Haematologica, volume 102, issue 7 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 Giugno 2017.

haematologica calendar of events

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

2nd EBMT International Transplant Course The European Group for Blood and Marrow Transplantation (EBMT) Chairs: M Mohty, J Kuball, R Duarte September 8-10, 2017 Barcelona, Spain

13 Educational Course of the Lymphoma Working Party on "Treatment of Malignant Lymphoma: State-of-the-Art and Role of Stem Cell Transplantation" The European Group for Blood and Marrow Transplantation (EBMT) Chairs: S Montoto, A Sureda, M Trneny September 21-22, 2017 Prague, Czech Republic th

EHA Scientific Meeting on Challenges in the Diagnosis and Management of Myeloproliferative Neoplasms Chairs: J Kiladjian and C Harrison October 12-14, 2017 Budapest, Hungary

EHA Tutorial on Biology and Management of Myeloid Malignancies October 20-22, 2017 Yerevan, Armenia

Russian Onco-Hematology Society's Conference on Malignant Lymphoma - Joint Symposium October 25-26, 2017 Moscow, Russian Federation

Turkish Society of Hematology - EHA Joint Symposium November 1 - 4, 2017 Antalya, Turkey

Argentinian Society of Hematology - EHA Joint Education Day November 17 - 18, 2017 Mar del Plata, Argentina

EHA Scientific Meeting on Shaping the Future of Mesenchymal Stromal Cells Therapy Chair: W Fibbe November 23-25, 2017 Amsterdam, The Netherlands

Calendar of Events updated on June 1, 2017

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

Table of Contents Volume 102, Issue 7: July 2017 Cover Figure Image generated by www.somersault1824.com.

Editorials 1131

Chronic myeloid leukemia: room for improvement? Michele Baccarani et al.

1133

Minimal residual disease in mantle cell lymphoma: are we ready for a personalized treatment approach? Simone Ferrero and Martin Dreyling

1137

Antithymocyte globulin and transplants for aplastic anemia Andrea Bacigalupo

Review Article 1139

Genetic background and evolution of relapses in aggressive B-cell lymphomas Darius Juskevicius et al.

Articles Hematopoiesis

1150

A new path to platelet production through matrix sensing Vittorio Abbonante et al.

Red Cell Biology & Its Disorders

1161

The endothelin B receptor plays a crucial role in the adhesion of neutrophils to the endothelium in sickle cell disease BĂŠrengĂ¨re Koehl et al.

Iron Metabolism & Its Disorders

1173

Imatinib and spironolactone suppress hepcidin expression Katarzyna Mleczko-Sanecka et al.

Coagulation & Its Disorders

1185

Venous thromboembolism is associated with graft-versus-host disease and increased non-relapse mortality after allogeneic hematopoietic stem cell transplantation Natasha Kekre et al.

Platelet Biology & Its Disorders

1192

Bleeding risk of surgery and its prevention in patients with inherited platelet disorders Sara Orsini et al.

Acute Myeloid Leukemia

1204

Epigenetically induced ectopic expression of UNCX impairs the proliferation and differentiation of myeloid cells Giulia Daniele et al.

Haematologica 2017; vol. 102 no. 7 - July 2017 http://www.haematologica.org/

haematologica Journal of the European Hematology Association Published by the Ferrata Storti Foundation 1215

A three-dimensional ex vivo tri-culture model mimics cell-cell interactions between acute myeloid leukemia and the vascular niche Laura J. Bray et al.

1227

Precision and prognostic value of clone-specific minimal residual disease in acute myeloid leukemia Pierre Hirsch et al.

Chronic Lymphocytic Leukemia

1238

Host virus and pneumococcus-specific immune responses in high-count monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia: implications for disease progression Ignacio Criado et al.

Non-Hodgkin Lymphoma

1247

Inhibition of monocarboxyate transporter 1 by AZD3965 as a novel therapeutic approach for diffuse large B-cell lymphoma and Burkitt lymphoma Richard A. Noble et al.

1258

Plasma Cell Disorders

1266

Monitoring multiple myeloma by quantification of recurrent mutations in serum Even Holth Rustad et al.

1273

Resistin induces multidrug resistance in myeloma by inhibiting cell death and upregulating ABC transporter expression Jianan Pang et al.

1281

Novel recurrent chromosomal aberrations detected in clonal plasma cells of light chain amyloidosis patients show potential adverse prognostic effect: first results from a genome-wide copy number array analysis Martin Granzow et al.

Cell Therapy & Immunotherapy

1291

Effect of antithymocyte globulin source on outcomes of bone marrow transplantation for severe aplastic anemia Natasha Kekre et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/102/7.toc

e241

Enhanced calreticulin expression in red cells of polycythemia vera patients harboring the JAK2V617F mutation MĂŠgane Brusson et al. http://www.haematologica.org/content/102/7/e241

e245

An extracellular matrix signature in leukemia precursor cells and acute myeloid leukemia Valerio Izzi et al. http://www.haematologica.org/content/102/7/e245

e249

NOTCH1 mutation, TP53 alteration and myeloid antigen expression predict outcome heterogeneity in children with first relapse of T-cell acute lymphoblastic leukemia Jana Hof et al. http://www.haematologica.org/content/102/7/e249

e253

Role of miR-15a/miR-16-1 and the TP53 axis in regulating telomerase expression in chronic lymphocytic leukemia Enrica Rampazzo et al. http://www.haematologica.org/content/102/7/e253

Haematologica 2017; vol. 102 no. 7 - July 2017 http://www.haematologica.org/

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

e257

Sphingosine kinase 1 participates in the activation, proliferation and survival of chronic lymphocytic leukemia cells María Belén Almejún et al. http://www.haematologica.org/content/102/7/e257

e261

Double autophagy stimulation using chemotherapy and mTOR inhibition combined with hydroxychloroquine for autophagy modulation in patients with relapsed or refractory multiple myeloma Emma C. Scott et al. http://www.haematologica.org/content/102/7/e261

e266

Both mucosal-associated invariant and natural killer T-cell deficiency in multiple myeloma can be countered by PD-1 inhibition Mérédis Favreau et al. http://www.haematologica.org/content/102/7/e266

e271

Salvage use of allogeneic hematopoietic stem cell transplantation after reduced intensity conditioning from unrelated donors in multiple myeloma. A study by the Plasma Cell Disorders subcommittee of the European Group for Blood and Marrow Transplant Chronic Malignancies Working Party Mohamad Sobh et al. http://www.haematologica.org/content/102/7/e271

Case Reports Case Reports are available online only at www.haematologica.org/content/102/7.toc

e275

Combination of ibrutinib and chemotherapy produced a durable remission in multiply relapsed diffuse large B-cell lymphoma leg type with mutant MYD88 and wildtype CD79 Andrew L. Deng et al. http://www.haematologica.org/content/102/7/e275

e278

Mutations in the adaptor-binding domain and associated linker region of p110δ cause Activated PI3K-δ Syndrome 1 (APDS1) Lucie Heurtier et al. http://www.haematologica.org/content/102/7/e278

Haematologica 2017; vol. 102 no. 7 - July 2017 http://www.haematologica.org/

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

The origin of a name that reflects Europe’s cultural roots.

Ancient Greek

aÂma [haima] = blood a·matow [haimatos] = of blood lÒgow [logos]= reasoning

Scientific Latin

haematologicus (adjective) = related to blood

Scientific Latin

haematologica (adjective, plural and neuter, used as a noun) = hematological subjects

Modern English

The oldest hematology journal, publishing the newest research results. 2015 JCR impact factor = 6.671

Haematologica, as the journal of the European Hematology Association (EHA), aims not only to serve the scientific community, but also to promote European cultural identify.

EDITORIALS Chronic myeloid leukemia: room for improvement? Michele Baccarani,1 Fabrizio Pane,2 Gianantonio Rosti,3 Domenico Russo4 and Giuseppe Saglio5 1

University of Bologna, and GIMEMA CML Working Party; 2University of Naples Federico II; 3Institute of Hematology "L. and A. Seràgnoli", S.Orsola-Malpighi University Hospital, University of Bologna; 4Chair of Hematology and Unit of Blood diseases and Bone Marrow Transplantation, Department of Clinical and Experimental Sciences, University of Brescia and Spedali Civili Brescia and 5 University of Torino, Italy E-mail: michele.baccarani@unibo.it

doi:10.3324/haematol.2017.166280

ollowing the introduction of imatinib and several other tyrosine kinase inhibitors (TKIs), the clinical scenario of Philadelphia chromosome-positive (Ph+) and BCRABL1-positive (BCR-ABL1+) chronic myeloid leukemia (CML) has changed almost completely.1 Although the number of studies is limited, and the follow up is short and sometimes defective, survival data are substantial; around 90% at 5 years, 89% at 6 years, 86% at 8 years, and 83-84% at 10 years2-10 (Table 1). Only 50% of deaths are due to the progression of leukemia, while 50% occur in remission and are due to other causes which occasionally include treatment-related toxicity and related complications. Relative survival analyses have concluded that the life expectancy of a patient with chronic phase (CP) CML, under proper TKI treatment, is now very close or almost identical to the life expectancy of nonleukemic, age-matched individuals.11,12 New drugs, both TKIs and non-TKIs, are emerging. New, more sophisticated molecular technologies are available. So, which problems remain to

be solved? Is there still room for improvement, and where? There is still room for improvement for patients with newly diagnosed accelerated phase (AP) and blastic phase (BP) CML, and to some extent also for patients with high-risk CP CML. The former account for 4% to 5% of all newly diagnosed patients.13 They respond to the TKI treatment, but the extent of response is inferior, and their ultimate outcome is still unclear.1 High-risk patients account for 10% to 25% of newly diagnosed CP CML patients, depending on which risk score is used, but even using Sokal, which is less selective and includes many more patients than Euro, European Treatment Outcome Study (EUTOS), and the new EUTOS long-term survival score,14 the outcome of these patients is inferior, with a reported survival of 83%-89% at 5-6 years and of 68% at 10 years7,10 (Table 1). In addition, the presence of clonal chromosome abnormalities in Ph+ cells (CCA/Ph+) at baseline, which occurs in 3% to 4% of patients, is a marker of an inferior outcome.1 It is believed that all these patients (AP, BP,

Table 1. A summary of the outcome of treatment with TKIs of newly diagnosed CP CML patients. Only the studies reporting on more than 200 patients, with a median follow-up observation of longer than 5 years, are listed. Notice the important differences in age and in the proportion of high-risk patients. MD Anderson, German CML IV and GIMEMA are academic studies. The Swedish data are taken from a population-based registry. IRIS, ENESTnd and DASISION are company-sponsored, registrative studies

Study

IRIS2,10

MDA6

GIMEMA5

ENESTnd8

(B)

German CML IV4,7 Ima (C)

First-line treatment

Ima(A)

No. pts Follow up, median Follow up, missing % Age, median High-risk % Progressions % Deaths, total % Deaths, leukemia % Deaths, other causes % OS 5 years % OS 6 years % OS 8 years % OS 10 years % OS low-risk % (at) OS high-risk % (at)

DASISION9

ENESTnd8

DASISION9

SWEDEN3

Ima (D)

Ima 400 OD

Nil 300 OD

Das 100 OD

(E)

553 10.9 y 20.1 50 y 18 Sokal 6.5 16.1 9.0 7.0 89 88 NR 83 90 (10y)

483 8.3 y NR (F) 7 Sokal NR 11.0 NR NR 93 NR NR 83 NR

1536 7.1 y 0.1 53 y 12 EUTOS 6.7 12.0 NR NR 90 NR 86 84 NR

559 6.3 y 4.1 52 y 22 Sokal 5.7 11.6 5.7 5.9 NR 89 NA NA 94 (6y)

282 5.5 y NR 46 y 28 Sokal 7.4 7.8 5.7 2.1 92 NR NA NA 100 (5y)

260 5.5 y NR 49 y 19 EURO 7.3 10.0 6.5 3.5 90 NR NA NA NR

282 5.5 y NR 47 y 28 Sokal 3.5 6.4 2.1 4.2 94 NR NA NA 97 (5y)

259 5.5 y NR 46 y 19 EURO 4.6 10.0 3.5 6.6 91 NR NA NA NR

717 5.1 y 17.8 60 y 32 Sokal NR NR NR NR 83 NA NA NA 95 (5y)

69 (10y)

83 (6y)

84 (5y)

89 (5y)

78 (5y)

(A) 400 mg once daily (OD). (B) Imatinib 400 mg OD (14% of patients), imatinib 400 mg twice daily (TD) (9%), imatinib 400 mg OD + pegylated interferon-α (33%), nilotinib (dose not specified) 21%, dasatinib (dose not specified) 22%. (C) imatinib 400 mg OD (26.4% of patients), imatinib 400 mg OD + interferon-α (28.2%), imatinib 400 mg OD + low dose cytarabine (10.3%), imatinib 400 mg OD after interferon-α (8.3%), imatinib 400 mg TD (27.3%). (D) imatinib 400 mg OD (76% of patients), imatinib 400 mg TD (24%). (E) No details, but TKIs for most patients. "A small and decreasing proportion of elderly patients received first-line treatment with the intention of palliation (i.e., not including TKIs or upfront HSCT)".3 (F) 40.8% of patients were less than 45 years old, 46.0% were 45 to 64 years old, 13.2% were ≥ 65 years old, no patient was ≥ 80 years old. NR: not reported; NA: not available; No.: number; pts: patients; OS: overall survival.

haematologica | 2017; 102(7)

1131

Editorials

high-risk CP, CCA/Ph+) can benefit more from second- or third-generation TKIs, and that some of them, in particular those of younger age, are candidates for allogeneic stem cell transplantation. However, much needed prospective, specifically addressed trials are lacking. But is there still room for improvement for the 80% to 90% of patients who become optimal responders and have a normal life expectancy? TKIs cannot prolong life, but their proper use can help to improve the quality of life without precluding the achievement of a remission that remains stable even after treatment discontinuation (treatment-free remission, TFR). Concerning the quality of life, many studies report that the side effects or toxic effects of TKIs were “manageable”, could be tolerated, and did not impel a change of TKI. Typically, such studies were designed to limit the switch from one TKI to another, and were analyzed more to define the tolerability profile of a specific TKI than to assess the quality of life of the patients. We suggest that in the case of so-called “manageable” side effects, mild or recurrent, that impair the daily life of a patient, more consideration should be given to a change of the drug and also to a change of the dose, provided that an optimal response is maintained. At times a patient-adapted policy, that is a policy adapted to side effects and to response, can be more convenient. For that purpose, trials may not be necessary, but a careful longterm observation will be important to control overall how patients adapt to side effects and to pick up on the socalled unexpected adverse events. TFR is an acronym that fully expresses the success of therapy, combining the perception of cure, a normal life expectancy, no treatment-related side effects and complications, independence from drugs, in addition to a proper use of the financial resources of health systems that are more and more challenged by the introduction of new and effective, but also expensive drugs.15-17 It is expected that a more widespread use of second-generation TKIs in firstline treatment would result in increased TFR. However, evidence is still missing as there are no studies reporting on the benefit (the TFR rate) and the cost (toxicity) of a policy of imatinib use vs. a policy of the use of second-generation TKIs, either in early or late first- or second-line therapy.17 Such studies require time, patience and resources, but are necessary in order to move from expectation to evidence. Gruppo Italiano Malattie EMatologiche dell'Adulto (GIMEMA) and the Haemato Oncology Foundation for Adults in the Netherlands (HOVON) are currently running a trial of first-line nilotinib vs. first-line imatinib therapy with a switch to nilotinib in the event of less than optimal response, with the TFR rate at 5 years being the primary endpoint.18 More than five years will be required to obtain answers from this study. Furthermore, the choice between continuing treatment indefinitely and no treatment at all may be challenged and alternative policies are currently being tested, either of partial, intermittent treatment,19,20 or of graded discontinuation.21 For the time being we must acknowledge that any recommendation concerning the policy of treatment regarding TFR is not yet evidence-based. In all likelihood, the best policy does not exist, and different policies may be successful in different situations. Notwithstanding the fact that the treatment and the 1132

modalities of treatment for TFR are still an issue for research, it should not be overlooked that treatment discontinuation and TFR are already a reality in practice, and are the goal of more and more patients. There are solid data showing that treatment discontinuation after five or more years of TKIs, and after one or more years of deep molecular response (MR), such as MR 4.0 (BCR-ABL1 ≤ 0.01% on the international scale), and particularly MR 4.5 (BCR-ABL1 ≤ 0.0032%, on the international scale) results in a rate of TFR of 50% or more, and that in the event of molecular relapse the resumption of treatment brings all patients back to molecular remission.15,16,22,23 In summary, once a deep MR is achieved, and provided that careful molecular monitoring is assured, no patient will die of leukemia because of treatment discontinuation, and 50% will enjoy a treatment-free life. Therefore, we suggest that both the possibilities and the problems of treatment discontinuation should be discussed not only with all the patients who fit the current, provisional eligibility criteria for discontinuation, but also with newly diagnosed patients.

References 1. Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia. Blood. 2013;2013;122(6):872-874. 2. Druker BJ, Guilhot F, O'Brien SG, et al. Five-year follow-up of patients receiving imatinib for chronic myeloid leukemia. N Engl J Med. 2006;355(23):2408-2417. 3. 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):1284-1292. 4. Hehlmann R, Muller MC, Lauseker M, et al. Deep molecular response is reached by the majority of patients treated with imatinib, predicts survival, and is achieved more quickly by optimized high-dose imatinib: results from the randomized CML-Study IV. J Clin Oncol. 2014;32(5):415-423. 5. Castagnetti F, Gugliotta G, Breccia M, et al. Long-term outcome of chronic myeloid leukemia patients treated frontline with imatinib. Leukemia. 2015;29:1823-1831. 6. Jain P, Kantarjian H, Alettar ML, et al. Long-term 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 patients data from five clinical trials. Lancet Haematol. 2015;2(3):e118-e128. 7. Kalmanti L, Saussele S, Lauseker M, et al. Safety and efficacy of imatinib in CML over a period of 10 years: data from the randomized CML-Study IV. Leukemia. 2015;29(5):1123-1132. 8. Hochhaus A, Saglio G, Hughes T, 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. 9. 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):23332340. 10. Hochhaus A, Larson RA, Guilhot F, et al. IRIS final analysis: long-term outcomes with imatinib treatment for CML. N Engl J Med. 2017, submitted. 11. Sasaki K, Strom SS, O'Brien S, et al. Relative survival in patients with chronic phase chronic myeloid leukaemia in the tyrosine-kinase inhibitor era: analysis of patients data from six prospective clinical trials. Lancet Haematol. 2015;2(5):e186-e193. 12. Bower H, Bjorkholm M, Dickman PW, Hoglund M, Lambert PC, Andersson T ML. Life expectancy of patients with chronic myeloid leukemia approaches the life expectancy of the general population. J Clin Oncol. 2016;34(24):2851-2857. 13. Hoffmann VS, Baccarani M, Hasford J, et al. The EUTOS populationbased registry: incidence and clinical characteristics of 2904 CML patients in 20 European countries. Leukemia. 2015;29(6):1336-1343. 14. Pfirrmann M, Baccarani M, Saussele S, et al. Prognosis of long-term sur-

haematologica | 2017; 102(7)

Editorials

15. 16. 17. 18.

19.

vival considering disease-specific death in patients with chronic myeloid leukemia. Leukemia. 2016;30(1):48-56. Hughes TP, Ross DM. Moving treatment-free remission into mainstream clinical practice in CML. Blood. 2016;128(1):17-23. Saussele S, Richter J, Hochhaus A, Mahon F-X. The concept of treatment-free remission in chronic myeloid leukemia. Leukemia. 2016;30(8):1638-1647. Baccarani M. Treatment-free remission in chronic myeloid leukemia: floating between expectation and evidence. Leukemia. 2017;31(4):10151016. SUSTRENIM (Sustained treatment-free remission in BCR-ABL+ chronic myeloid leukemia: a prospective study comparing nilotinib versus imatinib with switch to nilotinib in absence of optimal response). ClinicalTrials.gov: NCT02602314. Russo D, Martinelli G, Malagola M, et al. Effects and outcome of a policy of intermittent imatinib treatment in elderly patients with chronic

myeloid leukemia. Blood. 2013;121(26):5138-5144. 20. OPTkIMA (Phase-III randomized study to optimize TKIs multiple approaches and quality of life in elderly patients with Ph+ chronic myeloid leukemia and MR 3.0 / MR 4.0 stable molecular response). ClinicalTrials.gov:NCT02326311. 21. Clark R, Polydoros F, Apperley JF, et al. Chronic myeloid leukaemia patients with stable molecular responses (at least MR3) may safely decrease the dose of their tyrosine kinase inhibitor: data from the British Destiny study. Blood. 2016;128(22):938. 22. Mahon F-X, Richter J, Guilhot J, et al. Cessation of tyrosine kinase inhibitors treatment in chronic myeloid leukemia patients with deep molecular response: results of the Euro-Ski trial. Blood. 2016;128 (22):787. 23. Etienne G, Guilhot J, Rea D, et al. Long-term follow-up of the French Stop Imatinib (STIM1) study in patients with chronic myeloid leukemia. J Clin Oncol. 2017;35(3):298-305.

Minimal residual disease in mantle cell lymphoma: are we ready for a personalized treatment approach? Simone Ferrero1 and Martin Dreyling;2 on behalf of the European Mantle Cell Lymphoma Network 1

Division of Hematology, Department of Molecular Biotechnologies and Health Sciences, University of Torino, Italy and Department of Medicine III, Hospital of the University LMU München, Germany

E-mail: martin.dreyling@med.uni-muenchen.de

doi:10.3324/haematol.2017.167627

antle cell lymphoma (MCL) is nowadays recognized as a spectrum of diseases, characterized by significantly different treatment responses and outcomes. Some predictors of clinical and biological outcome have been established and validated over recent years, and these are either assessable at baseline (mainly MCL international prognostic indexes, Ki-67 proliferative index and genomic aberrations) or during treatment (functional imaging and minimal residual disease, MRD). MRD is defined as the minimal traceable persistence of lymphoma cells after a successful treatment. Many methods to monitor MRD have been published; however, the most sensitive and the most commonly used and best standardized approach in MCL is represented by the allele-specific oligonucleotide (ASO) quantitative polymerase chain reaction (qPCR) method.1 The most relevant prospective trials investigating the impact of MRD on MCL patient outcome are listed in Table 1A. The clinical role of MRD analysis in MCL is reflected according to four major aspects (Figure 1). MRD provides early feedback on the efficacy of the clearance of different induction regimens. The dynamics and stability of tumor shrinkage after treatment can currently be precisely tracked by MRD kinetics; these data might be useful as an early in vivo predictor of the anti-lymphoma effect of a new compound.6,9,10 Moreover, MRD can be used as a surrogate end point for progression-free survival (PFS) comparing the efficacy of different treatments in randomized trials, thus accelerating the development, and eventually the approval, of new drugs. For example, the superior outcome of the cytarabine-containing experimental arm of the “MCL Younger” phase III trial of the European MCL Network was heralded by a higher rate of MRD clearance many years before publication of the final results.5,15 MRD can provide an early prediction of disease recurrence. Even in the context of an incurable disease like MCL, the deepness of treatment response measured by MRD widehaematologica | 2017; 102(7)

ly reflects patient outcome in large, prospective trials.3-5 The predictive role of MRD analysis in MCL was confirmed in different patient subsets (both younger and elderly), treatment strategies (autologous transplantation and conventional immuno-chemotherapy), tissues (bone marrow and peripheral blood) and time points (end of induction and during maintenance treatment).4 MRD allows for risk stratification of patients after treatment. MRD describes the efficacy of therapy and presence of even minimal, resistant tumor clones; thus, this approach identifies patients at higher risk of recurrence after an apparently successful treatment. Actually, persistence of MRD positivity or recurrence after a transient MRD negativity precedes clinical relapse, with a median time lag of 18 months.16 MRD might drive pre-emptive treatment. As MRD positivity predicts upcoming clinical relapse, this approach can guide treatment tailoring, with the aim of preventing or delaying overt disease progression. In a number of prospective reports, a pre-emptive rituximab treatment of MRD positive patients was able to reconvert them to MRD negativity, with the possibility of also prolonging their PFS.8,17 Nevertheless, some limitations still hamper the widespread use of MRD analysis in clinical routine. The two major obstacles as far as methodology is concerned are the need for patient-specific primers and standardization issues. At present, the ASO-qPCR strategy relies upon either the clonal rearrangement of the IGH gene or the BCL-1/IGH rearrangement, derived from the t(11;14); both of these DNA sequences are unique for each B-cell clone, so individual primers are required for each patient to guarantee a reliable sensitivity. Thus a “one-fits-all” easy-to-use in vitro diagnostic medical device (IVD-kit) is not conceivable in MRD diagnostics so far, and access to an experienced and dedicated laboratory is mandatory. In addition, a rigorous standardization of the methods is essential in order to provide comparable results among different centers. Only selected laboratories across Europe 1133

Editorials

(http://www.euromrd.org).18 Only by standardized analyses in this laboratory network can reliable MRD results be obtained in large, international clinical trials, such as the

have so far been certified by the Euro MRD consortium, a standardization group regularly performing quality control rounds for MRD analysis in leukemia and lymphoma

Table 1A. Prospective clinical trials investigating the impact of minimal residual disease (MRD) on patients' outcome in mantle cell lymphoma (MCL) patients.

Study

Study population

Study features

Marker, method

Pott, 20062

MCL, <70 y and GLSG

Local protocol

IgH, qPCR

29 79

MCL <66 y

Prospective trial phase II, non-randomized Phase III, randomized

BCL1, IgH, N-PCR

Geisler, 20083

BCL1, IgH, qPCR

190

Pott, 20104

MCL, younger and elderly

Evaluable Sample analyzed patients

Study treatment

BM, PB, harvest

R-HDS +TBI + ASCT PB, BM R-maxiCHOP/RHD-AraC + ASCT + in vivo purging BM, PB, harvest R-CHOP +TBI + ASCT vs. R-CHOP/R-DHAP + R-HD-AraC +TBI + ASCT (younger); R-FC vs. R-CHOP (elderly) PB, BM R-CHOP +TBI + HD CTX + ASCT vs. R-CHOP/R-DHAP + AraC +TBI + Mel + ASCT

Hermine, 20165

MCL, <65 y

Phase III, randomized

Unspecified, qPCR

497

Albertsson-Lindblad, 20176 Liu, 20127

MCL, elderly

Phase I-II

BCL1, IgH, N-PCR

PB, BM

MCL, <69 y

Phase II, non-randomized

BCL1, IgH, qPCR

PB, BM

Kolstad, 20168

MCL, <66 y

Phase II, non-randomized

BCL1, IgH, N-PCR

183

Visco, 20169

MCL, elderly

Phase II

BCL1, IgH, N-PCR

PB, BM

R-BAC500

Zaja, 201710

MCL, >18 y

Phase II, BCL1, IgH, N-PCR, qPCR non-randomized

PB, BM

Phase II, non-randomized Phase III, randomized

IgH, NGS

PB, plasma

R2B + R2 consolidation + Len maintenance BR + R-HD-ARA-C

IgH, qPCR

178

PB, BM

Armand, 201611 Callanan, 201512

MCL, transplant eligible (42-69 y) MCL, <66 y

Gressin, 201413

MCL, elderly

Kaplan, 201514

MCL, <70 y

Phase II, IgH, qPCR non-randomized Phase II, BCL1, IgH, unspecified randomized

MRD evaluation or MRD impact on outcome

76 151

Len-BR x6 + Len maintenance R-HD-MTX + maxi-CHOP + ASCT + R maintenance R-maxiCHOP + ASCT ± RIT

PFS 92 vs. 21 m (P<0.001) PFS NR vs. 18 m (P<0.001) PFS 77% vs. 34% at 2 y FU (P=0.021)

EOI MRD neg: 47% vs. 79% (PB), 26% vs. 61% (BM) EOI MRD neg: 32% TTP at 3 y 82% vs. 48% (MRD at EOI) PFS 20 m vs. 142 m (MRD post-ASCT); median OS NR vs. 35 m No association between MRD status and PFS 36% of MRD negativization, predictive of PFS 93% MRD neg at EOT

R-DHAP +R-BEAM-ASCT ± R maintenance

MRD status pre-ASCT in BM and PB predicts longer PFS (P=0.0451, P=0.0016) PB, BM RiBVD MRD neg: 83% (PB) and 74% (BM) at 6 months Not available augCHOP+MTX+EAR+ 5-y PFS: 93% if MRD-neg CBV-ASCT+ bortezomib vs. 51% if MRD-pos maintenance vs. consolidation following induction therapy

Table 1B. Current MRD-driven trials in MCL.

Clinicaltrials.gov identifier

Short title characteristics

Patients'

Study features (n. of patients)

Estimated enrollment

Study treatment

Primary outcome

02354313

FIL_MCL0208

MCL , <65 y

Phase III, randomized

300

PFS

02896582

LyMa 101

MCL, <65 y

EA4151 ECOG group

MCL, transplant eligible

Phase II, single-arm Phase III, randomized

R-CHOP + HD-CTX + HDAra-C + BEAM and ASCT ± lenalidomide maintenance GA-DHAP + GA-BEAM + ASCT + GA maintenance induction therapy ± ASCT + R maintenance

Not available

689

MRD negativity at end of induction OS

MCL: mantle cell lymphoma; R-CHOP: rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone; TBI: total body irradiation; ASCT: autologous stem cell transplant; Mel: melphalan; EOI: end of induction; B-R: bendamustine, rituximab; Len: lenalidomide; R-DHAP: rituximab, cytarabine, dexamethasone, cisplatin; R-BAC500: rituximab, bendamustine, cytarabine; EOT: end of treatment; GLSG: German Lymphoma Study Group; HD-CTX: high-dose cyclophosphamide; HD-Ara-C: high-dose cytarabine; RIC: reduced intensity conditioning; HD-CHT: high-dose chemotherapy; RIT: radio-immunotherapy; RiBVD: rituximab, bendamustine bortezomib and dexamethasone; y: years; m: months; BM: bone marrow; PB: peripheral blood; vs.: versus; neg.: negative; pos: positive; PFS: progression-free survival; OS: overall survival.

1134

haematologica | 2017; 102(7)

Editorials

Figure 1. The clinical role of minimal residual disease analysis in mantle cell lymphoma.

recently launched TRIANGLE trial (EudraCT n. 2014001363-12), sponsored by the European MCL Network. Moreover, another limitation of the current technique is that 10-15% of patients still lack a reliable molecular marker for MRD. For the moment, in MCL, an IGH-based marker is available in approximately 70% of cases and a BCL-1/IGH marker in approximately 35-40%, with some overlapping cases.4 In particular, patients with low or absent bone marrow invasion often do not carry a marker and cannot, therefore, be analyzed for MRD. In addition, hypermutated IGH genes may hamper an optimal primer design. Finally, although the predictive role of MRD has been established in MCL,3-5 evidence for the usefulness of subsequent treatment tailoring based on the MRD results is unfortunately still scarce due to the lack of MRD-driven phase III trials in MCL.8,17 Moreover, no MRD data are available yet in the context of the new targeted treatments, such as the Bruton’s tyrosine kinase inhibitor ibrutinib. However, many technical innovations have recently been introduced in the MRD field, and these have the potential to overcome the issues of applicability and sensitivity described above. The droplet digital PCR (ddPCR), a 3rdgeneration, end point, quantitative PCR has been shown to provide comparable results to ASO-qPCR for MRD monitoring in MCL with the advantage that it is less labor intensive. Moreover, since it does not require a standard curve for tumor quantification, ddPCR might provide reliable and sensitive MRD results also in cases in which the classical approach has failed.19 Currently, a totally innovative haematologica | 2017; 102(7)

approach is represented by the application of next-generation sequencing (NGS) techniques to the MRD field. The LymphoSIGHTTM approach was first published for MRD detection in acute lymphoblastic leukemia,20 and was subsequently shown to be feasible also in MCL.21 Its main advantages rely on the fact that it does not require patientspecific reagents (being thus suitable for an IVD-kit), it can provide additional MRD targets for patients lacking a “classical” molecular marker, it should easily reach high sensitivity levels, and might overcome some false-negative results (e.g. deciphering the clonal evolution issues). However, until now, this promising NGS technology has been available as a commercial tool only in US. Nevertheless, many laboratories are currently implementing alternative NGSbased approaches for MRD and an international development and standardization effort is ongoing within the EuroClonality-NGS laboratory consortium (http://www.euroclonality.org/wp-content/uploads/2015/03/EuroClonalityNGS.pdf).22 Moreover, further promising NGS-based approaches for the identification of new molecular markers are being studied and might effectively provide an MRD target for each patient in the near future (Targeted Locus Amplification and Rapid Capture techniques).23,24 Finally, MRD targeting on plasmatic, circulating tumor DNA is extremely promising as a means to track lymphoma clones residing outside the peripheral blood or bone marrow compartments.25 Despite all of these encouraging data, a large-scale validation of each of these new technologies is required before their introduction into clinical practice. 1135

Editorials

Nowadays, MRD evaluation is mostly based on ASOqPCR in the most important ongoing European clinical trials and, as explorative investigation, as part of clinical routine in a few selected centers. However, for the moment, the published diagnostic and treatment guidelines discourage any clinical decision-making based on MRD results,26 mainly due to limitations of technical standardization and to the lack of prospective, randomized trials evaluating modified therapy according to MRD. Reliable and reproducible MRD data can currently be guaranteed only by the standardized methodological guidelines of the Euro MRD laboratory network. Therapeutic decisions based on MRD results obtained by different, as yet not validated techniques may even compromise the patient’s long-term outcome. Moreover, convincing data on the real usefulness of such an MRD-based treatment modulation are not yet available in lymphoma, with only retrospective evidence or evidence derived from a single, phase II trial available so far.8,17 Some MRD-driven phase II and III trials are ongoing in MCL and also in follicular lymphoma to assess the clinical impact of personalized treatment, with first results eagerly awaited in the near future (reviewed by Dogliotti and Ferrero27 and summarized for MCL in Table 1B). In conclusion, given that there is compelling evidence as to the predictive role of MRD in MCL, and ongoing standardization and methodological efforts are highly advanced, MRD analysis in MCL is already included in the majority of prospective trials. The next steps will consist of large trials investigating an MRD-driven tailored therapeutic approach (Table 1B). Thanks to the expected results and the development of rapidly evolving MRD techniques, we foresee that, in the near future, the huge research efforts over the last years will finally translate into personalized treatment strategies as part of clinical routine, leading to a further benefit for lymphoma patients in general.

References 1. Ferrero S, Drandi D, Mantoan B, Ghione P, Omedè P, Ladetto M. Minimal residual disease detection in lymphoma and multiple myeloma: impact on therapeutic paradigms. Hematol Oncol. 2011;29(4):167-176. 2. Pott C, Schrader C, Gesk S, et al. Quantitative assessment of molecular remission after high-dose therapy with autologous stem cell transplantation predicts long-term remission in mantle cell lymphoma. Blood. 2006;107(6):2271-2278. 3. Geisler CH, Kolstad A, Laurell A, et al. Long-term progression-free survival of mantle cell lymphoma after intensive front-line immunochemotherapy with in vivo-purged stem cell rescue: a nonrandomized phase 2 multicenter study by the Nordic Lymphoma Group. Blood. 2008;112(7):2687-2693. 4. Pott C, Hoster E, Delfau-Larue MH, et al. Molecular remission is an independent predictor of clinical outcome in patients with mantle cell lymphoma after combined immunochemotherapy: a European MCL intergroup study. Blood. 2010;115(16):3215-3223. 5. Hermine O, Hoster E, Walewski J, et al. Addition of high-dose cytarabine to immunochemotherapy before autologous stem-cell transplantation in patients aged 65 years or younger with mantle cell lymphoma (MCL Younger): a randomised, open-label, phase 3 trial of the European Mantle Cell Lymphoma Network. Lancet. 2016;38 8(10044):565-575. 6. Albertsson-Lindblad A, Kolstad A, Laurell A, et al. Lenalidomide-bendamustine-rituximab in untreated mantle cell lymphoma > 65 years, the Nordic Lymphoma Group phase I+II trial NLG-MCL4. Blood. 2016 Jun 27 [Epub ahead of print] 7. Liu H, Johnson JL, Koval G, et al. Detection of minimal residual disease following induction immunochemotherapy predicts progression free survival in mantle cell lymphoma: final results of CALGB 59909.

1136

Haematologica. 2012;97(4):579-585. 8. Kolstad A, Pedersen LB, Eskelund CW, et al. Molecular Monitoring after Autologous Stem Cell Transplantation and Preemptive Rituximab Treatment of Molecular Relapse; Results from the Nordic Mantle Cell Lymphoma Studies (MCL2 and MCL3) with Median Follow-Up of 8.5 Years. Biol Blood Marrow Transplant. 2017;23(3):428-435. 9. Visco C, Chiappella A, Nassi L, et al. Rituximab, bendamustine, and low-dose cytarabine as induction therapy in elderly patients with mantle cell lymphoma: a multicentre, phase 2 trial from Fondazione Italiana Linfomi. Lancet Haematol. 2017;4(1):e15-e23. 10. Zaja F, Ferrero S, Stelitano C, et al. Second-line rituximab, lenalidomide, and bendamustine in mantle cell lymphoma: a phase II clinical trial of the Fondazione Italiana Linfomi. Haematologica. 2017;102(5):e203-e206. 11. Armand P, Redd R, Bsat J, et al. A phase 2 study of RituximabBendamustine and Rituximab-Cytarabine for transplant-eligible patients with mantle cell lymphoma. Br J Haematol. 2016;173(1):89-95. 12. Callanan MB, Delfau M, Macintyre E, et al. Predictive Power of Early, Sequential MRD Monitoring in Peripheral Blood and Bone Marrow in Patients with Mantle Cell Lymphoma Following Autologous Stem Cell Transplantation with or without Rituximab Maintenance; Interim Results from the LyMa-MRD Project, Conducted on Behalf of the Lysa Group. Blood. 2015;126(23):338. 13. Gressin R, Callanan M, Daguindau N, Tempescul A, Carras S, Moles MP, et al. Frontline Therapy with the Ribvd Regimen Elicits High Clinical and Molecular Response Rates and Long PFS in Elderly Patients Mantle Cell Lymphoma (MCL); Final Results of a Prospective Phase II Trial By the Lysa Group. Blood. 2014;124(21):148. 14. Kaplan LD, Jung SH, Stock W, Bartlett NL, Pitcher B, Byrd JC. Bortezomib Maintenance (BM) Versus Consolidation (BC) Following Aggressive Immunochemotherapy and Autologous Stem Cell Transplant (ASCT) for Untreated Mantle Cell Lymphoma (MCL): CALGB (Alliance) 50403. Blood. 2015;126(23):337. 15. Pott C, Hoster E, Beldjord K, et al. R-CHOP/R-DHAP Compared to RCHOP Induction Followed by High Dose Therapy with Autologous Stem Cell Transplantation Induces Higher Rates of Molecular Remission In MCL: Results of the MCL Younger Intergroup Trial of the European MCL Network. Blood. 2010;116(21):965. 16. Pott C, Macintyre E, Delfau-Larue MH, et al. MRD Eradication Should be the Therapeutic Goal in Mantle Cell Lymphoma and May Enable Tailored Treatment Approaches: Results of the Intergroup Trials of the European MCL Network. Blood. 2014;124(21):147. 17. Ferrero S, Monitillo L, Mantoan B, et al. Rituximab-based pre-emptive treatment of molecular relapse in follicular and mantle cell lymphoma. Ann Hematol. 2013;92(11):1503-1511. 18. van der Velden VH, Cazzaniga G, Schrauder A, et al. Analysis of minimal residual disease by Ig/TCR gene rearrangements: guidelines for interpretation of real-time quantitative PCR data. Leukemia. 2007;21(4):604-611. 19. Drandi D, Kubiczkova-Besse L, Ferrero S, et al. Minimal Residual Disease Detection by Droplet Digital PCR in Multiple Myeloma, Mantle Cell Lymphoma, and Follicular Lymphoma: A Comparison with Real-Time PCR. JMD. 2015;17(6):652-660. 20. 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. 21. Ladetto M, Brüggemann M, Monitillo L, et al. Next-generation sequencing and real-time quantitative PCR for minimal residual disease detection in B-cell disorders. Leukemia. 2014;28(6):1299-1307. 22. Brüggemann M, Knecht H, Bartram J, et al. International multi-laboratory next-generation sequencing for MRD analysis in ALL. A pilot study by the Euroclonality-NGS Consortium. Haematologica. 2015;100(s1):P164:31. 23. Hottentot QP, van Min M, Splinter E, White SJ. Targeted Locus Amplification and Next-Generation Sequencing. Methods Mol Biol. 2017;1492:185-196. 24. Wren D, Walker BA, Brüggemann M, et al. Comprehensive translocation and clonality detection in lymphoproliferative disorders by nextgeneration sequencing. Haematologica. 2017;102(2):e57-e60. 25. Kurtz DM, Green MR, Bratman SV, et al. Noninvasive monitoring of diffuse large B-cell lymphoma by immunoglobulin high-throughput sequencing. Blood. 2015;125(24):3679-3687. 26. Dreyling M, Geisler C, Hermine O, et al. Newly diagnosed and relapsed mantle cell lymphoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2014;25 (Suppl 3):iii83-92. 27. Dogliotti I, Ferrero S. Personalized medicine in lymphoma: tailoring treatment according to minimal residual disease. Medical Research Archives, Vol. 5, Issue 5, May 2017.

haematologica | 2017; 102(7)

Editorials

Antithymocyte globulin and transplants for aplastic anemia Andrea Bacigalupo Istituto di Ematologia, Universitaâ&#x20AC;&#x2122; Cattolica del Sacro Cuore, Fondazione Policlinico Universitario A Gemelli, Rome, Italy E-mail: andrea.bacigalupo@unicatt.it

doi:10.3324/haematol.2017.171538

n this issue of Haematologica, Kekre and colleagues have completed a study on the use of antithymocyte globulin (ATG) as part of the conditioning regimen for severe aplastic anemia (SAA) patients, undergoing an allogeneic stem cell transplant (HSCT) from matched sibling (MSD) or unrelated (UD) donors.1 The Authors have found that rabbit ATG (thymoglobulin, Sanofi, France), as compared to horse ATG (ATGAM, Pfizer, USA) is associated with less acute graft-versus-host disease (GvHD), less chronic GvHD in MSD grafts, and improved survival in UD grafts, and recommend the use of rabbit ATG in the conditioning regimen for patients with SAA. The hypothesis for improved outcome of patients receiving rabbit ATG is that rabbit ATG is associated with more effective depletion of lymphocytes,2 and that rabbit ATG, but not horse ATG, enhances the number and function of regulatory T cells,3,4 which may be relevant in controlling GvHD and inducing tolerance. Horse ATG was part of the original protocol designed by Storb et al.5 for transplants from human leukocyte anti-

gen (HLA)-identical siblings in SAA, together with cyclophosphamide (CY) 200 mg/kg. This protocol (CY200-ATG), followed by the infusion of unmanipulated bone marrow (BM) cells, was introduced more than 4 decades ago and is still considered the standard of care for young patients grafted from MSDs,6 with survival exceeding 80% for these patients. However, several questions remain open, the first being whether we should continue to use ATG in 2017, especially as a prospective trial in 134 SAA patients failed to show a benefit for patients randomized to the horse ATG arm.7 The study by Kekre and et al.1 focused solely on patients receiving ATG, and thus we are missing the control arm. A registry-based study on HLAidentical sibling transplants (n=1886) showed a significant survival advantage for patients receiving ATG , irrespective of whether this was derived from horse or rabbit, both in univariate and multivariate analysis.8 Therefore the answer to this first question is that ATG should be part of the conditioning regimen for SAA grafted from identical siblings, and the study by Kekre and colleagues1 suggests

Figure 1. MSD transplants for SAA - EBMT 1999-2009. Survival in patients with acquired aplastic anemia grafted from HLA identical siblings with bone marrow as a stem cell source (left) or with G mobilized peripheral blood (right). A significant survival advanatage is seen, together with less acute and chronic GvHD.HLA: human leukocyte antigen; MSD: matched sibling donor; EBMT: European Group for Blood and Marrow Transplantation; ATG: antithymocyte globulin; aGvHD: acute graft-versus-host disease; cGvHD: chronic graft-versus-host disease.

haematologica | 2017; 102(7)

1137

Editorials

it should be rabbit ATG. The second question is whether this holds true for patients receiving either BM or peripheral blood (PB) as a stem cell source? The answer is again yes, as shown in Figure 1, derived from the European Group for Blood and Marrow Transplantation (EBMT) data set 1999-2009.8 The 10-year survival advantage is significant for both BM and PB, although it is more evident for patients receiving PB grafts , with a significant reduction of both acute and extensive chronic GvHD (Figure 1). Therefore a SAA patient grafted with BM from an identical sibling, with ATG in the conditioning regime, has an 86% chance of 10year survival, with an 8% risk of acute GvHD and a 2% risk of extensive chronic GvHD; the same patient receiving a PB graft without ATG in the conditioning regimen has a 65% chance of 10-year survival, with a 21% risk of acute GvHD and a 9% risk of extensive chronic GvHD (Figure 1). This information may be useful when designing the transplant strategy for a given patient. A multivariate Cox analysis confirms that no ATG (P=0.0004) and PB grafts (P<0.00001) are two independent negative predictors of survival, together with age, interval between diagnosis and transplant and a conditioning other than CY200.8 The third question is whether ATG has a beneficial effect on survival, also for those SAA patients grafted from UDs? Again using data from an EBMT study,9 the answer is yes: the 5-year survival rate of UD grafts with ATG (n=312) in 2005-2009 was 70%, compared to 52% for patients not given ATG in the conditioning regimen (n=198) (P=0.05). Once more this holds true in a multivariate Cox analysis, together with other independent predictors such as age, stem cell source and interval between diagnosis and treatment.9 UD transplants pose additional problems, especially with engraftment and GvHD, and for this reason conditioning regimens currently include fludarabine and a small dose of total body irradiation (TBI) to increase immune ablation of the host.10 The optimal dose of CY to be used in UD transplants, together with fludarabine, has also been the object of a recent study:11 CY is recommended at the dose of 50 to 100 mg/kg, and TBI at the dose of 2Gy in a single fraction.10,11 ATG would be used as a part of this conditioning regimen. If ATG is a positive predictor of survival in both MSD and UD grafts, it may be relevant to identify the optimal dose and timing. Unfortunately, we do not have prospective studies comparing these two important variables, and the study by Kekre et al.1 could not address this issue because the dose of ATG was not captured in the database. A conventional dose for thymoglobulin in transplants for SAA is between 5 and 7.5 mg/kg (total dose), administered in the 3 days before transplant; for ATGAM this would convert to a total dose of 120 mg/kg. In conclusion, allogeneic HSCT for patients with aplas-

1138

tic anemia should always include in vivo T-cell depletion, irrespective of patients' age, donor type and stem cell source. We now have evidence that rabbit ATG (thymoglobulin) seems to protect patients from GvHD, better than horse ATG. The study by Kekre et al.1 did not include patients receiving rabbit ATG-fresenius, so we do not have a comparison of this ATG with thymoglobulin and ATGAM. Another T cell depleting agent is alemtuzumab, which has been shown to achieve the same, if not superior protection against acute and chronic GvHD when compared to thymoglobulin in the transplant setting of SAA;12 this may be particularly relevant in older SAA patients. Given the unsatisfactory treatment of acute and chronic GvHD, and the lack of any benefit in SAA derived by the graft-versus-leukemia effect, particular care should be devoted to GvHD prophylaxis in patients with aplastic anemia. Stem cell source and in vivo T-cell depletion are modifiable transplant variables, and we now have evidence that the combination of BM stem cells and rabbit ATG provide optimal protection against GvHD, and may improve survival.

