0.003 0.005

0.13 (0.03- 0.49) 0.19 (0.06-0.61)

0.002 0.002

0.77 (0.66- 0.91) 0.83 (0.74-0.94)

Percentage cells per total lymphocytes – Multivariate analysis2 CD56bright NKreg cells Overall cGvHD Extensive cGvHD CD4+ T cells IFNγ Overall cGvHD Extensive cGvHD

0.02 0.01

0.16 (0.03- 0.74) 0.19 (0.05-0.69)

0.007 0.01

0.77 (0.64- 0.93) 0.84 (0.73-0.96)

1 The multivariate analysis based on logistic regression model was performed only on patients who had both values resulting in a decrease to N=75. 2The multivariate analysis based on the Cox proportional hazards model was performed only on patients who had both values resulting in a decrease to N=94. cGvHD: chronic graft-versus-host disease; IFN: interferon; NK: natural killer.

haematologica | 2017; 102(11)

1939

A. Kariminia et al.

the logistic regression analysis for both CD56bright NKreg cells with cGvHD (Online Supplementary Table S2; OR=0.54, P=0.02) and IFNγ+ CD4+ T cells with cGvHD (Online Supplementary Table S2, OR=0.93, P=0.001). Further analysis focused on these two populations (CD56bright NKreg cells and IFNγ+ CD4+ T cells) as outlined below.

Association of donor CD56bright NKreg cell composition with the development of cGvHD Significant associations between a lower percentage of donor CD56bright NKreg cells per total lymphocytes and development of any GvHD (aGvHD and/or cGvHD; P=0.003) as well as cGvHD only (Figure 2A and Table 2; OR=0.13; P=0.003) were found. Further analysis, limited to extensive cGvHD alone, was similar with logistic regression (Figure 2B, OR=0.19; P=0.005). A significant

association was also found with aGvHD status (P=0.02). We confirmed that the CD56bright NKreg cell population was the classic regulatory NK (NKreg) population by further evaluation for expression of CD335 (NKp46), CD336 (NKp44), and CD337 (NKp30) on all samples from the study population of G-BM and G-PB NKreg cells (Figure 2C). The expression of CD335 (NKp46) was found to be significantly higher in CD56bright NKreg cells with comparable expression of CD337 in both subpopulations consistent with the NKreg phenotype.21,22 From now on, we will refer to this population (CD3–, CD56bright/CD335bright/perforin–/granzyme B–/CD16+/-) as CD56bright NKreg cells. We evaluated whether cytomegalovirus (CMV) seropositivity of the donor impacted on the presence of CD56bright NKreg cells present, and found no significant difference in CD56 bright NKreg cells for CMV seropositive (0.59±0.50% CD56bright NKreg cells per total lymphocytes) versus CMV

Figure 1. Algorithm of sample analysis for immune population that correlate with the development of overall cGvHD. G-PB: G peripheral blood; cGvHD: chronic graft-versus-host disease; aGvHD: acute graft-versus-host disease; IFN: interferon; NK: natural killer; DC: dendritic cells; TRM: transplant related mortality.

1940

haematologica | 2017; 102(11)

NKreg cells Decrease Chronic GvHD in G-PB Transplants

seronegative (0.55±0.42%, P=0.65) donors. A ROC curve to predict the development of cGvHD by 2 years was calculated for the CD56bright NKreg population, and we found an area under the curve (AUC) of 0.85 (Figure 2D).

Association of donor IFNγ producing CD4+ T cells with the development of cGvHD Our group, and others, have previously shown that lower numbers of IFNγ producing cells were associated with cGvHD.23 Cytokine production was measured after

mitogen stimulation in vitro (PMA/Ionomycin). We identified a significant association between lower numbers of IFNγ-producing CD4+ T cell population (CD4+/ CD3+/IFNγ+/IL-4–/IL-17–) and development of any GvHD (either aGvHD and/or cGvHD; P=0.002) and cGvHD alone (Figure 3A and Table 2, OR=0.77, P=0.002). Further analysis, limited to extensive cGvHD alone, was similar for logistic regression (Figure 3B and Table 2, OR=0.83, P=0.001). A ROC analysis revealed an AUC of 0.88 (Figure 3C).

haematologica | 2017; 102(11)

Figure 2. Correlation of donor CD56bright NKreg cells infusion characteristics with acute and chronic GvHD. (A) Box and whisker plot of CD56bright NKreg cells percentage per total lymphocytes for overall cGvHD, which was calculated as the total nucleated cells gated on FSC/SSC plots, then the percentage of each subpopulation of interest was determined as a percentage of the total number of lymphocytes (based on forward and side scatter). The phenotype of CD56bright NK cells was CD56bright CD3-CD16-perforin-granzyme B. (B) The correlation of CD56bright NKreg cells with extensive cGvHD. (C) Representative dot plot showing CD56bright and CD56dim subpopulations of NKreg cells. Mononuclear cells were stained by appropriate conjugated mAbs for surface markers. The cells were then fixed and treated to be permeable, followed by intracellular staining (Online Supplementary Table S1). The data was acquired by LSRII equipped with four lasers. A minimum of 1x106 cells were acquired. Single cells were gated based on FSC-A vs. FSC-H and dead cells were excluded using fixable viability dye. Lymphocytes were gated as FSC lo and SSC lo. Upper left dot plot: NK cells were defined as CD3–CD56+ cells. Two populations are revealed based on expression of CD56, CD56bright (gate A) and CD56dim (gate B). CD56brightcells express a significantly higher level of activating receptor CD335 (NKp46) compared to CD56dim. CD56bright cells express a significantly lower level of molecules involved in killing, Granzyme B (bottom left dot plot) and Perforin (bottom right dot plot). (D) A ROC analysis was used to determine an ‘optimal’ cut point for correctly predicting the occurrence of overall chronic GvHD. ‘Optimal’ was defined as the point on the ROC curve with the shortest distance from the point (0, 1). The point (0, 1) represents the ideal, 100% sensitivity and 100% specificity. GvHD: graft-versus-host disease; NK: natural killer; AUC: area under the curve.

1941

A. Kariminia et al. A

Stimulated

Unstimulated

Figure 3. Correlation of donor IFNγ producing CD4+ T cells infusion characteristics with acute and chronic GvHD. (A) Box and whisker plot of the correlation of IFNγ producing CD4+ T cells percentage with overall cGvHD. The population was calculated as a subpopulation of CD3+ lymphocytes. PBMCs were stimulated with PMA (100ng/ml)/Ionomycin (1 mg/ml) in the presence of monensin for 6 hours. Unstimulated cells treated with monensin were used as control. At the end of incubation, the cells were harvested and surface markers were stained for CD3, CD4, and CD8, in addition to fixable viability to distinguish dead cells. Intracellular staining was performed to detect production of IL-17, IL-4 and IFNγ. The data was acquired using BD LSRII and analyzed by FlowJo v9. Hierarchical gating; 1: lymphocytes were selected based on FSC/SSC, 2: exclusion of dead cells, 3: selection of CD3+ T cells, and 4: determination of IFNγ producing CD4+ T-cells. The data is presented as % of CD4+IFNγ+ T cells per CD3+ lymphocytes. (B) Correlation of IFNγ producing CD4+ T cells with extensive cGvHD; (C) A representative dot plot of INFγ+ T helper cells. Mononuclear cells were seeded at a density of 1x106 per milliliter of culture medium (RPMI-1640 supplemented with 10% heat-inactivated FBS and 2mM l-glutamine), and the cells incubated in CO2 incubator providing 95% oxygen and 5% CO2 at 37 degrees centigrade. The cells were stimulated for 6 hours with 100ng/ml PMA and 1 mg/ml Ionomycin in presence of golgi inhibitor monensin (BD Biosciences; following manufacturers' instructions). As control, cells were cultured without stimulator, but received monensin. The cells were then harvested and surface staining was performed to detect CD3+CD4+ cells. Then the cells were fixed and treated to be permeable. Intracellular staining was performed to detect IFNγ, IL-4 and IL-17. At least 1x106 cells were acquired. Gating hierarchy; 1: single cells (FSC-A vs. FSC-H), 2: viable cells (fixable viability dye negative), and 3: lymphocytes 4-CD3+CD4+ cells. The percentage of Th1 (IFNγ+), Th17 (IL-17+) and Th2 (IL4+) were determined after setting quadrant based on unstimulated cells (upper row). Dot plots in lower row show cytokine expression after stimulation. D) A ROC analysis was used to determine an ‘optimal’ cut point for correctly predicting the occurrence of chronic GvHD. The optimal cut point for IFNγ producing CD4+ T cells/CD3+ T cells was 13.9%. (E) Logistic regression was performed with the two identified markers as predictors and overall cGvHD status as response. Each sample was given an estimated probability of overall cGvHD based on the fitted model. The ROC was generated by applying different probability thresholds. GvHD: graft-versus-host disease; IFN: interferon; AUC: area under the curve.

1942

haematologica | 2017; 102(11)

NKreg cells Decrease Chronic GvHD in G-PB Transplants

A multivariable logistic regression analysis was performed on samples with both CD56brightNKreg cells and IFNγ+ CD4+ T cells (N=94). The decrease from the original 121 to 94 patients was due to small cell numbers in donor samples and patients with relapse or death before cGvHD who were further removed (Table 2). We found that both IFNγ+ CD4+ T cells (OR=0.84; P=0.01) and CD56brightNKreg cells (OR=0.19; P=0.01; Table 2) maintained their significance, suggesting that each of these cell populations has some unique attributes that significantly relate to cGvHD status. We also found that the combination of CD56bright NKreg cells and IFNγ+ CD4+ T cells resulted in a higher ROC AUC of 0.91 (Figure 3D) than that of using each cell population alone.

Correlation of clinical factors and donor immune populations Each of the two cell markers was evaluated for any impact that the following clinical factors may have on their interpretation: clinical donor and recipient age, sex mismatch between donor and recipient, AML versus no AML, TBI versus no TBI, and presence or absence of aGvHD. Because all donors were related, 7/8 or 8/8 HLA

matches and received a myeloablative preparative regimen these variables were not evaluated. Only the IFNγ+ CD4+ T cell donor population correlated with a modest decrease in transplant related mortality (Online Supplementary Table S3).

Evaluation of the impact of donor immune populations on probability of G-BM and G-PB developing cGvHD The CBMTG 0601 trial was a prospective randomized non-blinded study comparing donor G-CSF stimulated marrow versus G-CSF stimulated peripheral blood. Specific, well-defined clinical endpoints, including cGvHD, were documented up to 2 years post-transplant; this allowed us to directly compare the impact of each cell population in marrow (G-BM) versus peripheral blood (G-PB) allografts. Using an interaction test, we evaluated whether either of the two populations were different in terms of their impact on G-BM versus G-PB. While the CD56bright NKreg cell population showed no significant impact on overall cGvHD in either donor source (Figure 4A; P=0.15), it did show a significant impact on the development of extensive cGvHD (Figure 4B; P=0.05) after G-PB transplantation compared to G-BM. By contrast, IFNγ+ CD4+ T cells appeared to have no significant impact

Figure 4. Impact of donor IFNγ producing CD4+ T cells and CD56bright NKreg cells infusion characteristics on cGvHD by donor source using G-PB or G-BM transplantation. (A) The estimated probability of overall cGvHD by treatment (donor source G-BM versus G-PB) as a function of the CD56bright cells per total lymphocytes. B) Estimated probability of extensive cGvHD as a function of CD56bright cells per total lymphocytes by donor source (G-PB or G-BM). C) Estimated probability of overall cGvHD as a function of donor IFNγ producing CD4+ T cells by donor source (G-PB or G-BM); D) Estimated probability of extensive cGvHD as a function of donor IFNγ producing CD4+ T cells by donor source (G-PB or G-BM). GvHD: graft-versus-host disease; IFN: interferon.

haematologica | 2017; 102(11)

1943

A. Kariminia et al.

on either G-BM or G-PB and later development of cGvHD (Figure 4C; P=0.58) or extensive cGvHD (Figure 4D; P=0.18).

Discussion G-CSF-mobilized peripheral blood apheresis donor product is used by a large number of BMT centers, despite the fact that it has a significantly higher rate of cGvHD compared to harvested bone marrow donor product. This major limitation could be minimized if the immune cellular component that influences the higher rate of cGvHD associated with G-PB were characterized. The CBMTG 0601 protocol comparing marrow versus apheresis peripheral blood donor product offered a unique opportunity to evaluate the impact of apheresis PB when both donor populations were treated with an identical dosing schedule of GCSF and collected with identical timing. Moreover, the study population were all adults (> 16 years of age) with related donors (HLA 8/8 and 7/8 matching); using a minimization randomization ensured that the recipient populations were matched for important contributing factors such as preparative regimen and underlying disease.17 A comprehensive evaluation of donor immune cell components that had previously been associated with the development of cGvHD allowed for an identification not only of populations associated with both donor sources, but, more importantly, of those associated with G-PB. We initially evaluated both G-BM and G-PB together and found two significant associations in two donor cell populations for both overall and extensive cGvHD. Both had inverse relationships with the development of overall cGvHD (IFNγ+ CD4+ T cells and CD56bright NKreg cells) suggesting regulatory functions. We then looked at the populations in a broader context to ensure that the observations were consistent, and found: a) no correlation with other factors except aGvHD in NKreg cell and TRM in the T cell population, and b) the association was consistent when a secondary analysis included patients that died or relapsed before 2 years after BMT. We also found that the inclusion of both populations together increased the association with cGvHD. Having looked at the two populations in the overall population, we subsequently looked at the impact on cGvHD in the two donor sources of graft (G-BM and GPB) separately. We found that the NKreg population had a proportionately greater impact on extensive cGvHD in GPB compared to G-BM. This controlled evaluation supports the importance of CD56bright NKreg cells as a suppressive immune population on cGvHD in related donor G-PB transplantation. It is now well established that strategies that impact the graft cellular composition at the time of transplant can impact the development of cGvHD many months later. As an example, in vivo depletion of T cell and B cells with either anti-thymocyte globulin or alemtuzumab, in vivo depletion of activated T cell and B cell populations using post transplantation cyclophosphamide, and ex vivo depletion of T and B cell populations in haploidentical transplants can reduce the cumulative incidence of cGvHD.24-30 A number of donor cell populations have been associated with the onset of cGvHD. These include T cells, B cells and dendritic cells.31-33 We have previously shown that activated B cells (CpG oligodeoxynucleotide (ODN) responsive, TLR9+) are associated with increased cGvHD, 1944

whereas regulatory T cells and IFNγ producing T cells are associated with decreased cGvHD.23,34 G-CSF administration may influence allograft cellular composition in marrow and peripheral blood products.35 One study found that the number of donor naïve and memory T cell subsets correlated with infections and aGvHD, and were impacted by whether the graft source was unstimulated marrow or G-CSF-stimulated apheresis donor product.21 The most comprehensive study was BMT CTN 0201,36 which evaluated the impact of donor G-PB versus unstimulated marrow as the donor product. The BMT CTN 0201 study analysis differed from our study in that their primary analysis focused on: a) unrelated donor sources, b) unstimulated marrow as the control rather than G-BM as in our analysis, and c) overall survival rather than cGvHD. They found that plasmacytoid dendritic cells (pDCs) and naïve T cells were associated with improved overall survival but not with cGvHD. Similar to our study, they found that the T cell content of the G-PB was higher than that of BM grafts. In spite of these differences, they found no increased incidence of cGvHD associated with donor graft CD8+ or CD4+ T cell populations, including those expressing CD45RA, CCR7, and CD62L, CD127, and Ki-67, for regulatory cells or for NKT cells. In the BMT CTN 0201 study, it appears that neither of the two populations identified in the current study, CD56bright NKreg cells or IFNγ producing CD4+ T cells, were included in their evaluations. Our study reports a strong association of CD56bright NKreg cells with a lower rate of cGvHD in both G-BM and G-PB. CD56bright NK cells were first described in 1992 as IL-2 responsive group with the high affinity IL-2 receptor.37 CD56bright NKp46 cells (NKregs) have been associated with lower GvHD in other small trials.38 The CD56bright NKreg population has abundant immunoregulatory cytokines, is located primarily in secondary lymphoid tissues, and has low cytotoxicity. The cytokine-secreting CD56bright CD16dim cells express high levels of inhibitory CD94/NKG2A complex, CD25, and CD117, recognize HLA-E but lack inhibitory major histocompatibility complex (MHC) class 1a allele specific KIRs.39 Unfortunately, KIR data was not collected as part of these studies and could not be further evaluated in these analyses. Expression of CD117 and NKp46 are typical for some populations of innate lymphocytes, associated with a lack of acute GVHD,40 which we observed in this study. Our group has previously shown an inverse relationship of CXCR3+ CD56bright NK cells with the onset of cGvHD in a large adult population,41 further supporting the important role of this population in cGvHD. Moreover, the impact of G-CSF on the induction of CD34+ progenitors for growth into an innate lymphoid effector population appears to be different in marrow versus PB.42 The other immune suppressive population that had an equal impact regardless of donor source, G-BM and G-PB, was the association of lower proportion IFNγ+ CD4+ T cells with cGvHD. We have previously observed that an increase in IFNγ was associated with a lower onset of late cGvHD in pediatric hematopoietic stem cell transplantation (HSCT) recipients, but had hypothesized that it would have been secreted by an NKreg population21 as opposed to a CD4+ T cell population. Murine models have shown that the role of IFNγ in GvHD appears to be variable depending on specific times post BMT, as early administration of recombinant IFNγ prevents CD4+ T cell–mediated GVHD.43 Support for this hypothesis haematologica | 2017; 102(11)