References 1. Kekre N, Zhang Ying, Zhan Mei-Jie, et al. Effect of antithymocyte globulin source on outcomes of bone marrow transplantation for severe aplastic anemia. Haematologica 2017; 102(7):1291-1298. 2. Scheinberg P, Fischer SH, Li L, et al. Distinct EBV and CMV reactivation patterns following antibody-based immunosuppressive regimens in patients with severe aplastic anemia. Blood. 2007;109(8):3219-3224. 3. Feng X, Kajigaya S, Solomou EE, et al. Rabbit ATG but not horse ATG promotes expansion of functional CD4+CD25highFOXP3+ regulatory T cells in vitro. Blood. 2008;111(7):3675-3683. 4. Lopez M, Clarkson MR, Albin M, et al. A novel mechanism of action for anti-thymocyte globulin: induction of CD4+CD25+Foxp3+ regulatory T cells. J Am Soc Nephrol. 2006;17(10):2844-2853. 5. Storb R, Thomas ED, Buckner CD, et al. Allogeneic marrow grafting for treatment of aplastic anemia. Blood. 1974;43(2):157-180. 6. Aljurf M, Al-Zahrani H, Van Lint MT, et al. Standard treatment of acquired SAA in adult patients 18â&#x20AC;&#x201C;40 years old with an HLA-identical sibling donor. Bone Marrow Transpl. 2013;48(2):178-179. 7. Champlin RE, Perez WS, Passweg JR, et al. Bone marrow transplantation for severe aplastic anemia: a randomized controlled study of conditioning regimens. Blood. 2007;109(10):4582-4585. 8. Bacigalupo A, Socie' G, Schrezenmeier H, et al. Bone marrow versus peripheral blood sibling transplants in acquired aplastic anemia: survival advantage for marrow in all age groups. Haematologica. 2012;97(8):1142-1148. 9. Bacigalupo A, SociĂŠ G, Hamladji RM, et al. Current outcome of HLA identical sibling vs. unrelated donor transplants in severe aplastic anemia: an EBMT analysis. Haematologica. 2015;100(5):696-702. 10. Deeg HJ, Amylon ID, Harris RE, et al. Marrow transplants from unrelated donors for patients with aplastic anemia: minimum effective dose of total body irradiation. Biol Blood Marrow Transplant. 2001;7(4):208-215. 11. Anderlini P, Wu J, Gersten I, et al. Cyclophosphamide conditioning in patients with severe aplastic anaemia given unrelated marrow transplantation: a phase 1-2 dose de-escalation study. Lancet Haematol. 2015;2(9):e367-375. 12. Marsh JC, Gupta V, Lim Z, et al. Alemtuzumab with fludarabine and cyclophosphamide reduces chronic graft-versus-host disease after allogeneic stem cell transplantation for acquired aplastic anemia. Blood. 2011;118(8):2351-2357.

haematologica | 2017; 102(7)

REVIEW ARTICLE

Genetic background and evolution of relapses in aggressive B-cell lymphomas Darius Juskevicius, Stephan Dirnhofer and Alexandar Tzankov

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Institute of Pathology & Medical Genetics, University of Basel and University Hospital Basel, Switzerland

ABSTRACT

elapses of aggressive B-cell lymphomas pose a higher risk to affected patients because of potential treatment resistance and usually rapid tumor growth. Recent advances, such as targeting Bruton tyrosine kinase, have provided promising results in small numbers of cases, but treatment for the majority of patients remains challenging and outcomes are generally poor. A number of recent studies have utilized state-of-the-art genomic technologies in an attempt to better understand tumor genome evolution during relapse and to identify relapse-specific genetic alterations. It has been found that in some settings (e.g. diffuse large B-cell lymphomas in immunocompromised patients, secondary diffuse large B-cell lymphomas as Richter transformations) a significant part of the recurrences are clonally-unrelated de novo neoplasms, which might have distinct genomic and drug sensitivity profiles as well as different prognoses. Similar to earlier findings in indolent lymphomas, genetic tumor evolution of clonally-related relapsing aggressive B-cell lymphomas is predominantly characterized by two patterns: early divergence from a common progenitor and late divergence/linear evolution from a primary tumor. The clinical implications of these distinct patterns are not yet clear and will require additional investigation; however, it is plausible that these two patterns of recurrence are linked to different treatment-resistance mechanisms. Attempts to identify drivers of relapses result in a very heterogeneous list of affected genes and pathways as well as epigenetic mechanisms and suggest many ways of how recurrent tumors can adapt to treatment and expand their malignant properties.

Haematologica 2017 Volume 102(7):1139-1149

Correspondence: alexandar.tzankov@usb.ch

Received: January 17, 2017. Accepted: May 3, 2017. Pre-published: May 29, 2017.

Introduction

doi:10.3324/haematol.2016.151647

Aggressive B-cell lymphomas are lymphoid neoplasms characterized by blastoid morphology and rapid tumor growth.1 As a whole, this group comprises multiple histologically and phenotypically defined diagnostic categories as well as â&#x20AC;&#x153;gray zoneâ&#x20AC;? categories. Despite their aggressive course, if properly diagnosed and treated, most aggressive B-cell lymphomas have favorable outcomes. For example, the most frequently occurring neoplasm within this category, diffuse large B-cell lymphoma (DLBCL), is treated by the now standard immunochemotherapy regimen, R-CHOP (rituximab, cyclophosphamide, hydroxydaunorubicin, oncovin, and prednisolone), and this treatment results in clinical remission or even cure of a large number of patients.2 However, approximately 30% to 40% of patients in whom initial therapy failed due to primary resistance (~15%) or relapse (~25%) have a poor prognosis and often succumb to disease.2,3 Treatments for such patients consist of more aggressive, high-dose chemotherapy regimens supported by autologous stem cell transplantation,4 which is associated with severe toxicity and risks of fatal infections. Therefore, only about 50% of relapsed patients can tolerate such intensive treatment, while others who are not eligible, because of advanced age or co-morbidities, are managed with experimental protocols or palliative care.5 Interestingly, while the addition of rituximab to first-line regimens improved therapy outcomes for the majority of DLBCL patients, its prior use is associated with

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1139

haematologica | 2017; 102(7)

ÂŠ2017 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.

1139

D. Juskevicius et al.

worse outcomes after salvage chemotherapy and autologous stem cell transplantation in the subgroup of early relapsing patients.6 The majority of DLBCL recurrences arise during the first 3 years after initial therapy, although up to 10% of patients relapse >5 years after their initial diagnosis.7 It has been hypothesized that resistance/relapse occurs either due to intrinsic tumor heterogeneity, or as a consequence of ongoing genomic tumor evolution under the selective pressure of therapy, or both.8 An initial genetic heterogeneity of the tumor can confer a propensity to relapse due to usually minor tumor cell population(s) that have a pre-existing resistance to treatment (e.g. due to stemness properties9), and that can outgrow following extermination of the dominant non-resistant (and often more proliferative) clone.10 On the other hand, genomic tumor evolution can occur after treatment if the initial clone is not completely eradicated.11 This evolution can potentially be driven by intrinsic genomic/genetic instability of the tumor itself or by the genomic/genetic instability caused by alkylating- and DNA intercalating agents, which are part of the chemotherapy regimen (cyclophosphamide and hydroxydaunorubicin in CHOP) and, finally, consolidated by selective pressures exerted by the treatment.12 The group of refractory/relapsed aggressive B-cell lymphomas constitutes an unmet challenge in terms of delineating the underlying biological mechanisms, identifying prognostic markers to identify such patients before starting treatment and, most importantly, developing effective treatments capable of curing more patients. In this review, we summarize the latest findings on ancestry, genetic evolution and determinants of relapses of aggressive B-cell lymphomas.

Clonally-unrelated “relapse” All B-cell lymphomas are clonal diseases resulting from the proliferation of a single B-cell (clone) that gives rise to a tumor. In fact, monoclonality is an important criterion to qualify a B-cell proliferation as a lymphoid neoplasm.13 While the definition of a cancer cell clone is a matter of ongoing debate in the field of solid tumors,14 it is widely accepted that lymphoid neoplasms sharing identical antigen-receptor gene rearrangements are clonally-related. This is because IG V(D)J recombination, generating a unique receptor gene sequence, occurs at an early stage of B-cell development, arguably earlier than genetic aberrations that initiate malignant proliferation appear, and is, therefore, irreversibly imprinted in the cell’s progeny. Thus, antigen receptor gene rearrangement represents a physiological marker of common origin,15 a feature that has been widely exploited in research and clinical settings to develop a broad variety of assay designs.16 Traditionally, second and subsequent occurrences of lymphoid neoplasms in the same patient after a period of complete remission are regarded as (clonally-related) outgrowths of the primary tumor.17–19 However, this concept has been challenged by studies of multiple subtypes of lymphomas showing that a significant percentage of secondary occurrences are clonally-unrelated neoplasms (summarized in Table 1). Although it has been suggested that clonally-unrelated relapses occur with different frequencies in different lymphoma subtypes,20 precise numbers are lacking. This can be attributed to small study 1140

cohorts, discrepancies in study designs and applied technologies, selection biases or any combination thereof. Several cohort-based studies showed that clonally-unrelated “relapses” occur in approximately 15-20% of DLBCL patients.11,21–23 Moreover, about 20% of DLBCL relapses emerging as Richter transformations had distinct IGH rearrangements compared to the initial chronic lymphocytic leukemia.24,25 In classical Hodgkin lymphoma, a higher frequency of approximately 35% was reported.26 Conversely, relapses of follicular lymphoma as well as their transformations to DLBCL are almost always clonallyrelated to the primary tumor.27–29 Collectively, these results show that the occurrence of clonally-unrelated lymphoma “relapses” at certain frequencies is specific to certain subtypes of B-cell neoplasms. This raises the following questions. (i) What are the factors that predispose individuals to develop a second independent lymphoid neoplasm? (ii) What are the clinical implications of such relapses? (iii) Can clonally-unrelated recurrences be predicted and, if so, how can current diagnostic procedures be adapted to facilitate their detection? A clonally-unrelated relapse, i.e. a second de novo lymphoid neoplasm in the same patient, cannot be explained by pure randomness, since the probability of this type of tumor is quite low in the general population (at an incidence of 19.5/100,000/year30 the probability of developing, e.g., a DLBCL twice within 15 years by chance is approximately 0.0009%). There must, therefore, be predisposing factors that increase the general risk of lymphoma in certain patients. Indeed, several such factors with various degrees of penetrance and incidence have been described, including inborn genetic properties,31,32 viral infections,33 immune suppression34,35 and exogenous exposure to mutagens.36–38 It is known that immunocompromised patients are more susceptible to lymphoma (e.g. post-transplant lymphoproliferative diseases in grafted patients, primary effusion lymphomas in patients with acquired immunodeficiency syndrome).39,40 Indeed, we and others have shown that clonally-unrelated DLBCL relapses can occur in patients with iatrogenic immunodeficiency due to transplantation.21,41 Additionally, some second de novo lymphoma recurrences in immunoproficient patients are frequently localized in immunologically-privileged anatomical sites (central nervous system, testes).21,42 In such cases, lack of immunosurveillance due to anatomical barriers could provide similarly favorable conditions for tumor development as acquired immunodeficiency; indeed, immunodeficiency can be reflected in the mutational background of DLBCL, as exemplified in a recent study showing that post-transplant DLBCL have fewer mutations in general, and do not harbor B2M mutations43 in particular, while B2M is mutated in approximately 12% of DLBCL occurring in immunocompetent patients.44 In addition to immunodeficiency, inherited or acquired genetic defects can further increase the risk of a second de novo lymphoma.31,45 It has recently been shown that somatic mutations of genes implicated in leukemo- and lymphomagenesis can be detected in the hematopoietic systems of healthy middle-aged and elderly individuals.46 This phenomenon was termed clonal hematopoiesis of indeterminate potential and was shown to increase the risk of hematopoietic cancers.46 Furthermore, it has been demonstrated that mice transplanted with hematopoietic stem cells purified from the bone marrow of patients with haematologica | 2017; 102(7)

Genetics of aggressive B-cell lymphoma relapses

Table 1. Summary of studies examining clonal relationships in lymphomas.

Year

Study

PubMed ID

Diagnosis of primary lymphoma, (n)

Diagnosis of relapse, (n)

Time to related relapse, years, range (median)

Time to unrelated relapse, years, range (median)

Method for testing clonal relationships

Unrelated relapses, (%)

1994

Matolcsy et al.

8118038

CLL (6)

DLBCL (6)

0-4 (1) months

1/6 (17%)

1996

Nishiuci et al.

8616769

DLBCL (4)

6-12 (9)

7-10 (8.5)

Southern blot, sequencing IG-FL/seq

3 4 5 6

1996 1996 1997 2002

Matolcsy et al. Lister et al. Campana et al. Libra et al.

8916960 8695860 9285551 12181049

2002

Fend et al.

11841433

2003

De Jong et al.

12649152

9 1-2 (2) NA NA 3-4 (3.5) 3-6 (4) 5 NA NA 4-17 (7)

NA 2 15 3.5 3 NA 7 3 2 6-8(6)

9 10 11 12 13 14 15 16

2003 2004 2004 2004 2007 2011 2011 2012

Heintel et al. Lossos et al. Timรกr et al. Thomas et al. Mao et al. Obermann et al. Rossi et al. Ganzel et al.

12581905 15372481 14671632 15291369 17895764 21737506 21266718 22659385

17 18 19

2013 2013 2013

Wobser et al. Jain et al. Geurts-Gieleet al.

24274426 23763914 23765542

20 21 22

2014 2014 2014

Pasqualucci et al. Okosun et al. Wei et al.

24388756 24362818 25179408

23 24

2014 2016

Jiang et al. Juskevicius et al.

25123191 27198204

2016

Lee et al.

26848863

26 27

2016 2016

Xiao et al. Wu et al.

27184478 27224912

DLBLC (3), Follicular small cleaved (1) Follicular mixed (1) Follicular mixed (1) FL (8) DLBCL (8) BL (1) BL (1) FL (1) BL (1) FL (3) FL (2), MCL (1) SLL (5) SLL (5) MCL (2) MCL (2) MCL (1) HL (1) CLL (1) HL (1) DLBCL (13) DLBCL (13) SMZL (1) BL (1) DLBCL (5) DLBCL (5) CLL (8) DLBCL (8) cHL(1) DLBCL (1) CLL (23) DLBCL (5) cHL (21), DLBCL (1) cHL (21), DLBCL (1) CLL (63) DLBCL (63) CLL (2) HL (2) HL (3) MZL (1) DLBCL (2) Low grade B (1) HL (1) MZL (1) MZL (1) CLL MCL DLBCL (6) DLBCL (4), HL (1), FL (1) FL (14) FL (10), DLBCL (4) PCFCL (1) PCFCL (1) MZL/MALT (7) MCL (3), MALT (1), CLL (1) DLCBL (1) MCL (3) MCL (3) CLL (2) CLL (2) NHL-NOS (1) NHL-NOS (1) WM (1) DLBCL (1) FL (12) DLBCL (12) FL (10) DLBCL (10) CLL (2) DLBCL (2) FL (1) FL (1) DLBCL (14) DLBCL (14) DLBCL (20) DLBCL (20) SLL (1) MCL (8) FL (3) DLBCL (5) CLL (33) MCL

DLBCL (1) MCL (8) FL(1) FL+DLBCL (2) DLBCL (5) HL (33) MCL

NA NA 0-4 (0) NA 0-1 (0) 1-10 (2.5) ? NA NA NA NA NA 6-11 (7) 5-16 (7) 7 6-13(6)

IG-FL/seq t(14;18) breakpoint analysis IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq, translocation analysis 2 IG-FL/seq 3-12(5) IG-FL/seq 1-2(2) IG-FL/seq 13 IG-FL/seq 0-4 (1.5) IG-FL/seq 2-13 (6.5) IG-FL/seq ? IG-FL/seq ? IG-FL ? IG-FL ? IG-FL 3 IG-FL 8 IG-FL/seq 0 IG-FL/seq NA IG-FL/seq NA IG-FL/seq NA IG-FL/seq

8-14 (14) 7 9 NA 0-9 (3.5) 1-19 (8) NA NA 0-13(3) 1-12(2)

NA NA NA 5 NA NA 0-18 2 2 3-8.5(7)

NA 1-6 (2.6) 0-3(1) 1-2 (2) NA 1-10 (3)

8 3 NA 5.6 NA NA

IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq WES, WGS IG-FL/seq IG-FL/seq IG-HTS IG-FL/seq, aCGH and targeted HTS IG-FL/seq IG-FL/seq IG-FL/seq IG-FL/seq IG-FL IG-HTS

2/4(50%) 0/1(0%) 1/8 (13%) 1/1 (100%) 1/1(100%) 1/3 (33%) 0/5 (0%) 1/2(50%) 1/1 (100%) 1/1 (100%) 3/13(23%) 1/1(100%) 5/5 (100%) 3/8 (38%) 1/1 (100%) 5/23 (22%) 8/22 (36%) 13/63 (21%) 2/2 (100%) 3/3 (100%) 1/1 (100%) 1/1 (100%) 1/1 (100%) 1/6 (17%) 0/14 (0%) 0/1 (0%) 0/7 (0%) 0/3 (0%) 0/2 (0%) 0/1 (0%) 1/1 (100%) 0/12 (0%) 0/10 (0%) 2/2 (100%) 1/1 (100%) 1/14 (7%) 3/20 (15%) 1/1 (100%) 1/7 (14%) 0/3 (0%) 1/5 (20%) 19/33 (58%) 0/11 (0%)

BL: Burkitt lymphoma; CLL: chronic lymphocytic lymphoma; cHL: classical Hodgkin lymphoma; DLBCL: diffuse large B-cell lymphoma; FL: follicular lymphoma; HL: Hodgkin lymphoma; HTS: high throughput sequencing; IG: immunoglobulin gene; IG-FL/seq: immunoglobulin fragment length/sequence analysis; IG-HTS: high throughput sequencing of immunoglobulin gene rearrangements; MALT: mucosa-associated lymphoid tissue; MCL: mantle cell lymphoma; MZL: marginal zone lymphoma; NA: not available; NHL: NOS: non-Hodgkin lymphoma not otherwise specified; PCFCL: primary cutaneous follicle center lymphoma; SLL: small lymphocytic lymphoma, SMZL: splenic marginal zone lymphoma; WES: whole exome sequencing; WGS: whole genome sequencing

haematologica | 2017; 102(7)

1141

D. Juskevicius et al.

chronic lymphocytic leukemia developed monoclonal or oligoclonal B-cell proliferations that displayed IGH rearrangements different from those of the original chronic lymphocytic leukemia clone.47 These data suggest that predisposing mutations may be present in precursor cells at a stage prior to the antigen-receptor gene rearrangements and terminal differentiation. The aggressive disease course and lack of efficacious treatments in relapses of lymphoid neoplasms result in high rates of morbidity and mortality. The clinical behavior of clonally-unrelated relapses remains controversial as different studies report contradictory outcomes. In our previous study, two of three patients with second de novo DLBCL were treated successfully and had long-lasting clinical remissions.21 In contrast, Lee et al. reported dismal outcomes in two cases of clonally-unrelated DLBCL relapses.23 Rossi et al. reported that 12 patients who developed DLBCL, via Richter transformation, which was clonally-unrelated to the primary chronic lymphocytic leukemia had better survival than 49 patients with clonally-related transformations.25 However, an earlier, smaller study found no difference in survival between 19 patients with clonally-related and five with clonally-unrelated Richter transformation.24 Currently, if the patient’s health condition allows, aggressive B-cell lymphoma recurrences are treated with aggressive secondary chemotherapy regimens and autologous stem cell transplantation because acquired tumor resistance is anticipated. However, clonally-unrelated relapses, effectively being de novo neoplasms, could be sensitive to standard first-line treatments since the clone that gave rise to the second neoplasm was potentially not yet present during the initial therapy, and, therefore, had neither evolved under selective treatment pressure nor developed therapy resistance. If that is the case, then the hazardous side effects of aggressive therapies could be avoided. Additionally, patients who are not eligible for secondline treatment and are diverted to palliative care could be eligible for milder treatment regimens with curative intent. This possibility of treating clonally-unrelated relapses with first-line therapies must be evaluated by specifically designed prospective clinical trials, and the cumulative data reviewed here suggest the rationale for such investigations. For the reasons described above, it would be useful to know which recurrences are most likely to be clonallyunrelated secondary de novo neoplasms. Time to relapse could be a useful indicator since unrelated relapses more often present as late recurrences (5 years or later after the initial disease).20,22 However, clonally-related relapses are also reported to occur long after the diagnosis of the primary tumor (up to 17 years later).22 In contrast, changes in morphology, phenotypic cell-of-origin conversions or viral infection status did not prove to be useful predictors of clonally-unrelated “relapses”.48 Ultimately, if therapy changes are envisioned for clonally-unrelated relapses, given their significant frequencies in some lymphoma subtypes (e.g. DLBCL, classical Hodgkin lymphoma), under certain circumstances (long interval, history of immunodeficiency or relapse in immunologically privileged site), determination of clonal relationship by genetic testing would be the most reliable solution.

Methodological considerations concerning clonality testing The multiplex polymerase chain reaction approach introduced almost two decades ago remains the gold stan1142

dard of clonality testing.49,50 With this technique, standardized sets of consensus primers are used to amplify rearranged IGH and IGL loci by polymerase chain reaction and, thereafter, lengths and sequences of the generated amplicons are investigated by capillary electrophoresis and Sanger sequencing.51 These methods provide conclusive results on clonality and/or clonal relationships in the vast majority of cases,49 but also have technical shortcomings that can lead to a lack of definitive output or misinterpretation.51,52 More recently, methods for high-throughput sequencing of B- and T-cell receptor genes have been developed, giving rise to a new field – immunogenomics.53–56 This new technique, also termed Ig-seq, allows qualitative and quantitative estimation of the entire immune repertoire of the investigated sample. In the context of clonality testing in lymphoid neoplasms this method can readily determine not only the sequence and the frequency of the dominant clone, but also provide information about less frequent clones “masked” in oligoclonal or polyclonal samples. This technology has already been advantageously used to provide answers about clonal relationships of recurrent tumors in cases in which results from traditional methods were hard to interpret and inconclusive.57 Ig-seq can also provide information about somatic hypermutations, which is useful in evolutionary analysis of emerging lymphoma populations.11,58 Furthermore, the high-throughput sequencing approach has been shown to be very sensitive, making it useful for monitoring minimal residual disease.59 Despite the aforementioned advantages, the introduction of Ig-seq into diagnostics and patient care has been slow. Like all current short read-based high-throughput sequencing techniques, Ig-seq is prone to sequencing mistakes, which can lead to an overestimation of receptor gene diversity or false interpretation of clonal relationship. Additionally, the nature of multiplex polymerase chain reactions can lead to amplification biases, with some rearrangements being preferred over others and causing incorrect estimation of clonotype frequencies. Multiple solutions have been suggested to deal with these shortcomings, including the use of unique molecular identifiers to label initial template molecules,60 synthetic spike-in controls to normalize biases and advanced sequencing library preparation methods.61,62 Data analysis represents an additional challenge that is slowing the broad implementation of this method. Ig-seq produces millions of reads, which have to be corrected for errors, aligned to reference sequences, normalized for biases and grouped into clonotypes according to the sequence similarity. To some avail, there are several publicly available software packages, which can handle these tasks.63–66 Analysis software is also offered by various commercial providers and can be purchased together with library preparation and sequencing kits. Irrespectively of which technology is used to determine clonality, there are still potential problems that have to be taken into account. First, one drawback of IG-based clonality testing is that the analysis is based on very precise locations in the genome (IGK, IGH and IGL loci on chr2, chr14 and chr22, respectively), and if these are disturbed by somatic changes in lymphoma cells (e.g. chromosomal deletions or somatic hypermutations at the primer binding sites), only partial or no results will be obtained, which makes interpretation unreliable or impossible. In such cases, other clonality markers, such as the presence of a haematologica | 2017; 102(7)

Genetics of aggressive B-cell lymphoma relapses

translocation with exactly identical junction regions in the primary and relapse tumors, can be utilized. However, recurring translocations only occur in some lymphomas (e.g. approximately 45% of DLBLC), and might not be applicable to the majority of samples. Alternatively, widescale genomic profiling would help to establish clonal relationship or lack thereof, as the presence or lack of shared copy number aberrations and/or somatic mutations between primary and relapse tumors are strongly indicative of clonally-related or -unrelated relapses, respectively, as previously shown in other types of cancers.67 Second, several studies have shown that genetic defects potentially increasing the risk of lymphomagenesis can occur in hematopoietic precursors prior to IG V(D)J recombination.21,47 This poses the following question: can subsequent lymphomas that bear distinct IG rearrangements but clearly share somatic mutations in genes implicated in lymphomagenesis be regarded as clonally related? In the sense of molecular oncology the answer is probably no, since lymphomagenesis is tightly associated with genetic events acquired due to aberrant effects of physiological mechanisms inherent to B-cell development that occur after IG rearrangements (“collateral damage” of somatic hypermutations, translocations related to classswitch recombination), and that shape clinically relevant tumor properties such as aggressiveness or treatment sensitivity. In the sense of pure tumor evolutionary biology, shared somatic aberrations, unless they occurred independently by chance, would be suggestive of a common lymphomagenesis and, even in the presence of independent IG rearrangements, bear witness to a common tumor derivation from a “primed” precursor cell pool.

Clonally-related relapses occur via two distinct genetic evolution patterns Until recently, lymphoma relapses, even after prolonged remission, were regarded as direct outgrowths of the primary tumor, assuming that all genetic alterations present in the primary tumor were directly passed to the relapse. Therefore, a relapse was considered to be a more evolutionarily advanced stage of the primary tumor. This was supported by stable detection of some genetic aberrations (e.g. classical drivers of lymphomagenesis such as BCL2 or CCND1 translocations) throughout the entire history of the individual disease. The recent advent of high-throughput genome-wide technologies enabled a broader look at the genetic composition of recurring tumor samples. Analysis of matched primary-relapse pairs allowed identification of shared and unique features between matched tumors inferring genetic tumor evolution. Previous studies in other types of cancer have unequivocally shown that the majority of tumors are heterogeneous, and that tumor growth occurs via (sub)clonal genome evolution.68 Multiple subsequent studies confirmed these findings in aggressive B-cell lymphomas and demonstrated in particular that relapses occur via at least two distinct patterns of clonal evolution.11,21,69–72 In one pattern, primary and relapse tumors share some common genetic alterations that attest to their clonal origin, but the majority of somatic mutations are unique to each tumor occurrence. This pattern suggests early divergence of tumor-founding clones and independent genetic evolution. In the other pattern, the relapsed clone bears the vast majority of genetic alterations of the primary haematologica | 2017; 102(7)

tumor, but can additionally possess variable numbers of relapse-specific mutations. Importantly, in the majority of such cases, individual tumor occurrences have a low number of private, usually small scale mutations, which provide evidence against linear tumor evolution and instead suggest very late divergence of primary and relapse clones. Similar genetic evolution patterns have previously been shown in the progression of transformed follicular lymphoma.27,28,73,74 Currently it is difficult to estimate which pattern is more frequent due to conflicting reports that reflect the use of different techniques as well as differences between the various types of aggressive lymphomas. Analyzing somatic hypermutation patterns within rearranged IGH VDJ sequences by Ig-seq in matched primary-relapse pairs of DLBCL, Jiang et al. found that relapses in six patients followed early divergent, and in seven patients late divergent evolution scenarios.11 In our previous study using targeted high-throughput sequencing and array comparative genomic hybridization, we found that six of 17 DLBCL relapses diverged early from the common progenitor, while 11/17 diverged late.21 In a recent study investigating mantle cell lymphoma by whole exome sequencing, Wu et al. found that most relapses followed the early-divergent/branching evolution pattern.70 Conversely, cytogenetic karyotyping of Burkitt lymphoma pairs showed primarily late divergent or direct tumor evolution.72

Implications of distinct genetic patterns in diffuse large B-cell lymphoma relapse The discovery of two distinct evolution patterns at relapse raises additional questions and provides the basis for new hypotheses concerning various aspects of lymphoma biology, such as considerations regarding tumor heterogeneity, timing of tumor growth and emergence of resistance. Testing these hypotheses could lead to new knowledge that would offer the potential for improved therapy.

Relapse patterns as markers of primary tumor genetic heterogeneity It has previously been shown in chronic lymphocytic leukemia that intratumoral heterogeneity might increase the risk of relapse or disease progression.75 Although intratumoral heterogeneity has been documented in DLBCL,76 this connection was not confirmed by a recent study in DLBCL showing that genome-wide methylation heterogeneity rather than genetic heterogeneity increased the risk of relapse.77 Although not yet clearly demonstrated, it is likely that the evolutionary pattern at relapse could potentially correlate with the degree of intratumoral heterogeneity of a pre-diagnostic or diagnostic tumor. Primary tumors that relapse via the late-divergent mode of evolution would be expected to be more homogenous and dominated by one clone. In contrast, primary tumors in cases in which relapse occurs via late-divergent/branching evolution are likely to have a more complex subclonal structure, with a dominant clone accompanied by minor clones that are the product of branching tumor evolution during early tumor development.

Implications for mechanisms of therapy resistance It is reasonable to speculate that the two different evolutionary patterns could represent two distinct mechanisms of therapy resistance (Figure 1). Late divergence 1143

D. Juskevicius et al.

could represent a spontaneous strategy of tumor survival under selective treatment pressure. In such case resistance is not pre-existing, but rather (rapidly) acquired by necessity. Mutagenic agents included in the chemotherapy regimen might create the genetic instability that is required for evolution to occur and to acquire resistance mutations essential for tumor survival. Already having all characteristics required for rapid tumor formation, such survivor clones would (more rapidly) replenish the tumor mass lost during initial treatment and form a relapse. In contrast, early divergence is obviously not occurring due to therapy pressure; rather it happens either due to limitations imposed by the surrounding microenvironment, such as the strength of immune response, the lack of nutrients, oxygen, etc., or due to intrinsic properties of the early tumor clone such as a higher degree of genomic instability due to defects in DNA repair, increased somatic hypermutation or aberrant mitotic mechanisms. The resistance to treatment would therefore occur as a purely stochastic consequence of clonal divergence. Genetic lesions that confer resistance to treatment might not be beneficial to the treatment-naĂŻve clone, which may proliferate slowly or stagnate while other, more efficient, clones thrive and lead to initial lymphoma manifestations. Despite the initially small contribution to mass and volume, once treatment is applied and the dominant clone is eradicated, the

former resistant subpopulation would survive and increase its growth potential by acquiring additional DNA lesions, and thus gradually evolve to tumor relapse.

Potential implications for treatment If the above model is feasible, and late divergent and early divergent relapses do indeed reflect acquired and pre-existing (intrinsic) resistance, correct identification of relapse patterns may have important implications for treatment, especially in the era of targeted treatment.78 It has been shown that intrinsic and acquired resistance occur principally via distinct cellular mechanisms determined by the specific types of selective pressures applied to developing tumor populations.79 Acquired resistance primarily occurs due to mutations in the targeted molecule. In such a scenario, the initial response to treatment is usually observed until the accumulation of resistance-conferring mutations. The effects of such mutations can be variable and include: (i) increased target expression (through inactivation of negative regulation loops or gene amplifications); (ii) compromised binding of drug to target (specific mutations at drug binding sites); and (iii) changes modulating target activity (mutations in target functional domains).8 In contrast, intrinsic resistance to targeted treatment is characterized by alterations that are initially present upstream or downstream of the target signaling pathway. Target inhibition is, therefore, ineffective since the

Figure 1. Schematic representation of two patterns of genetic evolution in clonally-related relapses of aggressive B-cell lymphomas. Different colors represent genetically distinct clones that possess private genetic alterations due to independent genome evolution. (A) In the early-divergent/branching scenario, the divergence occurs early in tumor development. The majority of subpopulations stagnate but one clone eventually acquires the effective combination of drivers, expands and gives rise to a heterogeneous primary tumor. The dominant population is exterminated by the treatment, however, an intrinsically resistant subclone exists and gradually gives rise to a relapse. (B) In the late-divergent/linear evolution scenario, the tumor initially possesses a strong driver combination. Therefore, the neoplastic cells grow fast and unrestricted, giving rise to a rather homogeneous primary tumor. Such a tumor is almost exterminated by the treatment but an acquired resistance emerges. The resistant subclone already has drivers of effective growth and rapidly replenishes the tumor mass giving rise to a more rapid relapse.

1144

haematologica | 2017; 102(7)

Genetics of aggressive B-cell lymphoma relapses

tumor has a pre-developed mechanism of circumventing the targets’ function by delivering important tumorigenic signals via alternative pathways. In such a scenario, the resistant cell population (possibly a minor clone in the context of the whole tumor, as discussed earlier) is unaffected by the therapy. Both intrinsic and acquired resistance can contribute to relapses. For example, one study showed that two types of mutations are responsible for acquired resistance to the Bruton tyrosine kinase (BTK) inhibitor ibrutinib in chronic lymphocytic leukemia: (i) mutations in BTK itself that change ibrutinib-mediated BTK inactivation from irreversible to reversible, and (ii) gain-of-function mutations in phospholipase gamma 2 (PLCγ2), which is a downstream phosphorylation target of BTK.80 BTK mutations are likely to represent acquired resistance, while PLCγ2 mutations most probably are related to pre-existing resistance in small subpopulations. This hypothesis is strongly supported by the detection of PLCγ2 mutation in diagnostic chronic lymphocytic leukemia samples at very low frequencies.80 It is known that intrinsic resistance to ibrutinib in DLBCL occurs due to a (pre-existing) L265P mutation in MYD88 and activating mutations in CARD11, both of which result in constitutive activation of nuclear factor-κB (NF-κB) signaling.81 PIM1 mutations were also suggested to contribute to intrinsic resistance to ibrutinib in activated B-cell DLBCL.82 Mechanisms of acquired resistance to ibrutinib in DLBCL are still unknown. In conclusion, information about early and late divergent patterns of genetic evolution at relapse can provide valuable insights into the potential type of treatment resistance in the recurrence, which can then aid in designing more efficaciously targeted treatment approaches.

Genetic events related to (clonally-related) recurrences Aggressive B-cell lymphoma relapses occur, at least in part, as a consequence of genetic evolution of the primary tumor. It is, therefore, conceivable that investigation of the exact changes that occur during this evolution will result in identification of genetic drivers of the process, prognostic as well as predictive markers and, it is to be hoped, “actionable mutations” that could be targeted to improve currently available treatments. Two primary experimental approaches are utilized for such identification. In one approach, matched pairs of primary and relapse tumors from the same patient are genetically profiled and compared, assuming that those events, which are relapse-specific or increase from low to high allelic frequency are related to recurrence. This approach has several advantages in that it allows: (i) studying only genetic alterations that occur specifically at relapse; (ii) identification of patient-specific shared somatic mutations that represent early drivers in tumorigenesis; (iii) exclusion of patientspecific single nucleotide polymorphisms that are not annotated in databases; and (iv) investigation of the genetic evolution of the tumors. However, paired samples with sufficient and good quality material are scarce because relapses are not routinely biopsied during the diagnostic process. Consequently, there are only a few of such studies and the overall number of cases analyzed is limited.10,11,21,71 In the alternative approach, unpaired relapsed/refractory tumors are analyzed and frequencies haematologica | 2017; 102(7)

of detected genetic alterations are compared to the respective frequencies in independent primary tumor collectives. This approach allows analysis of larger cohorts, but it has a significant drawback in that it is nearly impossible to accurately differentiate between mutations present at the primary stage and those that are relapse-specific. This problem can be illustrated by the ambiguous status of the MYD88 L265P mutation in DLBCL. In our previous study it was detected as a heterozygous early driver lesion present in both primary and relapsed tumors of the same patients.21 In contrast, an earlier DLBCL genome sequencing study reported that classic lymphomagenic events, such as mutations of MYD88, CD79A/B and EZH2, can also be subclonally present within primary DLBCL.76 The MYD88 mutation can, therefore, occur both as an early and a late event in DLBCL development, but this discrimination is highly problematic in cases in which only relapsed tumor samples are examined. The above mentioned shortcomings and the genuine heterogeneity of DLBCL potentially explain the limited overlap between five dedicated genomic studies on DLBCL relapses,10,11,21,71,83 and such studies in other types of aggressive lymphomas are even more limited (Table 2).69,70,72 In general, all studies of paired samples evidenced ongoing genetic evolution and acquisition of a variable number of additional mutations unique to the relapsed tumors or increased in frequency compared to the primary tumors through the process of clonal selection. A high diversity of alterations was observed, reflecting the many ways through which tumors adapt to the changed microenvironment after treatment. Pathway and gene ontology term enrichment analysis of the mutated genes at DLBCL relapse implicated activation of NF-κB, JAK-STAT, transmembrane receptor tyrosine kinase signaling and biological processes such as apoptosis, calcium channel activity, epigenetic regulation, cell cycle regulation and immunosurveillance (Table 2). On the background of profound variability, a small number of recurrent alterations were detected. Jiang et al. suggested that recurrences might be driven by mutations in immunosurveillance genes. Exome-sequencing revealed relapse-specific inactivating alterations in either CD58 or B2M, or both, in five of seven investigated pairs.11 Another study found that B2M and CD58 mutations were enriched in relapses of primary mediastinal B-cell lymphomas (PMBCL) and activated B-cell DLBCL, respectively.83 β2microglobulin (encoded by B2M) is necessary for the assembly and function of the major histocompatibility complex class I, which is involved in antigen presentation to cytotoxic T cells. CD58 is required for antigen adhesion and activation of NK and T cells.84 However, inactivating mutations in one or both of these genes were also frequently detected in primary DLBCL.85 Alterations in JAK-STAT pathway components were also implicated in DLBCL relapse. Morin et al. found that 36% of relapsed germinal center B-cell-like DLBCL and 38% of transformed follicular lymphomas bore STAT6 mutations of the DNA binding domain specifically affecting the D419 residue.10 Furthermore, this mutational status correlated with the overexpression of phosphorylated STAT6 levels in affected tumors, indicating that mutated STAT6 activates the JAK/STAT signaling cascade. Other dedicated studies, however, did not confirm the increased prevalence of this mutation at relapse, and clonal STAT6 D419 mutations were reportedly not rare in primary lym1145

D. Juskevicius et al.

phomas, including DLBCL.86–88 In addition to STAT6, inactivating mutations of SOCS1, a negative regulator of the JAK/STAT pathway, were frequently detected, predominantly in PMBCL relapses.83 However, it is not clear how relapse-specific these mutations are, since no paired

tumors were investigated in the study and SOCS1 mutations are frequently detectable in primary tumors (PMBCL and DLBCL).88 Another study found evidence of clonal selection of SOCS1 mutations at relapse, providing further confirmation that JAK/STAT signaling is important for

Table 2. Findings of the dedicated studies on aggressive B-cell lymphoma relapses.

Study

Study design

Genes implicated in Genes implicated early lymphomagenesis in relapse

Pathways and GO terms associated with relapse

Jiang et al.11

WES of 7 matched primaryrelapse DLBCL pairs and targeted HTS

KMT2D, EP300

Morin et al.10

WES of 38 rrDLBCL and NA their germline controls; Targeted HTS of 12 matched diagnostic samples; Frequency comparison to 138 primary DLBCL. WES of 14 rrDLBCL NA and their germline controls; Mutation frequencies compared to published mutational data of primary DLBCL.

Apoptosis (GO:0006915), transmembrane receptor tyrosine kinases (GO:0004714), calcium channel activity (GO:0005262), p53 binding (GO:0002039), B-cell proliferation (GO:0006915), focal adhesion (hsa04510), peptydyl-lysine acetylation (GO:0018394). JAK-STAT signaling, NF-κB signaling.

NF-κB signaling, sucrose degradation, tryptophan degradation, JAK-STAT signaling, meiosis, epigenetic regulation, S1P2 signaling.

TGF-β receptor activity.

Mareschal et al.83

Pan et al.77

Juskevicius et al.21

Melchardt et al.71

Beà et al.69

Wu et al.70

Aukema et al.72

Genome-wide methylation analysis of 13 matched primary-relapse DLBCL pairs and 7 non-relapsing DLBCL. aCGH and targeted HTS of 20 matched primary-relapse pairs and 20 primary non-relapsing DLBCL. Targeted HTS of 24 matched primary-relapse DLBCL pairs. WGS or WES of 29 primary MCL samples, 6 simultaneous tumors and 2 matched relapse samples. WES of 13 matched primary - relapse pairs of MCL

Cytogenetic karyotype analysis of 7 paired primary-relapse BL cases.

CD58, B2M, ARHGEF7, PLCB2, IL9R

TP53, FOXO1, MLL3, NFKBIZ, STAT6, MYC, MLL3, MPEG1, CCND3, MYD88, FOXO1, TNFRSF14, CARD11, B2M, MEF2B, ARID1A, TBL1XR1, NFKBIE, SOCS1, CD79B, BCL2 ABC-DLBCL: MYD88, CD58, TBL1XR1, IRF4 GCB-DLBCL: BCL2, MEF2B, DUSP2, NFKBIA, PIM1, KMT2D, GNA13 PMBCL: SOCS1, STAT6, TNAIFP3, B2M, ITPKB, NFKBIE SMAD6, ACVR2A, ACVR2B

Lesions specific to non-relapsing disease

KMT2D, MYD88, CD79B

KMT2D, MEF2B, BCL2 Gain chr10p15.3-13(GATA3, PRKCQ, MLLT10/AF10)

SOCS1

EP300, IRF8, MYD88

TP53, MCL1, ATM, FAT2, MYC, RB1, SMARCA4

ATM, CCND1, MLL2, KCNC2, KIAA1671, PCSK2, TNRC6, TRPM6

ABCA3, TLR2, TP53, WHSC1

t(11;14)(q13;q32)

PPM1D

t(8;14)(q24;q32)

Gains: 11q, 1q, 7q; Losses: 6q; 13q 17p; Genes affected: ATM, KMT2A, DLEU1, TP53, HIC1

p53 signaling, MAPK signaling, ErbB signaling, NF-κB signaling, Wnt signaling, DNA repair. NA

aCGH: array-comparative genomic hybridization; BL: Burkitt lymphoma; GO: gene ontology; HTS: high-throughput sequencing; MCL: mantle cell lymphoma; rrDLBCL: relapsed/refractory diffuse large B-cell lymphoma; WES: whole exome sequencing; WGS: whole genome sequencing. Genes found in at least two studies on the same entity are highlighted in bold.

1146

haematologica | 2017; 102(7)

Genetics of aggressive B-cell lymphoma relapses

recurrences.10 Importantly, SOCS1 is frequently mutated in primary DLBCL tumors88 and it was suggested that SOCS1 mutational status has a prognostic value: Schif et al. found that patients with primary tumors with truncating SOCS1 mutations had excellent overall survival, whereas those with non-foreshortening point mutations and minor consequences for the protein had poor overall survival.89 We found that SOCS1 mutations occurred only in non-relapsing primary DLBCL (5/20 cases), were absent in relapsing primary DLBCL (0/20),21 and were associated with excellent progression-free survival in uniformly treated (R-CHOP-14) cohort of patients with primary DLBCL.90 NF-κB pathway activation is well known in the pathogenesis of aggressive B-cell lymphomas, especially activated B-cell DLBCL. Two independent studies found identical 4 bp deletions of the NFKBIE gene that encodes an inhibitor of NF-κB signaling, IκBε, in three relapsing DLBCL and two PMBCL relapses.10,83 The same mutation was found by the authors in two additional cases of untreated DLBCL and was previously reported to occur in chronic lymphocytic leukemia, classical Hodgkin lymphoma and PMBCL.91–93 Additional NF-κB pathway-related genes shown to be enriched in relapses include NFKBIA and NFKBIZ, encoding IκBα and IκBζ, both of which have previously been implicated in the molecular pathogenesis of classical Hodgkin lymphoma and DLBCL, respectively.94,95 Inactivating aberrations (loss-of-function mutations and gene deletions) of TP53 and CDKN2A were also enriched in DLBCL relapses and are thought to be associated with resistance to chemotherapy.83 Defects in these two genes as well as gain-of-function mutations of NOTCH1 and MYC are most frequently implicated in Richter transformation, as excellently reviewed elsewhere.96 It has recently been demonstrated that, in addition to genetic mechanisms, changes of the epigenetic landscape can also play a role in DLBCL relapse.77,97 Pan et al. found that, compared to the primary tumors, relapses have significantly decreased intra-tumoral methylation heterogeneity.77 Furthermore, they showed that greater intratumoral methylation heterogeneity of primary tumors correlated with increased probability of relapse. Enrichment analysis of genes located in the differentially methylated regions between primary tumors and relapses suggested several deregulated pathways, such as antiapoptotic and tumor necrosis factor activity. In particular, genes associated with transforming growth factor-β

References 1. Okosun J, Cwynarski K. Non-Hodgkin Lymphoma: High Grade. In: Postgraduate Haematology. Oxford, UK: John Wiley & Sons, Ltd; 2015. p. 631–650. 2. 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. 3. Van Den Neste E, Schmitz N, Mounier N, et al. Outcome of patients with relapsed dif-

haematologica | 2017; 102(7)

receptor activity were overrepresented among the hypomethylated genes at relapse. Methylation-dependent transforming growth factor-β regulation has previously been implicated in DLBCL chemoresistance; Clozel et al. showed that SMAD1, a regulator of the transforming growth factor-β pathway, is more frequently hypermethylated in treatment-resistant DLBCL, and that demethylating agents administered prior to chemotherapy sensitize DLBCL and result in increased cytotoxic activity.97 Overall, recurrently altered genes and pathways in relapses overlap significantly with those implicated in the primary pathogenesis of aggressive B-cell lymphomas. It is, therefore, unclear whether they truly confer relapsespecific phenomena such as treatment resistance or merely reflect means of tumor clones to gain fitness. With regard to treatment resistance, it is thought to occur via two main mechanisms: (i) decreased drug accumulation or increased drug inactivation or (ii) biological pathways that evade cell death from applied agents; the majority of genetic alterations and pathways described in this review point towards the latter mechanism in DLBCL. To sum up, it is evident that considerable progress has been made in the last 5 years toward understanding relapses of aggressive B-cell lymphomas. Important insights have been gained from numerous studies examining different aspects of recurrences, from patterns of tumor evolution to recurrent genetic and epigenetic alterations. However, the substantial inter- and intra-tumoral heterogeneity, small study groups and varied study designs mean that we are just beginning to uncover the underlying mechanisms. Since there is an obvious lack of sufficient numbers of suitable specimens to conduct large-scale comprehensive studies on human samples, it could be worthwhile to pursue alternative strategies. One potential opportunity would be to conduct comprehensive studies of selected single cases at multiple, currently accessible levels (i.e. genome, epigenome, transcriptome, proteome, metabolome) in an attempt to develop a state-of-the-art individual tumor/tumor-pair molecular profile at great depth to identify actionable alterations. Another complementary strategy could be based on modeling DLBCL relapses in currently available or newly established animal models,98,99 in which potential relapse-causing events might be identified, functionally tested and subsequently screened by focused assays on human sample cohorts.

fuse large B-cell lymphoma who fail secondline salvage regimens in the International CORAL study. Bone Marrow Transplant. 2016;51(1):51–57. 4. Philip T, Guglielmi C, Hagenbeek A, et al. Autologous bone marrow transplantation as compared with salvage chemotherapy in relapses of chemotherapy-sensitive nonHodgkin’s lymphoma. N Engl J Med. 1995;333(23):1540–1545. 5. Friedberg JW. Relapsed/refractory diffuse large B-cell lymphoma. Hematology. 2011;2011(1):498–505. 6. Gisselbrecht C, Glass B, Mounier N, et al. Salvage regimens with autologous transplantation for relapsed large B-cell lymphoma in the rituximab era. J Clin Oncol. 2010;28(27):4184–4190.

7. Jain S, Shah N, Gregory S. Relapsed diffuse large B-cell lymphoma--10 years later. Clin Adv Hematol Oncol. 2011;9(9):704–708. 8. Asi K. Dominant mechanisms of primary resistance differ from dominant mechanisms of secondary resistance to targeted therapies. Crit Rev Oncol Hematol. 2016;97:178–196. 9. Gupta PB, Chaffer CL, Weinberg RA. Cancer stem cells: mirage or reality? Nat Med. 2009;15(9):1010–1012. 10. Morin RD, Assouline S, Alcaide M, et al. Genetic landscapes of relapsed and refractory diffuse large B-cell lymphomas. Clin Cancer Res. 2016;22(9):2290–2300. 11. Jiang Y, Redmond D, Nie K, et al. Deepsequencing reveals clonal evolution patterns and mutation events associated with relapse

1147

D. Juskevicius et al.

12.

13.

14.

15. 16.

17. 18.

19. 20.

21.

22.

23. 24.

25.

26.

27. 28.

29.

30.

1148

in B-cell lymphomas. Genome Biol. 2014;15(8):432. Kreso A, O’Brien CA, van Galen P, et al. Variable clonal repopulation dynamics influence chemotherapy response in colorectal cancer. Science. 2013;339(6119):543–548. Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, Thiele JWV. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues, Fourth Edition. IARC Press Lyon. 2008. Alizadeh AA, Aranda V, Bardelli A, et al. Toward understanding and exploiting tumor heterogeneity. Nat Med. 2015;21(8): 846–853. Alt FW, Oltz EM, Young F, et al. VDJ recombination. Immunol Today. 1992;13 (8):306–314. Gazzola A, Mannu C, Rossi M, et al. The evolution of clonality testing in the diagnosis and monitoring of hematological malignancies. Ther Adv Hematol. 2014;5(2): 35–47. Coiffier B. Treatment of aggressive nonHodgkin’s lymphoma. Semin Oncol. 1999;26(5 Suppl 14):12–20. Shioyama Y, Nakamura K, Kunitake N, et al. Relapsed non-Hodgkin’s lymphoma: detection and treatment. Radiat Med. 2000;18(6):369–375. Cappelaere P. Secondary non-Hodgkin’s lymphomas. Bull Cancer. 1998;85(3): 217–231. Geurts-Giele WR, Wolvers-Tettero IL, Dinjens WN, Lam KH, Langerak AW. Successive B-cell lymphomas mostly reflect recurrences rather than unrelated primary lymphomas. Am J Clin Pathol. 2013;140(1): 114–126. Juskevicius D, Lorber T, Gsponer J, et al. Distinct genetic evolution patterns of relapsing diffuse large B-cell lymphoma revealed by genome-wide copy number aberration and targeted sequencing analysis. Leukemia. 2016;30(12):2385–2395. de Jong D, Glas AM, Boerrigter L, et al. Very late relapse in diffuse large B-cell lymphoma represents clonally related disease and is marked by germinal center cell features. Blood. 2003;102(1):324–327. Lee SE, Kang SY, Yoo HY, et al. Clonal relationships in recurrent B-cell lymphomas. Oncotarget. 2016;7(11):12359–12371. Mao Z, Quintanilla-Martinez L, Raffeld M, et al. IgVH mutational status and clonality analysis of Richter’s transformation: diffuse large B-cell lymphoma and Hodgkin lymphoma in association with B-cell chronic lymphocytic leukemia (B-CLL) represent 2 different pathways of disease evolution. Am J Surg Pathol. 2007;31(10):1605–1614. Rossi D, Spina V, Deambrogi C, et al. The genetics of Richter syndrome reveals disease heterogeneity and predicts survival after transformation. Blood. 2011;117(12): 3391– 3401. Obermann EC, Mueller N, Rufle A, et al. Clonal relationship of classical Hodgkin lymphoma and its recurrences. Clin Cancer Res. 2011;17(16):5268–5274. Pasqualucci L, Khiabanian H, Fangazio M, et al. Genetics of follicular lymphoma transformation. Cell Rep. 2014;6(1):130–140. Okosun J, Bödör C, Wang J, et al. Integrated genomic analysis identifies recurrent mutations and evolution patterns driving the initiation and progression of follicular lymphoma. Nat Genet. 2014;46(2):176–181. Matolcsy A, Casali P, Warnke RA, Knowles DM. Morphologic transformation of follicular lymphoma is associated with somatic mutation of the translocated Bcl-2 gene. Blood. 1996;88(10):3937–3944. SEER Stat Fact Sheets: Non-Hodgkin

31.

32.

33.

34.

35. 36.

37.

38.

39.

40.

41.

42.

43.

44.

45. 46.

Lymphoma [Internet]. 2016 [cited 2017 Jan 2]. Available from: https://seer.cancer.gov/ statfacts/html/nhl.html Aricò M, Mussolin L, Carraro E, et al. NonHodgkin lymphoma in children with an associated inherited condition: a retrospective analysis of the Associazione Italiana Ematologia Oncologia Pediatrica (AIEOP). Pediatr Blood Cancer. 2015;62(10): 1782– 1789. Wang SS, Slager SL, Brennan P, et al. Family history of hematopoietic malignancies and risk of non-Hodgkin lymphoma (NHL): a pooled analysis of 10 211 cases and 11 905 controls from the International Lymphoma Epidemiology Consortium (InterLymph). Blood. 2007;109(8):3479–3488. Vannata B, Arcaini L, Zucca E. Hepatitis C virus-associated B-cell non-Hodgkin’s lymphomas: what do we know? Ther Adv Hematol. 2016;7(2):94–107. Bagg A, Dunphy CH. Immunosuppressive and immunomodulatory therapy-associated lymphoproliferative disorders. Semin Diagn Pathol. 2013;30(2):102–112. Ponce RA, Gelzleichter T, Haggerty HG, et al. Immunomodulation and lymphoma in humans. J Immunotoxicol. 2014;11(1):1–12. Kramer S, Hikel SM, Adams K, Hinds D, Moon K. Current status of the epidemiologic evidence linking polychlorinated biphenyls and non-Hodgkin lymphoma, and the role of immune dysregulation. Environ Health Perspect. 2012;120(8): 1067–1075. Alicandro G, Rota M, Boffetta P, La Vecchia C. Occupational exposure to polycyclic aromatic hydrocarbons and lymphatic and hematopoietic neoplasms: a systematic review and meta-analysis of cohort studies. Arch Toxicol. 2016;90(11):2643–2656. Merhi M, Raynal H, Cahuzac E, et al. Occupational exposure to pesticides and risk of hematopoietic cancers: meta-analysis of case-control studies. Cancer Causes Control. 2007;18(10):1209–1226. Filipovich AH, Mathur A, Kamat D, Shapiro RS. Primary immunodeficiencies: genetic risk factors for lymphoma. Cancer Res. 1992;52(19 Suppl):5465s–5467s. Brimo F, Michel RP, Khetani K, Auger M. Primary effusion lymphoma: a series of 4 cases and review of the literature with emphasis on cytomorphologic and immunocytochemical differential diagnosis. Cancer. 2007;111(4):224–233. Campana S, Corradini P, Astolfi M, et al. Analysis of the immunoglobulin heavychain gene rearrangement providing molecular evidence of second lymphoma in a patient in apparent relapse after autotransplantation. Bone Marrow Transplant. 1997;20(4):341–343. Lossos A, Ashhab Y, Sverdlin E, et al. Latedelayed cerebral involvement in systemic non-hodgkin lymphoma: a second primary tumor or a tardy recurrence? Cancer. 2004;101:1843–1849. Menter T, Juskevicius D, Alikian M, et al. Mutational landscape of B-cell post-transplant lymphoproliferative disorders. Br J Haematol. [Epub ahead of print] 2017. Forbes SA, Beare D, Gunasekaran P, et al. COSMIC: exploring the world’s knowledge of somatic mutations in human cancer. Nucleic Acids Res. 2015;43(Database issue):D805–D811. Skibola CF, Curry JD, Nieters A. Genetic susceptibility to lymphoma. Haematologica. 2007;92(7):960–969. Steensma DP, Bejar R, Jaiswal S, et al. Clonal hematopoiesis of indeterminate potential

47.

48.

49.

50.

51.

52.

53.

54. 55.

56.

57.

58.

59.

60.

61.

and its distinction from myelodysplastic syndromes. Blood. 2015;126(1):9–16. Kikushige Y, Ishikawa F, Miyamoto T, et al. Self-renewing hematopoietic stem cell is the primary target in pathogenesis of human chronic lymphocytic leukemia. Cancer Cell. 2011;20(2):246–259. Xiao W, Chen WW, Sorbara L, et al. Hodgkin lymphoma variant of Richter transformation: morphology, Epstein-Barr virus status, clonality, and survival analysis-with comparison to Hodgkin-like lesion. Hum Pathol. 2016;55:108–116. van Dongen JJ, Langerak AW, Bruggemann M, et al. Design and standardization of PCR primers and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936. Leukemia. 2003;17(12):2257–2317. Meier VS, Rufle A, Gudat F. Simultaneous evaluation of T- and B-cell clonality, t(11;14) and t(14;18), in a single reaction by a fourcolor multiplex polymerase chain reaction assay and automated high-resolution fragment analysis: a method for the rapid molecular diagnosis of lymphoproliferative disorders applicable to fresh frozen and formalinfixed, paraffin-embedded tissues, blood, and bone marrow aspirates. Am J Pathol. 2001;159(6):2031–2043. Langerak AW, Groenen PJTA, Brüggemann M, et al. EuroClonality/BIOMED-2 guidelines for interpretation and reporting of Ig/TCR clonality testing in suspected lymphoproliferations. Leukemia. 2012;26(10): 2159–2171. Langerak AW, van Dongen JJM. Multiple clonal Ig/TCR products: implications for interpretation of clonality findings. J Hematop. 2012;5(1–2):35–43. Boyd SD, Marshall EL, Merker JD, et al. Measurement and clinical monitoring of human lymphocyte clonality by massively parallel VDJ pyrosequencing. Sci Transl Med. 2009;1(12):12ra23. Robins H. Detecting and monitoring lymphoma with high-throughput sequencing. Oncotarget. 2011;2(4):287–288. He J, Wu J, Jiao Y, et al. IgH gene rearrangements as plasma biomarkers in nonHodgkin’s lymphoma patients. Oncotarget. 2011;2(3):178–185. Georgiou G, Ippolito GC, Beausang J, et al. The promise and challenge of high-throughput sequencing of the antibody repertoire. Nat Biotechnol. 2014;32(2):158–168. Appenzeller S, Gilissen C, Rijntjes J, et al. Immunoglobulin rearrangement analysis from multiple lesions in the same patient using next-generation sequencing. Histopathology. 2015;67(6):843–858. Carlotti E, Wrench D, Rosignoli G, et al. High Throughput sequencing analysis of the immunoglobulin heavy chain gene from flow-sorted B cell sub-populations define the dynamics of follicular lymphoma clonal evolution. PLoS One. 2015;10(9):e0134833. 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. Turchaninova MA, Davydov A, Britanova O V, et al. High-quality full-length immunoglobulin profiling with unique molecular barcoding. Nat Protoc. 2016;11(9): 1599–1616. Khan TA, Friedensohn S, de Vries ARG, et al. Accurate and predictive antibody repertoire profiling by molecular amplification finger-

haematologica | 2017; 102(7)

Genetics of aggressive B-cell lymphoma relapses

62.

63.

64.

65.

66.

67.

68. 69.

70.

71.

72.

73.

74.

printing. Sci Adv. 2016;2(3):e1501371– e1501371. Friedensohn S, Khan TA, Reddy ST. Advanced methodologies in high-throughput sequencing of immune repertoires. Trends Biotechnol. 2017;35(3): 203–214. Bolotin DA, Poslavsky S, Mitrophanov I, et al. MiXCR: software for comprehensive adaptive immunity profiling. Nat Methods. 2015;12(5):380–381. Gupta NT, Vander Heiden JA, Uduman M, et al. Change-O: a toolkit for analyzing large-scale B cell immunoglobulin repertoire sequencing data. Bioinformatics. 2015;31(20):3356–3358. Barak M, Zuckerman NS, Edelman H, Unger R, Mehr R. IgTree: creating immunoglobulin variable region gene lineage trees. J Immunol Methods. 2008;338(1–2):67–74. Kuchenbecker L, Nienen M, Hecht J, et al. IMSEQ--a fast and error aware approach to immunogenetic sequence analysis. Bioinformatics. 2015;31(18):2963–2971. Chen YJ, Yeh SH, Chen JT, et al. Chromosomal changes and clonality relationship between primary and recurrent hepatocellular carcinoma. Gastroenterology. 2000;119(2):431–440. Aparicio S, Caldas C. The implications of clonal genome evolution for cancer medicine. N Engl J Med. 2013;368(9):842–851. Beà S, Valdés-Mas R, Navarro A, et al. Landscape of somatic mutations and clonal evolution in mantle cell lymphoma. Proc Natl Acad Sci USA. 2013;110(45): 18250– 18255. Wu C, de Miranda NF, Chen L, et al. Genetic heterogeneity in primary and relapsed mantle cell lymphomas: impact of recurrent CARD11 mutations. Oncotarget. 2016;7(25):38180–38190. Melchardt T, Hufnagl C, Weinstock DM, et al. Clonal evolution in relapsed and refractory diffuse large B-cell lymphoma is characterized by high dynamics of subclones. Oncotarget. 2016;7(32):51494–51502. Aukema SM, Theil L, Rohde M, et al. Sequential karyotyping in Burkitt lymphoma reveals a linear clonal evolution with increase in karyotype complexity and a high frequency of recurrent secondary aberrations. Br J Haematol. 2015;170(6):814–825. Ruminy P, Jardin F, Picquenot J-M, et al. S mutation patterns suggest different progression pathways in follicular lymphoma: early direct or late from FL progenitor cells. Blood. 2008;112(5):1951–1959. Green MR, Gentles AJ, Nair R V, et al. Hierarchy in somatic mutations arising during genomic evolution and progression of

haematologica | 2017; 102(7)

75.

76.

77. 78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

follicular lymphoma. Blood. 2013;121(9): 1604–1611. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714–726. Morin RD, Mungall K, Pleasance E, et al. Mutational and structural analysis of diffuse large B-cell lymphoma using wholegenome sequencing. Blood. 2013;122(7): 1256–1265. Pan H, Jiang Y, Boi M, et al. Epigenomic evolution in diffuse large B-cell lymphomas. Nat Commun. 2015;6:6921. Roschewski M, Staudt LM, Wilson WH. Diffuse large B-cell lymphoma-treatment approaches in the molecular era. Nat Rev Clin Oncol. 2014;11(1):12–23. Bozic I, Nowak MA. Timing and heterogeneity of mutations associated with drug resistance in metastatic cancers. Proc Natl Acad Sci USA. 2014;111(45):15964–15968. Woyach JA, Furman RR, Liu T-M, et al. Resistance mechanisms for the Bruton’s tyrosine kinase inhibitor ibrutinib. N Engl J Med. 2014;370(24):2286–2294. Davis RE, Ngo VN, Lenz G, et al. Chronic active B-cell-receptor signalling in diffuse large B-cell lymphoma. Nature. 2010;463(7277):88–92. Kuo H-P, Ezell SA, Hsieh S, et al. The role of PIM1 in the ibrutinib-resistant ABC subtype of diffuse large B-cell lymphoma. Am J Cancer Res. 2016;6(11):2489–2501. Mareschal S, Dubois S, Viailly P-J, et al. Whole exome sequencing of relapsed/refractory patients expands the repertoire of somatic mutations in diffuse large B-cell lymphoma. Genes Chromosomes Cancer. 2016;55(3):251–267. Altomonte M, Gloghini A, Bertola G, et al. Differential expression of cell adhesion molecules CD54/CD11a and CD58/CD2 by human melanoma cells and functional role in their interaction with cytotoxic cells. Cancer Res. 1993;53(14):3343–3348. Challa-Malladi M, Lieu YK, Califano O, et al. Combined genetic inactivation of β2microglobulin and CD58 reveals frequent escape from immune recognition in diffuse large B cell lymphoma. Cancer Cell. 2011;20(6):728–740. Ritz O, Guiter C, Castellano F, et al. Recurrent mutations of the STAT6 DNA binding domain in primary mediastinal Bcell lymphoma. Blood. 2009;114(6): 1236– 1242. Yildiz M, Li H, Bernard D, et al. Activating STAT6 mutations in follicular lymphoma. Blood. 2015;125(4):668–679.

88. Dubois S, Viailly P-J, Mareschal S, et al. Next-generation sequencing in diffuse large B-cell lymphoma highlights molecular divergence and therapeutic opportunities: a LYSA study. Clin Cancer Res. 2016;22(12): 2919– 2928. 89. Schif B, Lennerz JK, Kohler CW, et al. SOCS1 mutation subtypes predict divergent outcomes in diffuse large B-cell lymphoma (DLBCL) patients. Oncotarget. 2013;4(1): 35–47. 90. Juskevicius D, Jucker D, Klingbiel D, et al. Mutations of CREBBP and SOCS1 are independent prognostic factors in diffuse large B cell lymphoma: mutational analysis of the SAKK 38/07 prospective clinical trial cohort. J Hematol Oncol. 2017;10(1):70. 91. Emmerich F, Theurich S, Hummel M, et al. Inactivating I kappa B epsilon mutations in Hodgkin/Reed-Sternberg cells. J Pathol. 2003;201(3):413–420. 92. Mansouri L, Sutton LA, Ljungstrom V, et al. Functional loss of IkappaBepsilon leads to NF-kappaB deregulation in aggressive chronic lymphocytic leukemia. J Exp Med. 2015;212(6):833–843. 93. Mansouri L, Noerenberg D, Young E, et al. Frequent NFKBIE deletions are associated with poor outcome in primary mediastinal B-cell lymphoma. Blood. 2016;128(23): 2666–2670. 94. Lake A, Shield LA, Cordano P, et al. Mutations of NFKBIA, encoding IkappaB alpha, are a recurrent finding in classical Hodgkin lymphoma but are not a unifying feature of non-EBV-associated cases. Int J Cancer. 2009;125(6):1334–1342. 95. 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. 96. Rossi D, Gaidano G. Richter syndrome: pathogenesis and management. Semin Oncol. 2016;43(2):311–319. 97. Clozel T, Yang SN, Elstrom RL, et al. Mechanism-based epigenetic chemosensitization therapy of diffuse large B-cell lymphoma. Cancer Discov. 2013;3(9): 1002– 1019. 98. Calado DP, Zhang B, Srinivasan L, et al. Constitutive canonical NF-κB activation cooperates with disruption of BLIMP1 in the pathogenesis of activated B cell-like diffuse large cell lymphoma. Cancer Cell. 2010;18(6):580–589. 99. Mandelbaum J, Bhagat G, Tang H, et al. BLIMP1 is a tumor suppressor gene frequently disrupted in activated B cell-like diffuse large B cell lymphoma. Cancer Cell. 2010;18(6):568–579.

1149

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Hematopoiesis

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1150-1160

A new path to platelet production through matrix sensing

Vittorio Abbonante,1,2 Christian Andrea Di Buduo,1,2 Cristian Gruppi,1,2 Carmelo De Maria,3 Elise Spedden,4 Aurora De Acutis,3 Cristian Staii,4 Mario Raspanti,5 Giovanni Vozzi,3 David L. Kaplan,6 Francesco Moccia,7 Katya Ravid8 and Alessandra Balduini1,2,6

1 Department of Molecular Medicine, University of Pavia, Italy; 2Laboratory of Biotechnology, IRCCS San Matteo Foundation, Pavia, Italy; 3Interdepartmental Research Center “E. Piaggio”, University of Pisa, Italy; 4Department of Physics and Astronomy, Tufts University, Medford, MA, USA; 5Department of Surgical and Morphological Sciences, University of Insubria, Varese, Italy; 6Department of Biomedical Engineering, Tufts University, Medford, MA, USA; 7Department of Biology and Biotechnology “Lazzaro Spallanzani”, University of Pavia, Italy and 8Department of Medicine and Whitaker Cardiovascular Institute, Boston University School of Medicine, MA, USA

ABSTRACT

Correspondence: alessandra.balduini@unipv.it

Received: December 2, 2016. Accepted: April 11, 2017. Pre-published: April 14, 2017.

egakaryocytes (MK) in the bone marrow (BM) are immersed in a network of extracellular matrix components that regulates platelet release into the circulation. Combining biological and bioengineering approaches, we found that the activation of transient receptor potential cation channel subfamily V member 4 (TRPV4), a mechano-sensitive ion channel, is induced upon MK adhesion on softer matrices. This response promoted platelet production by triggering a cascade of events that lead to calcium influx, β1 integrin activation and internalization, and Akt phosphorylation, responses not found on stiffer matrices. Lysyl oxidase (LOX) is a physiological modulator of BM matrix stiffness via collagen crosslinking. In vivo inhibition of LOX and consequent matrix softening lead to TRPV4 activation cascade and increased platelet levels. At the same time, in vitro proplatelet formation was reduced on a recombinant enzyme-mediated stiffer collagen. These results suggest a novel mechanism by which MKs, through TRPV4, sense extracellular matrix environmental rigidity and release platelets accordingly. Introduction

doi:10.3324/haematol.2016.161562 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1150 ©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.

1150

Megakaryocytes (MKs) reside within the bone marrow (BM), where they mature to extend proplatelets, at the end of which newly formed platelets are assembled and released into the bloodstream.1-4 Different extracellular matrix components (ECM) in the BM actively regulate megakaryopoiesis.5-7 In earlier studies, it was demonstrated that type IV collagen sustains proplatelet formation (PPF), opposite to type I collagen, which is a fundamental negative regulator of PPF through integrin α2β1/Rho/ROCK axis engagement.8-12 In addition, recent studies highlighted the direct involvement of PI3K/Akt and MAPK/ERK pathways in PPF.1317 Interestingly, ECM component stiffness was shown to be inversely correlated to MK maturation and PPF.8,10,18 Cell contact with ECM components occurs through integrins which transduce the signals from the ECM to the cell cytoskeleton.19-21 However, to sense the extracellular environment, integrins act in concert with different mechano-sensitive ion channels.22,23 Among these, the transient receptor potential cation channel subfamily V member 4 (TRPV4) is a membrane mechano-sensitive ion channel whose activation has been linked to activation of β1 integrin a major collagen-binding integrin receptor subunit.24-26 Interestingly, TRPV4 activity has been demonstrated to modulate PI3K/Akt and MAPK/ERK pathways in endothelial cells upon physical stimuli applied to the cell membrane.24 In the current study, we demonstrate a new mechanism by which MKs sense the environmental mechanics to regulate their haematologica | 2017; 102(7)

A new path to platelet production through matrix sensing

maturation and platelet production. On soft collagenous substrates, TRPV4 expressed on the MK surface was activated, inducing calcium influx, β1 integrin activation and internalization, with consequent Akt phosphorylation and proplatelet formation.

Methods In vitro megakaryocyte cultures Human CD34+ cells were isolated, separated and cultured, as described previously.27,28 All human samples were collected in accordance with the ethical committee of the IRCCS Policlinico San Matteo Foundation, Pavia, Italy, and the principles of the Declaration of Helsinki. For collagen receptor inhibition, MKs at day 13 of culture were incubated with 10 μg/mL anti-β1 integrin blocking antibody (Millipore, clone P5D2), 10 μg/mL anti-GPVI blocking antibody (a kind gift of Prof. Jandrot Perrus) or with 200 nM (125 ng/mL) Discoidin Domain Receptor 1 (DDR1)-IN-1 (Tocris), a selective DDR1 tyrosine kinase inhibitor, for one hour prior to being plated on type I or type IV collagen for three hours (h). For Akt inhibition experiments, MKs at day 13 of culture were treated with 10 μM Akt1/2 inhibitor (Sigma Aldrich) for 30 minutes (min) and then plated on collagens for 16 h for PPT evaluation. For treatment with the TRPV4 inhibitors (RN-1734, HC067047, Sigma Aldrich), MKs at day 13 of culture were incubated with vehicle or 10 μM of the indicated TRPV4 inhibitor for 30 min prior to being plated on collagens for 3 or 16 h. For treatment with the TRPV4 agonist (GSK1016790A, Sigma Aldrich), MKs at day 13 of culture were incubated or not with 10 μM GSK1016790A for 10 min prior to being plated on collagens for 3 h.

overnight at 4°C. Please refer to the Online Supplementary Methods for technical details.

Elastic modulus determination via atomic force microscope Elastic modulus maps were taken on an Asylum Research MFP-3D atomic-force microscope (AFM) (Asylum Research) using AC240TS-R3 cantilevers (Asylum Research) with a nominal spring constant of 2 N/m. Please refer to the Online Supplementary Methods for technical details.

[Ca2+]i measurements Megakaryocytes at day 13 of culture were harvested and plated onto substrate-coated coverslips in 24-well plates (1x105 cells/well). After 60 min at 37°C and 5% CO2, MKs were loaded with 4 μM fura-2 AM in physiological salt solution (PSS) for an additional 30 min and analyzed as previously described.17 Please refer to the Online Supplementary Methods for technical details.

Animals and in vivo treatment C57/BL6 wild-type mice were obtained from the Charles River Laboratories. Mice were housed at the animal facility of the Department of Physiology, section of General Physiology, University of Pavia (approval ref. n. 1302/2015). All mice were sacrificed according to the current European legal animal practice requirements. In vivo lysyl oxidase (LOX) inhibition by βaminopropionitrile (BAPN) treatment was carried out using a protocol modified from previous studies.35 Please refer to the Online Supplementary Methods for technical details.

In vivo bone marrow stiffness Evaluation of megakaryocyte spreading and proplatelet formation Evaluation of MK spreading and PPF onto collagens was performed as previously described.29 Please refer to the Online Supplementary Methods for technical details.

The mechanical properties of femoral BM from mouse treated with BAPN and controls were measured with a Zwick/Roell Z005 testing device (Zwick GmbH & Co.) equipped with a 10 N load cell as previously described.36 Please refer to the Online Supplementary Methods for technical details.

Immunoprecipitation and Western blotting

Flow cytometry

For immunoprecipitation and Western blotting analysis, cultured MKs and primary BM immunomagnetically-sorted MKs (CD41+; Biolegend) were collected, washed twice at 4°C, and lysed as previously described.30 For active β1 integrin staining, samples were not reduced. Please refer to the Online Supplementary Methods for technical details.

For the analysis of pAkt in BM MKs, femurs were flushed, red blood cells lysed with a 0.8% ammonium chloride solution, the remaining cells were washed by centrifugation with PBS and stained with anti-pAkt antibody (20 μL per test; BD Pharmingen) and anti-CD41 antibody (0.1 mg/mL; Biolegend) following the manufacturer's instructions. All MK samples, from the different tested sources, were routinely characterized as CD41+CD42b+ and CD3–CD4–CD8–CD11b–CD19–CD33– cells, using appropriate antibodies (Beckman Coulter Inc.). Please refer to the Online Supplementary Methods for technical details.

Immunofluorescence microscopy For cell immunofluorescence staining, 1x105 MKs at day 13 of culture were harvested and plated on collagen-coated coverslips. Adhering cells were washed, fixed, permeabilzed, and stained as previously described.29 Please refer to the Online Supplementary Methods for technical details.

Internalization assays For immunofluorescence internalization assay, MKs were treated as living cells with 15 μg/mL of the anti-β1 integrin mAb (Abcam) and acid washed before fixation as previously described.31,32 MKs were then fixed, permeabilized, and stained with the appropriate secondary antibody. Western blot internalization was evaluated as previously described.33 Please refer to the Online Supplementary Methods for technical details.

Silk film fabrication Silk films were produced as previously described.34 Type I or type IV collagen (25 μg/mL) were coated to the film surface haematologica | 2017; 102(7)

Bone marrow explant analysis Bone marrow explants were isolated and analyzed based on a previous protocol.37 Please refer to the Online Supplementary Methods for technical details.

Statistical analysis Values are expressed as mean±standard deviation (SD). t-test was used to analyze experiments. P<0.05 was considered statistically significant. All experiments were independently repeated at least three times. For RT-PCR and quantitative real-time PCR, tissue collection and immunohistochemistry, reticulated platelet analysis and LOX-mediated collagen crosslinking please refer to the Online Supplementary Appendix. 1151

V. Abbonante et al.

Results Type I and type IV collagen differently impact megakaryocyte adhesion To evaluate the impact of collagen stiffness on MK behavior, coverslips were coated with type I or type IV collagens and Young's modulus was evaluated by AFMbased nano-indentation.10,29 As we have previously demonstrated, type I collagen fibrils presented an elasticity more than 150 megapascal units (MPa), whereas material in early stage of self-aggregation had values in the order of 5-75 Mpa.10 On the contrary, more homogeneous values in the order of less than 10 MPa were found during the analysis of type IV collagen Young's moduli. Consistently, nano-indentation studies on live cells plated on the two different collagens after 3 h of adhesion revealed values ranging from 290 to 3600 Pa (weighted average 1036 Pa) in MKs plated on type I collagen, while values ranging from 300 to 1400 Pa (weighted average 667 Pa) were obtained with MKs plated on type IV collagen (Figure 1A). As a result of the different stiffness, MKs remained spread over a 16-h incubation on type I collagen, while they rearranged their cytoskeleton and extended proplatelets on type IV collagen (Figure 1B and C).

Megakaryocyte β1 integrin activation and internalization vary depending on the type of collagen

We hypothesized that specific collagen receptors may be responsible for the MK behavior observed on type I and type IV collagens. qRT-PCR analysis revealed that the most expressed collagen-interacting integrin domains were β1>α2, with DDR1 and GPVI expressed at a similar

level as α2 (Figure 2A). The most important reduction in Mk adhesion was obtained inhibiting β1 integrin, while DDR1 and GPVI inhibition did not significantly affect MK adhesion relative to untreated controls (Figure 2A).10,30 β1 integrin activation was studied in MKs plated on type I and type IV collagens at the 3 time points used for functional studies (3, 8 and 16 h), by employing a monoclonal antibody directed against epitopes in the 355-425 region (hybrid domain), whose expression reflects the activity of β1 integrin.38 An increase of β1 integrin activation was observed on type IV collagen, compared to type I collagen, after 3 and 8 h of adhesion, reaching a similar value under both conditions after 16 h (Figure 2B). Consistently, internalization assays showed that membrane distributed β1 integrin was reduced on MKs plated on type IV collagen as compared to type I collagen (Figure 2C).

Substrate stiffness modulates β1 integrin dynamics

To prove that β1 integrin activation and internalization were determined by collagen substrate stiffness, silk fibroin films were fabricated with imposed elasticity and were later coated with 25 μg/mL type I or type IV collagens.34 Silk films with two different ranges of elasticity were chosen on the basis of previous results: a softer silk film (≤10 MPa) was chosen as the optimum condition to maximize MK function, while a stiffer silk film (≥90 MPa) was chosen as a condition proven to decrease PPF.34 Importantly, it has been demonstrated that there are no specific cell-binding epitopes on the silk that would bias the outcomes.39 MKs extended proplatelets after 16 h of adhesion on both collagens coated on soft films, while they maintained the spread form, independently of the

Figure 1. Type I and type IV collagens differently regulate megakaryocyte (MK) adhesion. (A) Distribution of Young’s modulus values of live MKs plated on type I and type IV collagens (COLL). Three independent experiments were performed. (B) Human mature MKs, 1x105 per well in 24-well plates, were plated on type I and type IV collagen-coated coverslips for three different times. Cells were fixed and stained with anti-β1 tubulin (green) and tetramethylrhodamine (TRITC)-Phalloidin (red). Nuclei were counterstained with Hoechst (blue). Scale bar 50 μm; box scale bar 20 μm. (C) MKs exhibiting stress fibers were counted and presented as percentage of spread MK, while MKs extending proplatelets were counted and shown as percentage of proplatelet bearing MKs. Results shown in (C) are mean±standard deviation (SD) of six independent experiments. *P<0.05. Pa: Pascal units; h: hours.

1152

haematologica | 2017; 102(7)

A new path to platelet production through matrix sensing

collagen type, if plated on stiffer films (Figure 3A). Lower stiffness promoted higher activation of the β1 integrin in MKs plated on both collagen-type coated silk films, as demonstrated by western blot and immunofluorescence staining (Figure 3B and Online Supplementary Figure S1), with increased β1 integrin internalization (Figure 3C).

PI3K/Akt-dependent proplatelet formation on soft collagen substrate To understand the involvement of PI3K/Akt and MAPK/ERK signaling pathways in mediating the effect of soft collagen, MKs plated on type I and type IV collagen were lysed after 3, 8 and 16 h. Western blot analysis demonstrated an important difference in Akt and ERK activation (Figure 4A). Specifically, ERK phosphorylation peaked after 3 h on type I collagen, remaining higher up to 16 h compared to type IV collagen, while Akt phosphorylation appeared constantly higher in MKs plated on type IV compared to type I collagen. Considering favorable PPF on type IV collagen, these results suggested a role for Akt

haematologica | 2017; 102(7)

phosphorylation as a positive mediator of PPF. Consistently, treatment of MKs plated on type IV collagen with a specific Akt inhibitor resulted in a significant reduction of PPF (Figure 4A). The role of substrate elasticity in Akt and ERK phosphorylation was further explored by using the silk fibroin film approach. Akt phosphorylation increased on soft collagen substrates, while ERK phosphorylation was unaffected (Figure 4B). Finally, inhibition of Akt in MKs plated on type I collagen-coated soft films led to a reduction in PPF by approximately 80%, compared to untreated cells (Figure 4B). Accordingly, taking advantage of our system for platelet production through porous silk films,34 we showed that Akt inhibition significantly reduced the increase of platelet production obtained on soft collagen substrates (Online Supplementary Figure S2).

Megakaryocytes express functional mechano-sensitive TRPV4 ion channel It has been demonstrated that, in endothelial cells, upon mechanical stress, β1 integrin activation determines the

Figure 2. β1 integrin is strongly involved in mediating megakaryocyte (MK) adhesion on type I and type IV collagens. (A) Analysis of collagen (Coll) receptors. qRT-PCR analysis of different collagen receptors in human mature MKs [left; data are shown as mean±standard deviation (SD) of four independent experiments, relative to the housekeeping gene β2-microglobulin]. MK adhesion on type I and IV collagens was evaluated upon treatment with different collagen receptor blocking antibodies or blocking molecules (10 μg/mL anti-β1 integrin, 200 nM DDR1-IN-1, 10 μg/mL anti-GPVI) (right; data are shown as mean percentage of adhering MKs±SD of five independent experiments, relative to untreated cells). (B) Western blot analysis of active β1 integrin in MKs plated on type I and type IV collagens for different times. Total β1 integrin and actin were determined to show equal loading. The densitometry analysis of active β1 integrin level derived from three independent experiments is shown. (C) β1 integrin internalization assays. Western blot analysis of internalized β1 integrin after 1, 3 and 8 hours (h) of adhesion on type I and type IV collagens (top). To show equal loading actin was determined on total cell lysates. The first lane on the left represents the positive control of untreated MKs. The densitometry analysis of internalized β1 integrin level derived from four independent experiments is shown. Immunofluorescence analysis of internalized β1 integrin in MKs plated on type I and type IV collagens for 3 h (bottom). MKs were stained with antiβ1 integrin (green) and TRITCPhalloidin (red). Nuclei were counterstained with Hoechst (blue). Scale bar 10 μm. The densitometry analysis of staining intensities derived from four independent experiments is shown. A minimum of 20 cells per experiment were evaluated. Data of densitometry analysis are all expressed as mean±SD. *P<0.05. CTRL: untreated.

1153

V. Abbonante et al.

TRPV4 ion channel opening, which allows calcium influx inside the cell cytoplasm, supporting further β1 integrin engagement. Here, we show that human MKs expressed functional TRPV4 ion channels on their membrane (Figure 5A and B) and that, upon adhesion onto type IV collagen, human MKs stained with FURA2-AM display calcium oscillations of significant higher frequency and amplitude than those observed during adhesion on type I collagen. Inhibition of TRPV4 channels with a selective antagonist (RN1734) led to a significant decrease in the amplitude of calcium spikes only on type IV collagen, but not on type I collagen, thus suggesting that calcium flux through TRPV4 differs in MKs plated on the two collagen types (Figure 5C). The increased activation of TRPV4 on type IV collagen and on soft collagen substrates was also confirmed by immunoprecipitation of phosphorylated TRPV4 followed by western blot analysis (Online Supplementary Figure S3).40

β1 integrin activation, Akt phosphorylation and proplatelet formation on type IV collagen were sustained by TRPV4

To test whether TRPV4 channels mediate β1 integrin activation and Akt phosphorylation, both proteins were analyzed upon MK treatment with a specific TRPV4 inhibitor (RN-1734) on type IV collagen and with a specific TRPV4 agonist (GSK1016790A) on type I collagen.