NKreg cells Decrease Chronic GvHD in G-PB Transplants

comes from the fact that donors who have microsatellite polymorphisms with decreased IFNγ production have higher rates of cGvHD.44 In mouse models, high IFNγ production by NK T cells results in lower rates of cGvHD.45 Interestingly, IFNγ is not necessary for the development of GvHD in many murine GvHD models,46,2 and disease can progress despite a lack of IFNγ. The role of IFNγ+ CD4+ T cells in the induction of immune tolerance is not well understood. One mechanism may be that classic Th1 IFNγ+ CD4+ helper T cells induce immune tolerance via the activation of Th1 natural Treg (nTreg).47 Another possibility is that IFNγ inhibits donor T-cell expansion by promoting apoptosis and suppressing proliferation, thereby eliminating alloreactive T cells in GvHD tissues by interacting with recipient non-hematopoietic cells and upregulating programmed cell death (PD)-1L expression.48 A third possibility is that the IFNγ+ CD4+ T cell population represents a Th1 Treg population49 that is primed to progress to an IL10 producing Treg or Tr1 cell population. Lastly, IFNγ−licensed mesenchymal stem cells inhibit proliferation of activated T cells through both an indoleamine 2,3dioxygenase (IDO) and, possibly, PD-1 dependent manner.50 Whatever their role, this population requires further study in its potential to predict a later onset of cGvHD. One question is whether we could define a threshold of either CD56bright NKreg cells or INFγ+ CD4+ T cells which are required to be infused per Kg of the recipient. We found that the proportion in the donor product (cells per lym-

References 1. Lee SJ, Logan B, Westervelt P, et. al. Patientreported outcomes in 5-year survivors who received bone marrow vs. peripheral blood unrelated donor transplantation: long-term follow-up of a randomized clinical trial. JAMA Oncol. 2016;2(12):1583-1589. 2. Burns LJ, Logan BR, Chitphakdithai P, et. al. Recovery of unrelated donors of peripheral blood stem cells versus recovery of unrelated donors of bone marrow: a prespecified analysis from the Phase III Blood and Marrow Transplant Clinical Trials Network Protocol 0201. Biol Blood Marrow Transplant. 2016;22(6):1108-1116. 3. Anasetti C, Logan BR, Lee SJ, et. al. Blood and Marrow Transplant Clinical Trials Network. Peripheral-blood stem cells versus bone marrow from unrelated donors. N Engl J Med. 2012;367(16):1487-1496. 4. Couban S, Simpson DR, Barnett MJ, et. al. Canadian Bone Marrow Transplant Group A randomized multicenter comparison of bone marrow and peripheral blood in recipients of matched sibling allogeneic transplants for myeloid malignancies. Blood. 2002;100(5):1525-1531. 5. Chu R, Brazauskas R, Kan F, et. al. Comparison of outcomes after transplantation of G-CSF-stimulated bone marrow grafts versus bone marrow or peripheral blood grafts from HLA-matched sibling donors for patients with severe aplastic anemia. Biol Blood Marrow Transplant. 2011;17(7):1018-1024. 6. Nagafuji K, Matsuo K, Teshima T, et. al.

haematologica | 2017; 102(11)

9. 10.

11.

12.

phocytes) and not the infused number of cells per Kg for both CD56brightNK cells (P=0.64) and IFNγ+ CD4+ T cells (P=0.94) was of the greatest importance, suggesting that for regulatory cells there is a proportional relationship with other cell populations. Thus, focusing on the proportion of the regulatory cell populations such as CD56brightNK cells and IFNγ+ CD4+ T cells in relation to the total cells infused is more relevant as a strategy, rather than that of achieving a certain threshold dose. In summary, while controlling for the potential impact of G-CSF on marrow, our studies demonstrated that CD56bright NKreg cells had a much stronger impact on G-PB than on G-BM. This supports the conclusion that a lower proportion of CD56bright NKreg cells results in the high rate of cGvHD seen in G-PB, thus validating the development of strategies to increase the proportion of CD56bright NKreg cells after G-PB transplantation. Strategies could include alternative mobilization agents that selectively increase CD56bright NKreg cells, expansion ex vivo followed by adoptive transfer, and in vivo CD56bright NKreg cell expansion via the administration of low dose IL-2 after transplantation. Acknowledgments This study was undertaken by the CBMTG and funded by a grant from the United States National Cancer Institute (Principal Investigator: K.R. Schultz; Grant 1R01CA108752-01A2) and CBMTG. We gratefully acknowledge the recipients, donors and staff of the BMT Programs who participated in this study.

Peripheral blood stem cell versus bone marrow transplantation from HLA-identical sibling donors in patients with leukemia: a propensity score-based comparison from the Japan Society for Hematopoietic Stem Cell Transplantation registry. Int J Hematol. 2010;91(5):855-864. Eapen M, Logan BR, Confer DL, et. al. Peripheral blood grafts from unrelated donors are associated with increased acute and chronic graft-versus-host disease without improved survival. Biol Blood Marrow Transplant. 2007;13(12):1461-1468. Schrezenmeier H, Passweg JR, Marsh JC, et al. Worse outcome and more chronic GVHD with peripheral blood progenitor cells than bone marrow in HLA-matched sibling donor transplants for young patients with severe acquired aplastic anemia. Blood. 2007;110(4):1397-1400. Goldman J. Peripheral blood stem cells for allografting. Blood. 1995;85(6):1413-1415. Dhédin N, Prébet T, De Latour RP, et. al. Extensive chronic GVHD is associated with donor blood CD34+ cell count after G-CSF mobilization in non-myeloablative allogeneic PBSC transplantation. Bone Marrow Transplant. 2012;47(12):15641568. Mohty M, Bilger K, Jourdan E, et. al. Higher doses of CD34+ peripheral blood stem cells are associated with increased mortality from chronic graft-versus-host disease after allogeneic HLA-identical sibling transplantation. Leukemia. 2003;17(5):869-875. Vasu S, Geyer S, Bingman A, et. al. Granulocyte colony-stimulating factormobilized allografts contain activated

13.

14.

15.

16.

17.

immune cell subsets associated with risk of acute and chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2016; 22(4):658-668. Gallo S, Woolfrey AE, Burroughs LM, et. al. Marrow grafts from HLA-identical siblings for severe aplastic anemia: does limiting the number of transplanted marrow cells reduce the risk of chronic GvHD? Bone Marrow Transplant. 2016;51(12):15731578. Arpinati M, Chirumbolo G, Urbini B, et. al. Acute graft-versus-host disease and steroid treatment impair CD11c+ and CD123+ dendritic cell reconstitution after allogeneic peripheral blood stem cell transplantation. Biol Blood Marrow Transplant. 2004; 10(2):106-115. Ding L, Zhu H, Yang Y, et. al. The absolute number of regulatory T cells in unmanipulated peripheral blood grafts predicts the occurrence of acute graft-versus-host disease post haplo-identical hematopoietic stem cell transplantation. Leuk Res. 2017;56:13-20. Vela-Ojeda J, García-Ruiz Esparza MA, Reyes-Maldonado E, et. al. Clinical relevance of NK, NKT, and dendritic cell dose in patients receiving G-CSF-mobilized peripheral blood allogeneic stem cell transplantation. Ann Hematol. 2006;85(2):113120. Couban S, Aljurf M, Lachance S, et. al. Filgrastim-stimulated bone marrow compared with filgrastim-mobilized peripheral blood in myeloablative sibling allografting for patients with hematologic malignancies: a randomized Canadian Blood and

1945

A. Kariminia et al.

18.

19. 20. 21.

22.

23.

24.

25.

26.

27.

1946

Marrow Transplant Group Study. Biol Blood Marrow Transplant. 2016; 22(8):1410-1415. Przepiorka D, Weisdorf D, Martin P, et.al. Consensus conference on acute GVHD grading bone marrow transplant. 1995; 15:825-828. Sullivan K. Acute and chronic graft versus host disease in man. Int J Cell Cloning 1986;4 Suppl 1:42-93. Cox, D. R.; Oakes, D. 1984 Analysis of Survival Data. New York: Chapman & Hall. ISBN 041224490X. Yakoub-Agha I1, Saule P, Depil S, et. al. Comparative analysis of naïve and memory CD4+ and CD8+ T-cell subsets in bone marrow and G-CSF-mobilized peripheral blood stem cell allografts: impact of donor characteristics. Exp Hematol. 2007; 35(6):861-871. Shaw BE, Apperley JF, Russell NH, et al. Unrelated donor peripheral blood stem cell transplants incorporating pre-transplant invivo alemtuzumab are not associated with any increased risk of significant acute or chronic graft-versus-host disease. Br J Haematol. 2011;153(2):244-252. Rozmus J , Schultz KR, Wynne K, et, al. Early and late overall chronic graft-versushost disease (overall cGvHD) in children is characterized by different Th1/Th2 cytokine profiles: findings of The Children's Oncology Group Study (COG), ASCT0031. Biol Blood Marrow Transplant. 2011;17(12):1804-1813. Bacigalupo A, Lamparelli T, Barisione G, et. al. Gruppo Italiano Trapianti Midollo Osseo (GITMO). Thymoglobulin prevents chronic graft-versus-host disease, chronic lung dysfunction, and late transplant-related mortality: long-term follow-up of a randomized trial in patients undergoing unrelated donor transplantation. Biol Blood Marrow Transplant. 2006;12(5):560-565. Wolschke C, Zabelina T, Ayuk F, et. al. Effective prevention of GVHD using in vivo T-cell depletion with anti-lymphocyte globulin in HLA-identical or -mismatched sibling peripheral blood stem cell transplantation. Bone Marrow Transplant. 2014; 49(1):126-130. Devillier R, Granata A, Fürst S, et. al. Low incidence of chronic GVHD after HLA-haploidentical peripheral blood stem cell transplantation with post-transplantation cyclophosphamide in older patients. Br J Haematol. 2017;176(1):132-135. Kanakry CG, O'Donnell PV, Furlong T, et.al. Multi-institutional study of posttransplantation cyclophosphamide as single-agent graft-versus-host disease prophylaxis after allogeneic bone marrow transplantation using myeloablative busulfan and fludarabine conditioning. J Clin Oncol.

2014;32(31):3497-3505. 28. Carnevale-Schianca F, Caravelli D, Gallo S, et. al. Post-transplant cyclophosphamide and tacrolimus-mycophenolate mofetil combination prevents graft-versus-host disease in allogeneic peripheral blood hematopoietic cell transplantation from HLA-matched donors. Biol Blood Marrow Transplant. 2017;23(3):459-466. 29. Bashey A, Zhang X, Sizemore CA, et. al. Tcell-replete HLA-haploidentical hematopoietic transplantation for hematologic malignancies using post-transplantation cyclophosphamide results in outcomes equivalent to those of contemporaneous HLA-matched related and unrelated donor transplantation. J Clin Oncol. 2013; 31(10):1310-1316. 30. Li Pira G, Malaspina D, Girolami E, et. al. Selective Depletion of αβ T Cells and B Cells for Human Leukocyte AntigenHaploidentical Hematopoietic Stem Cell Transplantation. A Three-Year Follow-Up of Procedure Efficiency. Biol Blood Marrow Transplant. 2016;22(11):2056-2064. 31. Arai S, Sahaf B, Narasimhan B, et. al. Prophylactic rituximab after allogeneic transplantation decreases B-cell alloimmunity with low chronic GVHD incidence. Blood. 2012;119(25):6145-6154. 32. Delia M, Pastore D, Mestice A, et al. Outcome of allogeneic peripheral blood stem cell transplantation by donor graft CD3+/Tregs ratio: a single-center experience. Biol Blood Marrow Transplant. 2013; 19(3):495-499. 33. Nachbaur D, Kircher B. Dendritic cells in allogeneic hematopoietic stem cell transplantation. Leuk Lymphoma. 2005; 46(10):1387-1396. 34. She K, Gilman AL, Aslanian S, et al. Altered Toll-like receptor 9 responses in circulating B cells at the onset of pediatric chronic GVHD. Biol Blood Marrow Transplant. 2007;13(4):386-397. 35. Shier LR, Schultz KR, Imren S, et al. Differential effects of granulocyte colonystimulating factor on marrow- and bloodderived hematopoietic and immune cell populations in healthy human donors. Biol Blood Marrow Transplant. 2004;10(9):624634. 36. Waller EK, Logan BR, Harris WA, 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. 37. Baume DM, Robertson MJ, Levine H, Differential responses to interleukin 2 define functionally distinct subsets of human natural killer cells. Eur J Immunol. 1992;22(1):1-6. 38. Larghero J, Rocha V, Porcher R, et al.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

Association of bone marrow natural killer cell dose with neutrophil recovery and chronic graft-versus-host disease after HLA identical sibling bone marrow transplants. Br J Haematol. 2007;138(1):101-109. Karrich JJ, Cupedo T. Group 3 innate lymphoid cells in tissue damage and graft-versus-host disease pathogenesis. Curr Opin Hematol. 2016;23(4):410-415. Munneke JM, Björklund AT, Mjösberg JM, et al. Activated innate lymphoid cells are associated with a reduced susceptibility to graft-versus-host disease. Blood. 2014; 124(5):812-821. Kariminia A, Holtan SG, Ivison S, et al. Heterogeneity of chronic graft-versus-host disease biomarkers: the only consistent association is with CXCL10 and CXCR3+ NK cells. Blood. 2016;127(24):3082-3089. Moretta F, Petronelli F, Lucarelli B, et al. The generation of human innate lymphoid cells is influenced by the source of hematopoietic stem cells and by the use of G-CSF. Eur J Immunol. 2016;46(5):1271-1278. Bogunia-Kubik K, Mlynarczewska A, Wysoczanska B, Lange A. Recipient interferon-gamma 3/3 genotype contributes to the development of chronic graft-versushost disease after allogeneic hematopoietic stem cell transplantation. Haematologica. 2005;90(3):425-426. Baker J, Verneris MR, Ito M, Shizuru JA, Negrin RS. Expansion of cytolytic CD8(1) natural killer T cells with limited capacity for graft- versus-host disease induction due to interferon gamma production. Blood. 2001;97(10):2923-2931. Lu Y, Waller EK. Dichotomous role of interferon-gamma in allogeneic bone marrow transplant. Biol Blood Marrow Transplant. 2009;15(11):1347-1353. Fu J, Wang D, Yu Y, et al. T-bet is critical for the development of acute graft-versus-host disease through controlling T cell differentiation and function. J Immunol. 2015; 194(1):388-397. Hall BM, Tran GT, Verma ND, et al. Do natural T regulatory cells become activated to antigen specific T regulatory cells in transplantation and in autoimmunity? Front Immunol. 2013;4:208. Wang H, Yang YG. The complex and central role of interferon-γ in graft-versus-host disease and graft-versus-tumor activity. Immunol Rev. 2014;258(1):30-44. Cope A, Le Friec G, Cardone J, Kemper C. The Th1 life cycle: molecular control of IFN-γ to IL-10 switching. Trends Immunol. 2011;32(6):278-286. Chinnadurai R, Copland IB, Patel SR, Galipeau J. IDO-independent suppression of T-cell effector function by IFN-γ licensed human mesenchymal stromal cells. J Immunol. 2014;192(4):1491-1501.

haematologica | 2017; 102(11)