Blocking TRPV4 activity with RN-1734 on type IV collagen resulted in a significant decrease of β1 integrin activation and Akt phosphorylation (Figure 6A). Consistently, the inhibition of TRPV4 activity resulted in diminished PPF on type IV collagen (Figure 6B), thus confirming the active role of the mechano-sensitive TRPV4/β1 integrin/Akt axis in PPF. On the contrary, forcing TRPV4 activation with GSK1016790A resulted in increased β1 integrin activation, Akt phosphorylation (Figure 6C) and β1 integrin internalization in MKs plated on type I collagen (Figure 6D).

Lysyl oxidase-mediated collagen cross-linking inhibition increases PPF and platelet production in vivo A key regulator of collagen stiffness in vivo is the secreted enzyme LOX, which by oxidative deamination of lysine residues on collagen, leads to cross-linked aldehydes, contributing to a tight and stiffer ECM.41 To support the impact of collagen substrate elasticity on platelet production in vivo, mice were treated with a specific LOX inhibitor, which has been demonstrated to reduce tissue stiffness.42,43 Inhibition of LOX, by β-aminopropionitrile (BAPN) treatment, an irreversible inhibitor of LOX, diminished the bone marrow elastic modulus (Figure 7A) and reduced collagen fiber dimension in cortical bone (Online

Figure 3. Collagen substrate elasticity regulates β1 integrin dynamics and proplatelet formation. (A) Megakaryocyte (MKs) extending proplatelets were counted and shown as percentage of proplatelet bearing MKs. Data are expressed as mean percentage±standard deviation (SD) of eight independent experiments. (B) Western blot analysis of active β1 integrin in MKs plated on type I and type IV collagen coated soft and stiff silk fibroin films. Total β1 integrin and actin were determined to show equal protein loading. The densitometry analysis of active β1 integrin level derived from four independent experiments is shown. (C) β1 integrin internalization assays. Western blot analysis of internalized β1 integrin after 1, 3 and 8 hours (h) of adhesion on type I collagen-coated soft and stiff silk fibroin films (left). Actin, in total cell lysates, was determined to show equal loading. The first lane on the left represents the positive control of untreated MKs. The densitometry analysis of internalized β1 integrin level derived from four independent experiments is shown. Immunofluorescence analysis of internalized β1 integrin in MKs plated on type I collagen-coated soft and stiff silk fibroin films for 3 h (right). Cells were stained with anti-β1 integrin (green) and TRITC-Phalloidin (red). Nuclei were counterstained with Hoechst (blue). Scale bar 10 μm. The densitometry analysis of staining intensities, derived from four independent experiments is shown. A minimum of 20 cells per experiment was evaluated. Data of densitometry analysis are all expressed as mean±SD. *P<0.05, **P<0.01. CTRL: untreated.

1154

haematologica | 2017; 102(7)

A new path to platelet production through matrix sensing

Supplementary Figure S4). The reduction of bone marrow stiffness led to an increase in peripheral platelet count, while the count of the other blood cells was unchanged (Figure 7B and Online Supplementary Table S1). The percentage of newly released reticulated platelets was significantly increased in treated mice, confirming that the boost in platelet numbers was mainly due to newly formed platelets (Figure 7B). Of note, the number of BM MKs was not significantly affected by BAPN treatment, thus demonstrating that the increase in peripheral blood platelet counts was not due to a higher MK number (Online Supplementary Figure S5). Interestingly, Akt phosphorylation was augmented in BAPN-treated bone marrow MKs as compared to untreated control (CTRL) by means of immunofluorescence imaging of bone marrow sections and flow cytometry analysis of flushed bone marrow (Figure 7C). To study PPF within bone marrow in LOX-inhibited and control mice, fresh BM explants were examined by videomicroscopy (Online Supplementary Figure S6). Through this approach, MKs, recognized by their morphology, became visible at the periphery of the explant as round cells or as cells extending thick protrusions or proplatelets. Importantly, TRPV4 activation was analyzed in MKs explanted from these scenarios by evaluating the extent of TRPV4 phosphorylation, which

enhances TRPV4-mediated calcium inflow and may be employed as surrogate to monitor TRPV4 activation when the direct measurement of calcium signals is not feasible.40 Diminishing the bone marrow stiffness resulted in an increase of TRPV4 activation in MKs as demonstrated by western blot analysis of immunoprecipitated TRPV4 (Figure 8A). Proplatelet forming MKs at 3 and 8 h from the beginning of the experiment were quantified. The number of visible MKs per experiment was comparable in both conditions (CTRL 51+9 vs. BAPN 58+13 at 3 h; CTRL 63+14 vs. BAPN 72+17 at 6 h), while the percentage of proplatelet-forming MKs was significantly higher in LOXinhibited mice (Figure 8B and Online Supplementary Videos 1 and 2). Importantly, Akt inhibition in BM explants, from BAPN-treated mice, reverted the increase in proplatelet formation to a level lower than CTRL mice (Figure 8B and Online Supplementary Video 3). Finally, to further explore the direct dependence of PPF on collagen substrate crosslinking, and thus rigidity, PPF in MKs plated on type IV collagen pre-treated with recombinant LOXL2 protein was tested in the presence or not of BAPN. The incubation of type IV collagen with LOXL2 significantly reduced the percentage of proplatelet-forming MKs, while inhibition of LOXL2 with BAPN rescued PPF to a level comparable to type IV collagen alone (Online Supplementary Figure S7).

Figure 4. PI3K/Akt signaling mediates proplatelet formation on soft collagen (Coll) substrates. (A) Western blot analysis of Akt and ERK phosphorylation (pAkt and pERK) in MKs plated for 3, 8 and 16 hours (h) on type I and type IV collagens (left). Total Akt, ERK and actin were determined to show the equal loading. The densitometry analysis of pAkt and pERK level derived from four independent experiments is shown. Proplatelet formation assay in MKs plated for 16 h on type IV collagen in the presence or not of 10 μM Akt inhibitor (right). MKs extending proplatelet were counted and shown as percentage of proplatelet-bearing MKs. Data are expressed as mean percentage±standard deviation (SD) of five independent experiments. (B) Western blot analysis of pAkt and pERK in MKs plated for 3 h on type I and type IV collagen-coated soft and stiff silk fibroin films (left). Total Akt, ERK and actin were determined to show equal loading. The densitometry analysis of pAkt and pERK level derived from five independent experiments is shown. Proplatelet formation assay in MKs plated for 16 h on type I collagen-coated soft silk fibroin films in presence or not of 10 μM Akt inhibitor (right). MKs extending proplatelets were counted and expressed as percentage of proplatelet bearing MKs. Data are expressed as mean percentage±SD of five independent experiments. Data of densitometry analysis are all expressed as mean±SD. *P<0.05. CTRL: untreated.

haematologica | 2017; 102(7)

1155

V. Abbonante et al. A

Figure 5. Megakaryocytes (MKs) express a functional TRPV4 ion channel. (A) TRPV4 expression in mature MKs. RT-PCR analysis of TRPV4 (left). Western blot analysis of TRPV4 (right). (B) Functional characterization of TRPV4. Analysis of calcium signaling in response to 10 nM GSK1016790A (TRPV4 agonist) in FURA2-AM loaded MKs, pre-treated or not, for 60 minutes (min), with 10 μM of the indicated TRPV4 inhibitor (RN-1734 or HC067047) (left). GSK1016790A was added at 100 seconds (sec) from the beginning of the analysis. Data are representative of three independent experiments. A minimum of 40 MKs per experiment were evaluated. In absence of extracellular calcium (0Ca2+), treatment with 10 nM GSK1016790A did not elicit significant changes in calcium signaling in FURA2-AM-loaded MKs (right). After extracellular calcium restoration, in the presence of GSK1016790A, the increase in MKs fluorescence was indicative of extracellular calcium entry in MKs cytosol. GSK1016790A was added after 300 sec from the beginning of the experiment. Data are representative of three independent experiments. A minimum of 40 MKs per experiment were evaluated. (C) TRPV4 activity on type I and IV collagens. Number of calcium peaks per MKs plated on type I and type IV collagens in presence or not of 10 μM RN-1734. Time of analysis was 800 sec. Calcium peak amplitude in megakaryocytes plated on type I and type IV collagens in presence or not of 10 μM RN-1734 (left). Representative calcium oscillations in MKs plated on type I and type IV collagens (right). RN-1734 was added at 600 sec from the beginning of the analysis. Data in (C) refer to eight independent experiments. Data are expressed as mean±standard deviation (SD). *P<0.05, **P<0.01.

Figure 6. TRPV4 activity regulates β1 integrin dynamics and Akt phosphorylation. (A) Western blot analysis of active β1 integrin and phosphorylated Akt (pAkt) in megakaryocytes (MKs) plated on type IV collagen in the presence or not of 10 μM RN-1734 (RN). Total β1 integrin, Akt and actin were determined to show equal loading. The densitometry analysis of active β1 integrin level and pAkt derived from four independent experiments is shown. (B) Proplatelet formation assay in MKs plated on type IV collagen in presence or not of 10 μM RN-1734. Proplatelet bearing MKs were counted and shown as percentages of proplatelet-bearing MKs. Data are shown as mean percentage±standard deviation (SD) of six independent experiments. (C) Western blot analysis of active β1 integrin and pAkt in MKs plated on type I collagen in the presence or not of 10 nM GSK1016790A (GSK). Total β1 integrin, Akt and actin were determined to show equal loading. The densitometry analysis of active β1 integrin level and pAkt, derived from four independent experiments, is shown. β1 integrin internalization assays. (D) Immunofluorescence analysis of internalized β1 integrin in MKs plated on type I collagen in presence or not of 10 nM GSK1016790A (left). Cells were stained with anti-β1 integrin (green) and TRITC-Phalloidin (red). Nuclei were counterstained with Hoechst (blue). Scale bar 5 μm. The densitometry analysis of staining intensities derived from three independent experiments is shown. A minimum of 20 cells per experiment were evaluated. Western blot analysis of internalized β1 integrin in megakaryocytes plated on type I collagen in presence or not of 10 nM GSK1016790A (right). To show equal loading, actin was determined on total cell lysates. The densitometry analysis of internalized β1 integrin level derived from three independent experiments is shown. Data of densitometry analysis are all expressed as mean±standard deviation (SD). *P<0.05.

1156

haematologica | 2017; 102(7)

A new path to platelet production through matrix sensing

Discussion The signaling pathways that control platelet formation in the bone marrow matrix environment have remained quite elusive. This study demonstrates that calcium influx through TRPV4 channels stimulates integrin β1 activation upon MK adhesion on soft collagen substrate, promoting Akt phosphorylation and platelet formation. This previously unknown mechanism for platelet production adds new insights into the role of the ECM in regulating MK function,44-46 and determines the importance of ECMdependent calcium influx in regulating MK spreading and PPF.17 The BM microenvironment consists of various ECM components that interact with each other to form a structural framework that supports tissue organization and positional cues regulating megakaryopoiesis. In this structure, type I and IV are the most abundant collagens. Notably, type IV can support PPF,29 while type I collagen is the only ECM component known to inhibit this process.9,10,12 This inhibition is triggered by the tensile strength of fibrils in type I collagen that regulates cytoskeleton contractility of MKs through activation of the Rho-ROCK pathway and MLC-2 phosphorylation.8,10 Consistently, culturing MKs in soft methylcellulose (MC) hydrogels determines an increase of PPF through activation of the myosin IIA and MKL1 pathways.18 The current data showed that type I collagen displayed a significant higher stiffness than type IV collagen, as revealed by AFM analysis. Consistently, only MKs plated on the softer type IV collagen extended long and branched proplatelets. We hypothesized that specific collagen

receptors may regulate the different MK behavior on type I and type IV collagens. Among the collagen receptors, integrins have been extensively described to be major mechano-receptors, due to their ability to modulate the signals transmitted inside the cell according to the physical properties of the ligand they bind.47 In support of this hypothesis, the adhesion of MKs on both collagens was found to be mostly mediated by β1 integrin. However, β1 integrin was significantly more active and more internalized in MKs plated on type IV than on type I collagen during an 8-h incubation time. These data were in line with the notion that β1 integrin dynamics and internalizationre-cycling circle are regulated by matrix stiffness.48,49 Importantly, integrin detachment and internalization on soft substrate was previously shown in mesenchymal stromal cells, and was ascribed to the instability of the binding, and thus to the low stress level necessary for the binding rupture.50 In contrast, when the substrate stiffness increases, the relative number of nascent and retracting β1 integrin adhesions are reduced, and highly stable adhesions are promoted.51 In accordance, by using passive silk films with different rigidities, substrate stiffness and β1 integrin activation in MKs were shown to be inversely proportional, as softer substrates promoted β1 integrin activation and internalization. The increase in β1 integrin signaling that was observed on soft matrices, despite the higher rate of integrin internalization, is in line with recent work which demonstrated that endocytosis was necessary for full ECM-induced integrin-mediated signaling.52 PI3K/Akt and MAPK/ERK signaling pathways are known to be activated downstream β1 integrin and responsible for PPF regulation.13-17

Figure 7. Lysyl oxidase-mediated collagen cross-linking inhibition increases platelet (Plt) production in vivo. (A) Elastic modulus of untreated (CTRL) and β-aminopropionitrile (BAPN)-treated mice bone marrows (BM). Data refer to 4 different mice per group. BM from both femurs per mouse was analyzed. (B) Peripheral blood platelet count of untreated (CTRL) and BAPN-treated mice. Data refer to 10 mice per group. Reticulated platelet measurements in PB of CTRL and BAPN-treated mice. Data refer to 8 mice per group. (C) pAkt analysis in CTRL and BAPN-treated mouse MKs was performed by BM section immunofluorescence (left) and by flow cytometry (right). The densitometry analysis of pAkt immunofluorescence staining, of BM MKs, is derived from 4 mice per group. At least 30 MKs per section were analyzed. The flow cytometry analysis of pAkt phosphorylation was performed in CD41+ BM MKs. The mean fluorescence intensity (MFI) is shown. Data are expressed as mean±standard deviation (SD) of five independent experiments. Data of densitometry analysis are all expressed as mean±SD. *P<0.05.

haematologica | 2017; 102(7)

1157

V. Abbonante et al.

Interestingly, thrombocytopenia is one of the most frequent hematologic adverse events in patients treated with perifosine, an oral Akt inhibitor used in cancer treatment.53 Consistently, in our experiments, Akt and not ERK phosphorylation promoted MK maturation and PPF upon β1 integrin activation by soft substrates. These data prompted us to hypothesize that other mechano-sensitive molecules may be involved in the regulation of these signaling processes. The mechano-sensitive ion channels, a member of which is TRPV4, seemed to be the best candidates to explain the data. First, these ion channels detect and transduce external mechanical forces into electrical and/or chemical intracellular signals.22,54 Second, it was demonstrated that stretch-activated endothelial cells regulate Akt and ERK activation through a TRPV4- β1 integrin-dependent mechanism.24 Consistently, in the experiments MKs were shown to express functional TRPV4 ion channels

that were selectively activated only upon MK adhesion on soft substrates. Accordingly, TRPV4 channels present the unique ability to mediate an integrin-to-integrin signaling which is activated by mechanical forces transmitted by ECM components to β1 integrins. This results in the ultra-rapid (4 ms) alteration of the molecular conformation of TRPV4, which gates the channels and enables extracellular Ca2+ entry, thereby causing additional β1 activation.24,25,55 Overall, these in vitro studies lead us to posit a model by which a softer environment in the BM promotes the activation of the TRPV4 ion channel that, in turn, leads to further β1 integrin stimulation and Akt phosphorylation, culminating with PPF. In line with this contention, adult mice treated with a specific LOX inhibitor that reduces tissue stiffness42,43 promoted TRPV4, β1 integrin and Akt activation, with a consequent increase of platelet count in the

Figure 8. Lysyl oxidase-mediated collagen cross-linking inhibition increases proplatelet formation and TRPV4 activation ex vivo. (A) Western blot analysis of TRPV4 immunoprecipitated from bone marrow (BM) immunomagnetically-sorted megakaryocytes (MKs). A control sample (CTRL) was immunoprecipitated with an unrelated antibody (IgG). Membranes were probed with anti-PKC-substrates antibody (recognizing phosphorylated serine) and with anti-TRPV4 antibody. Actin staining on total cell lysates is shown to ensure equal loading. The densitometry analysis of the phosphorylated TRPV4/total TRPV4 ratio is shown. Data refer to three independent experiments. (B) Proplatelet formation assay in BM explants. MKs extending proplatelet were counted after 3 and 8 hours (h) from the beginning of the experiment, and the percentage of proplatelet-bearing MKs is shown. Data are expressed as mean percentage±standard deviation (SD). n=10 CTRL, 10 BAPN, 3 BAPN + AKT inhibitor (AKT INH). (C) Representative pictures of proplatelet-bearing MKs in BM explants. MKs were stained as living cells with anti-CD41-FITC antibody. Data of densitometry analysis are expressed as mean±SD. Scale bar = 50 μm upper pictures; 100 μm lower pictures. *P<0.05. PPF: proplatelet formation.

1158

haematologica | 2017; 102(7)

A new path to platelet production through matrix sensing peripheral blood. This occurred while MK numbers in the BM were not affected. Of note, an unexplained trend in increased platelet count was also reported earlier upon LOX inhibitor administration,35 although the dosage of inhibitor used in that study was low, and was tested in newborn to young mice. A recent study showed that upregulated LOX, at a level similar to that found in pathological BM fibrosis, led to increased platelet adhesion to monomeric collagen mediated via α2β1 collagen receptors.56 A similar increase in adhesion was observed with regard to LOX over-expressing MKs. Together with our current studies, we propose that LOX is capable of augmenting MK adhesion via collagen receptor activation, but whether these MKs effectively produce proplatelets or not also depends on LOX-regulated ECM stiffness and changes in TRPV4 activation. During life, ECM components are continuously rearranged to permit cell-tissue function. However, some patho-physiological conditions, such as aging or augmented bone marrow fibrosis determine an increase of ECM component rigidity, causing tissue function alterations.57,58 Interestingly, it is known that platelet count is inversely correlated with age and that in the late stages of primary myelofibrosis, when the BM is full of thick reticular fibers, thrombocytopenia occurs.59,60 Moreover, in these condi-

References 1. Machlus KR, Italiano JE. The incredible journey: From megakaryocyte development to platelet formation. J Cell Biol. 2013;201(6):785-796. 2. Patel SR, Richardson JL, Schulze H, et al. Differential roles of microtubule assembly and sliding in proplatelet formation by megakaryocytes. Blood. 2005; 106(13):4076-4085. 3. Junt T, Schulze H, Chen Z, et al. Dynamic visualization of thrombopoiesis within bone marrow. Science. 2007; 317(5845):1767-1770. 4. Malara A, Abbonante V, Di Buduo CA, Tozzi L, Currao M, Balduini A. The secret life of a megakaryocyte: emerging roles in bone marrow homeostasis control. Cell Mol Life Sci. 2015;72(8):1517-1536. 5. Hynes RO. The extracellular matrix: not just pretty fibrils. Science. 2009; 326(5957):1216-1219. 6. Gattazzo F, Urciuolo A, Bonaldo P. Extracellular matrix: a dynamic microenvironment for stem cell niche. Biochim Biophys Acta. 2014;1840(8):2506-2519. 7. Mouw JK, Ou G, Weaver VM. Extracellular matrix assembly: a multiscale deconstruction. Nat Rev Mol Cell Biol. 2014; 15(12):771-785. 8. Shin JW, Swift J, Spinler KR, Discher DE. Myosin-II inhibition and soft 2D matrix maximize multinucleation and cellular projections typical of platelet-producing megakaryocytes. Proc Natl Acad Sci USA. 2011;108(28):11458-11463. 9. Chen Z, Naveiras O, Balduini A, et al. The May-Hegglin anomaly gene MYH9 is a negative regulator of platelet biogenesis modulated by the Rho-ROCK pathway. Blood. 2007;110(1):171-179. 10. Malara A, Gruppi C, Pallotta I, et al. Extracellular matrix structure and nanomechanics determine megakaryocyte func-

haematologica | 2017; 102(7)

tions, other organs, such as the lung61 or the spleen,62 with a more favorable matrix environment, may become the site of platelet production. Based on these observations, and on the mechanism described in the current work, we propose that, beside the extensively described humoraldependent mechanisms of platelet production, the physical properties of the ECM components that fill the BM are crucial regulators of MK function and platelet production. Aknowledgments We thank Prof. Giampaolo Merlini for providing instruments for polarized light microscopy imaging, Prof. Elisabetta Dejana for helping in setting experiments for integrin internalization, Dr. Manuela Monti for providing instruments for confocal imaging, Dr. Franco Tanzi for providing instruments for calcium imaging, Dr. Jandrot Perrus for providing anti-GPVI antibody. Funding VA fellowship was funded by Collegio Ghislieri, Pavia progetto “Progressi in Biologia e Medicina”. This paper was supported by Cariplo Foundation (2013-0717), US National Institutes of Health (grant EB016041-01) to DK and AB and NHLBI HL80442 to KR. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

tion. Blood. 2011;118(16):4449-4453. 11. Chang Y, Auradé F, Larbret F, et al. Proplatelet formation is regulated by the Rho/ROCK pathway. Blood. 2007; 109(10):4229-4236. 12. Semeniak D, Kulawig R, Stegner D, et al. Proplatelet formation is selectively inhibited by collagen type I through Syk-independent GPVI signaling. J Cell Sci. 2016; 129(18):3473-3484. 13. Machlus KR, Johnson KE, Kulenthirarajan R, et al. CCL5 derived from platelets increases megakaryocyte proplatelet formation. Blood. 2016;127(7):921-926. 14. Bluteau D, Balduini A, Balayn N, et al. Thrombocytopenia-associated mutations in the ANKRD26 regulatory region induce MAPK hyperactivation. J Clin Invest. 2014; 124(2):580-591. 15. Mazharian A, Watson SP, Séverin S. Critical role for ERK1/2 in bone marrow and fetal liver-derived primary megakaryocyte differentiation, motility, and proplatelet formation. Exp Hematol. 2009;37(10):12381249.e1235. 16. Currao M, Balduini CL, Balduini A. High doses of romiplostim induce proliferation and reduce proplatelet formation by human megakaryocytes. PLoS One. 2013; 8(1):e54723. 17. Di Buduo CA, Moccia F, Battiston M, et al. The importance of calcium in the regulation of megakaryocyte function. Haematologica. 2014;99(4):769-778. 18. Aguilar A, Pertuy F, Eckly A, et al. Importance of environmental stiffness for megakaryocyte differentiation and proplatelet formation. Blood. 2016; 128(16):2022-2032. 19. Giancotti FG, Ruoslahti E. Integrin signaling. Science. 1999;285(5430):1028-1032. 20. Campbell ID, Humphries MJ. Integrin structure, activation, and interactions. Cold Spring Harb Perspect Biol. 2011; 3(3):a004994. 21. Schwartz MA. Integrins and extracellular

22.

23.

24.

25.

26.

27.

28.

29.

30.

matrix in mechanotransduction. Cold Spring Harb Perspect Biol. 2010;2(12): a005066. Gasparski AN, Beningo KA. Mechanoreception at the cell membrane: More than the integrins. Arch Biochem Biophys. 2015;586:20-26. Martinac B. Mechanosensitive ion channels: an evolutionary and scientific tour de force in mechanobiology. Channels (Austin). 2012;6(4):211-213. Thodeti CK, Matthews B, Ravi A, et al. TRPV4 channels mediate cyclic straininduced endothelial cell reorientation through integrin-to-integrin signaling. Circ Res. 2009;104(9):1123-1130. Matthews BD, Thodeti CK, Tytell JD, Mammoto A, Overby DR, Ingber DE. Ultra-rapid activation of TRPV4 ion channels by mechanical forces applied to cell surface beta1 integrins. Integr Biol (Camb). 2010;2(9):435-442. Jablonski CL, Ferguson S, Pozzi A, Clark AL. Integrin 1 1 participates in chondrocyte transduction of osmotic stress. Biochem Biophys Res Commun. 2014;445(1):184-190. Kröger N, Zabelina T, Alchalby H, et al. Dynamic of bone marrow fibrosis regression predicts survival after allogeneic stem cell transplantation for myelofibrosis. Biol Blood Marrow Transplant. 2014;20(6):812815. Abbonante V, Di Buduo CA, Gruppi C, et al. Thrombopoietin/TGF- 1 loop regulates megakaryocyte extracellular matrix component synthesis. Stem Cells. 2016; 34:1123-1133. Balduini A, Pallotta I, Malara A, et al. Adhesive receptors, extracellular proteins and myosin IIA orchestrate proplatelet formation by human megakaryocytes. J Thromb Haemost. 2008;6(11):1900-1907. Abbonante V, Gruppi C, Rubel D, Gross O, Moratti R, Balduini A. Discoidin domain receptor 1 protein is a novel modulator of

1159

V. Abbonante et al.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

1160

megakaryocyte-collagen interactions. J Biol Chem. 2013;288(23):16738-16746. Lampugnani MG, Orsenigo F, Gagliani MC, Tacchetti C, Dejana E. Vascular endothelial cadherin controls VEGFR-2 internalization and signaling from intracellular compartments. J Cell Biol. 2006;174(4):593-604. Lawson MA, Maxfield FR. Ca(2+)- and calcineurin-dependent recycling of an integrin to the front of migrating neutrophils. Nature. 1995;377(6544):75-79. Fabbri M, Di Meglio S, Gagliani MC, et al. Dynamic partitioning into lipid rafts controls the endo-exocytic cycle of the alphaL/beta2 integrin, LFA-1, during leukocyte chemotaxis. Mol Biol Cell. 2005; 16(12):5793-5803. Di Buduo CA, Wray LS, Tozzi L, et al. Programmable 3D silk bone marrow niche for platelet generation ex vivo and modeling of megakaryopoiesis pathologies. Blood. 2015;125(14):2254-2264. Eliades A, Papadantonakis N, Bhupatiraju A, et al. Control of megakaryocyte expansion and bone marrow fibrosis by lysyl oxidase. J Biol Chem. 2011;286(31):2763027638. Urciuolo A, Quarta M, Morbidoni V, et al. Collagen VI regulates satellite cell selfrenewal and muscle regeneration. Nat Commun. 2013;4:1964. Eckly A, Rinckel JY, Laeuffer P, et al. Proplatelet formation deficit and megakaryocyte death contribute to thrombocytopenia in Myh9 knockout mice. J Thromb Haemost. 2010;8(10):2243-2251. Luque A, Gómez M, Puzon W, Takada Y, Sánchez-Madrid F, Cabañas C. Activated conformations of very late activation integrins detected by a group of antibodies (HUTS) specific for a novel regulatory region (355-425) of the common beta 1 chain. J Biol Chem. 1996;271(19):1106711075. Zhou CZ, Confalonieri F, Jacquet M, Perasso R, Li ZG, Janin J. Silk fibroin: structural implications of a remarkable amino acid sequence. Proteins. 2001;44(2):119122. Fan HC, Zhang X, McNaughton PA.

41.

42.

43.

44.

45.

46.

47.

48. 49. 50.

51.

Activation of the TRPV4 ion channel is enhanced by phosphorylation. J Biol Chem. 2009;284(41):27884-27891. Papadantonakis N, Matsuura S, Ravid K. Megakaryocyte pathology and bone marrow fibrosis: the lysyl oxidase connection. Blood. 2012;120(9):1774-1781. Marturano JE, Xylas JF, Sridharan GV, Georgakoudi I, Kuo CK. Lysyl oxidasemediated collagen crosslinks may be assessed as markers of functional properties of tendon tissue formation. Acta Biomater. 2014;10(3):1370-1379. Mammoto A, Mammoto T, Kanapathipillai M, et al. Control of lung vascular permeability and endotoxin-induced pulmonary oedema by changes in extracellular matrix mechanics. Nat Commun. 2013;4:1759. Malara A, Currao M, Gruppi C, et al. Megakaryocytes contribute to the bone marrow-matrix environment by expressing fibronectin, type IV collagen, and laminin. Stem Cells. 2014;32(4):926-937. Ivanovska IL, Shin JW, Swift J, Discher DE. Stem cell mechanobiology: diverse lessons from bone marrow. Trends Cell Biol. 2015; 25(9):523-532. Choi JS, Harley BA. The combined influence of substrate elasticity and ligand density on the viability and biophysical properties of hematopoietic stem and progenitor cells. Biomaterials. 2012;33(18):4460-4468. Schwarz US, Gardel ML. United we stand: integrating the actin cytoskeleton and cellmatrix adhesions in cellular mechanotransduction. J Cell Sci. 2012;125(Pt 13):30513060. Pellinen T, Ivaska J. Integrin traffic. J Cell Sci. 2006;119(Pt 18):3723-3731. Bridgewater RE, Norman JC, Caswell PT. Integrin trafficking at a glance. J Cell Sci. 2012;125(Pt 16):3695-3701. Du J, Chen X, Liang X, et al. Integrin activation and internalization on soft ECM as a mechanism of induction of stem cell differentiation by ECM elasticity. Proc Natl Acad Sci USA. 2011;108(23):9466-9471. Doyle AD, Carvajal N, Jin A, Matsumoto K, Yamada KM. Local 3D matrix microenvironment regulates cell migration through

52.

53.

54.

55.

56.

57. 58. 59. 60.

61.

62.

spatiotemporal dynamics of contractilitydependent adhesions. Nat Commun. 2015; 6:8720. Alanko J, Mai A, Jacquemet G, et al. Integrin endosomal signalling suppresses anoikis. Nat Cell Biol. 2015;17(11):14121421. Richardson PG, Eng C, Kolesar J, Hideshima T, Anderson KC. Perifosine, an oral, anti-cancer agent and inhibitor of the Akt pathway: mechanistic actions, pharmacodynamics, pharmacokinetics, and clinical activity. Expert Opin Drug Metab Toxicol. 2012;8(5):623-633. Garcia-Elias A, Mrkonji S, Jung C, PardoPastor C, Vicente R, Valverde MA. The TRPV4 channel. Handb Exp Pharmacol. 2014;222:293-319. White JP, Cibelli M, Urban L, Nilius B, McGeown JG, Nagy I. TRPV4: Molecular Conductor of a Diverse Orchestra. Physiol Rev. 2016;96(3):911-973. Matsuura S, Mi R, Koupenova M, et al. Lysyl oxidase is associated with increased thrombosis and platelet reactivity. Blood. 2016;127(11):1493-1501. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche in cancer progression. J Cell Biol. 2012;196(4):395-406. Kurtz A, Oh SJ. Age related changes of the extracellular matrix and stem cell maintenance. Prev Med. 2012;54 Suppl:S50-56. Balduini CL, Noris P. Platelet count and aging. Haematologica. 2014;99(6):953-955. Cervantes F, Barosi G. Myelofibrosis with myeloid metaplasia: diagnosis, prognostic factors, and staging. Semin Oncol. 2005; 32(4):395-402. Wang Y, Hayes V, Jarocha D, et al. Comparative analysis of human ex vivogenerated platelets vs megakaryocyte-generated platelets in mice: a cautionary tale. Blood. 2015;125(23):3627-3636. Thiele J, Klein H, Falk S, Bertsch HP, Fischer R, Stutte HJ. Splenic megakaryocytopoiesis in primary (idiopathic) osteomyelofibrosis. An immunohistological and morphometric study with comparison of corresponding bone marrow features. Acta Haematol. 1992;87(4):176-180.

haematologica | 2017; 102(7)

ARTICLE

Red Cell Biology & Its Disorders

The endothelin B receptor plays a crucial role in the adhesion of neutrophils to the endothelium in sickle cell disease

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Bérengère Koehl,1* Pierre Nivoit,2* Wassim El Nemer,3* Olivia Lenoir,4* Patricia Hermand,5 Catia Pereira,6 Valentine Brousse,7 Léa Guyonnet,8 Giulia Ghinatti,9 Malika Benkerrou10, Yves Colin,11 Caroline Le Van Kim12# and Pierre-Louis Tharaux13#

1,3,5,6,11,12 Université Sorbonne Paris Cité, Université Paris Diderot, Inserm, INTS, Unité Biologie Intégrée du Globule Rouge, Laboratoire d’Excellence GR-Ex, France; 2,4,8,9,13 Inserm Paris Cardiovascular Centre (PARCC), Université Sorbonne Paris Cité, Université Paris Descartes & Laboratoire d’Excellence GR-Ex, France, 8Department of Infection and Immunity, Luxembourg Institute of Health, Luxembourg; 6Assistance Publique–Hôpitaux de Paris, Necker Hospital, Reference Centre of Sickle Cell Disease, France and 1,10 Assistance Publique–Hôpitaux de Paris, Robert Debré Hospital, Reference Centre of Sickle Cell Disease, France

*BK, PN, WEN and OL are equally contributing authors. #CLVK and PLT are equally contributing authors.

Haematologica 2017 Volume 102(7):1161-1172

ABSTRACT

lthough the primary origin of sickle cell disease is a hemoglobin disorder, many types of cells contribute considerably to the pathophysiology of the disease. The adhesion of neutrophils to activated endothelium is critical in the pathophysiology of sickle cell disease and targeting neutrophils and their interactions with endothelium represents an important opportunity for the development of new therapeutics. We focused on endothelin-1, a mediator involved in neutrophil activation and recruitment in tissues, and investigated the involvement of the endothelin receptors in the interaction of neutrophils with endothelial cells. We used fluorescence intravital microscopy analyses of the microcirculation in sickle mice and quantitative microfluidic fluorescence microscopy of human blood. Both experiments on the mouse model and patients indicate that blocking endothelin receptors, particularly ETB receptor, strongly influences neutrophil recruitment under inflammatory conditions in sickle cell disease. We show that human neutrophils have functional ETB receptors with calcium signaling capability, leading to increased adhesion to the endothelium through effects on both endothelial cells and neutrophils. Intact ETB function was found to be required for tumor necrosis factor α-dependent upregulation of CD11b on neutrophils. Furthermore, we confirmed that human neutrophils synthesize endothelin-1, which may be involved in autocrine and paracrine pathophysiological actions. Thus, the endothelin-ETB axis should be considered as a cytokine-like potent pro-inflammatory pathway in sickle cell disease. Blockade of endothelin receptors, including ETB, may provide major benefits for preventing or treating vaso-occlusive crises in sickle cell patients. Introduction Sickle cell disease (SCD) is a genetic hemoglobinopathy resulting from a unique mutation in the β-globin gene. SCD is characterized by hemolytic anemia, painful vaso-occlusive crises (VOC) and progressive organ failure. Although red blood cell dysfunction is the major contributor to disease development and progression, other types of cells, which are not affected by the genetic mutation (endothelial cells, leukocytes, platelets1,2), are also key actors in the pathophysiology of SCD. Several haematologica | 2017; 102(7)

Correspondence: caroline.le-van-kim@inserm.fr or pierre-louis.tharaux@inserm.fr Received: September 19, 2016. Accepted: March 30, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.156869 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1161 ©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.

1161

B. Koehl et al.

studies have highlighted the important role of polymorphonuclear neutrophils (neutrophils), both during an acute VOC3 and in the associated long-term morbidity and mortality.4 Interestingly, a high, steady-state, peripheral white cell count is a risk factor for both significant morbidity – stroke, pulmonary complications, nephropathy – and early SCD-related death.4-8 The central role of neutrophils in the pathophysiology of SCD has recently been explored.3,9 In vitro studies have shown that, compared to neutrophils from healthy controls, neutrophils from SCD patients have an increased expression of adhesion molecules,10-12 rendering them more susceptible to inflammatory stimuli.13 A relationship between clinical manifestations of SCD and the expression of adhesion molecules on neutrophils has also been reported.2,14 It is likely that activated neutrophils engage in a complex process of abnormal interactions between activated endothelial cells, platelets and circulating red blood cells contributing to decreased blood flow and to endothelial injury. This further accentuates erythrocyte sickling, neutrophil recruitment and tissue ischemia.9 Targeting the mechanisms of neutrophil-endothelial cell interactions would, therefore, represent a novel and potentially important therapeutic opportunity in SCD. Endothelin-1 (ET-1) is the most potent endogenous vasoconstrictor.15 It is released by activated endothelial16 and non-endothelial cells17 in response to hypoxia and reduced nitric oxide bioavailability in several animal models.18 The effects of ET-1 are mediated via two receptors, the ETA and ETB receptors.15 We previously found that mixed ETA/B receptor antagonism has profound effects on organ injury and mortality in a mouse model of SCD.19 In addition to inhibition of tonic ET-1-dependent vasoconstriction during experimental VOC, we also observed an unexpected but powerful inhibition of neutrophil recruitment in the lungs and kidneys although we could not link this effect to a direct action of ET-1 receptors on neutrophil-endothelial interactions. We, therefore, hypothesized that activation of ET receptors might promote a pathogenic pro-inflammatory role for neutrophils in SCD. In the present study, we combined intravital videomicroscopy of the microcirculation in a murine model of SCD with quantitative microfluidic fluorescence microscopy of human blood to investigate the involvement of ET receptors in the interaction of neutrophils with endothelial cells.

Methods Animal model Animals were used in accordance with the National Institutes of Health Guide for the care and use of laboratory animals (NIH publication n. 85-23) and the study protocol was approved by the French ministry of agriculture. SAD1 (SAD) Hbβ single/single hemizygous mice were used in this study. This strain harbors a recombinant hβ-globin gene construct expressing human hemoglobin SAD (A2β2SAD), which contains two mutations [Antilles (β23I) and D-Punjab (β121Q)] in addition to the βS6V mutation.19,20 This strain is bred on the C57BL/6J genetic background (with more than 30 backcrosses).

Intravital videomicroscopy: experimental protocol The complete protocol is described in the Online Supplementary Methods and illustrated in Online Supplementary Figure S1. ETA antagonist (BQ123, A.G. Scientific®, San Diego, 1162

CA, USA) or ETB antagonist (BQ788, Tocris Bioscience®, Bristol, UK) were injected 10 min before (10 mg/kg) and 3 h 15 min after (5 mg/kg) intrascrotal injection of 0.5 μg tumor necrosis factor (TNFα, R&D Systems, Minneapolis, MN, USA) and neutrophils were monitored by injection of labeled phycoerythrinconjugated anti-Ly6G antibody (BD Biosciences Pharmingen, Sparks, MD, USA). The neutrophil rolling flux fraction, adhesion density, adhesion efficiency and transmigration were measured from 2.5 h to 5 h after the TNFα injection at 30-min intervals.

Patients and healthy volunteers The study was approved by the Ethical Evaluation Committee for Biomedical Research Projects (CEERB) of Robert Debré Hospital, France (n. 2014/155). All patients were included during a scheduled medical consultation, at steady state. Blood samples were collected from 2- to 18-year old patients with SCD (SS genotype), and from young healthy adult donors (AA genotype) from the local blood bank (Etablissement Français du Sang). Patients in a chronic transfusion therapy program, those having received blood transfusion in the 3 months preceding enrollment and those treated with hydroxycarbamide (hydroxyurea) were excluded. Exceptions were made for the quantitative polymerase chain reaction analyses, as mentioned below.

Human neutrophil preparation Human neutrophils were isolated (98% purity) from fresh whole blood within 4 h after blood sampling with EDTA as anticoagulant, using a MACSxpress Neutrophil Isolation Kit followed by MACSxpress Erythrocyte Depletion Kit (Miltenyi Biotec, Paris, France). Details on the flow adhesion assays are provided in the Online Supplementary Methods.

Real-time quantitative polymerase chain reaction ET-1, ETA and ETB mRNA levels were determined by realtime quantitative polymerase chain reaction as described previously.21 Beta-2 microglobulin mRNA was used as a housekeeping gene The relative expression of ET-1, ETA and ETB mRNA in neutrophils was calculated following the ΔΔCt method,22 with the values obtained in HMEC-1 as a reference.

Flow cytometry assessment of adhesion molecules and intracellular Ca2+ measurements in neutrophils Three hours after treatment with TNFα with or without ET receptor antagonists, mice were anesthetized; blood was collected on EDTA by intracardiac puncture and immediately processed for flow cytometry staining for viability and CD11a, CD11b and CD62L surface expression on Ly5G+ neutrophils (see Online Supplementary Methods). Isolated human neutrophils were fluorescently labeled with fluo-4-acetoxymethyl ester (Fluo-4-AM; Invitrogen). Neutrophils were incubated with BQ123 or BQ788 (both from Sigma®, Saint-Quentin Fallavier, France) at 10 μmol/L for 15 min and analyzed immediately by flow cytometry (FACSCanto II, BD Biosciences, Sparks, MD, USA). After 2 min of signal acquisition, the analysis was interrupted for less than 10 s to add 100 nM of ET-1 (Sigma®) to the tube before the analysis was resumed.

Endothelin-1 immunoassay Isolated neutrophils were incubated at 37°C in RPMI medium supplemented with TNFα (10 ng/mL) for 2 h. Supernatants were frozen at -80°C until assay. The concentration of ET-1 was determined using a Quantikine® enzyme-linked immunosorbent assay (R&D systems. haematologica | 2017; 102(7)

ETB receptor in neutrophil recruitment in SCD

Flow adhesion assays The experimental protocol is described in the Online Supplementary Methods. Briefly, human neutrophils were stained in whole blood with anti-CD16 antibody together with BQ123 (1 μM) or BQ788 (1 μM). The time between drawing blood and starting the anti-CD16 staining was always less than 3 h and reproducible. If more than 3 h elapsed, samples were discarded. Whole blood samples were perfused through biochip channels containing HMEC-1 monolayers or coated recombinant proteins (P-selectin, VCAM-1 and ICAM-1).

Statistics Results are presented as means plus or minus the standard error of the mean (SEM), Data were analyzed using GraphPad Prism 6 software by ANOVA with the Tukey test, Wilcoxon rank test and Mann-Whitney test. P values less than 0.05 were considered statistically significant.

Results Acute ETA and ETB receptor antagonism does not alter leukocyte blood count or microvascular hemodynamics in wild-type and SAD mice In keeping with clinical SCD, SAD mice (n=5) showed an increase in circulating leukocytes with a neutrophil predominance compared to wild-type (WT) littermate mice, both at baseline and after TNFα administration (Online Supplementary Table S1). Treatment of TNFα-challenged animals with either BQ123 or BQ788 did not affect any of the peripheral cell counts. As changes in microvascular wall shear stress may influence neutrophil adhesion, we determined the effects of acute ETA and ETB receptor blockade on wall shear rate in cremasteric venules. There were no significant differences in venular wall shear rates between groups treated with TNFα + vehicle, TNFα + BQ123 and TNFα + BQ788 (Online Supplementary Table S2).

ETA and ETB receptors control neutrophil rolling in inflamed venules. In WT mice, neither BQ123 nor BQ788 affected neutrophil rolling flux throughout the time course of the experiment. As in WT mice, ETA antagonism with BQ123 had no effect on neutrophil rolling flux in SAD mice. In sharp contrast, ETB blockade with BQ788 profoundly inhibited neutrophil rolling flux in SAD mice, compared to that in the vehicle-treated SAD group, throughout the experiment (Figure 1A).

ETA and ETB receptor antagonism decreases neutrophil adhesion in inflamed venules from sickle mice Overall, in WT mice, neutrophil adhesion density was not influenced by BQ123 or BQ788 treatment. Over the time course of the study, adhesion decreased in all three groups of WT mice from about 670 adherent neutrophils/mm2 at 2 h 30 min after the inflammatory challenge to reach about 250 adherent neutrophils/mm2 at 5 h. In SAD mice, the inflammatory challenge with TNFα induced a marked increase in neutrophil adhesion compared to that in the WT counterparts (1000 versus 586 adherent neutrophils/mm2, P<0.05). The adhesion remained elevated and stable for 90 min and then decreased slightly. At the end of the experiment, adhesion haematologica | 2017; 102(7)

density in SAD mice challenged with TNFα was still increased compared to that in WT challenged mice (731 versus 271 adherent neutrophils/mm2, P<0.05). In SAD mice, both ETA and ETB receptor blockade markedly inhibited adhesion of neutrophils throughout the time course of the study with the degree of adhesion being similar to that observed in WT controls (Figure 1B). To define the neutrophil adhesion better, we calculated the efficiency of neutrophil adhesion as the ratio between adhering neutrophils and those that are estimated to be available to adhere (Figure 1C). We found that adhesion efficiency was, overall, unaffected by BQ123 and BQ788 in WT mice. In contrast, adhesion efficiency was markedly and significantly decreased in SAD mice treated with BQ788 compared to that in the TNFα and vehicle-treated SAD group throughout the time course of the experiment; no significant differences were observed in SAD mice treated with BQ123. Taken together, these data show a differential effect of ETA versus ETB receptor antagonism on neutrophil adhesion (Figure 1C-F).

ETA and ETB receptor antagonism decreases neutrophil transmigration in inflamed venules from sickle mice We counted neutrophils that had migrated into the extravascular space adjacent to the observed venules by optical sectioning (Online Supplementary Figure S1, Figure 1G,H). TNFα-induced emigration of neutrophils was significant in both groups of mice compared to the emigration in animals treated with vehicle (phosphate-buffered saline) but distinct kinetics were observed in SAD and WT animals. While neutrophil emigration reached a plateau at 3 h 30 min in the WT cremasteric microcirculation and subsequently resolved to baseline values, sickle SAD mice experienced a more sustained and more intense increase in emigration which was double that measured in WT after 4 h and lasted until the end of the experiments (Figure 1 G,H). In WT mice, TNFα-induced emigration was significantly decreased by BQ123 and BQ788 administration but only at early time points (∼2.3-fold decrease at 2 h 30 min, P<0.01 for BQ123, ∼3.5-fold decrease at 2 h 30 min and ∼2.1-fold decrease at 3 h for BQ788, P<0.001 versus vehicle-treated mice). In SAD mice that displayed more intense and prolonged emigration of neutrophils both ETA and ETB blockade significantly prevented trans-endothelial migration of neutrophils to tissues, especially at late points [BQ123: ∼2.3-fold decrease at 2 h 30 min (P<0.01 versus TNFα only-treated mice); BQ788: ∼3.5-fold decrease at 2 h 30 min (P<0.001 versus TNFα-treated mice) and ∼2.1fold decrease at 3 h (P<0.01 versus TNFα-treated mice)] (Figure 1H). Representative images indicating less neutrophil recruitment in the presence of ETA and ETB, particularly in SAD mice, are shown in Figure 2. All these experiments suggest that blocking ET receptors influences neutrophil recruitment in mice, especially in the context of sustained and intense neutrophil recruitment specific to experimental SCD.

ETA and ETB receptor antagonism reduces tumor necrosis factor α-induced high CD11b/Mac1 expression on neutrophils and does not alter cell viability.

CD11b, also known as Mac-1α or integrin αM chain, is part of the CD11b/CD18 heterodimer, also known as the C3 complement receptor. Increased adhesion has been linked to engagement of CD11b membrane expression on neutrophils from sickle mice.23 We, therefore, measured 1163

B. Koehl et al.

Figure 1. The rolling, adhesion and emigration of neutrophils in inflamed cremasteric venules are attenuated in tumor necrosis factor-alpha challenged wild-type and SAD mice treated with the selective endothelin receptor inhibitors BQ123 and BQ788. (A, B) Rolling neutrophil flux fraction in TNFα inflamed venules of WT and SAD mice. The number of rolling neutrophils per minute (flux) was counted and the neutrophil rolling flux fraction was determined by dividing the rolling neutrophil flux by the total neutrophil flux (all neutrophils passing through the venule). (C, D) Neutrophil adhesion density, defined as the number of adherent neutrophils per square millimeter, in TNFα inflamed venules after BQ123 and BQ788 treatment in WT and SAD mice. (E, F) Neutrophil adhesion efficiency, defined as the number of adhered neutrophils per square millimeter normalized by the number of circulating neutrophils passing through the venule. (G, H) Neutrophil emigration, defined as the number of emigrated neutrophils per square millimeter of extravascular space in inflamed venules after BQ123 and BQ788 treatment in WT and SAD mice. Emigrated neutrophils were visualized and quantified by optical sectioning and two-dimensional maximum intensity projection. Fifteen venules (5 mice) were analyzed in each group [12 venules and 3 mice for control with phosphate-buffered saline (PBS)]. Data are presented as mean ± SEM. Two-way ANOVA with the Tukey multiple comparison test. *P<0.05 compared with WT-TNFα; **P<0.01 compared with WT-TNFα; ***P<0.001 compared with WT-TNFα; ****P<0.0001 compared with WT-TNFα. †P<0.05 compared with SAD-TNFα. ††P<0.01 compared with SAD-TNFα. †††P<0.001 compared with SAD-TNFα. ††††P<0.0001 compared with SADTNFα.

1164

haematologica | 2017; 102(7)

ETB receptor in neutrophil recruitment in SCD

CD11b expression on neutrophils from SAD and WT mice without or after TNFα administration. In the latter case, we evaluated the effect of BQ123 and BQ788 antagonists 3 h after injection of TNFα or vehicle (phosphate-buffered saline), when ET receptor antagonists displayed inhibitory effects on neutrophil adhesion in intravital imaging experiments.

TNFα challenge induced a very significant upregulation of CD11b expression in neutrophils from both WT and SAD animals within 3 h (Online Supplementary Figure S2). This effect was markedly increased in neutrophils from SAD mice compared to that in WT mice (+40 %, P<0.05). Both ET receptor antagonists limited CD11b-associated mean fluorescence surface intensity and the proportion of

Figure 2. Kinetics of neutrophil recruitment. Representative images of cremasteric venules for each group witthout stimulation (administered phosphate - buffered saline, PBS) or after local TNFα stimulation associated or not with specific blocking of ETA or ETB receptors. White arrows indicate emigrated neutrophils. Scale bars: 10 μm.

haematologica | 2017; 102(7)

1165

B. Koehl et al.

neutrophils with high CD11b expression in neutrophils from TNFα-challenged SAD animals, whereas neither antagonist had any effect on CD11b expression in normal mice (Online Supplementary Figure S2). This suggests that involvement of the ET system in pro-inflammatory CD11b upregulation and neutrophil adhesion is specific to the SCD condition. No effect of either TNFα or the ET receptor antagonists was observed on the surface expression of CD11a and CD62L on neutrophils (Online Supplementary Figure S2) and no difference in cell viability (apoptosis or necrosis affected less than 3% of blood neutrophils) was observed between the different groups (data not shown). We next investigated the expression and the role of ET receptors on human neutrophils from SCD patients and healthy controls.

ETA and ETB mRNA expression in neutrophils from healthy volunteers and patients with sickle cell disease As very few studies report on the expression of ETB receptors in neutrophils from normal individuals24-26 and none in SCD patients, we first investigated the expression of the EDNRA and EDNRB genes encoding ETA and ETB, respectively, in endothelial cells (HMEC-1) and in human neutrophils from SCD patients (SS) and healthy individuals (AA). ETA and ETB mRNA are expressed both in endothelial cells and in human neutrophils and at similar levels in neutrophils isolated from healthy controls and SCD patients (Figure 3). To complete these results and confirm the presence of functional ET receptors on human neutrophils, we investigated the potential effect of neutrophil stimulation on the ET-1-mediated calcium response.

Endothelin-1 elicits an ETB-mediated calcium response in neutrophils from healthy controls and individuals with sickle cell disease

S3, preproET-1 mRNA is significantly expressed in neutrophils but at a much lower level than in endothelial cells. To confirm the ability of neutrophils to secrete ET-1, we measured the concentration of ET-1 in the supernatant of isolated human neutrophils activated or not by TNFα (Figure 5). We found detectable concentrations of ET-1 in the supernatant of isolated neutrophils at baseline (data not shown) and after stimulation with TNFα. The level of extracellular ET-1 in the presence of TNFα was higher in supernatants of neutrophils from SS individuals than those from AA healthy controls (0.27 ± 0.15 pg/mL versus 1.64 ± 0.45 pg/mL respectively; P=0.033) and from SS patients treated with hydoxycarbamide (0.20 ± 0.08 pg/mL; P=0.04) (Figure 5). These data highlight the ability of neutrophils to secrete ET-1, which may then activate endothelial cells and neutrophils themselves, with an autocrine and paracrine mechanism. Thus, we next investigated the effects of ET receptor antagonists on human neutrophil adhesion to endothelial cells.

ETB receptor antagonism prevents neutrophil adhesion to endothelial cells in controls and patients with sickle cell disease To mimic in vivo circulation accurately, we infused whole blood onto a layer of activated endothelial cells under conditions of laminar flow. Adhesion of neutrophils to TNFα-primed endothelium was similar in controls and SCD patients (Figure 6A). As shown in Figure 6B,C, ETA receptor antagonism with BQ123 did not affect this. Conversely, adhesion of BQ788-treated neutrophils to endothelial cells was decreased by 42% (± 26%) compared to the adhesion of untreated neutrophils (P=0.0008) (Figure 6B,C). These data indicate an unexpected stimulatory role for the ETB receptor in human neutrophil adhesion to endothelial cells, paralleling that observed in vivo in sickle mice.

Although ETA and ETB receptors have been shown to be expressed on many cell types, their presence on the surface of neutrophils and their signaling pathways are poorly understood. We investigated the activation of a calciumdependent signaling pathway in response to stimulation with ET-1 (Figure 4A,B). In the presence of ET-1 in neutrophils from both SCD individuals and healthy volunteers, we found a rapid increase in intra-cytoplasmic calcium concentration, which progressively decreased to steady state in less than 100 s. The same experiment was performed with neutrophils pre-incubated for 15 min with BQ123 or BQ788. Neither antagonist altered baseline values (Figure 4C). ETA antagonism did not prevent the ET-1-elicited calcium response (Figure 4A,C) whereas ETB receptor antagonism provoked a significant reduction in the calcium response after the addition of ET-1 both in AA and SS neutrophils (Figure 4A,C). These results indicate the presence of a functional ETB receptor on the surface of neutrophils. We next investigated the potential secretion of ET-1 by human neutrophils.

Endothelin-1 secretion by neutrophils from healthy individuals and patients with sickle cell disease We investigated the expression of preproET-1 mRNA in neutrophils from healthy controls and SCD patients, and compared this to expression levels in the endothelial HMEC-1 cell line. As shown in Online Supplementary Figure 1166

Figure 3. Endothelial cells and neutrophils from both healthy controls (AA), and SCD patients (SS) express ETA and ETB transcripts. Results of real time quantitative reverse transcriptase polymerase chain reaction are expressed as a comparison with ETA/B mRNA transcripts in endothelial cells (HMEC-1) with the ΔΔCT method. The reference gene used was the β2 microglobulin gene. Results are means (± SEM) of eight independent samples (4 controls, 4 patients) in duplicate.

haematologica | 2017; 102(7)

ETB receptor in neutrophil recruitment in SCD

ETB receptor antagonism prevents neutrophil adhesion to endothelial cells through effects on both endothelial cells and neutrophils

BQ123) compared to the untreated condition (P=0.02) (Figure 6D). To study the specific effects of ETA and ETB antagonists on leukocytes, we used a different protocol since it was not possible to rinse the whole blood and totally exclude contact between the antagonists and the endothelial cells. We therefore infused whole blood directly onto channels coated with purified adhesion proteins (P-selectin, VCAM-1 and ICAM-1). The results illustrated in Figure 6E show that neutrophil adhesion after incubation with BQ123 was not different to that in the untreated condition. In contrast, there was a 16% (±18%) decrease in the adhesion of neutrophils on recombinant adhesion molecules after incubation with BQ788 compared to the untreated sample (P=0.03) (Figure 6E). We could also rule

During the adhesion experiments described above, ET receptor antagonists were added to whole blood, then infused directly onto endothelial cells. We were not, therefore, able to exclude any potential effect of the pharmacological antagonists on endothelial cells. To discriminate between the effects of the antagonists on endothelial and neutrophil ET receptors, we performed the same experiments by incubating only endothelial cells with the two selective antagonists. We found a 28% (±33%) significant decrease in the adhesion of neutrophils following treatment of endothelial cells with BQ788 (but not with

Figure 4. Endothelin-1 elicits an ETB-mediated calcium response in neutrophils from healthy controls and individuals with sickle cell disease. (A) Representative images of the intracellular calcium measurements in isolated neutrophils: rapid increase of calcium concentration after addition of ET-1, similar effect on neutrophils previously incubated with BQ123, significant reduction of the ET-1-elicited calcium response in neutrophils previously incubated with BQ788. (B) Gating strategy for the comparison of the fluorescence intensity before and after addition of ET-1. (C) Comparison between the percentages of the Fluo4 AM-associated highly fluorescent sub-population (population “P7”) from seven independent samples (neutrophils from 4 healthy controls and 3 SCD patients) in the different conditions: (C) Percentages of the Fluo4 AM-associated highly fluorescent neutrophil sub-population (population “P7”) from AA individuals (green squares) and from SS patients (red circles) previously incubated with phosphate-buffered saline (unlabeled), BQ123, or BQ788. (D) Percentages of the Fluo4 AM-associated highly fluorescent neutrophil sub-population measured within the first 30 s after addition of ET-1 (period P3), after phosphate-buffered saline, BQ123 or BQ788. Paired comparisons were made. *P<0.05 vs. ET-1 + PBS.

haematologica | 2017; 102(7)

1167

B. Koehl et al.

out that ET-1 altered human and mouse neutrophil viability as neither exogenous ET-1 nor ETA or ETB antagonists altered the proportion of apoptotic and necrotic neutrophils in vitro and in vivo, respectively (data not shown). Overall these results indicate that activation of the ETB receptor promotes neutrophil adhesion through action on two compartments: endothelial cells and neutrophils.

Discussion Polymorphonuclear neutrophils play a major role in the pathogenesis of endothelial injury characteristic of conditions such as ischemia-reperfusion and SCD.9 Sequestration of neutrophils into the microvasculature is the essential initiating step in neutrophil recruitment in inflammation and this exacerbates vaso-occlusion in SCD.2 Neutrophils from SCD patients and mice are activated, and show increased adhesion.2,11 The identification of specific mediators of neutrophil recruitment in SCD is important given that the broad spectrum and non-specific inhibition of neutrophil-endothelium interactions may lead to an increased risk of bacterial infections and trigger VOC and death. Following studies indicating that the ET system is activated in SCD individuals,27-32 our group previously showed that ET receptor antagonism not only blunts experimental VOC-associated vasoconstriction but also substantially inhibits acute VOC-induced increases of the numbers of peripheral blood total leukocytes (and specifically neutrophils), neutrophil infiltration of the bronchoalveolar space, renal and pulmonary myeloperoxidase activity and organ damage.19 Furthermore, early studies suggested that ET-1 acts on neutrophils by increasing intracellular free calcium mobilization,33 N-formyl-L-methionyl- L-leucyl-Lphenylalanine-mediated superoxide anion production,34 or aggregation.35 We, therefore, hypothesized that the ET system may play a direct role in neutrophil-endothelial interactions in SCD. We found that ET receptors are involved in several steps of neutrophil microvascular recruitment in SCD mice. Whereas rolling adhesion in sickle SAD mice involves the ETB receptor alone, firm adhesion and postadhesive dynamic behavior with transmigration involve both the ETA and ETB receptors. We also demonstrated an unexpected stimulatory role for the ETB receptor in adhesion of human neutrophils to endothelial cells under laminar flow conditions. Our findings suggest a differential role for ET receptors between mice and humans given that we found no pro-adhesive role for the ETA receptor in human neutrophil adhesion, in contrast to the situation in sickle mice. Another feature was the more intense and prolonged neutrophil adhesion to endothelial cells in vivo in sickle mice compared to control mice, whereas no differences were found in human neutrophil adhesion in vitro. Our study with human samples may have lacked statistical power, given the greater variability in neutrophil adhesion compared to that measured in mice. This might be due to species specificities and greater genetic heterogeneity in humans than in mice, a factor that should be taken into account when translating animal studies into the clinical setting. One limitation to the in vitro experiments was the lack of significant matching of AA with SS individuals providing blood, due to ethical constraints. Differences in 1168

models caused by in vitro manipulation of human blood with phenotypic heterogeneity of cultured endothelial cells should also be considered. Moreover, chronic in vivo exposure of the endothelium to cytokines, products of hemolysis and blood cell micro-vesicles may alter endothelial phenotype and trigger expression of different patterns of adhesion molecules.36 Nevertheless, both murine and human models provide consistent evidence for a powerful anti-adhesive role for ETB receptor antagonism. Our work also challenges the common opinion that adhesion of neutrophils to endothelial cells is always increased in SCD subjects. In fact, Fadlon et al., using a different approach (counting the radioactivity associated with adherent extensively centrifuged, washed and radiolabeled isolated neutrophils bound to layers of human umbilical vein endothelial cells in a static condition), could not show any difference in adhesion of neutrophils from 25 control subjects and 25 patients out of crisis to untreated and TNF-treated endothelial cells.12 Thus, our measurements of in-flow adhesion of neutrophils in whole blood from six healthy control subjects and 12 patients are consistent with the seminal results of Fadlon et al. In contrast, peripheral blood neutrophils from the majority of SCD patients in crisis were more adhesive to cultured endothelial monolayers than neutrophils from patients out of crisis or from healthy control subjects.162 Likewise, TNFα−injected SCD mice exhibited significant increases in neutrophil adhesion compared to AS mice, as measured by multi-channel fluorescence intravital microscopy analyses.2,37 In retrospect it is interesting to note that despite the absence of anemia in young animals, the SAD mouse model still presents remarkable neutrophil features in common with other SCD mouse models such as the βS ‘‘BERK’’ and SS knock-in sickle strains, with high counts of circulating leukocytes and neutrophils at steady state and during experimental VOC, thus mimicking what is observed in humans,19 and exquisite sensitivity to TNFαinduced leukocyte adhesion.2,37 One of the advantages of the SAD strain is its robustness with a fully controlled

Figure 5. Endothelin-1 secretion by neutrophils from healthy controls and sickle cell disease patients at steady state. ET-1 concentration in supernatants of human neutrophils. Results represent means of 22 independent samples [9 healthy controls (AA), 9 SCD patients without hydroxycarbamide therapy (SS)]. Concentration is normalized for 107 neutrophils per condition. *P<0.05 vs. AA.

haematologica | 2017; 102(7)

ETB receptor in neutrophil recruitment in SCD

to ET-1 clearance42 and stimulation of nitric oxide production which is essential for termination of ET-1 signaling.45,46 Our novel findings demonstrate an important contribution of the ETB receptor to vascular inflammation in SCD. Here, we provide the first series of systematic investigations aimed at unravelling the role of ET receptors in the different phases of the neutrophil-endothelial interaction. They complement earlier work showing that endothelins are chemoattractants for neutrophils47 in vitro. The first in vivo studies used perfusion of exogenous ET-1; this led to a time- and dose-dependent sequential entrapment of platelets and neutrophils in the pulmonary circulation.48 Similarly, the multi-step recruitment of rabbit peritoneal neutrophils was stimulated by ET-1 and inhibited by a specific antagonist of the ETA receptor.49 These findings were recently confirmed and expanded, as treatment with a mixed ETA/B receptor antagonist, bosentan, and selective

congenic C57Bl6/J background and easy breeding thus enabling experiments without the use of fetal liver or bone marrow transplant. It has also proven its relevance as a model of vaso-occlusion20,38-40 and of typical chronic degenerative organ injury.20,38-41 Thus, the SAD mouse model displays a characteristic inflammatory microvascular disease. At least at the level of neutrophil adhesion to the endothelium, our study indicates that this model shares common pathways relevant to the human condition. Contrasting with the findings of previous studies, which used pharmacological blockade or genetic deficiency models42-45 to unravel physiological functions of the ETB receptor, our current data suggest that neutrophil endothelial recruitment is primarily dependent upon an unblocked ETB receptor. However, previous studies used both nonSCD animal models and human subjects at steady state, in which the primary role of endothelial ETB is to contribute

Figure 6. ETB but nor ETA blockade prevents adhesion of neutrophils from healthy individuals and patients with sickle cell disease on activated endothelial cells and recombinant adhesion proteins. (A) Adhesion of neutrophils to endothelial cells (HMEC-1) over a 30 min period in a microfluidic system at a shear stress of 1 dyn/cm2, expressed as median area of neutrophil-associated fluorescence per field comparing six healthy controls (AA) to 12 SCD patients (SS). (B) Representative photomicrographs of anti-CD16 Alexa 488 (in green)-labeled neutrophils that have gradually adhered to the endothelial cells in microfluidic channels. Three channels were infused with the same batch of whole blood. Conditions were: NT: untreated blood; BQ123: blood preincubated with 1 μM BQ123; BQ788: blood previously treated with 1 μM BQ788. (C) Relative change in neutrophil adhesion on activated HMEC-1 after 30 min of infusion of whole blood from seven healthy AA controls (green circles) and 14 SS SCD patients (red triangles) previously incubated with 1 μM BQ123 or 1 μM BQ788 compared with non-treated blood. (D) Relative change in neutrophil adhesion 30 min after infusion of whole blood from five healthy donors (green circles) and five SCD patients (red triangles) on activated HMEC previously incubated with BQ123 or BQ788 compared with untreated HMEC. (E) Neutrophil adhesion on recombinant P-selectin, VCAM-1 and ICAM-1 proteins 30 min after infusion of whole blood from four healthy donors (green circles) and four SCD subjects (red triangles) previously incubated with BQ123 or BQ788 compared with nontreated blood. *P<0,05, ***P<0,001 with paired, non-parametric t-test (Wilcoxon test) vs. untreated blood or HMEC.

haematologica | 2017; 102(7)

1169

B. Koehl et al.

ETA and ETB receptor antagonists (BQ-123 and BQ-788, respectively) inhibited ET-1 and ovalbumin-induced neutrophil migration to the peritoneal cavity, suggesting that ET-1, acting through both ETA and ETB, is an important mediator of neutrophil recruitment in adaptive inflammation.50 Further studies will be needed to explore these chemotactic actions of ET-1. It is intriguing that ET receptor blockers had a significantly more potent effect on neutrophil adhesion in sickle SAD mice than in their WT counterparts. This may be due to a potentially stronger chemotactic ET-1 concentration gradient emanating from the pathological endothelium, creating higher ET-1 concentrations in the SCD circulation. We hypothesize that SCD-specific vascular inflammation may prime neutrophils and cause increased neutrophil adhesion and trans-endothelial migration of neutrophils to tissues through increased local ET-1 availability from endothelial cells, neutrophils and, potentially, other, yet to be identified blood cells. Another complementary hypothesis could be that neutrophils from SCD individuals may differ in their ability to degrade local endothelialderived and autocrine-derived ET-1, as reported for human neutrophils in vitro.51 Our findings may also foster further studies to investigate the role of ET-3, a selective ETB agonist that may be induced in SCD.52 Increased potency of ET receptor blockers on neutrophil adhesion in SCD conditions may also be related to distinct patterns of ET receptor expression, although we did not observe this, at least in human neutrophils. Excess of ligand usually promotes desensitization of G-protein-coupled receptors such as ET receptors, a counter-regulatory mechanism that may be defective in SCD neutrophils or endothelial cells. We did not perform any comparative dose-response studies to evaluate neutrophil sensitivity to ET-1 in SCD and control subjects. Defective desensitization or SCD-specific bias in post-receptor signaling could cause increased ET receptor-dependent cell adhesion. This would be consistent with the more intense and prolonged adhesion of neutrophils to post-capillary venules induced by TNFα in sickle mice compared to that in their WT congenic controls. Thus, the SAD mouse model exhibits exquisite sensitivity to TNFα, reminiscent of TNF-induced leukocyte stasis in the cremasteric venules, which was fatal in a high percentage of mice transplanted with bone marrow from the severe βS ‘‘BERK’’ mouse model.2,37 The reason for this magnified response to TNFα is still unclear. Such induced blood stasis may trigger VOC in the context of infections, as classically reported in septic patients. The current studies set the stage for investigation of the ET system in neutrophil recruitment in other conditions beyond SCD. Indeed, early studies suggested a particular role for neutrophil recruitment in models of ischemiareperfusion injury53,54 a condition that is particularly relevant to SCD-associated VOC. Conflicting reports regarding the respective roles of the ETA and ETB receptors suggest that pathophysiological context matters. Overall, both receptor subtypes have been shown to be involved in undefined steps of neutrophil recruitment55,56 and our findings further demonstrate the potent effect of ET receptor antagonists on neutrophil adhesion and transmigration in SCD with a particular emphasis on the role of the ETB receptor. Our results with human neutrophils and cultured endothelial cells in microfluidic flow chambers indicate that the ETB receptor promotes neutrophil adhesion 1170

through activation at both the neutrophil and the endothelial cell surface. In endothelial cells, the exact mechanisms whereby recruitment of neutrophils is dependent on an unblocked ETB receptor remain to be clarified. As discussed above, more efficient coupling of the ETB receptor to adhesion molecules may be at play. On the other hand, ET-1 was shown to promote neutrophil aggregation (homotypic adhesion)35,57 as well as β2integrin-dependent attachment of neutrophils to cultured human and bovine vascular endothelial cells (heterotypic cell adhesion). We provide here the first direct evidence of functional ETB receptors coupled to calcium signaling in neutrophils from control and sickle cell subjects. Together with our results showing that selective ETB blockade alleviated adhesion of human neutrophils to adhesion molecules, this suggests that ETB signaling may trigger calciuminduced activation of integrins. Interestingly, increased adhesion has been linked to the higher membrane expression of CD11b/CD18 on neutrophils from SCD patients than on those from controls.13 Furthermore, blocking CD11b/Mac-1 using M1/70 antibody was sufficient to diminish the erythrocyte–leukocyte interaction significantly, preventing VOC and prolonging survival in SCD mice challenged by TNFα and surgical trauma.23 These studies highlighted the role of CD11b in the pathophysiology of acute VOC in SCD. It is worth noting that expression of CD18 and CD11b on the surface of neutrophils was also increased by ET-1 in vitro53,58,59 although it was not investigated which ET receptor was involved. In line with these results, administration of TNFα induced very significant upregulation of CD11b expression in neutrophils from both WT and SAD animals but this effect was markedly asymmetric with a further 50% increase in the sickle cell condition. Surprisingly, ET receptor antagonists limited CD11b-associated mean fluorescence surface intensity and the proportion of neutrophils with high CD11b expression in neutrophils from TNFα−challenged SAD animals. Moreover, ETA and ETB antagonists had no effect on CD11b expression in normal mice. This suggests that involvement of the ET system in pro-inflammatory CD11b upregulation and neutrophil adhesion is specific to the SCD condition. Moreover, the expression of the integrin is probably not the only regulatory mechanism involved, given that rapid alterations in β2 integrin conformation, and consequently ligand affinity,60 could also play a very important role in neutrophil adhesion. Surprisingly, and contrary to previous reports, we found no difference in CD62L expression in neutrophils from sickle mice compared to normal ones.9,11 An explanation for this phenomenon might be that, in vivo, strongly activated neutrophils may have left the circulation and are therefore underrepresented in the blood samples used for flow cytometry analysis. In summary, our current findings suggest that human neutrophils display functional ETB receptors with calcium signaling capability, leading to accentuated adhesion to the endothelium. Using our high-speed imaging platform, we observed that the ETB response promoted rolling and firm adhesion of neutrophils, resulting in significant subsequent migration of these cells through venular walls to tissue, specifically in sickle cell mice. Furthermore, we confirmed that human neutrophils synthesize ET-1, which may be involved in autocrine and paracrine pathophysiological actions, and may contribute to the higher haematologica | 2017; 102(7)

ETB receptor in neutrophil recruitment in SCD

plasma levels of ET-1 reported in SCD individuals. Thus, the ET-1-ETB axis should be considered as a cytokine-like potent pro-inflammatory pathway in SCD. If ET receptor antagonists are proven to be safe and effective in the prevention or treatment of acute VOC in the clinical setting, they should include anti-ETB potency. Such antagonists may provide major clinical benefits, improve quality of life, and prolong survival of SCD patients. Acknowledgments The authors thank the “Région Ile-de-France” for funding for the

References 14. 1. Kaul DK, Finnegan E, Barabino GA. Sickle red cell-endothelium interactions. Microcirculation. 2009;16(1):97-111. 2. Turhan A, Weiss LA, Mohandas N, Coller BS, Frenette PS. Primary role for adherent leukocytes in sickle cell vascular occlusion: a new paradigm. Proc Natl Acad Sci USA. 2002;99(5):3047-3051. 3. Manwani D, Frenette PS. Vaso-occlusion in sickle cell disease: pathophysiology and novel targeted therapies. Blood. 2013;122(24):3892-3898. 4. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sickle cell disease. Life expectancy and risk factors for early death. N Engl J Med. 1994;330(23):1639-1644. 5. Anyaegbu CC, Okpala IE, Akren'Ova YA, Salimonu LS. Peripheral blood neutrophil count and candidacidal activity correlate with the clinical severity of sickle cell anaemia (SCA). Eur J Haematol. 1998;60(4):267-268. 6. Castro O, Brambilla DJ, Thorington B, et al. The acute chest syndrome in sickle cell disease: incidence and risk factors. The Cooperative Study of Sickle Cell Disease. Blood. 1994;84(2):643-649. 7. Ohene-Frempong K, Weiner SJ, Sleeper LA, et al. Cerebrovascular accidents in sickle cell disease: rates and risk factors. Blood. 1998;91(1):288-294. 8. Wigfall DR, Ware RE, Burchinal MR, Kinney TR, Foreman JW. Prevalence and clinical correlates of glomerulopathy in children with sickle cell disease. J Pediatr. 2000;136(6):749753. 9. Zhang D, Xu C, Manwani D, Frenette PS. Neutrophils, platelets, and inflammatory pathways at the nexus of sickle cell disease pathophysiology. Blood. 2016;127(7):801809. 10. Okpala I, Daniel Y, Haynes R, Odoemene D, Goldman J. Relationship between the clinical manifestations of sickle cell disease and the expression of adhesion molecules on white blood cells. Eur J Haematol. 2002;69(3):135-144. 11. Lard LR, Mul FP, de Haas M, Roos D, Duits AJ. Neutrophil activation in sickle cell disease. J Leukoc Biol. 1999;66(3):411-415. 12. Fadlon E, Vordermeier S, Pearson TC, et al. Blood polymorphonuclear leukocytes from the majority of sickle cell patients in the crisis phase of the disease show enhanced adhesion to vascular endothelium and increased expression of CD64. Blood. 1998;91(1):266-274. 13. Lum AF, Wun T, Staunton D, Simon SI. Inflammatory potential of neutrophils

haematologica | 2017; 102(7)

15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

intravital microscopy platform (SESAME 2007) and supporting Giulia Ghinatti through a CORDDIM fellowship. This study was supported by a grant from Labex GR-Ex. The Labex GR-Ex, reference ANR-11-LABX-0051, is funded by the “Investissements d’avenir” program of the French National Research Agency, reference ANR-11-IDEX-0005-02. We also thank the Fondation pour la Recherche Médicale – FRM (ING20121226435 to PL Tharaux) for supporting Dr. P. Nivoit and charitable funding from LVMH (2011/RDB/018, 2013/RDB/028). Finally, we thank Anna Chipont for excellent technical assistance and Elizabeth Huc and the ERI970 team for state-of-the-art animal care.

detected in sickle cell disease. Am J Hematol. 2004;76(2):126-133. Benkerrou M, Delarche C, Brahimi L, et al. Hydroxyurea corrects the dysregulated Lselectin expression and increased H(2)O(2) production of polymorphonuclear neutrophils from patients with sickle cell anemia. Blood. 2002;99(7):2297-2303. Davenport AP, Hyndman KA, Dhaun N, et al. Endothelin. Pharmacol Rev. 2016;68(2):357-418. Phelan M, Perrine SP, Brauer M, Faller DV. Sickle erythrocytes, after sickling, regulate the expression of the endothelin-1 gene and protein in human endothelial cells in culture. J Clin Invest. 1995;96(2):1145-1151. Firth J, Ratcliffe J. Organ distribution of the three rat endothelin messengers RNAs and the effects of ischemia on renal gene expression. J Clin Invest. 1992;90(3):1023-1031. Kedzierski RM, Yanagisawa M. Endothelin system: the double-edged sword in health and disease. Annu Rev Pharmacol Toxicol. 2001;41:851-876. Sabaa N, de Franceschi L, Bonnin P, et al. Endothelin receptor antagonism prevents hypoxia-induced mortality and morbidity in a mouse model of sickle-cell disease. J Clin Invest. 2008;118(5):1924-1933. Trudel M, Saadane N, Garel MC, et al. Towards a transgenic mouse model of sickle cell disease: hemoglobin SAD. EMBO J. 1991;10(11):3157-3165. Brun M, Bourdoulous S, Couraud PO, Elion J, Krishnamoorthy R, Lapoumeroulie C. Hydroxyurea downregulates endothelin-1 gene expression and upregulates ICAM-1 gene expression in cultured human endothelial cells. Pharmacogenomics J. 2003;3(4):215-226. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods. 2001;25(4):402408. Hidalgo A, Chang J, Jang JE, Peired AJ, Chiang EY, Frenette PS. Heterotypic interactions enabled by polarized neutrophil microdomains mediate thromboinflammatory injury. Nat Med. 2009;15(4):384-391. Elferink JG, de Koster BM. Stimulation and inhibition of neutrophil chemotaxis by endothelin-3. J Cardiovasc Pharmacol. 1995;26 (Suppl 3):S142-144. Elferink JG, de Koster BM. The effect of endothelin-2 (ET-2) on migration and changes in cytosolic free calcium of neutrophils. Naunyn Schmiedebergs Arch Pharmacol. 1996;353(2):130-135. Elferink JG, de Koster BM. Modulation of human neutrophil chemotaxis by the endothelin-B receptor agonist sarafotoxin

27.

28.

29. 30.

31.

32.

33.

34.

35.

36.

37. 38. 39.

S6c. Chem Biol Interact. 1996;101(3):165174. Graido-Gonzalez E, Doherty JC, Bergreen EW, Organ G, Telfer M, McMillen MA. Plasma endothelin-1, cytokine, and prostaglandin E2 levels in sickle cell disease and acute vaso-occlusive sickle crisis. Blood. 1998;92(7):2551-2555. Werdehoff SG, Moore RB, Hoff CJ, Fillingim E, Hackman AM. Elevated plasma endothelin-1 levels in sickle cell anemia: relationships to oxygen saturation and left ventricular hypertrophy. Am J Hematol. 1998;58(3):195-199. Rybicki AC, Benjamin LJ. Increased levels of endothelin-1 in plasma of sickle cell anemia patients. Blood. 1998;92(7):2594-2596. Tharaux PL, Hagege I, Placier S, et al. Urinary endothelin-1 as a marker of renal damage in sickle cell disease. Nephrol Dial Transplant. 2005;20(11):2408-2413. Chaar V, Tarer V, Etienne-Julan M, Diara JP, Elion J, Romana M. ET-1 and ecNOS gene polymorphisms and susceptibility to acute chest syndrome and painful vaso-occlusive crises in children with sickle cell anemia. Haematologica. 2006;91(9):1277-1278. Lionnet F, Bachmeyer C, Stankovic K, Tharaux PL, Girot R, Aractingi S. Efficacy of the endothelin receptor blocker bosentan for refractory sickle cell leg ulcers. Br J Haematol. 2008;142(6):991-992. Lopez Farre A, Riesco A, Moliz M, et al. Inhibition by L-arginine of the endothelinmediated increase in cytosolic calcium in human neutrophils. Biochem Biophys Res Commun. 1991;178(3):884-891. Ishida K, Takeshige K, Minakami S. Endothelin-1 enhances superoxide generation of human neutrophils stimulated by the chemotactic peptide N-formyl-methionylleucyl-phenylalanine. Biochem Biophys Res Commun. 1990;173(2):496-500. Gomez-Garre D, Guerra M, Gonzalez E, et al. Aggregation of human polymorphonuclear leukocytes by endothelin: role of platelet-activating factor. Eur J Pharmacol. 1992;224(2-3):167-172. Camus SM, De Moraes JA, Bonnin P, et al. Circulating cell membrane microparticles transfer heme to endothelial cells and trigger vasoocclusions in sickle cell disease. Blood. 2015;125(24):3805-3814. Zhang D, Chen G, Manwani D, et al. Neutrophil ageing is regulated by the microbiome. Nature. 2015;525(7570):528-532. Trudel M, De Paepe ME, Chretien N, et al. Sickle cell disease of transgenic SAD mice. Blood. 1994;84(9):3189-3197. De Franceschi L, Brugnara C, RouyerFessard P, Jouault H, Beuzard Y. Formation of dense erythrocytes in SAD mice exposed

1171

B. Koehl et al.

40.

41.

42.

43.

44.

45.

46.

1172

to chronic hypoxia: evaluation of different therapeutic regimens and of a combination of oral clotrimazole and magnesium therapies. Blood. 1999;94(12):4307-4313. de Franceschi L, Baron A, Scarpa A, et al. Inhaled nitric oxide protects transgenic SAD mice from sickle cell disease-specific lung injury induced by hypoxia/reoxygenation. Blood. 2003;102(3):1087-1096. De Paepe ME, Trudel M. The transgenic SAD mouse: a model of human sickle cell glomerulopathy. Kidney Int. 1994;46(5): 1337-1345. Fukuroda T, Fujikawa T, Ozaki S, Ishikawa K, Yano M, Nishikibe M. Clearance of circulating endothelin-1 by ETB receptors in rats. Biochem Biophys Res Commun. 1994;199(3):1461-1465. Ohuchi T, Kuwaki T, Ling GY, et al. Elevation of blood pressure by genetic and pharmacological disruption of the ETB receptor in mice. Am J Physiol. 1999;276(4 Pt 2):R1071-1077. Pollock JS, Pollock DM. Endothelin and NOS1/nitric oxide signaling and regulation of sodium homeostasis. Curr Opin Nephrol Hypertens. 2008;17(1):70-75. Goddard J, Eckhart C, Johnston NR, Cumming AD, Rankin AJ, Webb DJ. Endothelin A receptor antagonism and angiotensin-converting enzyme inhibition are synergistic via an endothelin B receptormediated and nitric oxide-dependent mechanism. J Am Soc Nephrol. 2004;15(10):26012610. Goligorsky MS, Tsukahara H, Magazine H,

47.

48.

49. 50.

51.

52.

53.

Andersen TT, Malik AB, Bahou WF. Termination of endothelin signaling: role of nitric oxide. J Cell Physiol. 1994;158(3):485494. Wright CD, Cody WL, Dunbar JB Jr, Doherty AM, Hingorani GP, Rapundalo ST. Characterization of endothelins as chemoattractants for human neutrophils. Life Sci. 1994;55(21):1633-1641. Helset E, Lindal S, Olsen R, Myklebust R, Jorgensen L. Endothelin-1 causes sequential trapping of platelets and neutrophils in pulmonary microcirculation in rats. Am J Physiol. 1996;271(4 Pt 1):L538-546. Elferink JG, de Koster BM. Endothelininduced activation of neutrophil migration. Biochem Pharmacol. 1994;48(5):865-871. Zarpelon AC, Pinto LG, Cunha TM, et al. Endothelin-1 induces neutrophil recruitment in adaptive inflammation via TNFalpha and CXCL1/CXCR2 in mice. Can J Physiol Pharmacol. 2012;90(2):187-199. Sessa WC, Kaw S, Hecker M, Vane JR. The biosynthesis of endothelin-1 by human polymorphonuclear leukocytes. Biochem Biophys Res Commun. 1991;174(2):613618. Makis AC, Hatzimichael EC, Kolios G, Bourantas KL. Circulating endothelin-3 levels in patients with sickle cell disease during hydroxyurea treatment. Haematologica. 2004;89(3):360-361. Espinosa G, Lopez Farre A, Cernadas MR, et al. Role of endothelin in the pathophysiology of renal ischemia-reperfusion in normal rabbits. Kidney Int. 1996;50(3):776-782.

54. Farmer DG, Kaldas F, Anselmo D, et al. Tezosentan, a novel endothelin receptor antagonist, markedly reduces rat hepatic ischemia and reperfusion injury in three different models. Liver Transplant. 2008;14(12): 1737-1744. 55. Khimenko PL, Moore TM, Taylor AE. Blocked ETA receptors prevent ischemia and reperfusion injury in rat lungs. J Appl Physiol. 1996;80(1):203-207. 56. Ghandour S, Cetinel S, Kurtel H. Endothelin-3 induced mesenteric vasoconstriction and PMN infiltration in the rat small intestine: role of endothelin receptors. Regul Pept. 2004;119(1-2):125-131. 57. Lopez-Farre A, Caramelo C, Esteban A, et al. Effects of aspirin on platelet-neutrophil interactions. Role of nitric oxide and endothelin-1. Circulation. 1995;91(7):20802088. 58. Zouki C, Baron C, Fournier A, Filep JG. Endothelin-1 enhances neutrophil adhesion to human coronary artery endothelial cells: role of ET(A) receptors and platelet-activating factor. Br J Pharmacol. 1999;127(4):969979. 59. Fernandez-Patron C, Zouki C, Whittal R, Chan JS, Davidge ST, Filep JG. Matrix metalloproteinases regulate neutrophil-endothelial cell adhesion through generation of endothelin-1[1-32]. FASEB J. 2001;15(12): 2230-2240. 60. Kim M, Carman CV, Springer TA. Bidirectional transmembrane signaling by cytoplasmic domain separation in integrins. Science. 2003;301(5640):1720-1725.

haematologica | 2017; 102(7)

ARTICLE

Iron Metabolism & Its Disorders

Imatinib and spironolactone suppress hepcidin expression Katarzyna Mleczko-Sanecka,1,2* Ana Rita da Silva,1* Debora Call,1 Joana Neves,1 Nikolai Schmeer,1 Georg Damm,3,4 Daniel Seehofer3,4 and Martina U. Muckenthaler1

Department of Pediatric Oncology, Hematology and Immunology, University of Heidelberg and Molecular Medicine Partnership Unit, Heidelberg, Germany; 2 International Institute of Molecular and Cell Biology, Warsaw, Poland; 3Department of Hepatobiliary Surgery and Visceral Transplantation, University of Berlin, Germany and 4 Department of Hepatobiliary Surgery and Visceral Transplantation, University of Leipzig, Germany

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

*KM-S, ARdS and DC contributed equally to this manuscript.

Haematologica 2017 Volume 102(7):1173-1184

ABSTRACT

isorders of iron metabolism are largely attributed to an excessive or insufficient expression of hepcidin, the master regulator of systemic iron homeostasis. Here, we investigated whether drugs targeting genetic regulators of hepcidin can affect iron homeostasis. We focused our efforts on drugs approved for clinical use to enable repositioning strategies and/or to reveal iron-related side effects of widely prescribed therapeutics. To identify hepcidin-modulating therapeutics, we re-evaluated data generated by a genome-wide RNAi screen for hepcidin regulators. We identified 'druggable' screening hits and validated those by applying RNAi of potential drug targets and small-molecule testing in a hepatocytic cell line, in primary murine and human hepatocytes and in mice. We initially identified spironolactone, diclofenac, imatinib and Suberoylanilide hydroxamic acid (SAHA) as hepcidin modulating drugs in cellular assays. Among these, imatinib and spironolactone further suppressed liver hepcidin expression in mice. Our results demonstrate that a commonly used anti-hypertensive drug, spironolactone, which is prescribed for the treatment of heart failure, acne and female hirsutism, as well as imatinib, a first-line, lifelong therapeutic option for some frequent cancer types suppress hepcidin expression in cultured cells and in mice. We expect these results to be of relevance for patient management, which needs to be addressed in prospective clinical studies.