ARTICLE

Cell Therapy & Immunotherapy

Human leukocyte antigen-E mismatch is associated with better hematopoietic stem cell transplantation outcome in acute leukemia patients Chrysanthi Tsamadou,1,2 Daniel Fürst,1,2 Vladan Vucinic,3 Donald Bunjes,4 Christine Neuchel,1,2 Daphne Mytilineos,2 Martin Gramatzki,5 Renate Arnold,6 Eva Maria Wagner,7 Hermann Einsele,8 Carlheinz Müller,9,10 Hubert Schrezenmeier1,2 and Joannis Mytilineos1,2,10

Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Transfusion Service, Baden Wuerttemberg – Hessen, and University Hospital Ulm; 2 Institute of Transfusion Medicine, University of Ulm; 3Department of Hematology/Oncology, University of Leipzig; 4Department of Internal Medicine III, University of Ulm; 5Division of Stem Cell Transplantation and Immunotherapy, 2nd Department of Medicine, University of Kiel; 6Hematology/Oncology Department, Charité Campus Virchow-Klinikum, Berlin; 7Department of Internal Medicine III, Johannes Gutenberg-University Mainz; 8Department of Internal Medicine II, University Hospital Würzburg; 9ZKRD - Zentrale Knochenmarkspender-Register für Deutschland, German National Bone Marrow Donor Registry and 10DRST – German Registry for Stem Cell Transplantation, Ulm, Germany 1

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(11):1947-1955

ABSTRACT

he immunomodulatory role of human leukocyte antigen (HLA)-E in hematopoietic stem cell transplantation (HSCT) has not been extensively investigated. To this end, we genotyped 509 10/10 HLA unrelated transplant pairs for HLA-E, in order to study the effect of HLA-E as a natural killer (NK)-alloreactivity mediator on HSCT outcome in an acute leukemia (AL) setting. Overall survival (OS), disease free survival (DFS), relapse incidence (RI) and non-relapse mortality (NRM) were set as endpoints. Analysis of our data revealed a significant correlation between HLA-E mismatch and improved HSCT outcome, as shown by both univariate (53% vs. 38%, P=0.002, 5-year OS) and multivariate (hazard ratio (HR)=0.63, confidence interval (CI) 95%=0.48-0.83, P=0.001) analyses. Further subgroup analysis demonstrated that the positive effect of HLA-E mismatch was significant and pronounced in advanced disease patients (n=120) (5-year OS: 50% vs. 18%, P=0.005; HR=0.40, CI 95%=0.22-0.72, P=0.002; results from univariate and multivariate analyses, respectively). The study herein is the first to report an association between HLA-E incompatibility and improved post–transplant prognosis in AL patients who have undergone matched unrelated HSCT. Combined NK and T cell HLA-E-mediated mechanisms may account for the better outcomes observed. Notwithstanding the necessity for in vitro and confirmational studies, our findings highlight the clinical relevance of HLA-E matching and strongly support prospective HLAE screening upon donor selection for matched AL unrelated HSCTs. Introduction HSCT has long been established as an indispensable life-saving treatment, in particular against acute hematologic malignancies. 1 Despite the significant progress made in the last ten years, transplantation related mortality and graft-versus-host disease (GvHD) continue to substantially constrain the curative potential of HSCT, even in an HLA-matched context, underscoring the need to explore the role of other immune system-related genetic factors in HSCT.2 In this respect, a rather limited number of studies sought to investigate the effect of HLA-E on HSCT outcome, considering the significant immunomodulatory features of this molecule implicated in both innate and adaptive immunity.3,4 HLA-E, a member of haematologica | 2017; 102(11)

Correspondence: j.mytilineos@blutspende.de

Received: March 29, 2017. Accepted: September 4, 2017. Pre-published: September 7, 2017. doi:10.3324/haematol.2017.169805 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1947 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1947

C. Tsamadou et al.

the non-classical HLA-Ib family, is ubiquitously expressed on all nucleated cells, but at lower expression levels than the classical HLA-class I molecules.5 It is rather nonpolymorphic, with basically two functional forms of the protein found worldwide at similar prevalence rates,6 shares an almost identical structural pattern with its classical HLA-class I counterpart and is viewed as a surrogate marker for HLA-class I expression, as the leader sequences of the latter constitute its main peptide reservoir.7 Even though this prominent allelic variation derives from a single arginine to a glycine amino acid substitution at position 107 of the heavy chain α2 domain (HLAE*01:01 and HLA-E*01:03, respectively), the codominance of the two alleles in conjunction with their significantly different expression levels on cell surfaces imply functional differences which are yet to be fully understood.8-10 As a basic ligand to CD94/NKG2A,11 a robust inhibitory receptor found on the surface of NK cells and NK-like cytotoxic T lymphocytes (CTLs), the principal role of HLA-E is considered to be the protection of normal cells from aberrant NK killing. However, continuously arising data highlight that HLA-E may hold a much more multifaceted role in immune response by presenting “unconventional” peptides under stress conditions12,13 and by interacting with HLA-E-restricted CD8+ CTLs and regulatory T cells (Tregs) via their αβ T-cell receptors (TCRs) as well as with the activating CD94/NKG2C receptor on the surface of NK-cells and NK-like CTLs.14,15 Despite the evident role of HLA-E in immune response, no definite conclusions can be drawn from studies published thus far aiming to establish an association between HLA-E and HSCT outcome.16-24 The aim of the present study was to explore not only the role of HLA-E genotype but, primarily, the effect of HLA-E patient-donor compatibility on HSCT outcome, as the weak linkage disequilibrium between HLA-E and its classical HLA counterparts leads to a rather high rate of HLA-E mismatches among HLAA, -B, -C, -DRB1, and HLA-DQB1 allele-matched HSCT pairs.17,25 HLA-E as an NK-alloreactivity mediator is expected to have a more prominent role in an AL context where the graft-versus-leukemia effect (GvL) is of utmost relevance. Hence, we applied a study design including only adult AL patients who had undergone a 10/10 HLAmatched unrelated HSCT in order to evaluate the role of patient/donor HLA-E genotypes as well as of HLA-E matching status in HSCT outcome.

Methods Patients 509 adult patients diagnosed with AL, receiving their first allogeneic HSCT between 2002 and 2009 were included in the study. All patients were transplanted with 10/10 allele level HLA-A, -B, C, -DRB1, -DQB1-matched grafts, which were either bone marrow (BM) or peripheral blood stem cells (PBSCs). We included only those patients diagnosed with acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL) as well as undefined AL (undifferentiated, biphenotypic or secondary acute). Disease stages were assigned according to a previous report published by the European Society for Blood and Marrow Transplantation (EBMT) study group.26 Early disease stage included AML, AL, and ALL transplanted in first complete remission, intermediate disease

1948

stage was defined as AML and ALL in second complete remission or first relapse as well as AL transplanted in second complete remission. All other disease phases of AML, ALL and AL were characterized as advanced stage. All patients were treated with myeloablative (Mab) or reduced intensity conditioning (RIC).27,28 Recipient and donor consents for HLA typing and for the analysis of clinical data were obtained in accordance with the Declaration of Helsinki upon initiation of donor search and registration in the EBMT database, respectively. All clinical data were initially recorded in the EBMT ProMISE database and were subsequently provided to us by the German Registry for Stem Cell Transplantation (DRST), which is responsible for the clinical data management of the German patients’ subset. The study was approved by the ethical review board of the University of Ulm (project number: 263/09).

HLA-typing All patients and their respective donors were genotyped at high resolution level for the HLA-loci A, B, C, DRB1 and DQB1. HLA-DPB1 genotyping was performed retrospectively for all study subjects using stored DNA material. Permissiveness of DPB1-mismatches was assessed according to the TCE (T-cell epitope) algorithm.29 Additional testing for relevant non-expressed alleles was performed according to the National Marrow Donor Program confirmatory typing requirements.30

Killer Cell Immunoglobulin-Like Receptors (KIR) typing KIR-typing was performed using the commercially available “KIR Genotyping SSP Kit” from Life Technologies (Carlsbad, CA, USA). Donor KIR AA and Bx haplotypes were assigned as previously described.31

HLA-E typing All 509 patient-donor pairs were HLA-E high resolution genotyped. HLA-E specific primers were designed for complete Exon 2 and 3 sequencing analysis, allowing precise assignment of all known allelic variants. Allelic assignment was based on sequence data retrieved from the immunogenetics (IMGT)/HLA database.

Statistical analysis The cumulative estimates for the univariate analysis OS and DFS were obtained using the Kaplan-Meier method. For multivariate analyses Cox regression models were implemented. Competing risk analysis was used for the univariate analyses of NRM, RI and chronic (c)GvHD incidence, while competing risk regression models for stratified data were used for multivariate analyses. Acute (a)GvHD and severe infection incidence as well as prevalence of other causes of death are reported descriptively. Center effects were adjusted using a γ frailty term.32 Statistical models covered covariates in accordance with the previously published recommendations of the EBMT study group.28,33 In addition to these, patient and donor cytomegalovirus (CMV) serostatus, treatment with anti-thymocyte globulin (ATG), Karnofsky performance score (KPS) at time of transplantation, donor KIR haplotype (AA/Bx),31 patient C1/C2 KIR ligand status as well as HLA-DPB1 compatibility (based on T-cell epitope algorithm)29 were also evaluated. Missing data were treated as separate categories in multivariate analyses.26 A stepwise backward exclusion procedure was used for model selection.26,28 Statistical significance was set to a P-value≤0.05. All statistical analyses were performed using the open source program for statistical computing “R”, version 3.1.0. More section data available in Online Supplementary Material.

haematologica | 2017; 102(11)

Acute leukemia and HSCT: role of HLA-E matching

Table 1. Cohort characteristics.

Category Number of patients Number of transplantation centers Age category 18-29 30-39 40-49 50-59 60-69 70-79 Diagnosis AML ALL AL Disease stage Early Intermediate Advanced Conditioning regimen Myeloablative Reduced intensity Karnofsky performance score* KPS < 90 Missing data Stem cell source BM PBSC ATG Treatment* Yes No Missing data Patient-Donor CMV serostatus combination* neg neg neg pos pos neg pos pos Missing data Donor KIR Haplotype* Haplotype AA Haplotype Bx Missing data Patient C1/C2 KIR ligands C1 positive C1 negative HLA-DPB1 TCE mismatch* Permissive HvG non-permissive GvH non-permissive Missing data

Study cohort n(%)

HLA-E-matched n(%)

HLA-E-mismatched n(%)

509 21

320 20 (95.2)

189 20 (95.2)

0.61

97 (19.0) 71 (14.0) 89 (17.5) 129 (25.3) 108 (21.2) 15 (3.0)

62 (19.4) 32 (10.0) 62 (19.4) 82 (25.6) 75 (23.4) 7 (2.2)

35 (18.5) 39 (20.6) 27 (14.3) 47 (24.9) 33 (17.5) 8 (4.2)

313 (61.5) 132 (25.9) 64 (12.6)

196 (61.2) 85 (26.6) 39 (12.2)

117 (61.9) 47 (24.9) 25 (13.2)

237 (46.5) 152 (29.9) 120 (23.6)

147 (45.9) 92 (28.8) 81 (25.3)

90 (47.6) 60 (31.7) 39 (20.6)

345 (67.8) 164 (32.2)

215 (67.2) 105 (32.8)

130 (68.8) 59 (31.2)

0.78

98 (30.6) 189 (37.1)

66 (32.0) 114 (35.6)

32 (28.1) 75 (39.7)

0.50

32 (6.3) 477 (93.7)

18 (5.6) 302 (94.4)

14 (7.4) 175 (92.6)

0.54

252 (63.2) 147 (37.8) 110 (21.6)

157 (61.3) 99 (38.7) 64 (20.0)

95 (66.4) 48 (33.6) 46 (24.3)

0.31

126 (31.6) 45 (11.3) 102 (25.5) 126 (31.6) 110 (21.6)

81 (32.1) 31 (12.3) 61 (24.2) 79 (31.4) 68 (21.2)

45 (30.6) 14 (9.5) 41(27.9) 47 (32.0) 42 (22.2)

0.86

158 (31.3) 346 (68.7) 5 (0.98)

95 (30.0) 222 (70.0) 3 (0.94)

63 (33.7) 124 (66.3) 2 (1.0)

0.68

443 (87.0) 66 (13.0)

277 (86.6) 43 (13.4)

166 (87.8) 23 (12.2)

326 (64.3) 86 (17.0) 95 (18.7) 2 (0.4)

208 (65.2) 59 (18.5) 52 (16.3) 1 (0.3)

118 (62.8) 27 (14.4) 43 (22.8) 1 (0.5)

0.24

0.89

0.46

0.86

0.13

continued on next page

haematologica | 2017; 102(11)

1949

C. Tsamadou et al. continued in the previous page

Recipient-donor sex match Male-male Male-female Female-male Female-female Year of transplantation 2002-2005 2006-2009

228 (44.8) 46 (9.0) 168 (33.0) 67 (13.2)

143 (44.7) 26 (8.1) 108 (33.8) 43 (13.4)

85 (45.0) 20 (10.6) 60 (31.7) 24 (12.7)

127 (25.0) 382 (75.0)

77 (24.0) 243 (76.0)

50 (26.5) 139 (73.5)

0.80

0.62

*In compliance with the EBMT statistical guidelines, percentages for variables with missing data are presented with reference to the known data cases. AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; AL: acute leukemia not specified as AML or ALL (undifferentiated, biphenotypic or secondary acute); KPS: Karnofsky performance score; BM: bone marrow; PBSC: peripheral blood stem cells; ATG: anti-thymocyte globulin; CMV: cytomegalovirus; pos: positive; neg: negative. KIR: killer cell immunoglobulin-like receptor; HLA: human leukocyte antigen; TCE: T-cell epitope; HvG: Host vs. Graft; GvH: Graft vs. Host.

Table 2. Human leukocyte antigen (HLA)-E genotyping results.

HLA-E genotypes (n,%)

HLA-E allele frequencies (n,%) HLA-E* Patients Donors

01:01 567 (55.7) 572 (56.0)

01:03 451 (44.3) 445 (43.8)

01:07 1 (0.2)

01:01 157 (30.8) 151 (29.7)

01:01, 01:03 253 (49.7) 269 (52.8)

01:03 99 (19.5) 88 (17.3)

01:01, 01:07 1 (0.2)

Results

HLA-E incompatibility significantly improves OS, DFS and NRM

Patient characteristics

Analysis of OS, DFS and NRM with respect to HLA-E matching status between patients and donors revealed a significant favorable effect of HLA-E mismatch on these endpoints. As shown in Figure 1, patients transplanted with HLA-E-mismatched donors exhibit a significantly improved 5-year OS (53% vs. 38%, P=0.002), 5-year DFS (45% vs. 32%, P=0.007) and a significantly lower 5-year NRM (26% vs. 37%, P=0.006) when compared to cases receiving an HLA-E compatible graft. Multivariate analyses confirmed the above findings as the beneficial effect of HLA-E mismatch was statistically significant for all of the above HSCT outcome endpoints (OS: HR=0.63, CI 95%=0.48-0.83, P=0.001; DFS: HR=0.71, CI 95%=0.550.92, P=0.008; NRM: HR=0.63, CI 95%=0.43-0.91, P=0.015). Since better OS appeared to stem from lower NRM rates in the HLA-E-mismatched patient subgroup, we separately analyzed the prevalence rates of aGvHD and severe infection along with an overall cause of death analysis. Although Grade III-IV aGvHD rates were similar in the two groups (~10%), the death rate of 9% from GvHD in the HLA-E-matched group was substantially higher than the 5.8% found among HLA-E-mismatched patients. Furthermore, severe infection was reported in 17.2% of HLA-E-matched patients vs. 9.5% of HLA-Emismatched patients. Accordingly, infection-related mortality was higher in the HLA-E-matched group (10.9% vs. 7.9%). It should be noted that data on both aGvHD and cGvHD were incomplete for 9% (46/509) and 43% (217/509) of cases, respectively. No cause of death data were available for 2.1% of patients (11/509). With regard to cGvHD, presuming that missing values were most likely randomly distributed among HLA-E-matched and mismatched cases within our cohort, we decided to include this parameter in the statistical analysis. The analysis of

Patient cohort characteristics regarding HSCT outcome predictors and in relation to HLA-E matching status between patient and donor are summarized in Table 1. For the 509 patients included in the study, median post-transplant follow-up time was almost 5 years (4.97 years), while median patient age was 49 years (range: 18-74 years). Interestingly, 37.1% of the cases were HLA-E-mismatched, and as the P-values in Table 1 suggest, there was no biased distribution of HLA-E-matched and mismatched cases with regard to other parameters predictive for the outcome of HSCT which we evaluated.