Correspondence: martina.muckenthaler@med.uni-heidelberg.de or kmsanecka@iimcb.gov.pl Received: December 27, 2016. Accepted: April 5, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.162917

Introduction Hepcidin is a liver-expressed regulatory hormone that adjusts iron fluxes to body iron requirements.1 Conditions of high iron load or inflammation activate hepatic hepcidin production, thereby reducing plasma iron levels. Iron deficiency or augmented erythropoietic activity inhibit hepcidin expression to increase systemic iron availability. Disorders of iron metabolism belong to the most common diseases worldwide and are to a large extent attributed to excessive or insufficient levels of hepcidin. The frequent iron overload disorder hereditary hemochromatosis (HH) is caused by inappropriately low hepcidin expression.1 Likewise, hepcidin deficiency leads to iron accumulation in several subtypes of hereditary anemias (e.g. thalassemias), termed iron-loading anemias, that are hallmarked by ineffective yet augmented hematopoiesis.2 Of clinical importance, progressive tissue iron deposition, especially in the heart, represents the major cause of mortality in these disorders.3 Furthermore, alterations of iron levels or its misdistribution exacerbates pathologies of common acquired diseases such as chronic liver disease,4 diabetes,5 cardiovascular disease or neurodegenerative disorders.6 Conversely, the anemia of inflammation (AI), caused by iron retention within body iron reservoirs and hypoferremia, develops due to hepcidin induction by immune activation.7 The progression of the AI haematologica | 2017; 102(7)

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1173 ÂŠ2017 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.

1173

K. Mleczko-Sanecka et al.

increases morbidity rates, often prolongs time of hospitalization and is associated with a poorer prognosis in a variety of disease conditions.8 Current therapies for iron-related pathologies target symptoms of the disease rather than the underlying molecular mechanisms. Standard treatment of HH involves repetitive phlebotomies, which is not suitable for other patients with iron overload (e.g. for those suffering from concomitant anemias) and may be additionally compromised by suboptimal compliance.9 Treatment of choice for iron-loading anemias involves iron chelation, which, although effective, causes significant side effects.10 If combating the underlying pathology is not feasible, the current therapeutic options for the AI include blood transfusions, erythropoiesis stimulating agents (ESA) and parenteral iron administration.8 However, blood transfusions were shown to be associated with multiorgan failure in critically ill patients.11 ESA therapy for cancer-associated anemia increased incidence of tumor progression, cardiovascular pathologies and death,12 whereas the long-term safety of intravenous iron supplementations is still unknown. Because available therapies show limitations and adverse effects, there is substantial interest in identifying novel approaches that modify hepcidin levels or impact on its molecular function. Several novel strategies are currently under development, which include minihepcidins (hepcidin agonists), RNAi therapeutics targeting hepcidin regulators, anti-hepcidin drugs such as spiegelmers, anticalins or monoclonal antibodies, and non-coagulating heparins.13 Despite successful application of these agents in animal models, their clinical use may be challenging.13 Moreover, since these new therapeutics have not yet been approved for patients' therapy, their long-term safety and efficacy are unknown. The list of drugs licensed for therapy in patients consists of approximately 1900 small molecules, which are expected to impact on numerous molecular processes and for which a substantial amount of clinical data is available (http://www.fda.gov/). The spectrum of signals that control hepcidin expression extends beyond the originally identified bone morphogenetic proteins or inflammatory cues,1 and involves growth factors,14 erythroferrone,15 sex hormones,16 nutrient stimuli17 or retinoids.18 This demonstrates that multiple signaling events converge at the hepcidin promoter, which may imply that several drugs approved for clinical practice may modulate hepcidin transcription. Such a possibility opens avenues for the re-purposing of known therapeutics for the management of iron-related diseases,19 an approach which may offer an attractive alternative to both current and novel ironfocused therapies. Furthermore, the existence of a broad range of pathways that modulate hepcidin levels implies that disturbances of iron homeostasis may represent an overlooked side effect of licensed drugs. In this study, we reanalyzed data generated by a genome-wide RNAi screen for hepcidin regulators17 to identify putative drug targets and applied extensive small-molecule testing to identify drugs and drug targets that modulate hepcidin expression. We identified two widely prescribed drugs administered long term in patients, imatinib and spironolactone, as suppressors of hepcidin expression in primary hepatocytes and in mice. These findings may be of clinical relevance in that iron-related side effects of these therapeutics may be identified in further clinical studies. 1174

Methods Cell culture and drug treatment The human hepatocarcinoma cell line Huh7 was obtained from ATCC (Wesel, Germany). Human liver tissue was obtained from macroscopically healthy human liver tissue that remained from resected human liver of patients with primary or secondary liver tumors or benign local liver diseases. Informed consent of the patients for the use of tissue for research purposes was obtained according to the ethical guidelines of the Charite University Hospital Berlin. Protocols for isolation of human and murine primary hepatocytes (PHs) as well as culture conditions for all cell models used in the study are described in the Online Supplementary Appendix. If not specified, Huh7 cells and primary hepatocytes were treated with drugs in complete, serum-containing medium. The list of drugs used in the study is summarized in Online Supplementary Table S1.

Transfection of siRNAs and luciferase reporter constructs As reported previously,17 we reverse transfected Huh7 cells with 10 pmol siRNA (pooled or individual duplexes; Dharmacon, Online Supplementary Table S2) in a 96-well format. For RNAi of weakly-expressed candidate genes we applied reverse transfection in 24-well plates with 13x104 cells in each well, using 50 pmol of siRNA and 1.5 μl Dharmafect1 (Dharmacon). Cells were cultured for 72 hours prior to harvesting of total RNA. Luciferase reporter constructs that contain the full-length 2,762 bp hepcidin promoter (WT_2.7kb) or its mutant derivatives (STAT_BS_2.7kb and BMP_RE1_BMP_RE2_2.7kb) were described previously.20 Twenty-four hours after seeding Huh7 cells, these were transfected with reporter plasmids using Lipofectamine 2000 Transfection Reagent (2 μl/well; 200 ng of hepcidin promoter constructs, 20 ng of CMV-Renilla control plasmid). Twenty-four hours later cells were treated with drugs at the indicated concentrations and incubated for an additional 24 hours. Luciferase activity was measured as described before.21

Preparation of total RNA, reverse transcription, and quantitative real-time PCR analysis Isolation of total RNA and the protocols for reverse transcription (RT) and real-time qPCR were described previously.20 Total RNA extraction from mouse livers was performed using Trizol (Invitrogen). Total RNA extraction for the secondary RNAi assays in 96-well plates was performed using the QuickExtract™ RNA Extraction Kit (Epicentre Biotechnologies, Madison, WI, USA) according to the manufacturer’s instructions.17 RNA extraction from all other cell-based experiments was performed using RNeasy® kit (Qiagen). Sequences of the qPCR primers are shown in Online Supplementary Table S3. Expression levels of human GAPDH (glyceraldehyde-3-phosphate-dehydrogenase) or murine PpIb or Rpl19 were used as normalization controls.

Mice Experiments were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee of the EMBL Heidelberg. Male wild-type C57BL/6 mice were maintained with free access to food and water. At 11 weeks of age mice were administrated with the following drugs dissolved in drinking water for two weeks: imatinib (75 mg/kg/day), spironolactone (24 mg/kg/day), Suberoylanilide hydroxamic acid (SAHA; 50 mg/kg/day) and diclofenac (1 mg/kg/day). As described in detail in the Online Supplementary Appendix, the doses applied were chosen based on literature and U.S. Food and Drug Administration (FDA) haematologica | 2017; 102(7)

Drug targeting of hepcidin expression

approval documents to provoke desired biological responses without causing significant side effects. The drug doses used in mice correspond closely to the doses applied in human patients, based on the body surface area (BSA) dose escalation method.22,23 Since SAHA, spironolactone and diclofenac were difficult to dissolve in aqueous solutions, we increased their solubility in water by complexing with 2-hydroxypropyl-β-cyclodextrin (HOPβ-CD) (Online Supplementary Appendix for more details). Two different doses of HOPβ-CD were applied: 9 g/L for SAHA and diclofenac, and 1.1 mg/L for imatinib and spironolactone. As controls, mice administrated with an equivalent dose of HOPβ-CD dose were studied. Plasma levels of hepcidin peptide were measured using the ELISA kit from DRG Diagnostics according to the manufacturer’s protocol. Plasma iron levels were measured using SFBC and UIBC Iron Kits (Biolabor). Tissue non-heme iron content was measured using the bathophenanthroline method and calculated against dry weight tissue.24

Statistical analysis Statistical analyses were performed using Prism v.6 (GraphPad). Data from cell-based experiments are shown as a fold change ± 95% confidence intervals (CI) compared to control conditions. Data from mouse experiments are shown as mean ± SEM. One sample t-test or two-tailed Student’s t-test were used for estimation of statistical significance, respectively.

Results Identification of drug targets among putative hepcidin regulators We recently reported the outcome of an unbiased genome-wide RNAi screen that generated a comprehensive list of putative hepatic hepcidin regulators.17 We now further analyzed the hit list of this screen by utilizing ‘Manually Annotated Targets and Drugs Online Resource’ (MATADOR; http://matador.embl.de/), a resource of drug and drug-target interactions.25 Among 2159 putative hepcidin regulators, our bioinformatic analysis identified approximately 160 screening hits that are potentially sensitive to pharmacological manipulation. Importantly, the initial selection process was unbiased in terms of gene function, specific pathways or expression profiles. To further enrich for clinically relevant gene-drug pairs we applied literature mining to manually exclude biologically 'unspecific' hits (e.g. those involved in general drug metabolism or those that were only distally related to real drug targets). From the remaining 56 direct and specific drug targets, we excluded 19 genes which could not be reliably detected at transcript level in Huh7 cells, a cellular model used in the RNAi screen (data not shown).17

Secondary validation assays confirm six drug targets as hepcidin regulators Thirty-seven putative 'druggable' hepcidin modifiers were selected for RNAi-based validation in a 2-step manner. In a first step, we validated each drug target gene for its ability to modulate endogenous hepcidin mRNA levels upon siRNA-mediated knockdown with a pool of four single siRNAs. Thirteen genes, including 4 potential hepcidin suppressors and 9 potential activators were confirmed [Figure 1A; data for HDAC3 included also in Pasricha et al., (submitted manuscript)]. In a second step we individually applied four single siRNAs per gene and considered a gene as validated if at least two out of four haematologica | 2017; 102(7)

siRNAs significantly altered hepcidin mRNA expression (Figure 1B). By applying these criteria, we identified histone deacetylase 3 (HDAC3) and retinoic acid receptor beta (RARB) as hepcidin suppressors. These results are consistent with previous findings that HDAC inhibitors26 and retinoic acid18 exert suppressive effects on hepcidin expression. Our RNAi validation strategy further identified 4 additional, functionally diverse genes as hepcidin activators: membrane progestin receptor gamma (MPRG), NAD(P)H quinone dehydrogenase 2 (NQO2), prostaglandin I2 (prostacyclin) synthase (PTGIS) and phospholipase C beta 3 (PLCB3). MPRG acts in intestinal endocrine cells to regulate glucose homeostasis,27 NQO2 oxidoreductase catalyzes metabolic detoxification of quinones28 and regulates levels of the tumor suppressor p53,29 PTGIS synthesizes prostacyclin,30 a potent vasodilator and inhibitor of platelet aggregation, and PLCB3 generates the second messenger molecules diacylglycerol and inositol 1,4,5-trisphosphate involved in signal transduction.31 Beside the fact that PTGIS uses iron as a co-factor, roles of the identified genes in iron metabolism have not been reported.

Small-molecule testing identifies four approved therapeutics as hepcidin modulators in Huh7 cells To identify hepcidin-modifying drugs, we next retrieved those compounds from the MATADOR database that target the 13 genes validated by application of siRNA pools (Table 1 and Figure 1A). We did not further consider retinoic acid in our experiments as it was previously shown to affect hepcidin mRNA levels in Huh7 cells.18 Furthermore, administration of diazepam, the GABA receptor agonist, was not possible due to legal reasons. As shown in Table 1, most of the selected drugs are routinely used in clinical practice and are prescribed for a wide range of diseases. We next tested these drugs in a time- and dose-dependent manner. To establish the optimal incubation time, Huh7 cells were exposed to the selected small molecules for various time intervals, including 4, 8, 16 and 24 hours (data not shown). Subsequently, different drug doses were tested. In these experiments we identified five therapeutics as modulators of hepcidin mRNA levels. We show that diclofenac, an analgesic classified as a non-steroidal anti-inflammatory agent (NSAID) and the histone deacetylase inhibitor (HDACi) SAHA (Vorinostat) induce hepcidin mRNA expression in a dose-dependent manner (Figure 2A). In addition, the anti-hypertensive drug spironolactone and the pan-tyrosine kinase inhibitor imatinib, which is broadly applied in cancer therapies, suppress hepcidin mRNA expression in hepatoma Huh7 cells (Table 1 and Figure 2A). Of note, the potassium-sparing diuretic amiloride induced hepcidin expression in Huh7 cells (data not shown), but was excluded from our analysis due to its toxicity in primary hepatocyte cultures. We next wondered whether the drugs that modulate hepcidin expression act by involving well-established signaling pathways, the inflammatory JAK-STAT or the BMP pathway.1 To test this we utilized luciferase reporter constructs driven by the hepcidin full-length promoter and its mutant derivatives.20,32 We did not detect significant responses of the hepcidin promoter to SAHA, diclofenac and amiloride treatment (data not shown). By contrast, promoter activity was reduced following treatment with spironolactone and imatinib. The drug responses were 1175

K. Mleczko-Sanecka et al.

preserved upon mutation of the STAT-binding site in the hepcidin promoter, but were attenuated in promoter constructs lacking functional BMP-responsive elements (REs) 1 and 2. These data suggest that intact BMP signaling at the level of the hepcidin promoter contributes to the modulation of hepcidin transcription by these hepcidin-suppressing drugs.

levels consistently in murine and human hepatocytes, whereas spironolactone was only active in human PHs, where it potently suppressed hepcidin mRNA expression. Taken together, these data demonstrate that the response of hepcidin to the tested therapeutics is not restricted to hepatoma cells and is preserved in primary hepatocytes.

Identification of drug target - drug relationships Diclofenac, imatinib, SAHA and spironolactone affect hepcidin levels in primary hepatocyte cultures We next exposed murine and human primary hepatocytes (PHs) to the selected drugs (Figure 2B and C). Diclofenac, imatinib and SAHA altered hepcidin mRNA

The combination of RNAi and small-molecule testing identified 5 genes and 4 drugs that operate as hepcidin modulators (Table 2). Application of SAHA (Vorinostat), a broad-range HDAC inhibitor (HDACi), induced hepcidin mRNA expression in all three hepatocytic cell models

Figure 1. Two-step validation RNAi assays identify six novel hepcidin regulators. (A) Thirteen screening hits were validated by the analysis of endogenous hepcidin mRNA levels upon knockdown with pools of 4 individual siRNAs in Huh7 cells. RNAi of SMAD7, SMAD4 and STAT3 served as a control. (B) Six regulators of hepcidin mRNA expression were validated by at least two independent siRNAs. Shown are mRNA levels of hepcidin and target genes after RNAi. Results are presented as a fold change (Âą 95% CI) compared to samples transfected with scrambled siRNA. The mean of at least three independent experiments is shown. Significant changes are indicated by asterisks (*P<0.05, **P<0.005, ***P<0.001).

1176

haematologica | 2017; 102(7)

Drug targeting of hepcidin expression

(Figure 3A-C), supporting previous data that the HDACi trichostatin A induces hepcidin in hepatoma cells.26 Increased hepcidin mRNA levels in response to SAHA treatment of Huh7 cells correlated well with the hepcidin response to RNAi of HDAC3 (Figure 1), one target of SAHA. Further analysis showed that HDAC3 may be critical for hepcidin regulation, because Panobinostat (LBH589) and Trapoxin A, two pan-HDACi targeting HDAC3 within their inhibitory spectrum, induced hepcidin levels (Figure 3A). Furthermore, hepcidin mRNA levels remained unaltered upon treatment with those compounds that do not target HDAC3, including the HDAC8-specific inhibitor Compound 2 and Entinostat (MS-275), an inhibitor targeting preferentially HDAC1 (Figure 3A). Surprisingly, murine and human PHs exposed to the same panel of HDAC inhibitors showed a different pattern of hepcidin alteration (Figure 3B and C). These data suggest that contributions of specific HDACs to hepcidin suppression may differ between hepatoma cells and primary cell models, and that HDACs other than HDAC3 may be

involved in modulating hepcidin levels. This conclusion is supported by the finding that RNAi to Hdac3 in murine primary hepatocytes failed to alter hepcidin mRNA levels (data not shown). Independent of the specific HDAC involved in modulating hepcidin mRNA expression, these data, together with previous findings26 suggest that inhibition of histone deacetylation could represent an important intervention point for modulating hepcidin expression. Imatinib was selected from the MATADOR database, because it targets two tyrosine kinases (PDGFRB and TEC) identified in the RNAi screen as candidate hepcidin regulators (Table 1 and Figure 1A), a finding that we could not validate by further RNAi-based analysis (data not shown). Detailed literature mining however suggested NQO2 as a non-kinase target of imatinib.33 NQO2 was likewise identified as a hepcidin activator in our study. Interestingly, quercetin, a potent NQO2 inhibitor,34 attenuates hepcidin mRNA expression (Figure 3D), suggesting that NQO2 and its inhibitor imatinib may control hepcidin expression. It is of note that according to the experi-

Table 1. Drugs targeting hepcidin regulators identified by an RNAi screen.

Gene (drug target)

Drug

Description

Indication / main clinical use

RARB

Alitretinoin

Psoriasis, other skin diseases, promyelocytic leukemia

SLC38A3

Amiloride

Agonist of retinoic acid receptors, antineoplastic agent Potassium-sparing diuretic

HDAC3 GABRR1

Vorinostat (SAHA) Diazepam

Antineoplastic agent Agonist of gamma-aminobutyric acid (GABA)

NQO2

Dicumarol

OSBPL10

Spironolactone

PLCB3

Pentoxyfylline

PTGIS

Diclofenac

SLC22A8

Ranitidine

Anticoagulant vitamin K reductase inhibitor Potassium-sparing diuretic; antagonist of aldosterone Hemorrheologic, vasodilator agent Non-steroidal anti-inflammatory agent (NSAID) Histamine receptor antagonist

HMGCR

Lovastatin

Cholesterol-lowering agent HMG-CoA reductase inhibitor

PDGFRB TEC

Imatinib

Antineoplastic agent Protein kinase inhibitor

RNAi causes hepcidin suppression

Congestive heart failure, hypertension, Connâ&#x20AC;&#x2122;s syndrome Cutaneous T-cell lymphoma Epilepsy

RNAi causes hepcidin activation Oral anticoagulation Secondary hyperaldosteronism, renin deficiency, hypertension, Conn's syndrome Chronic occlusive arterial disease of the limbs Acute and chronic pain Peptic ulcer disease, gastroesophageal reflux disease Primary hypercholesterolemia, mixed dyslipidemia, prevention of coronary heart disease Philadelphia chromosome positive chronic myeloid leukemia (CML), gastrointestinal stromal tumors (GIST)

Depicted are those drugs that were selected based on the MATADOR25 database to target hepcidin regulators validated by RNAi (Figure 1). Alitretinoin and Diazepam were excluded from further analysis (please see main text for details). A short description of the therapeutics together with their main clinical application is indicated and was guided by information available in the 'DrugBank' database.36

haematologica | 2017; 102(7)

1177

K. Mleczko-Sanecka et al. mental data NQO2 is insensitive to dicumarol,34 a compound linked to NQO2 in the MATADOR database (Table 1) due to its similarity to the known dicumarol target NQO1. Hence, the lack of dicumarol effect on hepcidin expression (Table 2) is consistent with previous knowledge. Guided by information obtained from the MATADOR database, the analgesic drug diclofenac was selected to target PTGIS, a gene encoding the prostacyclin synthase. As

shown in Table 2, the results obtained from the analyses of this drug-drug target pair are inconsistent. However, diclofenac may not directly target PTGIS, but, as other NSAIDs, may predominantly inhibit cyclooxygenases (COX-1 and COX-2).35 Cyclooxygenases are the rate-limiting enzymes in the synthesis of prostaglandin H2, that is then further converted to the vasoactive prostacyclin by PTGIS.35 Importantly, treatment of cells with Aspirin, another member of the NSAID family known to inhibit

1178

Figure 2. Drugs identified as hepcidin modulators in Huh7 cells and primary hepatocytes. (A,C,D) Cells were exposed to increasing concentrations of individual drugs for the indicated time points. (B) Effects of 24-hour drug treatments (10 μM imatinib, 50 μM spironolacton) on hepcidin promoter activity in cells transfected with wild-type (WT_2.7) and mutant hepcidin promoter constructs (ST: STAT-BS mutant, B1/B2: double mutant of BMP-RE1 and 2). Results are presented as a fold change (± 95% CI) of hepcidin mRNA levels (A,C,D) or hepcidin promoter activity (Firefly/Renilla luciferase signal) (B) compared to vehicle-treated control cells. The mean of at least three independent experiments is shown. Significant changes are indicated by asterisks: (*P<0.05, **P<0.005, ***P<0.001).

haematologica | 2017; 102(7)

Drug targeting of hepcidin expression

COX-1 and to a lesser extent COX-2,35 failed to control hepcidin mRNA levels (data not shown). Taken together, these findings suggest that diclofenac controls hepcidin expression via a mechanism independent of PTGIS and cyclooxygenases. Spironolactone was selected for analysis, because it targets and antagonizes OSBPL10, a member of the oxysterol-binding protein-related proteins (ORP) family. Despite suppressing effects of spironolactone on hepcidin expression, OSBPL10 was not validated by our RNAi assays. More detailed analysis of an additional drug-target database 'DrugBank' (www.drugbank.ca)36 revealed that ORPs are only distally related to bona fide targets of spironolactone, aldosterone and androgen receptors.36-38 To further corroborate the role of spironolactone in hepcidin regulation and to address its mode of action, we treated human PHs with eplerenone, a highly specific blocker of the aldosterone receptor.36 Interestingly, we found a dose-dependent reduction of hepcidin mRNA levels in response to eplerenone supplementation (Figure 3E), suggesting that aldosterone, but not androgen signaling may cross-talk to hepcidin transcriptional control in primary hepatocytes.

Imatinib and spironolactone modulate hepcidin levels in vivo We next investigated whether the drugs that we identified in cellular assays (spironolactone, diclofenac, imatinib and SAHA) modulate hepcidin levels in mice, when administered in the drinking water. Drug doses were selected based on literature and FDA documentation (see Online Supplementary Appendix) in such a way that they would provoke the desired biological responses without causing significant side effects. Mice were treated for a period of two weeks, shown by previous studies to be suf-

ficient to affect systemic iron parameters.13 Generally, the applied drugs were well tolerated and did not affect the weight of the mice (data not shown). Diclofenac caused a mild decrease of red blood cell numbers, hemoglobin levels and hematocrit (Table 3), whereas blood parameters remained unaltered following treatment with all other drugs. Hepatic hepcidin mRNA levels remained unchanged upon administration of SAHA and diclofenac (Figure 4A), while treatment with imatinib and spironolactone significantly reduced hepcidin mRNA levels (Figure 4A). Spironolactone treatment of mice further decreased plasma hepcidin levels (Figure 4B). Hepatotoxicity unlikely contributes to diminished hepcidin expression, because plasma levels of the liver damage marker alanine aminotransferase (ALT) remained unaffected (Figure 4C). Imatinib and spironolactone treatment further caused a decrease of unsaturated iron binding capacity (UIBC), and an increase of transferrin saturation, without a significant change in serum iron levels (Figure 4D). Two weeks of drug supplementation was not sufficient to alter tissue iron levels (Online Supplementary Figure S1). Taken together, our data show that imatinib and spironolactone reduce hepcidin levels in cellular assays and mice, and mildly affect plasma iron parameters, suggesting that altered body iron levels may arise as a complication of long-term drug application. This will have to be tested in future studies.

Discussion The knowledge about processes and signaling mechanism which modulate hepcidin expression is constantly growing. We hypothesized that licensed therapeutics

Table 2. Summary of drug targets and drugs drugs identified as hepcidin modulators in hepatocytic cellular models.

Gene (drug target)

MPRG OSBPL 10 PLCB3 SLC22A8 PTGIS NQO2 NQO2 PDGFRB TEC HMGCR RARB HDAC3 GABRR1 SLC38A3

Hepcidin mRNA levels upon RNAi in Huh7 cells ↓ ns ↓ ns ↓ ↓ ↓ ns ns ns ↑

Drug

Drug effect on hepcidin in Huh7 cells

Candidate hepcidin activator Progesterone Spironolactone Pentoxifylline Ranitidine Diclofenac Dicumarol Imatinib Imatinib Imatinib Lovastatin Diazepam Candidate hepcidin repressor ↑ SAHA (Vorinostat) + other HDACi ns ns Amiloride

ns ↓ ns ns ↑ ns ↓ ns -

Drug effect on hepcidin in murine PH

Drug effect on hepcidin in human PH

↓↓

↑

↓

↓↓

↑

↑↑

↑

↓

toxic

Genes were considered as validated if at least 2 of the 4 siRNAs tested caused a significant change in hepcidin mRNA expression. Drugs were validated as hepcidin modulators using doses and time points shown in Figures 2 and 3 (see main text for details). Statistically significant repression or induction of hepcidin mRNA is marked with a downward arrow or upward arrow, respectively. Blank cells indicate that data are not available; ns: non-significant; PH: primary hepatocytes; HDACi: histone deacetylase inhibitors.

haematologica | 2017; 102(7)

1179

K. Mleczko-Sanecka et al.

affect some of these pathways and thus systemic iron levels. This assumption has two medically-relevant implications. First, it may open a possibility for drug re-positioning, a time- and cost-effective approach of finding new applications for compounds approved for clinical practice.19 Second, it may suggest iron-related side effects caused by FDA-approved drugs, which may not be detected by routine diagnostics.

We therefore reanalyzed data from our recently reported RNAi screen17 for hepcidin regulators to select genes that are 'druggable' hepcidin modulators. These genes were validated by RNAi and small-molecule testing was applied to identify the corresponding drugs as hepcidin modulators. This approach complements and extends previously reported drug screens in search for hepcidin regulators,39,40 in that it starts with a 'pre-selected' set of drugs

Figure 3. Refinement of drug target-drug relationships. (A-C) The indicated cell models were exposed to a panel of histone deacetylase (HDAC) inhibitors, SAHA (1 μM), Panobinostat (0.1 μM), Trapoxin A (0.08 μM), Entinostat (1 μM) and Compound-2 (250 nM) for 8 hours (h), in complete medium (A), or in serum-free medium (after a 24-h serum starvation) (B and C). (D) Huh7 cells were treated with increasing concentrations of quercetin for 24 h. (E) Human primary hepatocytes (PH) were exposed to increasing doses of eplerenone for 24 h. All results are presented as a fold change (± 95%CI) of hepcidin mRNA levels. The mean of at least three independent experiments is shown. Significant changes are indicated by asterisks: *P<0.05, **P<0.005, ***P<0.001.

1180

haematologica | 2017; 102(7)

Drug targeting of hepcidin expression

based on a comprehensive RNAi approach, applies timeand dose-response experiments, and involves validations in primary cells and mice. With this approach we identified four compounds that modulate hepcidin mRNA levels in primary hepatocyte models (Figure 2). Two of these, imatinib and spironolactone, further affected hepcidin levels in mice, causing a mild increase of transferrin saturation, without causing hepatotoxicity (Figure 4). Whether these relatively modest alterations of iron parameters affect tissue iron content during long-term therapy in patients requires further investigation.

Assessment of the potential for drug re-positioning The safety profile of imatinib excludes this compound from the possibility for drug re-positioning. Imatinib may cause peripheral edema, nausea, vomiting, rash, fatigue and abdominal pain.41,42 In addition, patients on imatinib may develop congestive heart failure due to severe left ventricular dysfunction.41

By comparison, spironolactone shows relatively minor adverse effects. It was approved in 1959 and initially was used as supplementary therapy for refractory hypertension.37Although it is well-tolerated, antiandrogenic effects of spironolactone were reported, mainly due to cross-reactivity with the androgen receptor and alterations of peripheral metabolism of testosterone.38 These properties of spironolactone were linked to undesirable effects in males (e.g. gynecomatia), but in women were exploited to successfully treat acne and hirsutism.43,44 In the latter case spironolactone therapy commonly caused menstrual irregularities and increased diuresis. These side effects were transient and usually not considered as sufficient for treatment discontinuation. Taken together, the safety profile of spironolactone, supported by its potential to affect hepcidin levels in vivo, make it promising for re-purposing for diseases hallmarked by inappropriately high hepcidin synthesis. Whether spironolactone can modulate hepcidin expression under disease-associated conditions, such as

Figure 4. Treatment with imatinib and spironolactone suppresses hepcidin in mice. Wild-type 11-week old mice received imatinib (75 mg/kg/day), spironolactone (24 mg/ kg/day), suberoylanilide hydroxamic acid (SAHA) (50 mg/kg/day) and diclofenac (1 mg/kg/day) for two weeks with drinking water. (A) Relative liver mRNA expression of hepcidin and (B) plasma hepcidin levels in drug-treated mice compared to respective controls. (C) Plasma alanine aminotransferase (ALT) levels for the indicated treatment groups. (D) Plasma iron content (SFBC), unsaturated iron binding capacity (UIBC) and transferrin saturation in response to drug administration. All data are presented as a meanÂąSEM from the indicated number of mice per treatment group. t-test, *P<0.05.

haematologica | 2017; 102(7)

1181

K. Mleczko-Sanecka et al. Table 3. Blood parameters in mice treated with hepcidin-modifying drugs. Control (n=5) Imatinib(n=5) Spironolactone (n=5) Control (n=4) Diclofenac (n=6) SAHA (n=4)

RBC (x1012/L)

HGB (g/dL)

HCT (%)

MCV (fl)

MCH (pg)

10.58±0.41 9.93±0.12 10.51±0.39 10.64±0.13 9.88±0.24* 10.20±0.09

16.79±0.61 16.26±0.24 16.48±0.56 16.58±0.14 15.22±0.31* 15.64±011

56.12±2.12 53.06±0.48 55.09±1.85 55.34±0.87 51.14±1.10* 52.16±0.29

53.00±0.32 53.40±0.24 52.60±0.24 51.75±0.25 52.00±0.26 51.25±0.25

15.88±0.09 16.38±0.27 15.68±0.20 15.59±0.09 15.42±0.14 15.33±0.07

RBC: red blood cell count; MCV: mean corpuscular volume; MCH: mean corpuscular hemoglobin. Data are shown as mean ± SEM and represent data from the indicated number of age-matched male mice per group t-test. *P<0.05.

inflammation, and whether prolonged spironolactonemediated hepcidin suppression affects tissue iron content remain to be tested.

Iron-related side effects caused by FDA-approved drugs Growing evidence suggests that a dysregulated systemic iron balance represents an important modifying factor for several common diseases, e.g. chronic liver disease, diabetes or atherosclerosis.4-6 Furthermore, epidemiological studies indicate that an increased body iron status may increase cancer risk.45 In this context it is of clinical relevance to identify iron-related side effects of frequently prescribed therapeutics. Imatinib represents a lifelong first-line therapeutic option in chronic myeloid leukemia, gastrointestinal stoma tumors and several other rare disorders (e.g. those hallmarked by platelet-derived growth factor aberrations).46 Here we report that imatinib treatment decreased hepcidin mRNA levels in vivo (Figure 4A). Whether this will cause imbalanced systemic iron levels under long-term treatment needs to be investigated. Interestingly, liver iron overload was reported in a patient treated with imatinib.47 It remained unclear whether in this single case a hepatotoxic response contributes to hepcidin suppression. Our data in mice however suggest that imatinib affects hepcidin transcription independently of liver damage (Figure 4A and C). It will be interesting to systematically analyze iron-related parameters, such as serum ferritin, transferrin saturation and hepcidin levels in patients treated with imatinib. In cases in which iron overload is observed in these patients it may contribute to progression of primary tumors due to pro-oxidative and nutritional roles of iron in cancer cells.45 Furthermore, iron overload in response to long-term imatinib treatment may cause heart iron deposition contributing to the reported cardiotoxicity of this drug.41 In such a case, a diet with limited iron content, phlebotomies47 or iron chelation therapies may be indicated to alleviate the severity of congestive heart failure observed in some patients on imatinib. Edema, a condition hallmarked by water and sodium retention in the body, is commonly observed in clinical practice.48 Congestive heart failure is one of the most lifethreatening causes of edema. Spironolactone was shown to combat hyperaldosteronism and subsequent edema and decreased morbidity and mortality rates in patients with severe heart failure.49 As described above, hepcidin deficiency upon long-term spironolactone therapy may cause iron overload also in the heart, which could aggravate heart disease and thus should be diagnosed and prevented. Another important indication for spironolactone is ascites, an accumulation of fluid in the abdominal cavity. 1182

Interestingly, ascites is tightly linked to liver cirrhosis,48 a condition in which hepcidin deficiency and increased iron levels were reported.50 Further suppression of hepcidin levels by spironolactone treatment (Figure 4A) may cause an even more pronounced dysregulation of iron homeostasis in cirrhotic patients, which may outweigh its beneficial effects on edema and thus provide a rationale for the application of alternative agents.

Mechanistic insights Our data further generate insight into the potential mode of action of imatinib and spironolactone. For them to modulate hepcidin expression, these drugs require preserved BMP signaling mediated by the BMP responsive elements located in the hepcidin promoter (Figure 2B).17 Imatinib displays a much broader range of inhibitory activity than originally described.33 Our data demonstrate that in the context of hepcidin regulation, imatinib may suppress its non-kinase target oxidoreductase NQO2,33 which was identified as a hepcidin activator in our study (Figure 1B). Consistently, quercetin, another agent that suppresses NQO2, also reduces hepcidin mRNA levels (Figure 3D). How NQO2 regulates hepcidin remains to be established. Previous studies suggested that it controls the protein levels of p53,29 a tumor suppressor that was reported as a hepcidin activator.51 Likewise, NQO2 may regulate the stability of other hepcidin activators. It is of note that we observe decreased hepcidin mRNA expression in response to quercetin which contrasts previously reported results showing hepcidin activation upon quercetin treatment.52 This discrepancy may arise from different cell lines used, drug doses applied or time periods of drug supplementation studied. Our results consistently reveal that two aldosterone receptor antagonists, spironolactone and eplerenone, suppress hepcidin in primary human hepatocytes (Figures 2D and 3E). This effect is unlikely to be explained by a crossreactivity of this drug with testosterone signaling, which is known to suppress hepcidin.16 Thus, our data suggest that aldosterone-mediated signaling in hepatocytes controls hepcidin expression, a finding that is consistent with a growing spectrum of extra-renal roles of aldosterone.53 Diclofenac was identified as hepcidin modulator in primary hepatocytes (Figure 2), but failed to affect hepcidin expression in mice (Figure 4A). Side effects of diclofenac therapy include hepatotoxicity, gastrointestinal ulcers and bleeding.54 According to reports accompanying the FDA approval for diclofenac, the dose used in this study was expected to cause mild anemia, but no other adverse effects. Consistently, we observe a mild reduction of hemoglobin levels, hematocrit and red blood cell counts in mice treated with diclofenac (Table 3). However, a moduhaematologica | 2017; 102(7)

Drug targeting of hepcidin expression

lation of iron homeostasis is unlikely to contribute to the mild anemic phenotype, which may rather arise from minor intestinal injury. Our work further demonstrates that SAHA and other HDAC inhibitors (HDACi) induce hepcidin expression in cultured cells and primary hepatocytes (Figure 2) but not in mice (Figure 4A). It is of note that the HDAC inhibitors tested modulate hepcidin expression to different degrees dependent on the in vitro model studied (Figure 3). Our data obtained in Huh7 cells suggest that HDAC3 may predominantly control hepcidin promoter accessibility (Figures 3A and 1B). However, further studies in murine and human PH suggested the involvement of other HDACs. Previous reports already identified HDACi as hepcidin modulators. Miura et al. showed that trichostatin A induces hepcidin in hepatoma cells, prevents hepatitis virus C-mediated hepcidin suppression and increases binding of the core transcriptional activators, STAT3 and CEBPÎą, to the hepcidin promoter.26 Furthermore, SAHA was independently identified as hepcidin-inducing agent by a drug screen for hepcidin regulators.39 Although in our hands SAHA did not alter hepcidin levels in vivo, it remains to be established whether different dosing or administration routes, or the employment of alternative HDACi

References 1. Hentze MW, Muckenthaler MU, Galy B, Camaschella C. Two to tango: regulation of Mammalian iron metabolism. Cell. 2010;142(1):24-38. 2. Papanikolaou G, Tzilianos M, Christakis JI, et al. Hepcidin in iron overload disorders. Blood. 2005;105(10):4103-4105. 3. Kremastinos DT, Farmakis D. Iron overload cardiomyopathy in clinical practice. Circulation. 2011;124(20):2253-2263. 4. Deugnier Y, Turlin B. Pathology of hepatic iron overload. World J Gastroenterol. 2007;13(35):4755-4760. 5. Rajpathak SN, Crandall JP, Wylie-Rosett J, Kabat GC, Rohan TE, Hu FB. The role of iron in type 2 diabetes in humans. Biochim Biophys Acta. 2009;1790(7):671-681. 6. Altamura S, Muckenthaler MU. Iron toxicity in diseases of aging: Alzheimer's disease, Parkinson's disease and atherosclerosis. J Alzheimers Dis. 2009;16(4):879-895. 7. Nemeth E, Rivera S, Gabayan V, et al. IL-6 mediates hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone hepcidin. J Clin Invest. 2004;113(9):1271-1276. 8. Weiss G, Goodnough LT. Anemia of chronic disease. N Engl J Med. 2005; 352(10):10111023. 9. Pietrangelo A. Hereditary hemochromatosis: pathogenesis, diagnosis, and treatment. Gastroenterology. 2010;139(2):393-408. 10. Al-Khabori M, Bhandari S, Al-Huneini M, Al-Farsi K, Panjwani V, Daar S. Side effects of Deferasirox Iron Chelation in Patients with Beta Thalassemia Major or Intermedia. Oman Med J. 2013;28(2):121-124. 11. Vincent JL, Baron JF, Reinhart K, et al. Anemia and blood transfusion in critically ill patients. JAMA. 2002;288(12):1499-1507. 12. Bohlius J, Schmidlin K, Brillant C, et al. Recombinant human erythropoiesis-stimulating agents and mortality in patients with

haematologica | 2017; 102(7)

13. 14.

15.

16.

17.

18.

19.

20.

21.

22.

modulates systemic iron parameters. According to the DrugBank database, HDACi are used predominantly in cancer therapy. Their hepcidin-inducing potential may thus be beneficial to limit iron availability and prevent disease progression.45 Nevertheless, such effects of HDACi may likely exacerbate cancer-related anemia, frequently observed in several malignancies.55 Taken together, our study investigated a large panel of approved drugs, and identified imatinib and spironolactone as modulators of hepcidin levels in primary hepatocytes and in mice. These data suggest that prolonged administration of these therapeutics may affect body iron balance, and possibly the course of an underlying disease. Such knowledge is of high relevance for patient care. It merits the study of potential iron-related side effects of these therapeutics in prospective clinical studies. Funding MUM acknowledges funding from the Deutsche Forschungsgemeinschaft (SFB1036) and the DietmarHopp Stiftung. KM-S acknowledges support from the University of Heidelberg (Walter Erb-Stiftung; D.10052215.03). GD and DS acknowledge support from the German Federal Ministry of Education and Research (BMBF) project Virtual Liver (0315741).

cancer: a meta-analysis of randomised trials. Lancet. 2009;373(9674):1532-1542. Ruchala P, Nemeth E. The pathophysiology and pharmacology of hepcidin. Trends Pharmacol Sci. 2014;35(3):155-161. Goodnough JB, Ramos E, Nemeth E, Ganz T. Inhibition of hepcidin transcription by growth factors. Hepatology. 2012; 56(1):291-299. Kautz L, Jung G, Valore EV, Rivella S, Nemeth E, Ganz T. Identification of erythroferrone as an erythroid regulator of iron metabolism. Nat Genet. 2014; 46(7):678-684. Latour C, Kautz L, Besson-Fournier C, et al. Testosterone perturbs systemic iron balance through activation of EGFR signaling in the liver and repression of hepcidin. Hepatology. 2013;59(2):683-694. Mleczko-Sanecka K, Roche F, da Silva AR, et al. Unbiased RNAi screen for hepcidin regulators links hepcidin suppression to proliferative Ras/RAF and nutrient-dependent mTOR signaling. Blood. 2014; 123(10):1574-1585. Tsuchiya H, Akechi Y, Ikeda R, et al. Suppressive effects of retinoids on ironinduced oxidative stress in the liver. Gastroenterology. 2009;136(1):341-350. Ashburn TT, Thor KB. Drug repositioning: identifying and developing new uses for existing drugs. Nat Rev Drug Discov. 2004; 3(8):673-683. Casanovas G, Mleczko-Sanecka K, Altamura S, Hentze MW, Muckenthaler MU. Bone morphogenetic protein (BMP)responsive elements located in the proximal and distal hepcidin promoter are critical for its response to HJV/BMP/SMAD. J Mol Med. 2009;87(5):471-480. Mleczko-Sanecka K, Casanovas G, Ragab A, et al. SMAD7 controls iron metabolism as a potent inhibitor of hepcidin expression. Blood. 2010;115(13):2657-2665. Reagan-Shaw S, Nihal M, Ahmad N. Dose

23.

24.

25.

26.

27.

28.

29.

30. 31. 32.

translation from animal to human studies revisited. Faseb J. 2008;22(3):659-661. Nair AB, Jacob S. A simple practice guide for dose conversion between animals and human. J Basic Clin Pharm. 2016;7(2):2731. Torrance JD, Bothwell TH. A simple technique for measuring storage iron concentrations in formalinised liver samples. S Afr J Med Sci. 1968;33(1):9-11. Gunther S, Kuhn M, Dunkel M, et al. SuperTarget and Matador: resources for exploring drug-target relationships. Nucleic Acids Res. 2008;36:D919-922. Miura K, Taura K, Kodama Y, Schnabl B, Brenner DA. Hepatitis C virus-induced oxidative stress suppresses hepcidin expression through increased histone deacetylase activity. Hepatology. 2008; 48(5):1420-1429. Flock GB, Cao X, Maziarz M, Drucker DJ. Activation of enteroendocrine membrane progesterone receptors promotes incretin secretion and improves glucose tolerance in mice. Diabetes. 2013;62(1):283-290. Vella F, Ferry G, Delagrange P, Boutin JA. NRH:quinone reductase 2: an enzyme of surprises and mysteries. Biochem Pharmacol. 2005;71(1-2):1-12. Khutornenko AA, Roudko VV, Chernyak BV, Vartapetian AB, Chumakov PM, Evstafieva AG. Pyrimidine biosynthesis links mitochondrial respiration to the p53 pathway. Proc Natl Acad Sci USA. 2010; 107(29):12828-12833. Stitham J, Midgett C, Martin KA, Hwa J. Prostacyclin: an inflammatory paradox. Front Pharmacol. 2011;2:24. Kadamur G, Ross EM. Mammalian phospholipase C. Annu Rev Physiol. 2013; 75:127-154. Verga Falzacappa MV, Vujic Spasic M, Kessler R, Stolte J, Hentze MW, Muckenthaler MU. STAT3 mediates hepatic hepcidin expression and its inflammatory

1183

K. Mleczko-Sanecka et al. stimulation. Blood. 2007;109(1):353-358. 33. Bantscheff M, Eberhard D, Abraham Y, et al. Quantitative chemical proteomics reveals mechanisms of action of clinical ABL kinase inhibitors. Nat Biotechnol. 2007;25(9):1035-1044. 34. Wu K, Knox R, Sun XZ, et al. Catalytic properties of NAD(P)H:quinone oxidoreductase-2 (NQO2), a dihydronicotinamide riboside dependent oxidoreductase. Arch Biochem Biophys. 1997;347(2):221-228. 35. Fosslien E. Cardiovascular complications of non-steroidal anti-inflammatory drugs. Ann Clin Lab Sci. 2005;35(4):347-385. 36. Wishart DS, Knox C, Guo AC, et al. DrugBank: a comprehensive resource for in silico drug discovery and exploration. Nucleic Acids Res. 2006;34:D668-672. 37. Ouzan J, Perault C, Lincoff AM, Carre E, Mertes M. The role of spironolactone in the treatment of patients with refractory hypertension. Am J Hypertens. 2002;15(4 Pt 1):333-339. 38. Corvol P, Michaud A, Menard J, Freifeld M, Mahoudeau J. Antiandrogenic effect of spirolactones: mechanism of action. Endocrinology. 1975;97(1):52-58. 39. Gaun V, Patchen B, Volovetz J, et al. A chemical screen identifies small molecules that regulate hepcidin expression. Blood Cells Mol Dis. 2014;53(4):231-240. 40. Patchen B, Koppe T, Cheng A, Seo YA, Wessling-Resnick M, Fraenkel PG. Dietary

1184

41.

42.

43. 44.

45. 46. 47.

48. 49.

supplementation with ipriflavone decreases hepatic iron stores in wild type mice. Blood Cells Mol Dis. 2016;60:36-43. Kerkela R, Grazette L, Yacobi R, et al. Cardiotoxicity of the cancer therapeutic agent imatinib mesylate. Nat Med. 2006; 12(8):908-916. van Oosterom AT, Judson I, Verweij J, et al. Safety and efficacy of imatinib (STI571) in metastatic gastrointestinal stromal tumours: a phase I study. Lancet. 2001; 358(9291):1421-1423. Cumming DC, Yang JC, Rebar RW, Yen SS. Treatment of hirsutism with spironolactone. JAMA. 1982;247(9):1295-1298. Shaw JC, White LE. Long-term safety of spironolactone in acne: results of an 8-year followup study. J Cutan Med Surg. 2002;6(6):541-545. Torti SV, Torti FM. Iron and cancer: more ore to be mined. Nat Rev Cancer. 2013;13(5):342-355. Jones RL, Judson IR. The development and application of imatinib. Expert Opin Drug Saf. 2005;4(2):183-191. Maiti B, Setrakian S, Daw HA. Hepatic iron overload, a possible consequence of treatment with imatinib mesylate: a case report. Cases J. 2009;2:7526. O'Brien JG, Chennubhotla SA, Chennubhotla RV. Treatment of edema. Am Fam Physician. 2005;71(11):2111-2117. Pitt B, Zannad F, Remme WJ, et al. The

50.

51.

52.

53.

54.

55.

effect of spironolactone on morbidity and mortality in patients with severe heart failure. Randomized Aldactone Evaluation Study Investigators. N Engl J Med. 1999;341(10):709-717. Tan TC, Crawford DH, Franklin ME, et al. The serum hepcidin:ferritin ratio is a potential biomarker for cirrhosis. Liver Int. 2012; 32(9):1391-1399. Weizer-Stern O, Adamsky K, Margalit O, et al. Hepcidin, a key regulator of iron metabolism, is transcriptionally activated by p53. Br J Haematol. 2007;138(2):253-262. Bayele HK, Balesaria S, Srai SK. Phytoestrogens modulate hepcidin expression by Nrf2: Implications for dietary control of iron absorption. Free Radic Biol Med. 2015;89:1192-1202. Baudrand R, Gupta N, Garza AE, et al. Caveolin 1 Modulates AldosteroneMediated Pathways of Glucose and Lipid Homeostasis. J Am Heart Assoc. 2016;5(10):e003845. Banks AT, Zimmerman HJ, Ishak KG, Harter JG. Diclofenac-associated hepatotoxicity: analysis of 180 cases reported to the Food and Drug Administration as adverse reactions. Hepatology. 1995;22(3):820-827. Bohlius J, Weingart O, Trelle S, Engert A. Cancer-related anemia and recombinant human erythropoietin-an updated overview. Nat Clin Pract Oncol. 2006; 3(3):152-164.

haematologica | 2017; 102(7)

ARTICLE

Coagulation & Its Disorders

Venous thromboembolism is associated with graft-versus-host disease and increased non-relapse mortality after allogeneic hematopoietic stem cell transplantation

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Natasha Kekre,1 Haesook T. Kim,2 Vincent T. Ho,3 Corey Cutler,3 Philippe Armand,3 Sarah Nikiforow,3 Edwin P. Alyea,3 Robert J. Soiffer,3 Joseph H. Antin,3 Jean M. Connors4* and John Koreth3*

1 Ottawa Hospital Research Institute, Ottawa, ON, Canada; 2Department of Biostatistics/Computational Biology, Dana-Farber Cancer Institute, Harvard School of Public Health, Boston, MA, USA; 2Division of Hematologic Malignancies, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA and 4Division of Hematology, Brigham & Womenâ&#x20AC;&#x2122;s Hospital, Harvard Medical School, Boston, MA, USA

*JMC and JK contributed equally to this work

Haematologica 2017 Volume 102(7):1185-1191

ABSTRACT

lthough venous thromboembolism rates and risk factors are well described in patients with cancer, there are limited data on the incidence, risk factors and outcomes of thrombosis after allogeneic stem cell transplantation, a curative therapy for patients with hematologic malignancies. We aimed to determine the incidence and risks associated with venous thrombosis in allogeneic stem cell transplants. We studied 2276 recipients of first transplant between 2002-2013 at our institution with a median follow up of 50 months (range 4-146). Using pharmacy records and subsequent chart reviews, 190 patients who received systemic anticoagulation for venous thrombosis were identified. The 1and 2-year cumulative incidence of all venous thrombotic events were 5.5% (95% confidence interval (CI) 4.6-6.5%) and 7.1% (95% CI 6.18.2%), respectively. There was no difference in age, sex, body mass index, diagnosis, disease risk index, conditioning intensity, donor type or graft source between transplant recipients with and without subsequent thrombosis. In multivariable models, both acute and chronic graft-versus-host disease were independently associated with thrombosis occurrence (Hazard ratio (HR)=2.05, 95% CI 1.52-2.76; HR=1.71, 95% CI 1.19-2.46, respectively). Upper extremity thrombosis differed from all other thromboses in terms of timing, risk factors and clinical impact, and was not associated with non-relapse mortality (HR=1.15; 95% CI 0.691.90), unlike all other thromboses which did increase non-relapse mortality (HR=1.71; 95% CI 1.17-2.49). In subgroup analysis evaluating conventional thrombosis predictors by comparing patients with and without thrombosis, a history of prior venous thrombosis was the only significant predictor. Venous thromboembolism has a high incidence after allogeneic stem cell transplant and is associated with graft-versus-host disease and non-relapse mortality. Introduction The care of patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT), a curative modality for advanced/aggressive hematologic malignancies, is highly complex, involving central venous catheters, conditioning chemotherapy, immune suppressive therapies for prophylaxis and treatment of graft-versushost disease (GvHD), donor graft infusions, and infection monitoring and treatment. Their medical acuity and complexity makes allogeneic HSCT recipients haematologica | 2017; 102(7)

Correspondence: john_koreth@dfci.harvard.edu

Received: January 11, 2017. Accepted: March 17, 2017. Pre-published: March 24, 2017. doi:10.3324/haematol.2017.164012 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1185 ÂŠ2017 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.

1185

N. Kekre et al.

potentially vulnerable to venous thromboembolism (VTE) despite eradication of the underlying malignancy, but this complication remains incompletely characterized. Studies in autologous HSCT recipients have described a VTE incidence between 3% and 23.5%1-3 but this incidence cannot be directly extrapolated to allogeneic HSCT recipients. For instance, autologous HSCT recipients are not at risk of GvHD and have a lower risk of hepatic veno-occlusive disease, both of which are associated with vascular disruption and prolonged systemic inflammation. Many autologous HSCT recipients are also increasingly exposed to thrombophilic medications such as lenalidomide for multiple myeloma (MM), which creates a VTE risk profile that is different from allogeneic HSCT recipients. Limited studies in allogeneic HSCT recipients have reported a wide range in the incidence of VTE, from 0.5% to 13%.4-10 The older studies are difficult to interpret as allogeneic transplant practices have changed over time. For example, a study that examined the risk of VTE in a cohort of over 400 HSCT recipients included patients receiving heparin as prophylaxis for veno-occlusive disease.11 This is no longer a common strategy and would have altered the incidence of VTE. A recent meta-analysis of 12 studies in allogeneic HSCT recipients estimated the cumulative incidence of VTE at 4% (95% CI 2–6%), but was fraught with significant heterogeneity between studies (I2=80).12 Although some studies have described an association between GvHD and the risk of VTE, the follow up was generally short, and insufficient to determine if the risk of VTE changes over time.10,13 The largest cohort study of 1514 HSCT recipients (including approximately 60% autologous transplants), reported that 4.6% of patients developed VTE.10 This study, however, had only 6 months of follow up after HSCT, making it difficult to establish an association between chronic GvHD and VTE. Furthermore, it included mostly myeloablative conditioning regimens, making it difficult to assess VTE incidence after reduced-intensity conditioning regimens. We therefore sought to rigorously examine the risk of VTE after first allogeneic HSCT in a large retrospective cohort of patients uniformly treated at a single center, describing VTE incidence, sites of involvement, risk associations and outcomes. The significant strength of this cohort at the Dana-Farber Cancer Institute/Brigham and Women’s Hospital is that patients remain under the care of their transplant physicians long-term, creating a cohort with extended follow up of outcomes after HSCT. In allogeneic HSCT it is also important to identify subgroups at the highest VTE risk that might benefit disproportionately from thromboprophylaxis, recognizing also the potential for harm associated with anticoagulation due to the higher risk of bleeding during thrombocytopenia and intestinal inflammation, as well as drug interactions with anticoagulants and GvHD medications. We therefore undertook further analyses in this cohort to determine if conventional VTE risk factors could identify a cohort of patients at particularly high risk for VTE after HSCT, who could potentially benefit from VTE prophylaxis.

Methods Patients We studied a retrospective cohort of 2276 patients who underwent first allogeneic HSCT between January 1, 2002 and 1186

December 31, 2013 at the Dana-Farber Cancer Institute/Brigham and Women’s Hospital. Patient, transplant, and outcome related factors were extracted from both the transplantation database and through medical chart review. Outpatient pharmacy records and subsequent medical chart review identified patients with VTE on systemic anticoagulation. This study was approved by the DanaFarber/Harvard Cancer Center institutional review board.

Definitions Venous thromboembolism sites were categorized as lower extremity deep vein thrombosis or pulmonary embolism (LE DVT/PE), upper extremity and other. LE DVT/PE included PE, symptomatic proximal or distal lower extremity DVT, and the combination of PE and DVT, requiring systemic treatment. Upper extremity VTE included any arm DVT with or without a concurrent central line or peripherally inserted central catheter in place. Other VTE included superior vena cava (SVC), pelvic, abdominal, or right-sided ventricular thrombosis requiring systemic treatment. VTE was defined as an event confirmed by radiologic imaging (ultrasound, computed tomography (CT), magnetic resonance imaging (MRI) or ventilation–perfusion (V/Q) scan) requiring systemic anticoagulation that occurred after stem cell infusion (ie., day 0 of HSCT). Systemic anticoagulation included low-molecular-weight heparin, warfarin, fondaparinux, and dabigatran. At our center, patients do not receive systemic thromboprophylaxis at the time of HSCT due to the expected fall in platelet count. If, however, patients are re-admitted with normal platelet counts and normal renal function, the standard of care at our center would include enoxaparin 40 mg subcutaneously daily or unfractionated heparin 5000 units subcutaneously three times a day for VTE prevention. For systemic therapy of VTE, practitioners at our center manage anticoagulation. If patients were on warfarin, they were managed by a central anticoagulation monitoring service. Information on the duration of anticoagulation and choice of anticoagulant was not obtained.

Statistical analysis Patient baseline characteristics were reported descriptively. Endpoints of interest were overall survival, progression-free survival, relapse, and non-relapse mortality. Overall survival was defined as the time from stem cell infusion to death from any cause. Patients who were alive or lost to follow up were censored at the time last seen alive. Progression-free survival was defined as the time from stem cell infusion to disease relapse, progression or death from any cause, whichever occurred first. Patients who were alive without disease relapse or progression were censored at the time last seen alive and progression-free. A cumulative incidence curve of VTE was constructed in the competing risks framework considering death without developing VTE as a competing event. All time to events were measured from the date of stem cell infusion. The analysis is composed of two cohorts: the entire study population (N=2276) for identifying patient and transplant-related factors that are associated with post-HSCT VTE; and a subgroup of patients (N=168) to additionally investigate known VTE risk factors in depth. For the entire study population (N=2276), univariable and multivariable Cox regression analysis was performed to examine whether occurrence of VTE was a risk factor for overall survival, progression-free survival, relapse, non-relapse mortality, and chronic GvHD. Risk factors considered in multivariable analysis included age, patient and donor sex combination, disease risk index (DRI),14 graft source, donor HLA type, comorbidity index,15 sirolimus use as GvHD prophylaxis, body mass index (BMI), disease type (myeloid vs. other), acute and chronic GvHD, and haematologica | 2017; 102(7)

VTE after allogeneic HSCT

occurrence of VTE. Occurrence of VTE and acute and chronic GvHD were analyzed as time-dependent covariates. GvHD had to occur prior to the VTE event to be considered as a risk factor for VTE. Cox models were stratified by conditioning intensity because conditioning intensity did not meet the proportional hazards assumption. To identify potential risk factors for VTE, univariable and multivariable Cox regression analysis was utilized. For the subgroup analysis (N=168), we included all recent VTE events over a fixed time interval from January 1, 2010 to December 31, 2012 (N=56), with a control cohort of 112 HSCT patients who were randomly selected from 721 patients without VTE in the same transplantation period in a 1:2 ratio between cases and controls. To preserve the ability to test the association between risk factors and VTE, the control cohort was a random subset of patients without VTE and not matched on the characteristics of VTE cases. There were no significant differences in age, patient sex, donor sex, sex mismatch, BMI, diagnosis, DRI, donor type, conditioning intensity, cell source, or cytomegalovirus (CMV) seropositive status between the randomly selected and nonselected controls (all P-values >0.5). Using this subset of patients with and without VTE, we additionally collected baseline conventional risk factors known to be associated with VTE: a past medical history of diabetes, dyslipidemia, hypertension, prior myocardial infarction, prior stroke, and prior VTE. From pharmacy records, we also collected information on the use of hormone replacement therapy (HRT) in women from the time of HSCT onwards. All analyses were done using SAS 9.3 (SAS Institute Inc., Cary, NC, USA), and R version 3.2.2 (The Comprehensive R Archive Network (CRAN) project). Multiplicity was not considered and the significance level was set to 0.05.

Results Incidence and Timing of VTE after HSCT Between January 1, 2002 and December 31, 2013, 2276 patients underwent first-time allogeneic HSCT. The median follow-up time was 50 months (range 4-146) among survivors. Of these patients, 190 (8.3%) developed VTE requiring systemic anticoagulation. The 1- and 2-year cumulative incidence of all VTE were 5.5% (95% CI 4.66.5%) and 7.1% (95% CI 6.1-8.2%), respectively, Figure 1. Amongst the 190 patients who developed VTE, 120 (62.3%) were LE DVT/PE (45 were PE, 65 were lower extremity DVT and 10 were both PE and DVT), 57 (30%) were upper extremity (48 had a catheter in situ at time of thrombosis, 9 had a catheter removed less than one month prior to arm DVT) and 13 (6.8%) were other VTE (4 of which were SVC thrombosis). Upper extremity VTE occurred at a median of 1.3 months (range 0.1-41.1) after HSCT, which was significantly shorter than LE DVT/PE (9.2 months, range 0.2-72.3) and other VTE (10.2 months, range 1.1-67.1), P<0.0001 (Figure 2).

Platelet count at Time of VTE Platelet count at time of VTE diagnosis in 176 patients was established by chart review. In 56 patients with an upper extremity VTE, the median platelet count at VTE diagnosis was 114.5 × 109/L (range 14-298), lower than for LE DVT/PE (N=108, median 142× 109/L, range 22-474) and other VTE (N=12, median 192.5 × 109/L, range 42-473), P=0.0046. The median platelet count at time of any VTE event was 130.5 × 109/L (range 14-474).

Risk Factors for VTE

Figure 1. Cumulative Incidence of VTE after HSCT. The 1-year cumulative incidence of any VTE event after HSCT was 5.5% and the 2-year cumulative incidence was 7.1%, with LE DVT/PE being the most common type of VTE. DVT/PE: deep vein thrombosis or pulmonary embolism; VTE: venous thromboembolism.

haematologica | 2017; 102(7)

Evaluating usual patient, disease and transplant variables, there was no difference in age, sex, BMI, diagnosis, DRI, conditioning regimen intensity, donor type (HLAmatched or HLA-mismatched, related or unrelated, or cord blood), or graft source (bone marrow or peripheral blood) between allogeneic HSCT recipients with and without VTE (Table 1). In a multivariable model to identify potential risk factors for VTE, both acute and chronic GvHD were independently associated with developing any type of VTE (HR=2.05, 95% CI 1.52-2.76, P<0.0001 and HR=1.71, 95% CI 1.19-2.46, P=0.0035, respectively), Table 2. As median time of VTE after HSCT and platelet count at time of VTE were significantly different for upper extremity VTE versus other types of VTE, we further performed multivariate models separating these types of VTE events. When examining upper extremity VTE only, conditioning regimen intensity, donor type and acute and chronic GvHD were risk factors for VTE. When upper extremity VTE was excluded, donor type (matched unrelated donor (MUD) vs. matched related donor (MRD), mismatched vs. MRD), myeloid disease and both acute and chronic GvHD were associated with all other VTE (Table 2). In patients who had a VTE after HSCT, 98 patients (51.6%) had active GvHD on therapy, 28 patients (14.7%) no longer had active GvHD but remained on therapy, 54 patients (28.4%) were on a GvHD prophylaxis regimen and 10 patients (5.3%) had no history of GvHD and were not on immunosuppressive therapy. The most common immunosuppressive regimens for patients with a VTE were tacrolimus and prednisone (N=35, 19.3%), tacrolimus and sirolimus (N=32, 17.7%) and tacrolimus, 1187

N. Kekre et al.

sirolimus and prednisone (N=26, 14.4%). In patients who developed GvHD after their first VTE, 8% had a subsequent VTE event. For patients with active or previous GvHD at the time of VTE, the most prevalent organ involved was the skin, including sclerodermatous GvHD (N=83, 65.9%), followed by gastrointestinal GvHD (N=37, 29.4%), ocular and/or oral GvHD (N=36, 28.6%) and liver GvHD (N=27, 21.4%). In order to compare conventional VTE risk factors, we performed a subset case control analysis between HSCT recipients with and without VTE. The only conventional risk factor associated with a risk of VTE after HSCT was a prior history of VTE before HSCT. The incidence of prior VTE was higher in patients who developed VTE after HSCT: only 10 out of 112 patients (8.9%) without VTE after HSCT had a prior VTE event, while 12 out of 56 patients (21.4%) with VTE after HSCT had a prior VTE event (P=0.03, Table 3).

Clinical Outcomes In multivariable analysis, any VTE event was associated with increased non-relapse mortality (HR=1.47; 95% CI 1.08-2.00; P=0.015), but not with relapse (HR=0.88, 95% CI 0.63-1.23; P=0.47), progression-free survival (HR=1.13, 95% CI 0.90-1.41; P=0.29) or overall survival (HR=1.05,

95% CI 0.84-1.31; P=0.65). Evaluating by site of VTE, upper extremity VTE alone did not impact any clinical outcome including non-relapse mortality. LE DVT/PE and other VTE remained significantly associated with nonrelapse mortality (HR=1.71; 95% CI 1.17-2.49; P=0.005). Table 4 summarizes the multivariable models.

Discussion Venous thromboembolism in cancer is well described, but the incidence and risk factors after allogeneic HSCT, a presumptively curative therapy for aggressive hematologic malignancies, is not well defined. While treating the underlying malignancy with allogeneic HSCT can reduce the risk of VTE, the study herein reports that the incidence of VTE post-HSCT remains high. In the largest allogeneic HSCT cohort evaluated to date with a median follow up of over 4 years, we demonstrated a very high incidence of VTE of 8.3%. We found that acute and chronic GvHD were independent risk factors for developing VTE. A prior history of VTE was also significantly associated with developing VTE in this cohort of patients. Lower extremity DVT and PE, as well as VTE in unusual locations such as splanchnic veins, was associated with an increased risk

Table 1. Baseline Characteristics of Patients with and without VTE after HSCT.

N (%)

No VTE (N=2086)

VTE (N=190)

Age, median (range) 52 (17, 74) 53 (20, 73) 0.1 Male 1220 (58.5%) 111 (58.4%) 1 Body mass index >= 30 582 (27.9%) 59 (31.1%) 0.36 Male patient with Female donor 475 (22.8%) 41 (21.6%) 0.79 Diagnosis Acute leukemia 937 (44.9%) 71 (37.4%) 0.1 Lymphoma 411 (19.7%) 44 (23.1%) Multiple myeloma 59 (2.8%) 3 (1.6%) Other 679 (32.6%) 66 (34.7%) Donor Type Matched unrelated 1031 (49.4%) 100 (52.6%) 0.71 Matched related 746 (35.8%) 66 (34.7%) Mismatched unrelated 283 (13.6%) 21 (11.1%) Mismatched related 26 (1.2%) 3 (1.6%) Disease Risk Index Low 381 (18.7%) 36 (19.4%) 0.57 Intermediate 1030 (50.7%) 102 (54.8%) High 545 (26.8%) 43 (23.1%) Very high 77 (3.8%) 5 (2.7%) Conditioning Intensity Myeloablative 875 (41.9%) 83 (43.7%) 0.65 Reduced intensity 1211 (58.1%) 107 (56.3%) Graft Source Bone marrow 161 (7.7%) 11 (5.8%) 0.17 Peripheral blood 1785 (85.6%) 174 (91.6%) Umbilical cord 136 (6.5%) 5 (2.6%) Type of VTE Lower extremity DVT/PE --120 (62.3%) --Upper Extremity VTE 57 (30%) Other VTE 13 (6.8%) HSCT: hematopoietic stem cell transplantation; VTE: venous thromboembolism: DVT/PE: deep vein thrombosis or pulmonary embolism. 1188

haematologica | 2017; 102(7)

VTE after allogeneic HSCT

of non-relapse mortality (HR=1.71). Although allogeneic HSCT recipients have a 2.5% incidence of upper extremity VTE (which is associated with central venous catheters), thrombosis at this site was not associated with increased non-relapse mortality. Upper extremity VTE represents a unique cohort of patients with VTE, as evidenced by the shorter time to VTE after HSCT which reflects that these events were due to central venous

catheter placement routinely used during conditioning chemotherapy and HSCT at our center. The lower platelet count observed with upper extremity VTE also likely reflects the early timeframe when VTE was diagnosed in this cohort, reflecting that these patients were still undergoing platelet recovery after HSCT. Prior to this study, the largest cohort examining VTE in HSCT recipients by Gerber and colleagues reported that

Table 2. Multivariable Model of HSCT Variables Associated with VTE.

Outcome All VTE

LE DVT/PE and Other VTE

Upper Extremity VTE

Variable

HR (95% CI)

Acute GvHD Chronic GvHD MUD (6/6) vs. MRD Mismatched vs. MRD Myeloid vs. non- myeloid Acute GvHD Chronic GvHD RIC vs MAC MUD (6/6) vs. MRD Acute GvHD Chronic GvHD

2.05 (1.52-2.76) 1.71 (1.19-2.46) 1.66 (1.12-2.47) 1.86 (0.98-3.52) 0.66 (0.47-0.93) 1.89 (1.32-2.71) 1.94 (1.23-3.05) 0.46 (0.26-0.81) 0.48 (0.27-0.84) 2.13 (1.14-3.97) 3.39 (1.17-9.85)

<.0001* 0.0035* 0.012* 0.057 0.019* 0.0005* 0.0044* 0.0075* 0.0109* 0.0173* 0.025*

*Indicates statistical significance. MUD: matched unrelated donor; MRD: matched related donor. RIC: reduced intensity conditioning; MAC: myeloablative conditioning; HSCT: hematopoietic stem cell transplantation; HR: hazard ratio; CI: confidence interval; VTE: venous thromboembolism; LE DVT/PE: lower extremity deep vein thrombosis or pulmonary embolism; GvHD: graft-versus-host disease

Table 3. Subgroup analysis of Conventional VTE Risk Factors in HSCT Recipients.

N (%) Diabetes Dyslipidemia Hypertension Prior myocardial infarction Prior VTE Prior stroke Smoking status Current smoker Ex-smoker Non-smoker HRT (women only)

No VTE (N=112)

VTE (N=56)

7 (6.3%) 29 (25.9%) 27 (24.1%) 5 (4.5%) 10 (8.9%) 0

5 (8.9%) 9 (16.1%) 13 (23.2%) 2 (3.6%) 12 (21.4%) 0

0.54 0.17 1 1 0.03* N/A

12 (10.7%) 37 (33%) 63 (56.3%) 13 (25%)

2 (3.6%) 17 (30.4%) 37 (66.1%) 6 (37.5%)

0.23

0.35

*Indicates statistical significance. HRT: hormone replacement therapy; HSCT: hematopoietic stem cell transplantation; VTE: venous thromboembolism.

Figure 2. Timing of VTE after HSCT. Each dot represents one VTE event after HSCT. Upper extremity VTE occurred at a median of 1.3 months after HSCT, whereas LE DVT/PE occurred at a median of 9.2 months and other VTE at a median of 10.2 months (upper extremity versus LE DVT/PE and other, P<0.0001). DVT/PE: deep vein thrombosis or pulmonary embolism.

haematologica | 2017; 102(7)

1189

N. Kekre et al. Table 4. Multivariable Analysis of VTE and HSCT Clinical Outcomes.

HR (95% CI)

Overall Survival

Progression-free Survival

Non-relapse Mortality

Relapse

All VTE 1.05 (0.84-1.31) 1.13 (0.90-1.41) 1.47 (1.08-2.00)* 0.88 (0.63-1.23) LE DVT/PE and Other VTE 1.18 (0.90-1.55) 1.24 (0.93-1.65) 1.71 (1.17-2.49)* 0.87 (0.56-1.36) Upper Extremity 0.86 (0.59-1.25) 0.98 (0.69-1.39) 1.15 (0.69-1.90) 0.89 (0.54-1.46) *Indicates statistical significance. HR: hazard ratio; CI: confidence interval; HSCT: hematopoietic stem cell transplantation; VTE: venous thromboembolism; LE DVT/PE: lower extremity deep vein thrombosis or pulmonary embolism.

4.6% of patients developed VTE,10 but that study differs significantly from our analysis as it includes patients treated with autologous transplant who were excluded from our cohort. Gerberâ&#x20AC;&#x2122;s study also included mainly myeloablative conditioning, which differs from our report in which about 60% of patients received reduced intensity conditioning. One could hypothesize that myeloablative conditioning may increase the incidence of VTE due to more vascular endothelial damage, veno-occlusive disease (VOD), and organ toxicity, but this was not the case in our multivariate models for VTE risk. Conditioning intensity was only associated with upper extremity VTE, perhaps because patients receiving myeloablative conditioning are more likely to have an indwelling central line for a longer time. In the study herein, GvHD was independently associated with VTE after HSCT. We postulate that GvHD does indeed induce a pro-inflammatory state which likely makes patients more prone to VTE. As this was a database study, we were limited by the data collected and did not have enough information regarding GvHD therapy or immobilization. Of note, we did see a high incidence of skin GvHD (65%) in patients who developed VTE which could be due to a decreased mobility of limbs and possible impact on venous return. This is merely hypothesis generating at this point and as such requires a prospective analysis of VTE after HSCT. The limitations of our study are related to the constraints of database research and our reliance on outpatient pharmacy records to identify patients on anticoagulation who were then assessed for VTE. The incidence of VTE captured may therefore be an underestimation of the true rate of VTE. Patients who had an inpatient thrombotic episode but were not treated as an outpatient would have been missed, although this number is likely to be small, as most patients receive a minimum of 3 months of anticoagulation which is generally longer than hospital admissions for HSCT recipients. Patients who had contraindications to anticoagulation, such as thrombocytopenia, would also have been missed. While our overall incidence of VTE at 8.3% is high, this may be an underestima-

References 1. Barlogie B, Tricot G, Anaissie E, et al. Thalidomide and hematopoietic-cell transplantation for multiple myeloma. N Engl J Med. 2006;354(10):1021-1030. 2. Lonial S, Kaufman J, Tighiouart M, et al. A phase I/II trial combining high-dose melphalan and autologous transplant with bortezomib for multiple myeloma: a dose- and

1190

tion given the aforementioned limitations in identifying all cases, reinforcing the message that VTE is a prevalent and concerning complication after allogeneic HSCT. Variables that could not be formally addressed in this analysis included incidence of VOD and cause of death. In an effort to assess for VTE risk factors, we evaluated both standard HSCT and VTE related variables, and also evaluated sirolimus use in the multivariate analysis. Patients receive sirolimus either as GvHD prophylaxis or treatment. There is literature suggesting that the use of sirolimus is associated with vascular disruption, possibly leading to more vascular complications such as VTE.16-18 In our analysis, we did not find an association between sirolimus and VTE incidence, suggesting that GvHD itself, and not treatment with sirolimus, is the risk factor for developing VTE. We do however recognize that the number of patients in each of the GvHD prophylaxis/therapeutic categories is small, and thus this analysis is likely not sufficiently powered to draw any firm conclusions regarding the impact of immunosuppressive medications on the risk of VTE in HSCT. Although we documented an increased risk of VTE after HSCT, it is difficult to make recommendations about routine thromboprophylaxis without further study. Current CHEST guidelines recommend thromboprophylaxis for hospitalized patients with a VTE risk greater than 10%,19 while the International Society on Thrombosis and Haemostasis (ISTH) guidelines recommend that patients with a Khorana risk score of 3 or more and a VTE risk of about 7%, receive thromboprophylaxis,20 as the high rates of VTE justify the risk of bleeding from anticoagulation in these patient groups. In our population, the cumulative 1year and 2-year incidence of VTE in all allogeneic HSCT patients was 5.5% and 7.1%, respectively, which would be within the range of the current recommendations. For patients with normal platelet counts, no additional bleeding risk, active GvHD and prior history of VTE, thromboprophylaxis could be considered. Carefully designed prospective interventional trials of thromboprophylaxis after allogeneic HSCT are needed, targeting the subsets of patients that we have identified as being at the highest risk of VTE.

schedule-finding study. Clin Cancer Res. 2010;16(20):5079-5086. 3. Krishnan A, Pasquini MC, Logan B, et al. Autologous haemopoietic stem-cell transplantation followed by allogeneic or autologous haemopoietic stem-cell transplantation in patients with multiple myeloma (BMT CTN 0102): a phase 3 biological assignment trial. Lancet Oncol. 2011;12(13): 1195-1203.

4. de Lima M, Anagnostopoulos A, Munsell M, et al. Nonablative versus reduced-intensity conditioning regimens in the treatment of acute myeloid leukemia and high-risk myelodysplastic syndrome: dose is relevant for long-term disease control after allogeneic hematopoietic stem cell transplantation. Blood. 2004;104(3):865-872. 5. Russell JA, Duan Q, Chaudhry MA, et al. Transplantation from matched siblings using

haematologica | 2017; 102(7)

VTE after allogeneic HSCT

once-daily intravenous busulfan/fludarabine with thymoglobulin: a myeloablative regimen with low nonrelapse mortality in all but older patients with high-risk disease. Biol Blood Marrow Transplant. 2008; 14(8):888895. Maris MB, Sandmaier BM, Storer BE, et al. Unrelated donor granulocyte colony-stimulating factor-mobilized peripheral blood mononuclear cell transplantation after nonmyeloablative conditioning: the effect of postgrafting mycophenolate mofetil dosing. Biol Blood Marrow Transplant. 2006;12(4): 454-465. Gonsalves A, Carrier M, Wells PS, McDiarmid SA, Huebsch LB, Allan DS. Incidence of symptomatic venous thromboembolism following hematopoietic stem cell transplantation. J Thromb Haemost. 2008;6(9):1468-1473. Azik F, Gokcebay DG, Tavil B, Isik P, Tunc B, Uckan D. Venous thromboembolism after allogeneic pediatric hematopoietic stem cell transplantation: A single-center study. Turk J. Haematol. 2015;32(3):228-233. Labrador J, Gonzalez-Rivero J, Monroy R, et al. Management patterns and outcomes in symptomatic venous thromboembolism following allogeneic hematopoietic stem cell transplantation. A 15-years experience at a

haematologica | 2017; 102(7)

single center. Thromb Res. 2016; 142:52-56. 10. Gerber DE, Segal JB, Levy MY, Kane J, Jones RJ, Streiff MB. The incidence of and risk factors for venous thromboembolism (VTE) and bleeding among 1514 patients undergoing hematopoietic stem cell transplantation: implications for VTE prevention. Blood. 2008;112(3):504-510. 11. Pihusch R, Salat C, Schmidt E, et al. Hemostatic complications in bone marrow transplantation: a retrospective analysis of 447 patients. Transplantation. 2002;74(9): 1303-1309. 12. Zahid MF, Murad MH, Litzow MR, et al. Venous thromboembolism following hematopoietic stem cell transplantation-a systematic review and meta-analysis. Ann Hematol. 2016;95(9):1457-1464. 13. Labrador J, Lopez-Anglada L, Perez-Lopez E, et al. Analysis of incidence, risk factors and clinical outcome of thromboembolic and bleeding events in 431 allogeneic hematopoietic stem cell transplantation recipients. Haematologica. 2013;98(3):437443. 14. 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):3664-3371. 15. Sorror ML, Maris MB, Storb R, et al.

16.

17.

18.

19.

20.

Hematopoietic cell transplantation (HCT)specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. Guba M, Yezhelyev M, Eichhorn ME, et al. Rapamycin induces tumor-specific thrombosis via tissue factor in the presence of VEGF. Blood. 2005;105(11):4463-4469. Schuetze SM, Zhao L, Chugh R, et al. Results of a phase II study of sirolimus and cyclophosphamide in patients with advanced sarcoma. Eur J Cancer. 2012; 48(9):1347-1353. Xu L, Qin Y, Huang J, et al. Effect of rapamycin-induced tumor vessel thrombosis combined with docetaxel in non-smallcell lung cancer. Anticancer Drugs 2013; 24(4):406-414. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous thromboembolism: American College of Chest Physicians evidence-based clinical practice guidelines (8th Edition). Chest J. 2008;133(6 Suppl):381S453S. Khorana AA, Otten HM, Zwicker JI, Connolly GC, Bancel DF, Pabinger I. Prevention of venous thromboembolism in cancer outpatients: guidance from the SSC of the ISTH. J Thromb Haemost. 2014; 12(11):1928-1931.

1191

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Platelet Biology & Its Disorders

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1192-1203

Bleeding risk of surgery and its prevention in patients with inherited platelet disorders

Sara Orsini,1 Patrizia Noris,2 Loredana Bury,1 Paula G. Heller,3 Cristina Santoro,4 Rezan A. Kadir,5 Nora C. Butta,6 Emanuela Falcinelli,1 Ana Rosa Cid,1 Fabrizio Fabris,8 Marc Fouassier,9 Koji Miyazaki,10 Maria Luisa Lozano,11 Pamela Zúñiga,12 Claire Flaujac,13 Gian Marco Podda,14 Nuria Bermejo,15 Remi Favier,16 Yvonne Henskens,17 Emmanuel De Maistre,18 Erica De Candia,19 Andrew D. Mumford,20 Gul Nihal Ozdemir,21 Ibrahim Eker,22 Paquita Nurden,23 Sophie Bayart,24 Michele P. Lambert,25 James Bussel,26 Barbara Zieger,27 Alberto Tosetto,28 Federica Melazzini,2 Ana C. Glembotsky,3 Alessandro Pecci,2 Marco Cattaneo,14 Nicole Schlegel29 and Paolo Gresele1; on behalf of the European Hematology Association - Scientific Working Group (EHA-SWG) on thrombocytopenias and platelet function disorders

Department of Medicine, Section of Internal and Cardiovascular Medicine, University of Perugia, Italy; 2Department of Internal Medicine, IRCCS Policlinico S. Matteo Foundation, University of Pavia, Italy; 3Hematología Investigación, Instituto de Investigaciones Médicas Alfredo Lanari, Universidad de Buenos Aires, CONICET, Argentina; 4La Sapienza University of Rome, Italy; 5Haemophilia Centre and Haemostasis Unit, Royal Free Hospital, London, UK; 6Unidad de Hematología, Hospital Universitario La Paz-IDIPaz, Madrid, Spain; 7Unidad de Hemostasia y Trombosis, Hospital Universitario y Politecnico La Fe, Valencia, Spain; 8Clinica Medica 1 - Medicina Interna CLOPD, Dipartimento Assistenziale Integrato di Medicina, Azienda-Ospedale Università di Padova and Dipartimento di Medicina, Università di Padova, Italy; 9 Consultations d'Hémostase – CRTH, CHU de Nantes, France; 10Department of Hematology, Kitasato University School of Medicine, Sagamihara, Japan; 11Servicio de Hematología y Oncología Médica, Hospital Universitario Morales Meseguery Centro Regional de Hemodonación, IMIB-Arrixaca, Universidad de Murcia, Murcia 30003 and Grupo de Investigación CB15/00055 del Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), Instituto de Salud Carlos III (ISCIII), Madrid, Spain; 12 Department of Hematology-Oncology, School of Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile; 13Service d'Hématologie Biologique Cochin Hospital, Paris, France; 14Medicina III, ASST Santi Paolo e Carlo, Dipartimento di Scienze della Salute, Università degli Studi di Milano, Italy; 15Department of Hematology, Hospital San Pedro de Alcántara, Cáceres, Spain; 16Assistance Publique-Hôpitaux de Paris, Armand Trousseau Children’s Hospital, French Reference Centre for Inherited Platelet Disorders, Paris, France; 17Hematological Laboratory, Maastricht University Medical Centre, Maastricht, The Netherlands; 18Department of Biology and Haematology, Centre Hospitalier Universitaire Dijon, France; 19Hemostasis and Thrombosis Unit, Institute of Internal Medicine, Policlinico Agostino Gemelli-Università Cattolica Sacro Cuore, Rome, Italy; 20School of Clinical Sciences, University of Bristol, UK; 21Cerrahpasa Medical Faculty, Pediatric Hematology Department, Istanbul, Turkey; 22Gülhane Military Medical Faculty, Pediatric Hematology Department, Ankara, Turkey; 23Reference Centre for Platelet Disorders, Bordeaux University Hospital Centre, Rythmology and Cardiac Modeling Institute (LIRYC), Xavier Arnozan Hospital, Pessac, France; 24Centre Régional de Traitement des Hémophiles, Centre Hospitalier Universitaire de Rennes, France; 25 st 1 Division of Hematology, Department of Pediatrics, Children's Hospital of Philadelphia, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PN, USA; 26Department of Pediatrics, Division of Hematology, Weill Cornell Medicine, New York, NY, USA; 27Department of Pediatrics and Adolescent Medicine, University Medical Center Freiburg, Germany; 28Hematology Department, S. Bortolo Hospital, Vicenza, Italy and 29Centre de Référence des Pathologies Plaquettaires (CRPP), Service d'Hématologie Biologique, CHU Robert Debré, AP-HP, Paris, France 1

Correspondence: paolo.gresele@unipg.it

Received: November 25, 2016. Accepted: April 4, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.160754 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1192 ©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.

1192

ABSTRACT

xcessive bleeding at surgery is a feared complication in patients with inherited platelet disorders. However, very few studies have evaluated the frequency of surgical bleeding in these hemorrhagic disorders. We performed a worldwide, multicentric, retrospective study to assess the bleeding complications of surgery, the preventive and therapeutic approaches adopted, and their efficacy in patients with inherited platelet disorders: the Surgery in Platelet disorders And Therapeutic Approach (SPATA) study. We rated the outcome of 829 surgical procedures carried out in 423 patients with well-defined forms of inherited haematologica | 2017; 102(7)

Surgery in inherited platelet disorders

platelet disorders: 238 inherited platelet function disorders and 185 inherited platelet number disorders. Frequency of surgical bleeding was high in patients with inherited platelet disorders (19.7%), with a significantly higher bleeding incidence in inherited platelet function disorders (24.8%) than in inherited platelet number disorders (13.4%). The frequency of bleeding varied according to the type of inherited platelet disorder, with biallelic Bernard Soulier syndrome having the highest occurrence (44.4%). Frequency of bleeding was predicted by a pre-operative World Health Organization bleeding score of 2 or higher. Some types of surgery were associated with a higher bleeding incidence, like cardiovascular and urological surgery. The use of pre-operative pro-hemostatic treatments was associated with a lower bleeding frequency in patients with inherited platelet function disorders but not in inherited platelet number disorders. Desmopressin, alone or with antifibrinolytic agents, was the preventive treatment associated with the lowest bleedings. Platelet transfusions were used more frequently in patients at higher bleeding risk. Surgical bleeding risk in inherited platelet disorders is substantial, especially in inherited platelet function disorders, and bleeding history, type of disorder, type of surgery and female sex are associated with higher bleeding frequency. Prophylactic pre-operative pro-hemostatic treatments appear to be required and are associated with a lower bleeding incidence.

Introduction Inherited platelet disorders (IPDs) are a heterogeneous group of bleeding diseases of variable clinical severity associated with a reduction of platelet number (inherited platelet number disorders, IPNDs) and/or function (inherited platelet function disorders, IPFDs). Spontaneous hemorrhages are mainly mucocutaneous and rarely serious, while the hemorrhagic risk of trauma or surgery is not well defined.1-3 Excessive bleeding at surgery is a feared complication of IPDs and is empirically prevented or treated with platelet transfusions, antifibrinolytic agents, desmopressin, or recombinant activated factor VII (rFVIIa), although evidence of the effectiveness of these measures is mostly anecdotal.4-7 Two recent international collaborative studies have assessed the delivery-associated bleeding risk and pregnancy outcome in a large series of patients with welldefined forms of IPNDs or IPFDs. These studies have shown that delivery-related maternal bleeding was more frequent in IPDs than in healthy pregnant women, and that the degree of thrombocytopenia and history of severe bleeding were predictive of delivery-related hemorrhagic risk.8,9 Although guidelines for the management of bleeding and of invasive procedures in patients with platelet disorders and/or thrombocytopenia have been generated,10,11 they were not based on objective data on the incidence of surgery-related bleeding, and thus were necessarily rather generic. Indeed, very few studies on surgery in IPDs have been carried out. One retrospective study including 44 children with mild bleeding disorders undergoing adeno-tonsillar procedures, 27 of whom had an unspecific platelet function disorder, concluded that prophylactic treatment with desmopressin and tranexamic acid is effective in preventing perioperative bleeding.12 Another retrospective study in 113 patients with congenital hemostatic disorders undergoing general surgery or endoscopic procedures, including 5 with platelet disorders, showed low morbidity and mortality rates with desmopressin pre-treatment.13 With regard to well defined IPFDs, surgery outcome was described mostly in case reports of patients with haematologica | 2017; 102(7)

Glanzmann thrombasthenia (GT), Bernard Soulier syndrome (BSS), or Hermansky-Pudlak syndrome (HPS). Platelet transfusion for major surgery and antifibrinolytics for minor invasive procedures were reported as effective prophylactic measures for GT.14-16 Surgery-related bleeding, when this occurred, was successfully treated with rFVIIa17-19 or platelet transfusions.15 Platelet transfusions, alone or in combination with antifibrinolytics20-22 or desmopressin,22 prevented bleeding in BSS patients, while platelet transfusions, alone23 or in combination with rFVIIa,24 were used for HPS patients. However, no conclusions on the rate of bleeding complications and its prevention can be drawn from these studies. Concerning the prevention and treatment of surgical bleeding, the largest experience reported so far is the recent evaluation of rFVIIa effectiveness and safety in 96 GT patients from an international observational registry, showing that rFVIIa, administered alone or together with platelet transfusions and/or antifibrinolytics, was effective for both minor and major surgery.25 With regard to IPNDs, platelet transfusions and, more recently, eltrombopag have been reported to successfully prevent bleeding without side-effects in a few MYH9related disease (MYH9-RD) patients.26-28 The aim of the Surgery in Platelet disorders And Therapeutic Approach (SPATA) study was to evaluate the bleeding complications associated with surgical procedures, the therapeutic approaches adopted for the prevention and treatment of hemorrhage, and their efficacy in a large series of IPD patients diagnosed according to welldefined, standardized criteria and undergoing a wide range of invasive procedures. Here we report the results of the analysis of the outcome of 829 surgical interventions carried out in 423 patients with IPD.