HLA-E genotyping results A summary of the HLA-E genotyping results is displayed in Table 2. The HLA-E allele frequencies found were in accordance with those previously reported for Caucasian populations,6,17,25 confirming the codominant prevalence of the two basic allelic forms of HLA-E. No differences were identified regarding the distribution of the HLA-E allelic variants between patients and donors.

HLA-E*01:03, 01:03 patient genotype is not associated with better HSCT outcome Our results do not confirm the findings of previously published studies regarding the positive impact of patient HLA-E*01:03, 01:03 genotype on HSCT outcome. On the contrary, HLA-E*01:03, 01:03 patients in our cohort had worse OS, DFS and NRM rates compared to the patients carrying the two other genotypes as shown in the multivariate analysis (OS: HR=1.45, CI 95%=1.00-2.10, P=0.05; DFS: HR=1.47, CI 95%=1.04-2.07, P=0.03; NRM: HR=1.74, CI 95%=1.09-2.78, P=0.02). Of note, this finding did not reach statistical significance in any of the univariate models (data not shown). 1950

haematologica | 2017; 102(11)

Acute leukemia and HSCT: role of HLA-E matching

Table 3. Univariate and multivariate analyses as to the effect of human leukocyte antigen (HLA)-E mismatch on HSCT outcome.

Endpoints Overall Survival 1 year 3 year 5 year Disease free survival 1 year 3 year 5 year Non-relapse mortality 1 year 3 year 5 year Relapse incidence 1 year 3 year 5 year cGvHD incidence 6 months 12 months 24 months

Univariate Analysis HLA-E-matched 0.59(0.53-0.65) 0.42(0.36-0.49) 0.38(0.32-0.44)

HLA-E-mismatched 0.67(0.61-0.75) 0.57(0.50-0.65) 0.53(0.46-0.62)

0.51(0.46-0.58) 0.36(0.31-0.43) 0.32(0.27-0.39)

Multivariate Analysis P

CI 95%

0.002

0.63

0.48-0.83

0.001

0.59(0.52-0.66) 0.51(0.44-0.59) 0.45(0.38-0.53)

0.007

0.71

0.55-0.92

0.008

0.27(0.22-0.32) 0.36(0.30-0.41) 0.37(0.31-0.43)

0.19(0.14-0.26) 0.22(0.16-0.29) 0.26(0.19-0.33)

0.006

0.63

0.43-0.91

0.015

0.25(0.20-0.31) 0.32(0.27-0.38) 0.35(0.29-0.41)

0.25(0.19-0.32) 0.31(0.24-0.38) 0.34(0.26-0.41)

0.84

1.02

0.73-1.43

0.90

0.34(0.27-0.42) 0.39(0.31-0.46) 0.39(0.32-0.47)

0.24(0.17-0.32) 0.28(0.21-0.37) 0.32(0.24-0.40)

0.102

0.70

0.47-1.04

0.074

Number of patients included in the analyses, n=509. Omitted observations due to missing data for overall survival (OS)=6, disease free survival (DFS)=4, non-relapse mortality=6, relapse incidence= 20 and cGvHD incidence=217. Statistical significance is marked in italics. Hazard ratio (HR) values for survival endpoints (Overall survival and Disease free survival) refer to the risk of death and/or relapse as measured in the analyses for these endpoints. cGvHD: chronic graft-versus-host disease.

the cumulative probability of cGVHD revealed a tendency toward association between HLA-E mismatch and less cGvHD. However, given the admittedly high number of missing data, these results should be interpreted with caution. All results for both univariate and multivariate analyses are summarized in Table 3. After stepwise backward exclusion procedure used for model selection, patient age, disease stage, diagnosis, CMV serostatus compatibility, ATG treatment and patient HLA-E haplotype were integrated as significant clinical predictors in our multivariate analyses.

Advanced disease acute leukemia patients benefit the most from HLA-E-mismatched unrelated 10/10 HLA matched HSCT Exploratory controls for potential interactions between HLA-E matching status and other clinical predictors revealed an association between the â&#x20AC;&#x153;HLA-E mismatch effectâ&#x20AC;? and advanced disease stage. For this reason we extended our analysis by dividing patients into an advanced (n=120) and a non-advanced disease (n=389) group, with the latter including patients in early or intermediate disease stage. Both univariate and multivariate analyses for OS, DFS and NRM revealed a much stronger effect of HLA-E mismatch in the advanced disease group compared to the early/intermediate stage patients. The 5-year survival rates were markedly improved in advanced disease patients who received HLA-E disparate grafts (OS: 50% vs. 18%, P=0.005; DFS: 40% vs. 12%, P=0.002), as likewise depicted by the Kaplan-Meier curves in Figure 2. NRM was also notably lower among these haematologica | 2017; 102(11)

patients (32% vs. 55%, P=0.038, Figure 2). Multivariate analyses confirmed the above findings for all three endpoints in advanced disease patients (OS: HR=0.40, CI 95%=0.22-0.72, P=0.002; DFS: HR=0.42, CI 95%=0.250.72, P=0.001; NRM: HR=0.44, CI 95%=0.20-0.95, P=0.036). Additionally, HLA-E mismatch in advanced disease patients was associated with markedly higher rates of none or mild (grade 0-I) aGvHD (66.7% vs. 56.8%) and lower rates of grade II-IV aGvHD (7.7% vs. 12.3%). Moreover, 14.8% of HLA-E-matched patients died due to severe GvHD compared to only 2.6% of HLA-E-mismatched cases. No significant differences were observed on account of severe infection prevalence between the two groups (21.0% of HLA-E-matched vs. 17.9% of HLAE-mismatched cases). Interestingly, infection-related mortality was higher in the HLA-E-mismatched group (17.9% vs. 12.3%). Possible subjectivity involved in the reporting of only one cause of death in the case of concomitant fatal conditions may account for this discordance. It should be underscored that no aGvHD data were available in 17.5% (21/120) of cases, while cause of death data were incomplete for 2.5% (3/120) of advanced disease patients. The effect of HLA-E mismatch in non-advanced disease patients, albeit noticeable, did not reach statistical significance for any of the endpoints in either univariate or multivariate analyses. No significant differences were identified in this subset of patients with respect to aGvHD rates and GvHD-related death. However, there was a marked difference observed regarding severe infection prevalence with 15.9% in HLA-E-matched cases vs. 7.3% in HLA-Emismatched ones, likewise regarding infection-related mortality rates (10.5% vs. 5.3% in HLA-E-matched and 1951

C. Tsamadou et al.

mismatched cases, respectively). Cause of death data were missing for 2% (8/389) of early/intermediate disease patients. All results for both univariate and multivariate analyses and for both patient subgroups are listed in Tables 4 and 5, respectively.

Discussion

No differences in RI were observed with respect to HLA-E matching status. Moreover, advanced as well as non-advanced disease patients exhibited similar relapse rates regardless of HLA-E matching status to their donor. The results for RI are summarized in Tables 3-5.

The immunomodulatory role of HLA-E and its implication in both innate and adaptive immunity has long been investigated and established.4 Its impact, however, on HSCT remains markedly elusive, as there are only a relatively few number of studies with small and heterogeneous cohorts to be found in the literature;3 most of which have aimed at establishing a correlation between certain patient HLA-E genotypes and HSCT outcome. The study herein is, to our knowledge, the first to report a favorable effect of HLA-E incompatibility in an AL-matched unrelated HSCT setting. Our data suggest significantly improved

HLA-E mismatch has no effect on relapse incidence rates

C C

Figure 1. Hematopoietic stem cell transplantation outcome with respect to human leukocyte antigen (HLA)-E matching status in acute leukemia patients, n=509. (A) Overall survival (P=0.002); (B) Disease free survival (P=0.007) and (C) Non-relapse mortality (P=0.006) curves, respectively, of patients transplanted with HLA-E-matched donors (black line) versus patients transplanted with HLA-E-mismatched donors (red line).

1952

Figure 2. Hematopoietic stem cell transplantation outcome with respect to human leukocyte antigen (HLA)-E matching status in advanced disease patients, n=120. (A) Overall survival (P=0.005); (B) Disease free survival, (P=0.002) and (C) Non-relapse mortality (P=0.038) curves, respectively, of patients transplanted with HLA-E-matched donors (black line) versus patients transplanted with HLA-E-mismatched donors (red line).

haematologica | 2017; 102(11)

Acute leukemia and HSCT: role of HLA-E matching

overall and disease free survival rates as well as lower NRM in adult AL patients transplanted with 10/10 HLAmatched unrelated donors when grafts received were HLA-E disparate. No effect was found in relation to relapse incidence. Other confounding factors putatively responsible for this observation were excluded, as HLA-Ematched and mismatched pairs had no significant differences from one another with respect to other known HSCT outcome predictors33 (Table 1). In previous studies which investigated the role of HLA-E compatibility in HSCT outcome, Fürst et al. did not observe any association between HLA-E mismatch and HSCT outcome, while the results of Harkensee et al. suggested a negative impact of HLA-E incompatibility on survival.20,23 These two studies, however, were designed on a different basis, hence the results are not comparable. The cohort of Fürst et al., apart from its significantly smaller size (n=116), was heterogeneous in terms of diagnoses, which for reasons that will be analyzed subsequently, may be of fundamental importance. The Harkensee et al. study rationale was performed in an HLA-mismatched setting and its primary goal was to establish associations between various nonHLA genetic factors and HSCT outcome for HLA disparate transplant pairs. Previous studies16-19,21,22 reported lower transplantation related mortality, less severe bacterial infection rates as well as lower relapse and severe GvHD incidences in patients with the HLA-E*01:03 genotype. We could not confirm these associations. In our multivariate models, where patient HLA-E genotype was a significant covariate, patient HLA-E*01:03 homozygosity was, in fact, correlated with inferior outcome. However, it must be acknowledged that any comparison between these studies and ours is not applicable, as some of them

included HSCT from related or HLA-E-matched donors,16,18,22 and cohorts in all of them were not only significantly smaller in size but also heterogeneous with regard to diagnoses.16-19,21,22 According to our findings, HLA-E mismatch appears to confer its beneficial effect through dampening of NRM. On account of this, two very interesting observations are of note. First, that HLA-E mismatch seems to differentially impact patients according to their disease stage, and secondly, that a putatively combined mechanism may account for the overall beneficial effect, as the lower NRM rates in advanced disease patients appear to be prevalently related with lower GvHD rates, whereas in early/intermediate disease patients there is better control of infection. As far as the first observation is concerned, our results clearly suggest a much stronger impact of HLA-E mismatch on advanced disease patients’ outcome (Tables 4, 5). In fact, the results within this subgroup of patients drive the findings in the entire study cohort since they clearly reach significance, while the effect of HLA-E mismatch in the larger group of early/intermediate disease patients, although visible, does not reach statistical significance. This is most likely due to the different “baseline” prognostic odds of the two subgroups.34 It is well known that HLA-E is an important modulator of NK-cytotoxicity, as it constitutes the main ligand to the CD94/NKG2A/C group of NK receptors.11 According to the murine model proposed by Olson et al., early posttransplant NK alloreactivity could be associated with better OS rates due to lower GvHD incidence and NRM.35 The fact that CD94/NKG2A/C receptors are the first to appear on freshly reconstituted NK cells immediately following HSCT, strengthens the assumption that this “HLA-

Table 4. Univariate analysis of advanced vs. non-advanced patients with respect to human leukocyte antigen (HLA)-E mismatch.

Early/intermediate disease patients n=389† HLA-E-matched HLA-E-mismatched

Endpoints Overall Survival

Advanced disease patients n=120* HLA-E-matched HLA-E-mismatched

1 year 3 year 5 year Disease free survival 1 year 3 year 5 year Non-relapse mortality 1 year 3 year 5 year Relapse incidence 1 year 3 year 5 year cGvHD incidence 6 months 12 months 24 months

0.32(0.23-0.46) 0.20(0.12-0.33) 0.18(0.11-0.31)

0.57(0.43-0.76) 0.54(0.39-0.74) 0.50(0.35-0.71)

0.005

0.68(0.62-0.74) 0.50(0.43-0.57) 0.45(0.38-0.53)

0.72(0.65-0.80) 0.60(0.52-0.69) 0.56(0.48-0.65)

0.27(0.18-0.40) 0.16(0.09-0.28) 0.12(0.06-0.24)

0.52(0.38-0.72) 0.52(0.38-0.72) 0.40(0.25-0.62)

0.002

0.60(0.53-0.67) 0.43(0.37-0.51) 0.39(0.33-0.47)

0.60(0.52-0.69) 0.51(0.43-0.60) 0.46(0.38-0.56)

0.48(0.36-0.59) 0.55(0.42-0.66) 0.55(0.42-0.66)

0.32(0.17-0.48) 0.32(0.17-0.48) 0.32(0.17-0.48)

0.038

0.20(0.15-0.26) 0.26(0.20-0.32) 0.29(0.23-0.36)

0.16(0.11-0.23) 0.18(0.12-0.25) 0.20(0.14-0.27)

0.083

0.31(0.20-0.43) 0.33(0.22-0.45) 0.37(0.25-0.49)

0.24(0.10-0.41) 0.24(0.10-0.41) 0.35(0.16-0.54)

0.60

0.24(0.18-0.30) 0.29(0.23-0.35) 0.32(0.26-0.39)

0.25(0.18-0.33) 0.30(0.23-0.28) 0.32(0.24-0.40)

0.86

0.50(0.30-0.67) 0.50(0.30-0.67) 0.50(0.30-0.67)

0.13(0.03-0.30) 0.13(0.03-0.30) 0.13(0.03-0.30)

0.009

0.31(0.24-0.39) 0.36(0.29-0.44) 0.37(0.29-0.45)

0.27(0.19-0.36) 0.32(0.23-0.41) 0.36(0.27-0.46)

0.60

P 0.071

0.26

*Advanced disease patients, n=120. Omitted observations due to missing data for overall survival=2, disease free survival=1, non-relapse mortality=2, relapse incidence= 14 and cGvHD incidence=69. †Early/intermediate disease patients, n=389. Omitted observations due to missing data for overall survival=4, disease free survival=3, non-relapse mortality=4, relapse incidence= 6 and cGvHD incidence=148. Statistical significance is marked in italics. cGvHD: chronic graft-versus-host disease.

haematologica | 2017; 102(11)

1953

C. Tsamadou et al.

Table 5. Multivariate analysis of advanced vs. non-advanced patients with respect to HLA-E mismatch.

Endpoints Overall Survival Disease free survival Non-relapse mortality Relapse incidence cGvHD incidence

Advanced disease patients n=120* HR CI 95% P 0.40 0.42 0.44 1.10 0.18

0.22-0.72 0.25-0.72 0.20-0.95 0.50-2.43 0.05-0.65

0.002 0.001 0.036 0.81 0.008

Early/intermediate disease patients n=389† HR CI 95% P 0.75 0.85 0.72 1.05 0.86

0.55-1.04 0.63-1.15 0.46-1.12 0.72-1.55 0.56-1.31

0.088 0.29 0.14 0.80 0.48

*Advanced disease patients, n=120. Omitted observations due to missing data for overall survival=2, disease free survival=1, non-relapse mortality=2, relapse incidence= 14 and cGvHD incidence=69. †Early/intermediate disease patients, n=389. Omitted observations due to missing data for overall survival=4, disease free survival=3, non-relapse mortality=4, relapse incidence= 6 and cGvHD incidence=148. Hazard ratio (HR) values for survival endpoints (Overall survival and Disease free survival) refer to the risk of death and/or relapse as measured in the analyses for these endpoints. cGvHD: chronic graft-versus-host disease.

E effect”, at least as far as the “dampening” of GvHD is concerned, could be NK-mediated.36,37 Numerous studies have highlighted the prominent effect of peptide specificity in peptide/HLA-E (pHLA-E) complexes as to the affinity and intensity of HLA-E interactions with its corresponding NK receptors, namely the inhibitory CD94/NKG2A and the activating CD94/NKG2C.9-10,38-43 The impact of HLA-E polymorphism, with respect to the NK “licensing” process, has not yet been investigated and as such remains elusive. Given the apparent ability of CD94/NKG2 receptors to discriminate different pHLA-E constellations through differential binding affinity, however, it is plausible to assume that during their “licensing” phase NK cells may be educated and tuned according to “self” pHLA-E patterns. Moreover, it has been shown that under abnormal conditions (e.g., infection, stress or tumorigenesis) HLA-E molecules are able to present “unconventional” peptides, generating pHLA-E complexes that go unnoticed by the dominant inhibitory CD94/NKG2A receptor, while on certain occasions they instigate activating signals through the CD94/NKG2C receptor.12 This in turn may lead to exacerbated NK activation. According to our hypothesis model, in an advanced-stage AL setting, aggravated stress conditions, heavier leukemia-cell burden and further alterations due to advanced leukemogenesis44 may lead to an enhanced NK-mediated attenuation of T cell alloreactivity.45 This, in succession, could explain the significantly lower GvHD related mortality observed in advanced disease patients. As previously mentioned, cause of death analysis in advanced and non-advanced disease patients revealed two potential mechanisms implicated - at a different degree according to disease stage - in a significant reduction of NRM rates. The decrease of GvHD-related death in advanced disease patients, as discussed above, may be NK-mediated. The reduction of fatal infection-related death in non-advanced disease patients, on the other hand, is more likely to be T cell-mediated, as it has been reported that HLA-E-restricted αβ T cells may play a significant role in the control of viral as well as bacterial infections (CMV, Epstein-Barr virus (EBV), human immunodeficiency virus (HIV), M.tuberculosis, S. typhi etc.).14 Given the role of HLA-E allelic variation in the specificity of HLA-E bound peptides, the ability of HLA-E to bind pathogen-derived peptides13 and the importance of peptide specificity in TCR recognition of pHLA-E complexes,14 it is plausible to presume that in an HLA-E-mismatched context, the chances of pathogen-specific HLA1954

E-restricted T cells to encounter the right pHLA-E constellation may be significantly higher due to a theoretically extended pHLA-E repertoire on account of HLA-E disparity. In an infection setting, “unconventional” pHLA-E complexes can be presented by both donor antigen presenting cells (APCs) and patient infected cells, hence pathogenspecific donor HLA-E-restricted T cells are more likely to encounter an immune-response-instigating pHLA-E pattern.14,43 These two independent mechanisms probably act synergistically but to a different degree according to disease stage. The differences observed in the two subgroups may be the result of NK interference in the T cell-mediated infection control potential in advanced disease patients on the one hand, and the less intense NK activation in early/intermediate disease patients due to lighter disease burden on the other. Significant limitations of our study are the incompleteness of the data regarding significant clinical parameters, such as aGvHD, cGvHD, type of infection and CMV reactivation, which would allow for a much more thorough and precise understanding of the way in which HLA-E mismatch exerts its beneficial effect on NRM and OS. Despite these drawbacks, however, the size and homogeneity of our cohort with respect to diagnosis, type of donor and HLA compatibility, certainly justify further investigation with larger confirmatory cohorts and functional in vitro studies. Considering that AL patients constitute the majority of all HSC-transplanted patients, and that even 10/10 HLA-matched unrelated transplant pairs have about 30-40% chance to be HLA-E disparate, our data support future integration of HLA-E compatibility as an additional clinical predictor, which ought to be considered upon selection of an optimal donor in an AL setting. Even though our findings, from a statistical point of view, did not confirm the effect of HLA-E mismatch in “early/intermediate disease” patients, we suspect, on account of our hypothesis model, that all AL patients, albeit to a different degree, could benefit from HLA-E disparate grafts. Future larger independent cohort studies, such as that of our ongoing CIBMTR IB16-01 project with more than 1500 AL patients enrolled, which may or may not confirm these results, will undoubtedly show the way. Funding The authors would like to thank the Deutsche José Carreras Leukämie-Stiftung e.V. (Grant No. DJCLS 11/10 and R 15/19) and the German Red Cross Blood Transfusion Service, BadenWuerttemberg – Hessen for financially supporting this work. haematologica | 2017; 102(11)

Acute leukemia and HSCT: role of HLA-E matching

References 1. Appelbaum FR. Haematopoietic cell transplantation as immunotherapy. Nature. 2001;411(6835):385-389. 2. Petersdorf EW. The major histocompatibility complex: a model for understanding graftversus-host disease. Blood. 2013;122(11): 1863-1872. 3. Wieten L, Mahaweni NM, Voorter CE, Bos GM, Tilanus MG. Clinical and immunological significance of HLA-E in stem cell transplantation and cancer. Tissue Antigens. 2014;84(6):523-535. 4. Sullivan LC, Clements CS, Rossjohn J, Brooks AG. The major histocompatibility complex class Ib molecule HLA-E at the interface between innate and adaptive immunity. Tissue Antigens. 2008;72(5):415424. 5. Braud V, Jones EY, McMichael A. The human major histocompatibility complex class Ib molecule HLA-E binds signal sequence-derived peptides with primary anchor residues at positions 2 and 9. Eur J Immunol. 1997;27(5):1164-1169. 6. Grimsley C, Ober C. Population genetic studies of HLA-E: evidence for selection. Hum Immunol. 1997;52(1):33-40. 7. Lee N, Goodlett DR, Ishitani A, Marquardt H, Geraghty DE. HLA-E surface expression depends on binding of TAP-dependent peptides derived from certain HLA class I signal sequences. J Immunol. 1998;160(10):49514960. 8. Ulbrecht M, Couturier A, Martinozzi S et al. Cell surface expression of HLA-E: interaction with human beta2-microglobulin and allelic differences. Eur J Immunol. 1999;29(2):537-547. 9. Maier S, Grzeschik M, Weiss EH, Ulbrecht M. Implications of HLA-E allele expression and different HLA-E ligand diversity for the regulation of NK cells. Hum Immunol. 2000;61(11):1059-1065. 10. Strong RK, Holmes MA, Li P et al. HLA-E allelic variants. Correlating differential expression, peptide affinities, crystal structures, and thermal stabilities. J Biol Chem. 2003;278(7):5082-5090. 11. Braud VM, Allan DS, O'Callaghan CA et al. HLA-E binds to natural killer cell receptors CD94/NKG2A, B and C. Nature. 1998;391(6669):795-799. 12. Kraemer T, Celik AA, Huyton T et al. HLA-E: presentation of a broader peptide repertoire impacts the cellular immune esponse-implications on HSCT outcome. Stem Cells. Int. 2015;2015:346714. 13. Lampen MH, Hassan C, Sluijter M et al. Alternative peptide repertoire of HLA-E reveals a binding motif that is strikingly similar to HLA-A2. Mol Immunol. 2013;53(12):126-131. 14. Joosten SA, Sullivan LC, Ottenhoff TH. Characteristics of HLA-E restricted T-cell responses and their role in infectious diseases. J Immunol Res. 2016;2016:2695396. 15. Pietra G, Romagnani C, Moretta L, Mingari MC. HLA-E and HLA-E-bound peptides: recognition by subsets of NK and T cells. Curr Pharm Des. 2009;15(28):3336-3344. 16. Tamouza R, Busson M, Rocha V et al. Homozygous status for HLA-E*0103 confers protection from acute graft-versushost disease and transplant-related mortality in HLA-matched sibling hematopoietic stem cell transplantation. Transplantation. 2006;82(11):1436-1440.

haematologica | 2017; 102(11)

17. Tamouza R, Rocha V, Busson M et al. Association of HLA-E polymorphism with severe bacterial infection and early transplant-related mortality in matched unrelated bone marrow transplantation. Transplantation. 2005;80(1):140-144. 18. Danzer M, Polin H, Proll J et al. Clinical significance of HLA-E*0103 homozygosity on survival after allogeneic hematopoietic stem-cell transplantation. Transplantation. 2009;88(4):528-532. 19. Ludajic K, Rosenmayr A, Fae I et al. Association of HLA-E polymorphism with the outcome of hematopoietic stem-cell transplantation with unrelated donors. Transplantation. 2009;88(10):1227-1228. 20. Furst D, Bindja J, Arnold R et al. HLA-E polymorphisms in hematopoietic stem cell transplantation. Tissue Antigens. 2012; 79(4):287290. 21. Hosseini E, Schwarer AP, Jalali A, Ghasemzadeh M. The impact of HLA-E polymorphisms on relapse following allogeneic hematopoietic stem cell transplantation. Leuk Res. 2013;37(5):516-519. 22. Hosseini E, Schwarer AP, Ghasemzadeh M. The impact of HLA-E polymorphisms in graft-versus-host disease following HLA-E matched allogeneic hematopoietic stem cell transplantation. Iran J Allergy Asthma Immunol. 2012;11(1):15-21. 23. Harkensee C, Oka A, Onizuka M et al. Single nucleotide polymorphisms and outcome risk in unrelated mismatched hematopoietic stem cell transplantation: an exploration study. Blood. 2012; 119(26):6365-6372. 24. Hosseini E, Schwarer AP, Ghasemzadeh M. Do human leukocyte antigen E polymorphisms influence graft-versus-leukemia after allogeneic hematopoietic stem cell transplantation? Exp Hematol. 2015; 43(3):149157. 25. Geraghty DE, Stockschleader M, Ishitani A, Hansen JA. Polymorphism at the HLA-E locus predates most HLA-A and -B polymorphism. Hum Immunol. 1992;33(3):174-184. 26. Iacobelli S. Suggestions on the use of statistical methodologies in studies of the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant. 2013;48 Suppl 1:S1-37. 27. Bacigalupo A, Ballen K, Rizzo D et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):1628-1633. 28. EBMT.org [Internet]. European Group for Blood and Marrow Transplantation. MEDAB forms manual. A guide to the completion of the EBMT HSCT MED-AB forms. appendix iii. [Updated 2015 August 18; cited 2015 September 12]. Available from: http://www.ebmt.org/Contents/DataManagement/Registrystructure/MEDABdatacollectionforms/Documents/MEDABFormsManual.pdf. Last accessed 27.09. 2017. 29. 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. 30. bioinformatics.bethematchclinical.org [Internet]. National Marrow Donor Program. NMDP Policy for HLA Confirmatory Typing Requirements for Unrelated Adult Donors and Patients. [Updated 2015 January; cited 2015 September 12]. Available from:

31.

32.

33. 34. 35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

https://bioinformatics.bethematchclinical.or g/policies/. Last accessed 27.09.2017. Cooley S, Trachtenberg E, Bergemann TL et al. Donors with group B KIR haplotypes improve relapse-free survival after unrelated hematopoietic cell transplantation for acute myelogenous leukemia. Blood. 2009; 113(3):726-732. CRAN.R-project.org. Therneau T. A Package for Survival Analysis in S. R package version 2.38. [updated 2015 July 02; cited 2015 September 12].Available from: https://cran.r-project. org/web/ packages/ survival/index.html. Last accessed 27.09. 2017 Gratwohl A. The EBMT risk score. Bone Marrow Transplant. 2012;47(6):749-756. Freedman LS. Tables of the number of patients required in clinical trials using the logrank test. Stat Med. 1982;1(2):121-129. Olson JA, Leveson-Gower DB, Gill S et al. NK cells mediate reduction of GVHD by inhibiting activated, alloreactive T cells while retaining GVT effects. Blood. 2010;115(21):4293-4301. Shilling HG, McQueen KL, Cheng NW et al. Reconstitution of NK cell receptor repertoire following HLA-matched hematopoietic cell transplantation. Blood. 2003; 101(9):37303740. Picardi A, Mengarelli A, Marino M et al. Upregulation of activating and inhibitory NKG2 receptors in allogeneic and autologous hematopoietic stem cell grafts. J Exp Clin Cancer Res. 2015;34:98. Hoare HL, Sullivan LC, Clements CS et al. Subtle changes in peptide conformation profoundly affect recognition of the non-classical MHC class I molecule HLA-E by the CD94-NKG2 natural killer cell receptors. J Mol Biol. 2008;377(5):1297-1303. Houchins JP, Lanier LL, Niemi EC, Phillips JH, Ryan JC. Natural killer cell cytolytic activity is inhibited by NKG2-A and activated by NKG2-C. J Immunol. 1997;158(8): 3603-3609. Vales-Gomez M, Reyburn HT, Erskine RA, Lopez-Botet M, Strominger JL. Kinetics and peptide dependency of the binding of the inhibitory NK receptor CD94/NKG2-A and the activating receptor CD94/NKG2-C to HLA-E. EMBO J. 1999;18(15):42504260. Kaiser BK, Barahmand-Pour F, Paulsene W et al. Interactions between NKG2x immunoreceptors and HLA-E ligands display overlapping affinities and thermodynamics. J Immunol. 2005;174(5):2878-2884. Llano M, Lee N, Navarro F et al. HLA-Ebound peptides influence recognition by inhibitory and triggering CD94/NKG2 receptors: preferential response to an HLAG-derived nonamer. Eur J Immunol. 1998; 28(9):2854-2863. Celik AA, Kraemer T, Huyton T, Blasczyk R, Bade-Doding C. The diversity of the HLA-Erestricted peptide repertoire explains the immunological impact of the Arg107Gly mismatch. Immunogenetics 2016;68(1):2941. Palmisano GL, Contardi E, Morabito A et al. HLA-E surface expression is independent of the availability of HLA class I signal sequence-derived peptides in human tumor cell lines. Hum Immunol. 2005;66 (1):1-12. Hu B, He Y, Wu Y et al. Activated allogeneic NK cells as suppressors of alloreactive responses. Biol Blood Marrow Transplant. 2010;16(6):772-781.

1955

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Complications in Hematology

Ferrata Storti Foundation

HIF1A is a critical downstream mediator for hemophagocytic lymphohistiocytosis

Rui Huang,1,2 Yoshihiro Hayashi,1,3 Xiaomei Yan,1 Jiachen Bu,1,4 Jieyu Wang,1 Yue Zhang,1,5 Yile Zhou,1 Yuting Tang,1,6 Lingyun Wu,1 Zefeng Xu,5 Xin Liu,4,7 Qianfei Wang,4,7 Jianfeng Zhou,6 Zhijian Xiao,5 James P. Bridges,8 Rebecca A. Marsh,9 Kejian Zhang,10 Michael B. Jordan,9 Yuhua Li2 and Gang Huang1

RH and YH contributed equally to this work and GH and YHL contributed equally to this study as joint senior authors

Division of Experimental Hematology and Cancer Biology, Cincinnati Children’s Hospital Medical Center, OH, USA; 2Department of Hematology, Zhujiang Hospital, Southern Medical University, Guangzhou, China; 3Laboratory of Oncology, School of Life Sciences, Tokyo University of Pharmacy and Life Sciences, Japan; 4Key Laboratory of Genomic and Precision Medicine, Beijing Institute of Genomics, Chinese Academy of Sciences, China; 5 State Key Laboratory of Experimental Hematology, Institute of Hematology and Blood Diseases Hospital, Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin, China; 6Department of Hematology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China; 7 University of Chinese Academy of Sciences, Beijing, China; 8Perinatal Institute, Division of Pulmonary Biology, Cincinnati Children’s Hospital Medical Center, OH, USA; 9Division of Bone Marrow Transplantation and Immune Deficiency, Cincinnati Children’s Hospital, OH, USA and 10Division of Human Genetics, Cincinnati Children’s Hospital Medical Center, OH, USA 1

Haematologica 2017 Volume 102(11):1956-1968

ABSTRACT

Correspondence: gang.huang@cchmc.org

Received: June 20, 2017. Accepted: August 24, 2017. Pre-published: August 31, 2017. doi:10.3324/haematol.2017.174979 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/11/1956 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

1956

emophagocytic lymphohistiocytosis (HLH) is a life-threatening syndrome characterized by overwhelming immune activation. A steroid and chemotherapy-based regimen remains as the first-line of therapy but it has substantial morbidity. Thus, novel, less toxic therapy for HLH is urgently needed. Although differences exist between familial HLH (FHL) and secondary HLH (sHLH), they have many common features. Using bioinformatic analysis with FHL and systemic juvenile idiopathic arthritis, which is associated with sHLH, we identified a common hypoxia-inducible factor 1A (HIF1A) signature. Furthermore, HIF1A protein levels were found to be elevated in the lymphocytic choriomeningitis virus infected Prf1–/– mouse FHL model and the CpG oligodeoxynucleotide-treated mouse sHLH model. To determine the role of HIF1A in HLH, a transgenic mouse with an inducible expression of HIF1A /ARNT proteins in hematopoietic cells was generated, which caused lethal HLH-like phenotypes: severe anemia, thrombocytopenia, splenomegaly, and multi-organ failure upon HIF1A induction. Mechanistically, these mice show type 1 polarized macrophages and dysregulated natural killler cells. The HLH-like phenotypes in this mouse model are independent of both adaptive immunity and interferon-γ, suggesting that HIF1A is downstream of immune activation in HLH. In conclusion, our data reveal that HIF1A signaling is a critical mediator for HLH and could be a novel therapeutic target for this syndrome.

Introduction Hemophagocytic lymphohistiocytosis (HLH) is a syndrome of overwhelming immune activation characterized by several clinical features, such as high fever, multi-lineage cytopenia, splenomegaly, and hyperferritinemia.1-3 In primary HLH patients, various mutations in genes related to the granule-dependent cytotoxicity pathway in T/natural killer (NK) cells have been identified.4 CD8+ T cells and interferon gamma (IFN-γ) have been shown to play critical roles in the pathogenesis of primary HLH, the onset of which is usually triggered by viral infection.5 On the other hand, secondary HLH (sHLH) has a heterogeneous etiology, which is often haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

triggered by systemic viral infection, autoimmune disorder, or hematologic malignancies.3,6 In contrast to primary HLH, cytotoxic activity of T/NK cells is not always decreased in sHLH.7 Macrophage activation syndrome (MAS), a form of sHLH in the context of rheumatic disease that is especially associated with systemic juvenile idiopathic arthritis (sJIA),8 is associated with aberrant tolllike receptor (TLR)-induced gene expression patterns.9 Despite reports that primary and sHLH have different genetic aberrations and distinct cytotoxic activities, the key features for HLH (cytopenias and a unique presentation of extreme inflammation) remain common.7 Thus, we hypothesized that there are common underlying downstream mediators for HLH phenotype development. Identifying these mediators for HLH will not only help to understand the disease, but also lead to the development of better therapies. Hypoxia-inducible factor (HIF), which was originally discovered as a critical transcription factor for optimal cellular adaptation to hypoxia, actually plays an important role in immune response, both in hypoxia and normoxia conditions10,11 HIF consists of heterodimers, HIF-α subunit and ARNT. ARNT is stably expressed in many cell types, while HIF-α subunits are expressed differently in tissues and cells. So far, three isoforms of HIF-α subunits, HIF1A, HIF2A, and HIF3A, have been documented.12 HIF1A is widely expressed in both innate and adaptive immune cells,13-16 while HIF2A/EPAS1 expression is limited in certain immune cell types,17,18 and the HIF3A expression pattern has not been so well studied. Many stimuli or factors that have a function in the immune response lead to HIF1A protein accumulation and activation independent of the hypoxic regulation. Bacteria, lipopolysaccharide (LPS), and tumor necrosis factor-α (TNF-α) have been well documented to induce HIF1A accumulation in macrophages, thereby boosting their microbicidal capacity.19,20 T-cell receptor ligation in T cells increases HIF1A transcription and HIF1A protein accumulation. Moreover, cytokines, such as IL-2, induce HIF1A accumulation in CD8+ T cells enhancing their cytotoxic function.10 There is growing evidence to indicate the potential role of HIF1A signaling in immune activation. To identify the key downstream mediators for HLH, we first performed a bioinformatic analysis of published microarray expression datasets of familial hemophagocytic lymphohistiocytosis (FHL) and sJIA.9,21 We found that the HIF1A signature, which refers to a group of HIF1A-induced genes, was enriched in both FHL and sJIA patients’ peripheral blood mononuclear cells (PBMCs). Then, we confirmed that HIF1A protein expression was significantly increased in two widelyused HLH mouse models. Furthermore, in vivo data from transgenic mice show that activation of HIF1A in hematopoietic cells results in HLH-like phenotypes. Our study suggests that the HIF1A signaling pathway is a critical pathological downstream mediator for HLH development.

Methods Cell line and mice Murine cell line Raw264.7 was purchased from ATCC. Cells were cultured in DMEM medium in a 5% CO2 incubator at 37°C, and subcultured every 2-3 days. Rag1–/–, Ifngr–/–, Ifng–/–, Prf1–/–, Vav1haematologica | 2017; 102(11)

Cre, and Rosa26-LSL-rtTA mice were purchased from Jackson Laboratory. Transgenic HIF1A-TPM mice were kindly provided by Professor James Bridge from Cincinnati Children’s Hospital Medical Center (CCHMC).22 Mice, with the genotype of Ncr1iCre, were kindly provided by Professor Eric Vivier from the French INSERM Laboratory.23 LCMV-infected Prf1–/– mouse model and the repeated CpG-treated mouse model were generated as previously described.5,24,25 All animal studies were performed according to an approved Institutional Animal Care and Use Committee protocol and federal regulations.

Statistical analysis Data were analyzed by Prism 6.0 (GraphPad Software). P<0.05 was considered significant. Continuous variables were analyzed by using Student t-test or one-way ANOVA. Mice survival was estimated using the Kaplan-Meier method. Information concerning the antibodies and reagents, bioinformatic analysis, flow cytometry, ELISA, western blot, histology, generation of bone marrow-derived macrophages, and the assay used to determine the capacity of macrophages to engulf erythroblasts is reported in detail in the Online Supplementary Appendix.

Results Activated HIF1A signaling in FHL and sJIA patients Immune activation is coupled with cytokine signaling and transcriptional changes. To investigate the network of transcription factors in HLH pathogenesis, we took a bioinformatic approach and analyzed two published microarray datasets of patients with FHL and sJIA.9,21 Since there are no available sHLH transcript profile data, we utilized the microarray data of sJIA patients which are more likely to be complicated by macrophage activation syndrome, a subtype of sHLH in the context of rheumatoid diseases.26-28 We performed unbiased TF-target enrichment analysis29 and found that HIF1A, NF-κB, GATA1, and STAT1 are the common immune-related transcription factors in both the FHL and sJIA datasets (Figure 1A and B and Online Supplementary Figure S1A and B). Among these transcription factors, HIF1A is of particular interest as it regulates not only the up-regulated genes but also the down-regulated genes in HLH, indicating that HIF1A might be a critical mediator in HLH development. HIF1A is known to be regulated in both transcriptional and post-translational levels. In the FHL dataset, HIF1A mRNA is increased in the FHL patients compared to healthy donors; however, there is no significant difference in HIF1A expression between FHL patients with and those without a genetic diagnosis (Online Supplementary Figure S1C and D). At the same time, in the sJIA dataset, there is no significant change in HIF1A expression at the mRNA level between patients and healthy donors (Online Supplementary Figure S1E). To investigate enrichment of the HIF1A signature in HLH, we also utilized another bioinformatic approach of gene set enrichment analysis (GSEA) to analyze these two datasets, and revealed that the HIF1A signature is significantly enriched in both the FHL and sJIA PBMCs datasets (Figure 1C and D). However, there is no significant difference in HIF1A signature enrichment between FHL patients with and those without a genetic diagnosis (Online Supplementary Figure S1F). There are 258 leading edge genes (LEGs) in the FHL dataset and 214 LEGs in the sJIA 1957

R. Huang et al.

dataset with 108 overlapping common LEGs (Figure 1E). Gene ontology analysis showed that these overlapping common LEGs are related to blood coagulation, chemotaxis, glycolysis, oxygen species metabolic process, platelet activation, immune response, and cytokines (Figure 1E). These results further suggest that HIF1A may play a key role in regulating downstream targets in both primary HLH and sHLH patients.

1958

Elevated HIF1A protein in LCMV-infected Prf1–/– and CpG-treated mouse HLH models Several mouse models that recapitulate primary HLH, sHLH, or MAS have been reported.7 We determined whether HIF1A signaling is activated in established HLH mouse models. The LCMV-infected perforin-deficient (Prf1–/–) mouse model is a well-known HLH mouse model that recapitulates biallelic perforin mutation patients with

Figure 1. HIF1A signature is enriched in familial hemophagocytic lymphohistiocytosis (FHL) and systemic juvenile idiopathic arthritis (sJIA) patients. (A and B) Heatmaps showing differentially expressed genes with more than a 1.5-fold difference in expression comparing peripheral blood mononuclear cells (PBMCs) from FHL patients (n=11) with healthy donors (n=33) based on a published microarray dataset (GSE26050) (A) and from sJIA patients (n=17) with healthy donors (n=30) based on a published microarray dataset (GSE7753). (B) To the left of the heatmap are top predicted transcription factors using transcription factor-target enrichment analysis using the Go-Elite algorithm in AltAnalyze software. The common predicted transcription factors in both FHL and sJIA datasets are marked in red. (C and D) Gene set enrichment analysis (GSEA) plot showing an increase in gene expression of HIF1A-induced genes in FHL microarray dataset (C) and sJIA dataset (D). Up-regulated genes (fold change >2.0) in HIF1A overexpressed human cord blood (CB) CD34+ cells serve as the HIF1Ainduced genes (Genomic Spatial Event database; GSE 54663). Normalized enrichment score (NES), P-value, and false discovery rate (FDR) are shown. (E) Venn diagram showing the overlap of the leading edge genes from GSEA comparing FHL and sJIA datasets. Tabular data showing gene ontology (GO) analysis to the overlapping leading edge genes.

haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

pathogen infection (Figure 2A).5 After challenging them with LCMV, the Prf1–/– mice quickly developed anemia and thrombocytopenia (Figure 2B). CD8+ T cells activate macrophages via secreting IFN-γ in this mouse model. We measured HIF1A expression levels in Gr1–CD115–F4/80+SSClow spleen macrophages by using flow cytometry (Figure 2C)30 and found that HIF1A levels in spleen macrophages were significantly increased in LCMV-infected Prf1–/– mice compared to the control mice (Figure 2D). We also identified profound type 1 polarized macrophages in the spleen (Figure 2E and F) and bone marrow (data not shown) from the LCMV-infected Prf1–/– mice.

Repeated injections of TLR9 ligand CpG oligodeoxynucleotides (ODN) causes sHLH in wild-type (WT) mice,24 which mimics sHLH features (Figure 2G). After injecting CpG five times into WT mice, CpG-treated-mice developed anemia and thrombocytopenia (Figure 2H). Similar to what was observed in the primary HLH model, we found that HIF1A expression in the spleen macrophages was also increased in CpG-treated mice (Figure 2I). Type 1 polarization of macrophages in spleen (Figure 2J and K) and bone marrow (data not shown) was observed in CpGtreated mice. Taken together, these data suggest that HIF1A protein expression is elevated both in primary HLH and sHLH mouse models.

haematologica | 2017; 102(11)

Figure 2. HIF1A protein in macrophages is elevated in hemophagocytic lymphohistiocytosis (HLH) mouse models. (A) Schematic diagram depicting lymphocytic choriomeningitis virus (LCMV)-infected –/– Prf1 HLH mouse model. Perforin-defi–/– cient (Prf1 ) mice were infected with or without LCMV-WE of 200 plaque-forming units (PFU) via intraperitoneal injection. Mice were sacrificed and analyzed on day (d)14 after inoculation. (B) Hemoglobin (Hb) and platelets (PLT) of LCMV-infected –/– mice. (C) Flow or non-infected Prf1 cytometry gating strategy showing splenic macrophages identified as Gr1-CD115SSClowF4/80+ from indicated mice. (D) Flow cytometry histogram plot showing HIF1A level in splenic macrophages in indicated mice. The tinted gray histogram represents –/– an LCMV-infected Prf1 mouse. The blue histogram represents a non-infected –/– Prf1 mouse. Plot is representative of 4 independent intracellular staining. (E) Flow cytometry dot plot showing CD80 and CD206 expression in splenic macrophages in LCMV-infected or non-infected mice. Plots are representative of 4 independent stainings. (F) Quantitative analysis of percentage of CD80+CD206– macrophages in total splenic macrophages in indicated mice. (G) Schematic diagram showing repeated CpG-treated HLH mouse model. CpG (75 mg) or PBS was injected i.p. to wild-type mice on days 0, 2, 4, 6, 8. On d9, mice were sacrificed and analyzed. (H) Hb and PLT of CpG or PBS treated mice on d9. (I) Flow cytometry histogram plot showing HIF1A level of splenic macrophages in indicated mice. The tinted gray histogram represents a CpG-treated mouse. The blue histogram represents a PBS-treated mouse. Plot is representative of 3 independent experiments. (J) Flow cytometry dot plots showing CD80 and CD206 expression in splenic macrophages in CpG or PBS-treated mice. Plots are representative of 3 independent experiments. (K) Quantitative analysis of percentage of CD80+CD206– macrophages in total splenic macrophages in CpG or PBS-treated mice. **P<0.01, ***P<0.001 versus control. Individual symbols each represent one mouse. PFU: plaque-forming units.

1959

R. Huang et al.

Figure 3. Induction of HIF1A/ARNT allele in hematopoietic cells in C57BL/6 mice develops features of hemophagocytic lymphohistiocytosis (HLH). (A) Schematic diagram of inducible HIF1A/ARNT transgenic mouse and HIF1A triple point mutant (TPM) under the Vav1-Cre driver. (B) HIF1A protein expression in c-Kit+ and c-Kitâ&#x20AC;&#x201C;cells from bone marrow of Vav1-Cre/TPM mice and control mice shown by western blot. (C) Flow cytometry histogram plot showing HIF1A level in splenic macrophages gated on Gr1-CD115-F4/80+cells, T cells gated on CD3+cells, B cells gated on B220+cells. The tinted gray histogram represents a Vav1-Cre/wild-type (WT) mouse, the blue histogram represents a Vav1-Cre/TPM mouse. (D-I) Vav1-Cre/TPM mice and control mice were administrated with doxycycline, sacrificed and analyzed on day (d)8. Hemoglobin (Hb) (D), platelet (PLT) (E), and bone marrow (BM) cellularity of one femur (F) were shown. (G-I) Representative plot of spleen (SP) (G), quantitative analysis of spleen weight (H) and representative plot of liver (I) were shown. (J and K) Liver sections of Vav1-Cre/TPM mice and control mice on d4 were stained by H&E. (J) Representative plots at an original magnification of Ă&#x2014;400. Infiltrates are marked with black arrow. (K) Representative scattered dot plots indicating infiltration of CD11b+ myeloid cells (gated on CD45+ cells) into liver. (L) Serum ferritin levels of Vav1-Cre/TPM mice and control mice were measured by ELISA at end point of survival. (M) Kaplan-Meier analysis of survival of Vav1-Cre/TPM mice (n=32) and control mice (n=26). (N) Relative mRNA expression of selected genes from the common leading edge genes of GSEA of FHL and sJIA datasets was measured in PBMCs from Vav1-Cre/TPM mice (n=3) and control mice (n=3) by qRT-PCR. Data are representative of 3 independent experiments. Depicted data are from at least 3 independent experiments. Individual symbol in dot plots each represents one mouse. **P<0.01, ***P<0.001 versus control for all experiments. GSEA: gene set enrichment analysis; sJIA: systemic juvenile idiopathic arthritis.

1960

haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

I Figure 4. Unaffected T-cell populations and dysregulated natural killer (NK) cells in inducible triple point mutant (TPM) mice. Vav1-Cre/TPM and control mice were administrated with doxycycline and analyzed on day (d)8 after doxycycline induction. (A and B) Percentage of CD8+ cells in CD3+ T cells (A) and absolute number of CD8+ T cells (B) in spleen (SP). (C) Left: representative flow cytometry scatter dot plots showing IFN-γ production in splenic CD8+ T cells (gated on CD3+CD8+ cells). Splenocytes from Vav1-Cre/TPM and control mice were harvested and re-stimulated with PMA, ionomycin, and monensin for 5 hours to determine the capacity of IFN-γ production by intracellular staining of flow cytometry. Right: quantitative analysis of IFN-γ level in splenic CD8+ T cells. (D) Representative flow cytometry scattered dot plots showing the frequency of total NK cells (CD3–CD115–B220–NK1.1+) and mature NK cells (CD3-CD115-B220-NK1.1+ DX5+) in spleen of Vav1-Cre/TPM and control mice. (E and F) Quantitative analysis of frequency (E, left) and absolute number (E, right) of total splenic NK cells and frequency (F, left) and absolute number (F, right) of mature splenic NK cells from flow cytometry data of (D). (G) Splenocytes were separated in indicated mice and re-stimulated as mentioned in (B). Left: representative flow cytometry scatter dot plots showing IFN-γ production in splenic NK cells (gated on CD3-CD115-B220-NK1.1+ cells). Right: quantitative analysis of IFN-γ production in splenic NK cells from flow cytometry data of (G). (H) Quantitative analysis of CD107a expression of splenic NK cells (gated on CD3-CD115-B220NK1.1+ cells). (I) Quantitative analysis of NKp46 expression of splenic NK cells (gated on CD3-CD115-B220-NK1.1+ cells). Depicted data are representative of 3 independent experiments. Individual symbol in dot plots each represents one mouse. *P<0.05, **P<0.01, ***P<0.001 versus control for all experiments. #: absolute number; ns: not significant.

haematologica | 2017; 102(11)

1961

R. Huang et al.

Inducible activation of HIF1A is sufficient for developing HLH-like phenotypes in C57BL/6 background mice

include P402A, P564A and N803A, which prevent degradation and facilitate transcriptional activation of HIF1A.22 Thus, Vav1-Cre/TPM mice have stabilized and constitutively active HIF1A protein (Figure 3A) in hematopoietic cells after administration of doxycycline. Vav1-Cre mice without the TPM allele (Vav1-Cre/WT) served as control. We confirmed an increase in HIF1A protein level in both c-Kit positive and negative cells (Figure 3B) by western blot and in individual cell lineages by flow cytometry (Figure 3C) in Vav1-Cre/TPM mice compared to control mice. Importantly, the level of HIF1A in macrophages in

To determine the significance of HIF1A signaling activation in HLH development in vivo, we generated transgenic mice with inducible HIF1A/ARNT expression in hematopoietic cells. We combined the Vav1-Cre allele, Rosa26-loxp-stop-loxp (LSL) reverse-tetracycline-controlled transactivator (rtTA) allele, and triple point mutation (TPM) HIF1A /wild-type ARNT alleles (tet-onTPM/ARNT) (Vav1-Cre/TPM). Triple point mutations

1962

Figure 5. Dendritic cells (DCs) are slightly changed but macrophages are strongly polarized in triple point mutation (TPM)-induced mice. Vav1Cre/TPM and control mice were administrated with doxycycline, DCs and macrophages were analyzed. (A) Representative flow cytometry scatter dot plots showing splenic DCs from Vav1-Cre/TPM and control mice on day (d)4 after doxycycline induction. Plasmacytoid DCs (pDCs) are identified as NK1.1 – CD11c + B220 + CD11b – CD8 cells. CD11b+ conventional DCs (cDCs) are identified as NK1.1CD11c+B220-CD11b+CD8-cells. CD8+ cDCs are identified as N K 1 . 1 – C D 1 1 c + B 2 2 0 CD11b–CD8+cells. (B and C) Quantitative analysis of frequency (B) and absolute number (C) of pDCs, CD11b+ cDCs, and CD8+ cDCs in spleen cells on d4. (D-I) Quantitative analysis of frequency and absolute number of total macrophage in spleen (SP) (D) and bone marrow (BM) (G). Representative flow cytometry CD80/CD206 scatter dot plots of splenic macrophages (E) and BM macrophages (H). Quantitative analysis of percentage of CD80+CD206– macrophages in spleen (F) and BM (I). (J) Schematic diagram depicting assay measuring the ability of polarized macrophages to engulf erythroblast. BMDM from wild-type mice were incubated with IFN-γ for 48 hours (h) to polarize towards type 1, or with IL-4 for 24 h to polarize towards type 2, followed by co-culturing with erythroblasts for 12 h. (K) Representative flow cytometry CD80/CD206 scatter dot plots (left) of cultured BMDM. Cytospin was made followed by Giemsa staining. Engulfment of erythroblasts by polarized BMDM was observed under the microscope (right). (L) Quantitative analysis of percentage of macrophages engulfing erythroblasts of total macrophages. Depicted data are representative of at least 3 independent experiments. Individual symbol in dot plots each represents one mouse. *P<0.05, ***P<0.001 versus control for all experiments. #: absolute number; ns: not significant; Mf: macrophage.

haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

TPM mice is comparable to that in CpG-injected HLH mice (Online Supplementary Figure S2). After doxycycline administration, TPM mice with the pure C57BL/6 background quickly developed severe anemia and thrombocytopenia (Figure 3D and E). Bone marrow cellularity was dramatically reduced in the TPM mice (Figure 3F and Online Supplementary Figure S3A). We did not find a blockade of erythropoiesis in the bone marrow and spleen from Vav1-Cre/TPM mice (Online Supplementary Figure S4), indicating a cell extrinsic mechanism for quick progression of anemia and a decrease in bone marrow cellularity. Consistent with the diagnostic criteria for HLH, TPM mice showed splenomegaly (Figure 3G and H). Normal splenic follicular architecture was disrupted in the TPM mice (Online Supplementary Figure S3B). Liver dysfunction is commonly observed in HLH patients. Indeed, TPM mice had substantial inflammatory cells infiltrated into the liver (Figure 3I). Flow cytometric analysis revealed that most of these cells were CD11b+ myeloid cells (Figure 3J). However, we failed to find robust hemophagocytosis in the cytospins or sections of bone marrow,

spleen, or liver (data not shown). High levels of serum ferritin, which is also one of the diagnostic criteria for HLH, was observed in the TPM mice in comparison with the control mice (Figure 3K). Furthermore, several inflammatory cytokines, such as IL-6, IL-12, and IFN-γ, were increased in the serum from the TPM mice (Online Supplementary Figure S5). All of the mice succumbed within three weeks (Figure 3L). We further confirmed that several genes related to chemokine, macrophage activation, and glycolysis (which are the common LEGs of HIF1A signature in FHL and sJIA datasets), were elevated in the Vav1-Cre/TPM mice (Figure 3M and Online Supplementary Figure S6). Taken together, inducible expression of stabilized and active HIF1A with ARNT gives rise to HLH-like phenotypes in pure C57BL/6 background mice.

Unaffected T-cell populations and dysregulated NK cells in induced TPM mice Given that robust activation of CD8+ T cells is observed in the primary HLH mouse model, we first determined the T-cell populations in TPM mice. However, no significant

Figure 6. Adaptive immunity is not required for triple point mutation (TPM)-induced anemia, thrombocytopenia, and macrophage polarization. Rag1 /Vav1–/– –/– –/– –/– –/– Cre/LSL/TPM (Rag1 /TPM) mice were generated. Rag1 /Vav1-Cre/LSL/WT (Rag1 /WT) mice served as their control. Rag1 /TPM, Rag1 /WT, Vav1-Cre/TPM, Vav1-Cre/WT mice were administrated with doxycycline. (A and B) Representative flow cytometry scatter dot plots showing percentage of T, B cells (A) and natural –/– killer (NK) cells (B) in peripheral blood (PB) in Rag1 and wild-type mice. (C and D) Mice were administrated with doxycycline and analyzed on day (d)8 after doxycycline induction. Hemoglobin (Hb), platelets (PLT) (C) and spleen (SP) weight (D) are shown. (E) Representative flow cytometry CD80/CD206 scatter dot plots of –/– –/– splenic macrophages. (F) Kaplan-Meier analysis of survival of Rag1 /TPM (n=8), Rag1 /WT (n=8), Vav1-Cre/TPM (n=32), Vav1-Cre/WT (n=26) mice. Statistical –/– –/– –/– analysis showed that there was a significance between Rag1 /TPM and Rag1 /WT mice (P<0.0001), but no significance between Rag1 /TPM and Vav1-Cre/TPM mice (P>0.05). Depicted data are representative of 3 independent experiments. Individual symbol in dot plots each represents one mouse. ***P<0.001 versus control for all experiments. ns: non-significance. –/–

haematologica | 2017; 102(11)

1963

R. Huang et al.

change in the frequency and absolute number of CD8+ T cells was observed in the TPM mice (Figure 4A and B). We also measured IFN-γ production in CD8+ T cells and did not find a significant difference between the TPM mice and the control mice (Figure 4C), indicating that CD8+ T cells may not play a key role in HIF1A induced HLH-like phenotypes. Natural killer cell activity is important for immune homeostasis. NK cell defect is one of the critical features of primary HLH. Reduced NK cell number or impaired NK cell function has been reported in some of the sHLH patients. Interestingly, we found a significant reduction in the number of total NK cells and DX5+ mature NK cells in the spleen (Figure 4D-F), peripheral blood and bone marrow (data not shown) from the TPM mice. Importantly, IFN-γ production in NK cells was impaired in the TPM mice (Figure 4G). However, cell surface CD107a expression in NK cells was comparable between the TPM mice and the control mice (Figure 4H), suggesting no major defect in degranulation of cytotoxic granules in TPM mice. It has been reported that hypoxia may lead to a reduction in NKp46 expression, an NK cell activating receptor, in vitro.31 However, there was no significant difference in NKp46 expression between the TPM mice and the control mice (Figure 4I). These data suggest that TPM mice have quantitative and functional dysregulation in NK cells. To determine whether the impairment of NK cells is due to intrinsic or extrinsic NK cell factors, we generated Ncr1-iCre /LSL/TPM (Ncr1-iCre/TPM) mice. Using GFP reporter, we confirmed that Ncr1-iCre is specifically expressed in NK cells (Online Supplementary Figure S7A and C).23 Surprisingly, after NK cell specific TPM induction, we did not find any changes in the NK cell number or differentiation pattern compared to the control mice (Online Supplementary Figure S7B and D), indicating that the NK cell dysregulation in Vav1-Cre/TPM mice may be due to a non-autonomous cellular mechanism.

Slightly changed dendritic cells but strongly polarized Type 1 macrophages in induced TPM mice Since a minor fraction of dendritic cells (DCs) persistently present antigens and drive T cells in the primary HLH mouse model, we measured DC population in the TPM mice. There was an increase in the number of plasmacytoid DCs (pDCs) in the early stage but not in the number of conventional DCs (cDCs) (Figure 5A-C). However, at a later time, there was no significant difference in numbers of pDCs and cDCs between the TPM mice and the control mice (Online Supplementary Figure S8). Macrophage activation by diverse triggers is a common feature in HLH. We found type 1 polarization of macrophages in both primary and sHLH models; thus, we investigated the macrophage population in TPM mice. The number of macrophages was significantly increased in the spleen, but not in the bone marrow in TPM mice (Figure 5D and G). More importantly, the macrophages were polarized toward type 1 in both bone marrow and spleen from the TPM mice (Figure 5E, F, H and I). To further determine whether type 1 polarized macrophages are able to phagocytose erythrocytes and cause anemia, we cultured erythroblasts with IFN-γ-polarized type 1 bone marrow-derived macrophages (BMDMs) or IL-4-polarized type 2 BMDMs and found that only type 1 macrophages, but not type 2 macrophages, engulfed erythroblasts (Figure 5J-L and Online Supplementary Figure S9). Taken 1964

together, our data suggest that HIF1A signaling activation causes type 1 macrophage polarization, which might also contribute to engulfment of erythroblasts and cause anemia in the HLH disease scenario.

Adaptive immune cells are not required for TPM-induced HLH phenotypes Since there was no change in the frequency of the total and IFN-γ producing CD8+ T cells in the TPM mice, we further investigated the role of the lymphocytes in TPMinduced HLH phenotypes. We crossed Vav1-Cre/TPM mice with recombination activation gene 1 (Rag1)-deficient (Rag1–/–/) mice that lack T cells and B cells (Figure 6AB) and generated Rag1–/–/Vav1-Cre/TPM mice. Rag1–/–/Vav1-Cre/WT mice served as control. After administration of doxycycline, Rag1–/–/Vav1-Cre/TPM developed similar anemia, thrombocytopenia, and splenomegaly as the Vav1-Cre/TPM mice (Figure 6C and D). Type 1 macrophage polarization was also observed in Rag1–/–/Vav1-Cre/TPM mice (Figure 6E). Survival of Rag1–/–/Vav1-Cre/TPM mice was not prolonged compared to Vav1-Cre/TPM mice (Figure 6F). These data indicate that adaptive immunity is not essential for TPM-induced HLH phenotypes and non-lymphoid cells are sufficient to mediate disease progression in TPM mice.

Genetically blocking IFN-γ signaling could not rescue TPM-induced phenotypes except to partly rescue the anemia IFN-γ is a critical factor upstream of HIF1A for type 1 macrophage polarization and is essential for disease development in the LCMV-infected Prf1–/– HLH mouse model. Thus, we determined the role of IFN-γ signaling in TPMinduced HLH. We generated Ifngr–/–/Vav1-Cre /TPM (Ifngr–/–/TPM) mice. Ifng–/–/Vav1-Cre /WT (Ifngr–/–/WT) mice served as control. Interestingly, TPM-induced anemia was partially rescued in Ifngr deficient mice (Figure 7A). However, the Ifngr–/–/TPM mice still developed severe thrombocytopenia (Figure 7B) and all of them succumbed to disease, but with a prolonged latency compared to the Vav1-Cre/TPM mice (Figure 7C). Flow cytometric analysis revealed that TPM-induced type 1 polarization of macrophages was not blocked in Ifngr-deficient mice (Figure 7D). We also generated Ifng–/–/Vav1-Cre /TPM (Ifngr–/–/TPM) mice and found similar results (Online Supplementary Figure S10) indicating that HIF1A-induced HLH-like phenotypes in TPM mice are independent of IFN-γ ligand and receptor. It is likely that IFN-γ ligand and receptor are the upstream of HIF1A signaling, and HIF1A activation itself could lead to type 1 macrophage polarization to some degree even without IFN-γ ligand and receptor. To determine whether IFN-γ could induce HIF1A signaling and cause type 1 polarization in macrophages, we treated the mouse macrophage cell line, Raw264.7 cells, with IFN-γ. We found that the level of HIF1A protein and the expression of known HIF1A target genes, including the critical macrophage polarization gene Nos2, are significantly increased in the IFN-γ treated Raw264.7 cells compared to the control (Figure 7E and F). We also cultured BMDMs from TPM mice and induced TPM expression in vitro. We found an increase in mRNA expression of macrophage polarization-related gene (Nos2), glycolysisrelated genes (Hk2, Pfkfb3), and also other HIF1A direct target genes (Adm) (Figure 7G and H). These genes were similarly activated by IFN-γ when treated with Raw264.7 haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

Figure 7. Genetic deletion of IFN-γ receptor cannot rescue the triple point mutation (TPM) induced hemophagocytic lymphohistiocytosis (HLH)-like phenotypes. (A-D) Ifngr–/–/Vav1-Cre/LSL/TPM (Ifngr-/-/TPM) mice were generated. Ifngr–/–/Vav-Cre/LSL/WT [Ifngr–/–/wild-type (WT)] mice served as their control. Ifngr–/–/TPM, Ifngr–/–/WT, Vav1-Cre/TPM, Vav1-Cre/WT mice were administrated with doxycycline. (A and B) Hemoglobin (Hb) (A) and platelets (PLT) (B) on day (d)8 after doxycycline induction. (C) Kaplan-Meier analysis of survival of Ifngr–/–/TPM (n=8), Ifngr–/–/WT (n=8), Vav1-Cre/TPM (n=32), Vav1-Cre/WT (n=26) mice. Statistical analysis showed that there was a significant difference between Ifngr–/–/TPM and Ifngr–/–/WT mice (P<0.0001), and between Ifngr–/–/TPM and Vav1-Cre/TPM mice (P<0.05). (D) Representative flow cytometry CD80/CD206 scatter dot plots of splenic macrophages in the indicated mice. (E) Cell lysates of IFN-γ (20 IU/mL or 100 IU/mL) as final concentration) or Cocl2-treated Raw264.7 cells were analyzed for HIF1A and β-Actin by western blot. (F) Raw264.7 cells were treated with or without IFN-γ (100 IU/mL as final concentration) for 24 hours (h). mRNA expression of the indicated genes was shown by qRT-PCR. (G) Schematic diagram showing Vav1-Cre/TPM mice-derived BMDM turning on expression of HIF1A/ARNT following addition of doxycycline in vitro (top). HIF1A protein level in doxycycline-treated or untreated Vav1Cre/TPM mice-derived BMDM measured by western blot (bottom). (H) Relative mRNA expression of the indicated genes was shown by qRT-PCR. (I) Working model for the role of HIF1A in HLH. *P<0.05, **P<0.01, ***P<0.001 versus control for all experiments.

haematologica | 2017; 102(11)

1965

R. Huang et al.

cells (Figure 7F). The evidence suggests that HIF1A signaling is the downstream of IFN-γ signaling and that activation of HIF1A signaling could activate and polarize macrophages which results in some key features of HLHlike phenotypes. Activation of HIF1A signaling in combination with activation of other pathways, such as NF-κB and STAT1, might lead to the complete complex HLH phenotypes as seen in humans (Figure 7I).

Discussion Although there are differences in etiology and pathological immune response between FHL patients and sHLH patients, they have similar clinical manifestations and the same features of hyper-inflammation and hyper-immune response.2,7 Identification of key common mediators in HLH may help to explore novel therapeutic targets that would have a wider application in HLH patients. In the present study, we identified that the HIF1A signature is enriched in both FHL and sJIA patients, and its protein level is elevated in primary and sHLH mouse models. Induction of HIF1A signaling in hematopoietic cells in vivo results in HLH-like phenotypes. This indicates that HIF1A is a critical mediator for HLH. HIF1A is reported to be involved in inflammation and immune response.10,12,32 In line with this, our bioinformatic analysis of FHL and sJIA datasets revealed that HIF1A might have a wide regulatory effect in HLH pathogenesis, which may be related to regulation of chemotaxis, cytokines, immune response, glycolysis, blood coagulation, and apoptosis, as indicated by the GO analysis. Notably, it is evident that FHL patients have a stronger signature than sJIA patients. This could be due to the independent processes of these two array datasets from both groups. There is also a possibility that the strong genetic component of FHL leads to this discrepancy. However, we did not observe a significant difference in the HIF1A signature between FHL patients with and those without a genetic diagnosis in this dataset. It is hard to completely rule out the possibility that genetic mutation drives HIF1A signature since only mutations of PRF1, UNC13D, and STX11 were tested in this FHL microarray dataset.21 Patients without a genetic diagnosis may still carry disease-causing mutations. Future studies are needed to clarify whether the strong genetic background could have additional effects on the gene signatures. Nonetheless, the stronger signature in FHL patients as compared to the sJIA patients could be the underlying difference between these two diseases, and it is possible that the involvement of distinct cell types causes this difference. T cells, NK cells, macrophages, and DCs are all involved in the FHL pathogenesis, while T cells and NK cells are less involved in the sJIA pathogenesis.33 Thus, in microarray datasets of PBMCs, FHL may show a stronger signature than sJIA. Macrophage activation/polarization is a common feature in HLH mouse models irrespective of their specific etiology. Macrophages have been reported to switch their metabolism from oxidative phosphorylation towards glycolysis upon pro-inflammatory stimuli by the upregulation of HIF1A.34,35 Indeed, we found that HIF1A was stabilized in macrophages in both the LCMV/Prf1–/– model and the CpG model. Our data are consistent with other reports that numerous stimuli, such as IFN-γ, TNF-α, CpG, and LPS, are able to increase the HIF1A protein level 1966

in macrophages.10 These cytokines and TLR ligation might co-operate to increase HIF1A protein levels in HLH. Although several studies have showed an HIF1A deficiency in myeloid cells leads to impaired inflammatory responses, the effect of activation of HIF1A in hematopoietic cells in vivo remains unclear. Here, we show that induction of HIF1A in hematopoietic cells in vivo is lethal and gives rise to some HLH-like phenotypes, such as severe anemia, thrombocytopenia, splenomegaly, liver damage, ferritinemia, and macrophage activation, suggesting that HIF1A is a critical mediator in HLH. We are also aware that HIF1A-induced phenotypes cannot recapitulate all the manifestations seen in human HLH patients. This could be due to the fact that other transcription factors which co-operate with HIF1A activation to generate the overt HLH phenotypes are required. Our analysis of the transcription factor network sheds light on other key transcription factors in HLH development which could help in future research. Defective CD8+ T cells are regarded as the driver in the primary HLH mouse model.5 However, the HLH-like phenotypes in the TPM mouse model are not dependent on lymphocytes since activation of HIF1A also causes HLH phenotypes in Rag1–/– mice. This indicates that non-lymphocytes contribute to the HLH-like phenotypes. Here, we observed that HIF1A activation leads to macrophage type 1 polarization in vivo. Our in vitro data and other reports revealed that HIF1A can up-regulate Nos2, IL-6, IL-1β, CXCR4 and, glycolysis-related genes which might account for the in vivo type 1 polarization.32 The role of type 1 polarization of macrophages for anemia in HLH is still unclear. Our in vitro data show that IFN-γ-polarized type 1 macrophages engulf erythroblasts, which is consistent with our earlier report36 that IFN-γ acts directly on macrophages resulting in hemophagocytosis, leading to a consumptive anemia in vivo. There is also evidence that hemophagocytes express type 2 polarized macrophage markers such as CD206 or CD163, and exhibit expression profiles similar to resting splenic macrophages.37,38 This discrepancy in distinct macrophage activation type may be due to a different subpopulation of macrophage, or to various different etiological scenarios. Although IFN-γpolarized type 1 macrophages phagocytose erythroblasts in our in vitro experiment setting, robust phagocytosis was not observed in the TPM mice, which suggests that type 1 polarization of macrophages induced by HIF1A transgene is not sufficient to induce hemophagocytosis in vivo.24,39,40 Other additional factors, such as blocking IL-10 signaling or involvement of DCs may be required for phagocytosis in vivo. Future investigation will be needed to verify these possibilities. IFN-γ is a potent stimulator for type 1 macrophage polarization, and plays a central role in a large proportion of HLH patients and in FHL mouse models; however, it has also been seen that it is not essential in some of the sHLH mouse models.39,41 Our study also showed that TPM-induced HLH-like phenotypes are independent of IFN-γ. In fact, our in vitro data and other reports showed that HIF1A is a downstream effector of IFN-γ, and activation of HIF1A in macrophages leads to the increase in type 1 polarization-related genes, such as Nos2, and other glycolysis-related genes. However, partial rescue of anemia was observed in TPM mice with IFN-γ deficiency. IFN-γ may also have HIF1A-independent mechanisms that affect erythropoiesis, as has been reported.42 haematologica | 2017; 102(11)

HIF1A is a mediator for HLH

In summary, our study suggests that HIF1A is a common critical downstream mediator for HLH. We propose that HIF1A activation as the consequence of systemic inflammation, cytokine storm, or ligation of TLR may contribute to HLH development. Thus, HIF1A might be a promising therapeutic target for HLH intervention. Acknowledgments The authors would like to thank Eric VIVIER who kindly provided the transgenic mice. Funding This work was supported by a Pilot Grant of The HLH Center of Excellence at Cincinnati Childrenâ&#x20AC;&#x2122;s Hospital Medical Center

References 1. Janka GE, Lehmberg K. Hemophagocytic lymphohistiocytosis: pathogenesis and treatment. Hematology Am Soc Hematol Educ Program. 2013;2013:605-611. 2. Jordan MB, Allen CE, Weitzman S, Filipovich AH, McClain KL. How I treat hemophagocytic lymphohistiocytosis. Blood. 2011;118(15):4041-4052. 3. Lehmberg K, Nichols KE, Henter JI, et al. Consensus recommendations for the diagnosis and management of hemophagocytic lymphohistiocytosis associated with malignancies. Haematologica. 2015;100(8):9971004. 4. Filipovich AH, Chandrakasan S. Pathogenesis of Hemophagocytic Lymphohistiocytosis. Hematol Oncol Clin North Am. 2015;29(5):895-902. 5. Jordan MB, Hildeman D, Kappler J, Marrack P. An animal model of hemophagocytic lymphohistiocytosis (HLH): CD8+ T cells and interferon gamma are essential for the disorder. Blood. 2004;104(3):735-743. 6. Lehmberg K, Sprekels B, Nichols KE, et al. Malignancy-associated haemophagocytic lymphohistiocytosis in children and adolescents. Br J Haematol. 2015;170(4):539-549. 7. Brisse E, Wouters CH, Matthys P. Advances in the pathogenesis of primary and secondary haemophagocytic lymphohistiocytosis: differences and similarities. Br J Haematol. 2016;174(2):203-217. 8. Emile JF, Abla O, Fraitag S, et al. Revised classification of histiocytoses and neoplasms of the macrophage-dendritic cell lineages. Blood. 2016;127(22):2672-2681. 9. Fall N, Barnes M, Thornton S, et al. Gene expression profiling of peripheral blood from patients with untreated new-onset systemic juvenile idiopathic arthritis reveals molecular heterogeneity that may predict macrophage activation syndrome. Arthritis Rheum. 2007;56(11):3793-3804. 10. Palazon A, Goldrath AW, Nizet V, Johnson RS. HIF transcription factors, inflammation, and immunity. Immunity. 2014;41(4):518528. 11. Greer SN, Metcalf JL, Wang Y, Ohh M. The updated biology of hypoxia-inducible factor. EMBO J. 2012;31(11):2448-2460. 12. Cummins EP, Keogh CE, Crean D, Taylor CT. The role of HIF in immunity and

haematologica | 2017; 102(11)

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

(to GH), and a grant from Histiocytosis Association Research Grant Program (to GH), Southern Medical University Basic Research Grant Program (No. QD2016N016 to RH), Natural Science Foundation of Guangdong Province, China (No. 2017A030310112 to RH), National Natural Science Funds of China (No. 81300392 to JW, No.81370611 to ZFX, No. 81470338 to YZ, No. 81470297, and No. 81770129, to GH, and No. 81530008, and No. 81470295 to ZJX, No. 81570173 to XL), Tianjin science and technology projects (13JCYBJC42400 to YZ). We would like to acknowledge the assistance of the Research Flow Cytometry Core in the Division of Rheumatology at Cincinnati Childrenâ&#x20AC;&#x2122;s Hospital Medical Center. All flow cytometric data were acquired using equipment maintained by the Research Flow Cytometry Core.

inflammation. Mol Aspects Med. 2016;4748:24-34. Cramer T, Yamanishi Y, Clausen BE, et al. HIF-1alpha is essential for myeloid cellmediated inflammation. Cell. 2003;112(5):645-657. Walmsley SR, Chilvers ER, Thompson AA, et al. Prolyl hydroxylase 3 (PHD3) is essential for hypoxic regulation of neutrophilic inflammation in humans and mice. J Clin Invest. 2011;121(3):1053-1063. Jantsch J, Chakravortty D, Turza N, et al. Hypoxia and hypoxia-inducible factor-1 alpha modulate lipopolysaccharide-induced dendritic cell activation and function. J Immunol. 2008;180(7):4697-4705. McNamee EN, Korns Johnson D, Homann D, Clambey ET. Hypoxia and hypoxiainducible factors as regulators of T cell development, differentiation, and function. Immunol Res. 2013;55(1-3):58-70. Imtiyaz HZ, Williams EP, Hickey MM, et al. Hypoxia-inducible factor 2alpha regulates macrophage function in mouse models of acute and tumor inflammation. J Clin Invest. 2010;120(8):2699-2714. Doedens AL, Phan AT, Stradner MH, et al. Hypoxia-inducible factors enhance the effector responses of CD8(+) T cells to persistent antigen. Nat Immunol. 2013;14(11):1173-1182. Blouin CC, Page EL, Soucy GM, Richard DE. Hypoxic gene activation by lipopolysaccharide in macrophages: implication of hypoxia-inducible factor 1alpha. Blood. 2004;103(3):1124-1130. Albina JE, Mastrofrancesco B, Vessella JA, Louis CA, Henry WL Jr, Reichner JS. HIF-1 expression in healing wounds: HIF-1alpha induction in primary inflammatory cells by TNF-alpha. Am J Physiol Cell Physiol. 2001;281(6):C1971-1977. Sumegi J, Barnes MG, Nestheide SV, et al. Gene expression profiling of peripheral blood mononuclear cells from children with active hemophagocytic lymphohistiocytosis. Blood. 2011;117(15):e151-160. Bridges JP, Lin S, Ikegami M, Shannon JM. Conditional hypoxia inducible factor1alpha induction in embryonic pulmonary epithelium impairs maturation and augments lymphangiogenesis. Dev Biol. 2012;362(1):24-41. Narni-Mancinelli E, Chaix J, Fenis A, et al. Fate mapping analysis of lymphoid cells

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

expressing the NKp46 cell surface receptor. Proc Natl Acad Sci USA. 2011; 108(45):18324-18329. Behrens EM, Canna SW, Slade K, et al. Repeated TLR9 stimulation results in macrophage activation syndrome-like disease in mice. J Clin Invest. 2011; 121(6):2264-2277. Das R, Guan P, Sprague L, et al. Janus kinase inhibition lessens inflammation and ameliorates disease in murine models of hemophagocytic lymphohistiocytosis. Blood. 2016;127(13):1666-1675. Grom AA, Villanueva J, Lee S, Goldmuntz EA, Passo MH, Filipovich A. Natural killer cell dysfunction in patients with systemiconset juvenile rheumatoid arthritis and macrophage activation syndrome. J Pediatr. 2003;142(3):292-296. Ravelli A, Grom AA, Behrens EM, Cron RQ. Macrophage activation syndrome as part of systemic juvenile idiopathic arthritis: diagnosis, genetics, pathophysiology and treatment. Genes Immun. 2012;13(4):289-298. Villanueva J, Lee S, Giannini EH, et al. Natural killer cell dysfunction is a distinguishing feature of systemic onset juvenile rheumatoid arthritis and macrophage activation syndrome. Arthritis Res Ther. 2005;7(1):R30-37. Emig D, Salomonis N, Baumbach J, Lengauer T, Conklin BR, Albrecht M. AltAnalyze and DomainGraph: analyzing and visualizing exon expression data. Nucleic Acids Res. 2010;38(Web Server issue):W755-762. Chow A, Lucas D, Hidalgo A, et al. Bone marrow CD169+ macrophages promote the retention of hematopoietic stem and progenitor cells in the mesenchymal stem cell niche. J Exp Med. 2011;208(2):261-271. Balsamo M, Manzini C, Pietra G, et al. Hypoxia downregulates the expression of activating receptors involved in NK-cellmediated target cell killing without affecting ADCC. Eur J Immunol. 2013; 43(10):2756-2764. Lin N, Simon MC. Hypoxia-inducible factors: key regulators of myeloid cells during inflammation. J Clin Invest. 2016; 126(10):3661-3671. Mellins ED, Macaubas C, Grom AA. Pathogenesis of systemic juvenile idiopathic arthritis: some answers, more questions.

1967

R. Huang et al. Nat Rev Rheumatol. 2011;7(7):416-426. 34. Gordan JD, Thompson CB, Simon MC. HIF and c-Myc: sibling rivals for control of cancer cell metabolism and proliferation. Cancer Cell. 2007;12(2):108-113. 35. Liu L, Lu Y, Martinez J, et al. Proinflammatory signal suppresses proliferation and shifts macrophage metabolism from Myc-dependent to HIF1alpha-dependent. Proc Natl Acad Sci USA. 2016; 113(6):1564-1569. 36. Zoller EE, Lykens JE, Terrell CE, et al. Hemophagocytosis causes a consumptive anemia of inflammation. J Exp Med. 2011; 208(6):1203-1214. 37. Canna SW, Costa-Reis P, Bernal WE, et al.

1968

Brief report: alternative activation of lasercaptured murine hemophagocytes. Arthritis Rheum. 2014;66(6):1666-1671. 38. McCoy MW, Moreland SM, Detweiler CS. Hemophagocytic macrophages in murine typhoid fever have an anti-inflammatory phenotype. Infect Immun. 2012; 80(10):3642-3649. 39. Canna SW, Wrobel J, Chu N, Kreiger PA, Paessler M, Behrens EM. Interferon-gamma mediates anemia but is dispensable for fulminant toll-like receptor 9-induced macrophage activation syndrome and hemophagocytosis in mice. Arthritis Rheum. 2013;65(7):1764-1775. 40. Ohyagi H, Onai N, Sato T, et al. Monocyte-

derived dendritic cells perform hemophagocytosis to fine-tune excessive immune responses. Immunity. 2013; 39(3):584-598. 41. Avau A, Mitera T, Put S, et al. Systemic juvenile idiopathic arthritis-like syndrome in mice following stimulation of the immune system with Freund's complete adjuvant: regulation by interferon-gamma. Arthritis Rheumatol. 2014;66(5):13401351. 42. Lin FC, Karwan M, Saleh B, et al. IFNgamma causes aplastic anemia by altering hematopoietic stem/progenitor cell composition and disrupting lineage differentiation. Blood. 2014;124(25):3699-3708.

haematologica | 2017; 102(11)