Methods Study population This study was promoted by the Scientific Working Group (SWG) on Thrombocytopenias and Platelet Function Disorders of the European Hematology Association (EHA). The Institutional Review Board of the co-ordinating center (CEAS Umbria, Italy) approved the study, each center complied with local ethical rules, 1193

S. Orsini et al.

and all patients or their legal representatives signed a written informed consent. Participating investigators were asked to review their records and extract data on surgery and invasive procedures carried out in patients with IPDs over recent years [median 8; interquartile range (IQR) 4-18 years] and to obtain additional data directly from the surgeon who carried out the intervention or, if this was not possible, from the patient or his/her relatives. Only patients with a definite diagnosis of IPD confirmed according to well-defined laboratory and/or molecular genetic criteria6,8-9 (Online Supplementary Tables S1 and S2) were eligible for the study. IPDs were subdivided into IPNDs, when low platelet count was the main phenotypic characteristic (e.g. MYH9-RD), and IPFDs, when platelet dysfunction was the dominant phenotypic feature independently of platelet count (e.g. autosomal dominant GT-variant) (Table 1). Patients with acquired platelet disorders of any etiology were excluded. All types of surgical procedures, including invasive diagnostic procedures (e.g. angiography, endoscopic and tissue biopsies) and dental extractions, were included. Caesarian sections

were excluded because these had been analyzed in a previous study.8,9 Surgical procedures were categorized post hoc as major surgery, minor invasive procedures and dental procedures according to the following criteria: a) major (any procedure in which a body cavity was entered, a mesenchymal barrier was crossed, a facial plane was opened, an organ was removed or normal anatomy was altered); b) minor invasive (any operative procedure in which only skin, mucous membranes or superficial connective tissue were manipulated, gastroscopy, colonoscopy and similar); and c) dental (i.e. extraction, abscess removal, apicectomy and similar).25

Classification of bleeding Bleeding history was assessed by the World Health Organization (WHO) bleeding assessment scale29 and, when available, by the International Society of Thrombosis and Hemostasis (ISTH) bleeding score scale.30 Severity of surgical bleeding was defined according to three different criteria: 1) the Bleeding Academic Research Consortium (BARC) classification, consider-

Table 1. Diagnosis and features of study subjects.

IPFD Glanzmann thrombasthenia Primary secretion defect Bernard-Soulier syndrome (biallelic) Delta granule deficiency Hermanskyâ&#x20AC;&#x201C;Pudlak syndrome Gray platelet syndrome Autosomal dominant GT-variant Familial platelet disorder and predisposition to acute myelogenous leukemia Defects in 2-adrenergic receptor Defects in collagen receptors P2Y12 deficiency Platelet-type von Willebrand disease Defect of thromboxane A2 receptor Scott syndrome CalDAG-related platelet disorder Combined alpha-delta granule deficiency TOTAL

IPND MYH9-related disease ANKRD26-related thrombocytopenia Bernard-Soulier syndrome (monoallelic) ACTN1-related thrombocytopenia X-linked thrombocytopenia Thrombocytopenia with absent radii Paris-Trousseau thrombocytopenia Congenital amegakaryocytic thrombocytopenia FLNA-related thrombocytopenia TOTAL

N (% of total)

Females (%)

Age (median-IQR)

Platelet count (x109/L) (median-IQR)

WHO (median-IQR)

89 (37.4) 46 (19.3) 17 (7.1) 17 (7.1) 11 (4.6) 10 (4.2) 10 (4.2) 8 (3.4)

53.9 56.5 58.8 82.3 54.5 30 90 50

33 (19-47) 29 (14-46) 34 (21-49) 50 (37-59) 31 (24-50) 31 (21-58) 45 (42-53) 38 (24-51)

245 (172-297) 219 (160-272) 33 (20-39) 197 (141-232) 261 (228-321) 65 (56-85) 85 (62-110) 102 (93-139)

3 (2-4) 1.5 (1-3) 3 (2-3) 2 (1.75-3) 2 (1-2.5) 2 (2-2) 1 (0.25-2) 2 (1.75-2.25)

6 (2.5) 5 (2.1) 5 (2.1) 4 (1.7) 3 (1.3) 3 (1.3) 2 (0.8) 2 (0.8) 238 (56.3% of 423)

33.3 60 40 100 66.7 100 50 50 58

19 (11-43) 30 (23-59) 36 (27-78) 41 (31-52) 33 (29-52) 74 (58-75) 54 (53-56) 52 (44-61) 36 (20-50)

178 (146-239) 184 (126-256) 190 (183-195) 130 (113-168) 166 (129-169) 340 (-)* 267 (251-284) 85 (78-92) 191.5 (113-266)

1 (0.25-1.75) 1 (0-2) 2 (2-2) 3 (2.75-3) 2 (2-2.5) 1.5 (0.75-3) 3 (3-3) 0 (0-0) 2 (1-3)

N (% of total)

Females (%)

Age (median-IQR)

Platelet count (x109/L) (median-IQR)

Median (IQR)

75 (40.5) 39 (21.1) 34 (18.4) 17 (9.2) 7 (3.8) 5 (2.7) 4 (2.2) 2 (1.1) 2 (1.1) 185 (43.7% of 423)

57.3 41 61.7 76.4 0 40 25 50 100 53.5

41 (31-56) 48 (38-67) 49 (42-60) 43 (35-51) 29 (16-36) 26 (17-30) 12 (7-16) 9 (6-11) 40 (33-47) 43 (32-59)

38 (22-50)1 40 (29-54)2 82 (61-105)3 93 (72-118)4 34 (24-34) 20 (16-31) 52 (18-115) 61 (61-61) 34 (29-38)5 50 (30-81)

1 (0-2) 1 (0-2) 1 (0-2) 1 (0-2) 1 (1-2) 2 (0-2) 0.5 (0-1) 1 (1-1) 1 (0.5-1.5) 1 (0-2)

Platelet count microscopic (x109/L): 155 (36-72); 250 (33-70); 391 (82-121.5); 4110 (93-129); 555.5 (47.25-63.75). *Only one count available. IPFD: inherited platelet function disorders; IPND: inherited platelet number disorders; N: number; IQR: interquartile range; WHO: World Health Organization; GT: Glanzmann thrombasthenia.

1194

haematologica | 2017; 102(7)

Surgery in inherited platelet disorders

ing as excessive any bleeding with a BARC 2 or more;31 2) a subjective evaluation from the surgeon or, when not available, from the patient;8,9 and 3) duration (from less than six hours to more than three days), considered as excessive when more than six hours. Procedures associated with excessive bleeding according to any of the above three criteria were classified as any excessive bleeding (AEB). When available, maximal drop of hemoglobin after surgery (g/dL) was registered. Outcome of treatment of surgical bleeding was classified as successfully controlled, not responsive or re-bleeding. Not responsive was an excessive bleeding episode that the treatment(s) applied were not able to stop. Re-bleeding indicates a new episode of bleeding occurring at a later time point after the procedure.

Statistical analysis Data are reported as medians and 25th-75th percentiles (IQR) when continuous and as counts and percentages when categorical.

Logistic regression was used to assess the association between patients' or surgery characteristics with bleeding outcome. The Ď&#x2021;2 test and Cochran-Armitage's Trend Test were used to compare categorical data. R software (R Foundation for Statistical Computing, Vienna, Austria; www.R-project.org) was used for all analyses. Two-sided P<0.05 was considered statistically significant.

Results Patientsâ&#x20AC;&#x2122; characteristics Four hundred and twenty-three patients (age 2-91 years; median 40; IQR 23.7-54; 56% females), with 25 different forms of IPDs (16 IPFDs and 9 IPNDs) enrolled by 49 centers across 17 countries underwent a total of 829 surgical procedures. Two hundred and thirty-eight (56.3%; median age 36 years; 58% females), for a total of 455 proce-

Table 2. Characteristics of surgical procedures according to diagnosis.

IPFD

N (% of total)

Glanzmann thrombasthenia 182 (40) Primary secretion defect 76 (16.7) Bernard-Soulier syndrome (biallelic) 36 (7.9) Delta granule deficiency 36 (7.9) Hermanskyâ&#x20AC;&#x201C;Pudlak syndrome 22 (4.8) Autosomal dominant GT-variant 22 (4.8) Gray platelet syndrome 17 (3.7) Defects in collagen receptors 16 (3.5) Familial platelet disorder and 13 (2.8) predisposition to acute myelogenous leukemia Defects in 2-adrenergic receptor 9 (1.9) P2Y12 deficiency 9 (1.9) Defect of thromboxane A2 receptor 5 (1.1) Platelet-type von Willebrand disease 5 (1.1) Scott syndrome 3 (0.7) CalDAG-related platelet disorder 2 (0.4) Combined alpha-delta granule deficiency 2 (0.4) TOTAL 455 (54.9% of 829)

IPND

N (% of total)

MYH9-related disease 148 (39.6) ANKRD26-related thrombocytopenia 89 (23.8) Bernard-Soulier syndrome (monoallelic) 74 (19.8) ACTN1-related thrombocytopenia 39 (10.4) X-linked thrombocytopenia 9 (2.4) Thrombocytopenia with absent radii 6 (1.6) Paris-Trousseau thrombocytopenia 5 (1.3) Congenital amegakaryocytic thrombocytopenia 2 (0.5) FLNA-related thrombocytopenia 2 (0.5) TOTAL 374 (45.1% of 829)

Age at surgery (median-IQR)

Major surgery (%)

Minor surgery (%)

Dental procedure (%)

31 (13-50) 26 (8-40) 30 (16-49) 50 (35-55) 16 (12-24) 23 (11-33) 60 (22-69) 10 (7-47) 28 (20-48)

43.9 69.7 52.8 72.2 63.6 63.6 52.9 56.2 61.5

15.4 10.5 19.4 13.9 18.2 0 23.5 12.5 23.1

40.7 19.7 27.8 13.9 18.2 36.4 23.5 31.2 15.4

36 (22-44) 73 (72-76) 24 (21-30) 34 (30-38) 74 (73-76) 30 (28-31) 31 (15-52)

44.4 33.3 80 40 66.7 50 0 54.5

22.2 66.7 0 40 33.3 0 50 16

33.3 0 20 20 0 50 50 29.5

Age at surgery (median-IQR)

Major surgery (%)

Minor surgery (%)

Dental procedure (%)

28 (12-47) 35 (19-50) 34 (20-51) 30 (9-44) 20 (17-26) 16 (10-27) 1 (1-4) 8 (6-11) 23 (17-28) 31 (14-47)

64.9 70.8 67.6 66.7 44.4 50 80 50 0 66

6.7 2.2 12.2 10.3 0 33.3 0 50 0 7.5

28.4 26.9 20.3 23.1 55.5 16.7 20 0 100 26.5

IPD: inherited platelet disorders; IPFD: inherited platelet function disorders; IPND: inherited platelet number disorders; N: number; AEB: any excessive bleeding; IPFD: inherited platelet function disorders; IPND: inherited platelet number disorders; N: number; IQR: interquartile range; GT: Glanzmann thrombasthenia.

haematologica | 2017; 102(7)

1195

S. Orsini et al.

dures, had an IPFD and 185 (43.7%; median age 43 years; 53.5% females), for a total of 374 procedures, an IPND. Diagnosis and baseline characteristics are reported in Table 1. In order of frequency, IPFDs were GT, primary secretion defect, biallelic BSS (bBSS), Î´ granule deficiency, HPS, Gray platelet syndrome (GPS) and autosomal dominant GT-variant; IPNDs were, in order, MYH9-related disorder, ANKRD26-related thrombocytopenia, monoallelic BSS (mBSS), ACTN1-related thrombocytopenia. Bleeding history was on average mild (WHO grade median 2, IQR 1-3), but 25% of patients had a WHO grade 3 (79.8% of which were IPFD) and 3.3% a WHO grade 4 (64.3% of which were IPFD) (Online Supplementary Figure S1). Thrombocytopenia in IPNDs was on average mild (median 68x109/L; IQR 30-81x109/L) but 50% of patients had a platelet count less than 50x109/L and 25% less than 30x109/L. The ISTH bleeding assessment tools (BAT) bleeding score was available for 143 patients (33.5%), with a median score of 3 in the overall IPD population (IQR 1-7), 6 (IQR 2-11.25) for IPFD (n=89), and 1 (IQR 01) for IPND (n=54).

Type of surgery and prophylactic treatments Procedures were: 59.7% major surgeries, 28.1% dental, and 12.2% minor invasive. Among major surgeries, the most frequent procedures were abdominal (15.2%), otorhinolaryngological (12.8%), gynecological (6.8%), and orthopedic (6.8%). Median age at surgery was 31 years for both IPFDs (IQR 15-52) and IPNDs (IQR 14-47.5). In IPND, median platelet count at surgery was 56x109/L (IQR 40-93). Table 2 summarizes the characteristics of surgical procedures according to IPD diagnosis.

An anti-hemorrhagic prophylactic pre-operative treatment was administered in 57.2% of the procedures, in particular in 80.6% of procedures in patients with IPFDs and in 20.6% of procedures in IPNDs. Frequency of use of pre-operative prophylaxis was independent of the type of surgery; in fact, for IPFDs, 77.4% of major surgeries, 87.7% of minor invasive procedures, and 82.8% of dental procedures were treated; for IPNDs these were 28.7%, 35.7%, and 26.3%, respectively. In contrast, administration of a prophylactic treatment was positively correlated with pre-operative WHO bleeding score (CochranArmitage trend test P<0.0001) (Figure 1A). Moreover, among IPNDs, the use of prophylactic pro-hemostatic treatment was negatively associated with platelet count quartiles (Cochran-Armitage trend test P<0.0001) (Figure 1B). Most frequently used prophylactic treatments were platelet transfusions (26.8% for IPFDs and 20.6% for IPNDs), followed by antifibrinolytic agents (11.2% for IPFDs and 4% for IPNDs), a combination of both (12.3% for IPFDs and 1.3% for IPNDs), desmopressin (8.1% for IPFDs and 1.1% IPNDs), desmopressin with antifibrinolytic agents (7.9% for IPFDs and 0% in IPND), and rFVIIa (5.9% for IPFDs and 0% for IPNDs) (Table 3). The use of prophylactic treatments according to diagnosis is reported in Table 4. Platelet transfusions consisted of fresh platelets from random donors in 97.1% of cases, 34.2% of which were HLA-matched, and of cryopreserved platelets in 2.9% of the procedures. In IPNDs, platelet transfusions were used more frequently in patients with platelet counts of 50x109/L or lower (OR 2.73, 95%CI: 1.63-4.61; P=0.0001)

Table 3. Prophylactic treatments according to disease category and type of surgery.

Prophylaxis

N (%)

None 355 (42.8) Any 474 (57.2) PT 199 (24) AA 66 (8) PT+AA 61 (7.4) DDAVP 41 (4.9) DDAVP+AA 36 (4.3) FVIIa 27 (3.3) Other 17 (2.1) PT+AA+DDAVP 8 (1) FVIIa+AA 7 (0.8) PT+ Other 4 (0.5) AA+Other 3 (0.4) PT+DDAVP 2 (0.2) FVIIa+AA+ Other 1 (0.1) PT+DDAVP+ Other 1 (0.1) PT+FVIIa 1 (0.1) TOTAL 829

IPD % AEB 19.7 19.6 26.3 15.1 16.1 7.3 8.3 18.5 35.3 0 14.3 50.0 33.3 0 0 0 0 19.66

N (%)

IPFD % AEB

88 (19.3) 367 (80.7) 122 (26.8) 51 (11.2) 56 (12.3) 37 (8.1) 36 (7.9) 27 (5.9) 11 (2.4) 8 (1.7) 7 (1.5) 4 (0.9) 3 (0.7) 2 (0.4) 1 (0.2) 1 (0.2) 1 (0.2) 455

40.9 20.9 31.4 17.6 17.5 8.1 8.3 18.5 45.4 0 14.3 50.0 33.3

24.84

IPND N (%) % AEB 267 (71.4) 12.7 107 (28.6) 14.9 77 (20.5) 18.2 15 (4) 6.7 5 (1.3) 0 4 (1.1) 0

6 (1.6)

374

16.7

13.37

Major surgery N (%) % AEB 232 (46.9) 263 (53.1) 137 (27.7) 15 (3.0) 30 (6.1) 17 (3.4) 31 (6.3) 7 (1.4) 12 (2.4) 6 (1.2) 1 (0.2) 3 (0.6) 1 (0.2) 2 (0.4)

1 (0.2) 495

20.8 22.8 29.2 6.7 20.0 17.6 3.2 14.3 33.3 0 100 66.7 100 0

0 21.86

Minor surgery N (%) % AEB 27 (26.7) 74 (73.3) 33 (32.7) 12 (11.9) 8 (7.9) 8 (7.9)

21.4 24.3 33.3 33.3 25.0 0

9 (8.9) 2 (2.0)

11.1 0

1 (1.0)

101

23.53

Dental procedures N (%) % AEB 96 (41.2) 137 (58.8) 29 (12.4) 39 (16.7) 23 (9.9) 16 (6.9) 5 (2.1) 11 (4.7) 3 (1.3) 2 (0.9) 5 (2.1) 1 (0.4)

16.7 10.9 3.6 12.8 8.3 0 40.0 27.3 66.7 0 0 0

1 (0.4)

233

13.30

IPD: inherited platelet disorders; IPFD: inherited platelet function disorders; IPND: inherited platelet number disorders; N: number; AEB: any excessive bleeding; PT: platelet transfusion; AA: antifibrinolytic agents; DDAVP: desmopressin; FVIIa: activated factor VII; Other: cryoprecipitate, fibrin-glue, fibrinogen, fresh frozen plasma, intravenous immunoglobulin, local hemostatic agent, suture, local tranexamic acid (for IPFD); Eltrombopag, fresh frozen plasma, intravenous immunoglobulin, local hemostatic agent, prophylactic surgical hemostasis (for IPND).

1196

haematologica | 2017; 102(7)

Surgery in inherited platelet disorders

and in those undergoing major surgery (OR 1.94, 95% CIs 1.10-3.42; P=0.02), while in IPFDs they were used more frequently in patients with a WHO of 2 or higher (OR 2.22, 95%CI: 1.37-3.59; P=0.001) and in those undergoing major surgery (OR 1.45, 95%CI: 1-2.12; P=0.051).

Bleeding outcome Excessive bleeding after surgery occurred in 163 of the surgical procedures in the overall IPD population (19.7%), when assessed by any of the pre-defined criteria (i.e. AEB), i.e. 1 episode of bleeding every 5 procedures. Excessive bleeding occurred in 15.3% of the procedures when scored as higher than 2 by the BARC classification, in 15.3% when assessed by subjective evaluation, and in 10.5% when defined by a duration of more than six hours (Online Supplementary Figure S2). When the two populations were analyzed separately, AEB was reported almost twice as frequently in IPFDs, who suffered AEB in 24.8% of the procedures, than in IPNDs, for whom AEB was reported in 13.4% of the procedures. The drop in hemoglobin was 1.29 g/dL (95%CI: 0.89-1.68) for IPFD (n=83), and 0.53 g/dL (95%CI: 0.170.89) for IPND (n=30) (P=0.005). Among IPFDs, AEB was reported more frequently in bBSS (44.4%), familial platelet disorder / acute myeloid leukemia (FPD/AML) (30.7%), GT (29.1%), HPS (27.3%), GPS (23.5%), and autosomal-dominant GT-variant (22.7%) than in the overall IPFD population. bBSS and GT patients received pre-operative prophylactic treatment in 97.2% and 90.6% of procedures, respectively, emphasizing the perceived high surgical bleeding risk associated with these disorders. Among IPNDs, AEB was reported more frequently in MYH9-RD patients (15.5%) than in the overall IPND population. The distribution of bleeding outcomes in the individual IPDs is shown in Online Supplementary Table S3. Frequency of excessive bleeding in the overall IPD population was 21.8% after major surgery, 23.8% after minor invasive procedures, and 13.3% after dental procedures, (for IPFDs, 28.6%, 28.8%, and 15.7%, respectively, and for IPNDs, 15%, 10.7%, and 10.1%). In the overall IPD population, AEB was reported more frequently after cardiovascular surgery (47.1%), followed by urological (34.2%), gynecological (26.8%), otorhinolaryngological (24.5%), plastic (21.4%), eye (20%), and abdominal (19.8%) surgery. Incidence of bleeding in the different IPD populations according to the type of surgery is shown in Table 5. Any excessive bleeding, occurring during the first procedure was a risk factor for the recurrence of bleeding during a second (IPFD: OR 6.7, 95%CI: 2.3-18.9, P=0.0004; IPND: OR 3.82, 95%CI: 1.09-13.4, P=0.0357) or third (IPFD: OR 10, 95%CI: 1.1-90.6, P=0.04; IPND: OR 27, 95%CI: 2.2324.9, P=0.0094) procedure.

into major, minor invasive and dental (Figure 2B). Among IPFDs, a higher frequency of post-surgical bleeding was observed in females (Table 6). To exclude the possibility that this might be due to the relatively high bleeding rate associated with gynecological procedures, we reevaluated the association between sex and AEB after excluding gynecological procedures. Results confirmed that female sex is a risk factor for post-surgical bleeding in IPFDs [OR 1.70 (1.05-2.77); P=0.03]. A higher frequency of post-surgical bleeding was observed in some disorders, like bBSS, GT and HPS (Online Supplementary Table S4). Finally, in IPND, a platelet count below the median (68x109/L) and an age over 70 were associated with a significantly higher frequency of post-surgical bleeding (Table 4).

Efficacy of prophylactic treatments The use of a prophylactic anti-hemorrhagic preparation was associated with a markedly reduced frequency of surgical bleeding in IPFDs (OR 0.38; 95%CI: 0.23-0.63) (Table 6). Indeed, AEB was reported more frequently in patients

Characteristics associated with post-surgical bleeding A significant association was found between the frequency of AEB and clinical bleeding history assessed by the WHO bleeding score, both in the overall IPD population (WHO 1=OR 2.6, 95%CI: 1.0-6.9; WHO 2=OR 5.7, 95%CI: 2.4-13.6; WHO 3=OR 11.2, 95%CI: 4.7-26.7; WHO 4=OR 17.5, 95%CI: 5.6-54.7) and in the two populations analyzed separately (Table 6 and Figure 2A). The association between the WHO bleeding score and AEB was also maintained when procedures were subdivided haematologica | 2017; 102(7)

Figure 1. Frequency of prophylactic treatments according to World Health Organization (WHO) bleeding score and platelet count. (A) Use of prophylactic pre-operative treatments according to pre-operative WHO bleeding score in the overall inherited platelet disorder (IPD) population, in inherited platelet function disorder (IPFDs) and inherited platelet number disorder (IPNDs). (B) Use of prophylactic pre-operative treatments according to pre-operative platelet count quartiles (x109/L) (microscopic) in IPNDs.

1197

S. Orsini et al.

not receiving a prophylactic treatment than in those receiving it (40.9% vs. 21%; P<0.01). Post-surgical bleeding was lowest in desmopressin-treated patients (AEB in 8.1% of procedures, OR 0.13; 95%CI: 0.04-0.45, as compared with no treatment), followed by desmopressin and antifibrinolytic agents (8.3%, OR 0.13; 95%CI: 0.04-0.46), antifibrinolytic agents alone (17.6%, OR 0.31; 95%CI: 0.13-0.71), antifibrinolytic agents and platelet transfusions (17.8%, OR 0.31; 95%CI: 0.14-0.70), and rFVIIa (18.5%, OR 0.33; 95%CI: 0.11-0.95). Platelet transfusions alone, the majority of which were, however, given to patients with WHO grade 3 or higher (86.8%) and/or undergoing major surgery (68.4%), were not associated with a lower frequency of AEB (31.1%).

Information about platelet transfusion modalities was obtained for 123 out of 276 procedures in which these were used. The median platelet transfusion dose was 4 units (2-8), significantly higher in IPFDs (5, range 4-8) than in IPNDs (2.5, range 1-4) (P=0.0007). The median time of administration before the procedure was 1 hour (ranging from a few minutes to 3 days). Among GT patients, the dose of platelet transfusions was recorded in 58 procedures: AEB was reported in 12 of them (20.7%) and occurred more frequently when the amount of platelets transfused was less than 6 units (9 of 12, 75%) than when it was more than 6 units (3 of 12, 25%) (P=0.04 by Ď&#x2021;2 test). Moreover, information about platelet refractoriness and/or antiplatelet antibody positivity was obtained for 42 GT

Table 4. Prophylactic treatments according to diagnosis.

IPFD

Prophylatics/total (% of treated)

PT (%)

AA (%)

PT+AA (%)

DDAVP (%)

DDAVP+AA (%)

FVIIa (%)

Other (%)

155/182 (90.7) 62/76 (85.5) 34/36 (97.2) 25/36 (83.3) 11/22 (50) 6/22 (36.4) 12/17 (70.6) 6/16 (37.5) 6/13 (61.5)

54 (34.8) 9 (14.5) 26 (76.5) 2 (8) 4 (36.4) 2 (33.3) 8 (66.7) (0) 4 (66.7)

30 (19.3) 9 (14.5) 1 (2.9) 3 (12) 1 (9.1) 2 (33.3) 2 (16.7) 1 (16.7) 1 (16.7)

35 (22.6) 6 (9.7) 3 (8.8) 6 (24) 1 (9.1) 2 (33.3) (0) (0) 1 (16.7)

6 (3.9) 15 (24.2) 1 (2.9) 5 (20) 4 (36.4) (0) (0) 2 (33.3) (0)

(0) 23 (37.1) 2 (5.9) 9 (36) (0) (0) 1 (8.3) (0) (0)

24 (15.5) (0) (0) (0) 1 (9.1) (0) (0) 1 (16.7) (0)

6 (3.9) (0) 1 (2.9) (0) (0) (0) 1 (8.3) 2 (33.3) (0)

7/9 (77.8) 7/9 (77.8) 3/5 (60) 1/5 (40) 3/3 (100) 2/2 (100) 0/2 (0) 455 (80.7)

3 (42.9) 5 (71.4) (0) 1 (100) 2 (66.7) 2 (100)

1 (14.3) (0) (0) (0) (0) (0)

1 (14.3) (0) 1 (33.3) (0) (0) (0)

2 (28.6) 1 (14.3) 1 (33.3) (0) (0) (0)

(0) (0) 1 (33.3) (0) (0) (0)

(0) (0) (0) (0) 1 (33.3) (0)

(0) 1 (14.3) (0) (0) (0) (0)

122

Prophylaxed/total (% of treated)

PT (%)

AA (%)

PT+AA (%)

DDAVP (%)

DDAVP+AA (%)

FVIIa (%)

Other (%)

42 (79.2) 18 (75) 7 (53.8)

8 (15.1) 1 (0.4) 5 (38.5)

1 (0.4)

3 (12.5)

Glanzmann thrombasthenia Primary secretion defect Bernard-Soulier syndrome (biallelic) Delta granule deficiency Hermanskyâ&#x20AC;&#x201C;Pudlak syndrome Autosomal dominant GT-variant Gray platelet syndrome Defects in collagen receptors Familial platelet disorder and predisposition to acute myelogenous leukemia Defects in 2-adrenergic receptor P2Y12 deficiency Defect of thromboxane A2 receptor Platelet-type von Willebrand disease Scott syndrome CalDAG-related platelet disorder Combined alpha-delta granule deficiency TOTAL

IPND

MYH9-related disease 53/148 (35.8) ANKRD26-related thrombocytopenia 24/89 (27) Bernard-Soulier syndrome (monoallelic) 13/74 (17.6) ACTN1-related thrombocytopenia 0/39 (0) X-linked thrombocytopenia 6/9 (66.7) Thrombocytopenia with absent radii 4/6 (66.7) Paris-Trousseau thrombocytopenia 5/5 (100) Congenital amegakaryocytic thrombocytopenia 2/2 (100) FLNA-related thrombocytopenia 0/2 (0) TOTAL 374 (28.6)

2 (33.3) 3 (75) 3 (60) 2 (100) 77

3 (50) 1 (25) 1 (20)

3 (5.6) 1 (0.4) 1 (7.7) 1 (16.7)

1 (20)

IPFD: inherited platelet function disorder; PT: platelet transfusion; AA: antifibrinolytic agents; DDAVP: desmopressin; FVIIa: activated factor VII; IPND: inherited platelet number disorder; prophylactic treatments with sample size <10 not shown. Other: cryoprecipitate; fibrin-glue, fibrinogen, fresh frozen plasma, intravenous immunoglobulin, local hemostatic agent, suture, local tranexamic acid (for IPFD), Eltrombopag, fresh frozen plasma, intravenous immunoglobulin, local hemostatic agent, prophylactic surgical hemostasis (for IPND).

1198

haematologica | 2017; 102(7)

Surgery in inherited platelet disorders

patients undergoing 143 invasive procedures in which platelet transfusions were given as prophylaxis. The bleeding rate (AEB) was 23.3% in those without and 37.5% in those with a history of platelet refractoriness or anti-platelet antibodies (P=not significant, ns). In contrast to IPFDs, prophylactic treatments did not seem to modify surgery-related bleeding frequency in IPNDs, since, indeed, AEB was reported in 12.7% of the procedures carried out without preparation (34 of 267) and in 14.9% of the procedures carried out with pre-operative prophylaxis (16 of 107). But in fact, antifibrinolytic agents

were associated with a lower post-surgical bleeding frequency in IPNDs (AEB in 6.7% of procedures), while other treatments were not. In particular, platelet transfusions were not associated with lower post-surgical bleeding, although it must be considered that they were mainly given to patients undergoing major surgery.

Treatment of bleeding and outcome Surgical procedures followed by AEB received an emergency treatment in 86.7% of the cases (98 of 113) for IPFDs (platelet transfusions 60.2%, antifibrinolytic agents

Table 5. Incidence of bleeding in the different inherited platelet disorder populations according to the type of surgery.

Procedure DENTAL PROCEDURES MINOR SURGERY: Cyst/abscess drainage Central catheter placement Hemorrhoidectomy Invasive procedure Colonoscopy Gastroscopy Biopsy TOTAL minor surgery MAJOR SURGERY: Thoracic surgery Cardiovascular surgery Urological surgery Neurological surgery Gynecological surgery Otorinolaringological surgery Plastic surgery Eye surgery Abdominal surgery Orthopedic surgery Breast surgery Dermatologic surgery TOTAL major surgery TOTAL

IPFD vs. IPND 2 P

IPD N (% AEB)

IPFD N (% AEB)

IPND N (% AEB)

233 (13.3)

134 (15.7)

99 (10.1)

5 (60) 11 (36.4) 8 (25) 25 (20) 25 (20) 11 (18.2) 17 (17.6) 102 (23.5)

3 (66.7) 9 (44.4) 3 (33.3) 15 (26.7) 21 (23.8) 10 (20) 13 (23.1) 74 (28.4)

2 (50) 2 (0) 5 (20) 10 (10) 4 (0) 1 (0) 4 (0) 28 (10.7)

ns ns ns ns ns ns ns ns

2 (50) 17 (47.1) 38 (34.2) 7 (28.6) 56 (26.8) 106 (24.5) 14 (21.4) 25 (20) 126 (19.8) 56 (12.5) 13 (7.7) 34 (5.9) 494 (21.9) 829 (19.7)

2 (50) 9 (77.8) 24 (37.5) 5 (20) 31 (35.5) 49 (24.5) 4 (25) 12 (41.7) 52 (30.8) 30 (16.7) 7 (14.3) 22 (9.1) 247 (28.7) 455 (24.8)

8 (12.5) 14 (28.6) 2 (50) 25 (16) 57 (24.6) 10 (20) 13 (0) 74 (12.2) 26 (7.7) 6 (0) 12 (0) 247 (15) 374 (13.4)

ns 0.02 ns ns ns ns ns 0.03 0.01 ns ns ns 0.0003 0.0001

IPD: inherited platelet disorder; IPFD: inherited platelet function; IPND: inherited platelet number disorder; vs.: versus; N: number: AEB: any excessive bleeding; ns: not significant.

Table 6. Univariate and multivariate logistic analyses of factors associated with surgical bleeding.

IPFD (n=455) Univariate analysis Multivariate analysis OR (95% CI) OR (95% CI) Female WHO bleeding score Grade 0 Grade 1 Grade 2 Grade 3 Grade 4 Platelet count <68x109/L Age â&#x2030;Ľ 70 years Prophylaxis

1.8 (1.1-2.9)

1.28 (2.13-0.76)

1.00 0.91 (0.17-4.97) 3.87 (0.87-17.17) 6.62 (1.52-28.89) 10 (1.8-55.53) -

1.00 1.29 (0.23-7.35) 4.96 (1.08-22.82) 8.47 (1.83-39.13) 19.23 (3.23-114.55) -

0.38 (0.23-0.63)

0.24 (0.14-0.43)

IPND (n=374) Univariate analysis Multivariate analysis OR (95% CI) OR (95% CI) -

1.00 4.27 (1.34-13.57) 5.16 (1.69-15.79) 10.9 (3.28-36.25) 20.44 (3.38-123.53) 2.04 (1.01-4.12) 1.84 (1.009-3.37) -

1.00

IPFD: inherited platelet function; IPND: inherited platelet number disorder; OR: Odds Ratio; CI: Confidence Interval; WHO: World Health Organization.

haematologica | 2017; 102(7)

1199

S. Orsini et al.

17.3%, other 11.2%, FVIIa 6.1%), and in 62% for IPNDs (31 of 50) (other 41.9%, platelet transfusions 38.7%, and antifibrinolytic agents 19.3%). Treatment of bleeding according to disorder and type of surgery is reported in Online Supplementary Table S4. Successful control was obtained in 73.4% of IPFDs and in 58% of IPNDs. The most effective treatments were antifibrinolytic agents (88.2% of bleedings controlled; 15 of 17), platelet transfusions (83% of bleedings controlled; 49 of 59), and other treatments (fresh frozen plasma, stitches, ice, compression and dressing; 100% of bleedings controlled, 10 of 10) for IPFDs, while for IPNDs, these were platelet transfusions (100% of bleedings controlled, 12 of 12), followed by other treatments (surgical hemostasis, packing, compression, stitches; 92.3% of bleedings controlled, 12 of 13), and antifibrinolytic agents (83.3% of bleedings controlled, 5 of 6). In 21 procedures carried out in 18 patients (12.9% of the procedures with AEB) outcome of bleeding was unfavorable (19 re-bleeding, 2 not responding to treatment). Of these, 80.9% occurred in IPFD (12 GT, 1 bBSS, 1 primary secretion defect, 1 CalDAG-related platelet disorder, 1 GPS, 1 defect of TP receptor) and 19.1% in IPNDs (3 MYH9-RD, 1 ANKRD-26 related thrombocytopenia). Among these patients, 1 patient had a WHO grade 0, 9 patients a WHO grade 2, 10 patients a WHO grade 3, and 1 patient a WHO grade 4. The majority of these procedures (n=15) were major surgeries, but 3 were colonoscopy with polypectomy, 2 dental procedures, and one an enteroscopy for an angiodysplasia; in 76.2% of them, prophylaxis had been administered before surgery. Re-bleeding was treated mainly with platelet transfusions and/or antifibrinolytic agents and resolved in all but 2 cases, a 32-year old man with GT undergoing partial lung resection for recurrent severe hemoptysis and a 51-year old man with MYH9-RD undergoing endovascular treatment of an intracranial aneurysm for whom the outcome was death.

and relevant clinical information on the bleeding risk associated with IPFDs. In fact, while a high bleeding risk of GT and bBSS is generally considered acceptable, the high frequency of surgery-associated bleeding in HPS, FPD/AML, GPS and autosomal-dominant GT-variants was unexpected because they are commonly considered to be nonsevere. However, it can not be excluded that AEB in these patients was due to the infrequent use of a pre-operative prophylactic treatment, probably consequent to the assessment of these disorders as 'mild'. Other IPFDs, too, like platelet type-von Willebrand disease (PT-VWD), TXA2 receptor defect and CalDAG-related platelet disorder, suffered frequent post-surgical bleedings, although this observation can only be considered anecdotal given the low number of patients enrolled. On the other hand, other IPFDs showed a low bleeding risk, like collagen

Discussion Although IPDs are conventionally considered to be rare, at least 14,000 patients each year undergo investigations worldwide for a suspected IPD and over 5600 new diagnoses are made.2 Therefore, cases in which a surgeon may have to deal with a patient with an IPD are not unexpected events, and knowledge of the bleeding risks associated with the distinct invasive procedures and the different platelet disorders may be of great help in guiding surgical management. Our study of 829 surgical procedures in 423 patients represents the largest experience reported so far on surgery in patients with IPDs. The study shows that the frequency of excessive bleeding associated with surgery in patients with IPD is substantial, varying from 9.9% to 19.7%, depending on the definition used. In particular, it is striking that bleeding frequency is almost double in platelet function disorders compared with thrombocytopenias, ranging from 15.4% to 24.8%, depending on the definition used. The frequency of AEB appeared to be especially high for some disorders, including bBSS, FPD/AML, GT and HPS, with up to 44.4% of the procedures resulting in excessive bleeding. Some of the findings of the present study provide novel 1200

Figure 2. Post-surgical bleeding in inherited platelet disorders (IPD). (A) Incidence of any excessive bleeding (AEB) in the overall IPD population, in inherited platelet function disorders (IPFDs) and in inherited platelet number disorder (IPNDs) according to pre-surgical World Health Organization (WHO) bleeding score. (B) Incidence of AEB in the different procedures according to pre-surgical WHO bleeding score.

haematologica | 2017; 102(7)

Surgery in inherited platelet disorders receptor defects, δ-granule deficiency, and primary secretion defects. Given that pre-operative prophylaxis was administered in only 37.5% of the procedures in patients with collagen receptor defects, these seem, indeed, to be mild disorders. Inherited platelet number disorders were associated with infrequent surgical bleeding, ranging from 5.4% to 13.3%, depending on the definition, with no significant differences observed among the different disorders. Given that 21% of the enrolled patients were GT, and that a prophylactic treatment was administered in 90.7% of the procedures carried out in this patient subgroup, the impact of this on the overall analysis was evaluated by excluding the procedures carried out in GT. AEB in the remaining IPFD population was 21.9%, not significantly different from the 24.8% observed in the total IPFD population (P=ns, χ2). The same procedure was applied to IPNDs by excluding MYH9-RD; AEB in the remaining IPND population was 11.9%, with no difference from the 13.4% of the total IPND population (P=ns, χ2 test). In IPNDs, 68x109/L platelets was the threshold below which bleeding rate increased significantly; a value very similar to that previously identified as predictive of bleeding at childbirth.8 In the overall IPD population, bleeding history was highly predictive of surgical bleeding. In fact, a WHO grade of 2 or higher was associated with a more than 4fold increase in bleeding rate. Moreover, bleeding occurring after the first surgical procedure strongly predicted the rate of bleeding in subsequent procedures. Of note, bleeding tendency was higher in IPFDs than in IPNDs, with 79.8% of patients with a WHO grade 3 and 64.3% with a WHO grade 4 being IPFD.8,9,32 Similarly, in the subgroup of patients for whom the ISTH BAT bleeding score was available, this was highly predictive of postsurgical bleeding, with an ISTH BAT of 6 or more associated with a strongly increased bleeding risk. Some types of surgery were associated with higher bleeding risk, like cardiovascular or urological surgery. On the other hand, also minor invasive procedures were associated with bleeding, suggesting that prophylactic measures also need to be applied to procedures such as gastroscopy with biopsy. Administration of an anti-hemorrhagic prophylactic treatment was associated with a reduced bleeding frequency in the IPFD population but not in IPND. This seems to be reflected in the current practice, likely based on expert consultation, given that, in our study, the vast majority of patients with IPFDs received prophylaxis (80.6%), while patients with IPND only occasionally received prophylaxis (20.6%). The apparent lower efficacy of pre-operative prophylaxis in IPND patients may derive from the lower absolute bleeding risk in this subpopulation, with consequent lower statistical power to detect reduced bleeding in subjects receiving prophylaxis, and/or by the use of milder prophylactic measures (lower dosage, shorter duration, etc.) due to the perception of a lower bleeding risk. On the other hand, the choice of the preventive measures did not appear to be always appropriate, since platelet transfusions, the most frequently used prophylactic treatment, was shown to be poorly effective, with bleeding occurring in 30.1% of the procedures in which they were used. These data are in agreement with previous findings in haematologica | 2017; 102(7)

pregnancy.8,9 However, it should be considered that platelet transfusions were most frequently used in patients with severe bleeding disorders and/or undergoing major surgery, and that the mode of administration of platelet transfusions was rather heterogeneous, and possibly sometimes incongruous. These observations suggest that the way platelet transfusions are employed (amount, type, timing) is often inappropriate.33,34 The most effective prophylactic treatment was desmopressin, alone or in combination with antifibrinolytic agents, while antifibrinolytic agents used alone were less effective. In particular, desmopressin was used as prophylaxis in 88 procedures (10.3% minor invasive, 26.1% dental, and 63.6% major procedures), only 6 of which were followed by AEB (7%), 4 (66.7%) after major surgeries, and 2 (33.3%) after dental procedures. Interestingly, 31.9% (28 of 88) of patients in whom desmopressin was used had a severe bleeding history (WHO grade 3) and only 3 of them suffered AEB, supporting the efficacy of this pro-hemostatic treatment for IPFDs, even for the more severe conditions. Activated recombinant factor VII, an approved treatment for GT, was seen to be a good prophylactic measure, in line with previous results.25 In our cohort of patients, rFVIIa alone was used as a prophylaxis in 36 patients, 32 of whom were GT; 55% of these patients had a WHO grade 1 or 2, 33% had a WHO grade 3 and 2.7% had a WHO grade 4. rFVIIa was used in cases of minor invasive procedures (27.8%), dental procedures (47.2%), and major procedures (25%). These data suggest that rFVIIa is efficacious also for severe cases and when used alone. Although the current study does not provide any data on the safety of the prophylactic measures, previous experience suggests that they are generally well tolerated. Mild adverse effects of desmopressin may include headache, nausea and hypotension, although sometimes more serious side effects, such as hyponatremia and renal dysfunction, may occur. After antifybrinolytic agents allergic or anaphylactic reactions and sometimes headache may occur. Finally, pro-hemostatic agents, and in particular recombinant FVIIa, may predispose to thrombotic complications; however, the latter are relatively rare and depend on the thrombotic risk profile of the patient and the procedure.5,35 Finally, treatment of surgical bleeding was successful in most IPFD cases (73.4%), and in a slightly lower number of IPND cases (58%). Our study has several limitations. First, it is retrospective, with all the inaccuracies in data collection that this may imply. However, the submitted questionnaire was strongly structured, with mandatory fields and predefined possible replies, ensuring a high degree of standardization; moreover, the large number of patients and procedures collected strengthen the conclusions. Second, a comparative population of normal subjects undergoing the same invasive procedures would have provided a better quantification of the excess surgical bleeding risk. This was indeed planned in the study protocol, but it was impossible to collect enough control cases. However, for the few surgical procedures in healthy controls collected, frequency of post-surgical bleeding was strikingly lower (1 of 34; 3%), and similar to the estimated hemorrhagic complications rate of surgery (1.4% to 6%) in otherwise healthy subjects.30,36 Third, for specific disorders, such as 1201

S. Orsini et al. CalDAG-related platelet disorder, combined α/δ granule deficiency, and Scott syndrome, the exact bleeding risk could not be estimated, because only a few procedures were available, reflecting their rarity. However, our results provide a first useful hint of the surgical bleeding phenotype of these forms. Fourth, we do not have information about possible side effects of the pro-hemostatic procedures employed, or about other concomitant factors that may have increased the risk of surgical bleeding (e.g. blood pressure, acquired coagulopathy, VWD, abnormalities of whole cell count). In conclusion, our study shows that surgery-related bleeding risk is substantial in IPDs, especially in IPFDs, that the bleeding history, some specific disorders and female sex are predictors of the bleeding risk, and that some types of invasive procedures are at particularly high risk. Importantly, prophylactic treatment is associated with a significant reduction of the bleeding frequency in IPFDs. Appendix: Study collaborators Gabriella Mazzucconi, Ematologia, Università Sapienza, Roma (Italy); Omamurhomu Otomewo, Haemophilia Centre and Haemostasis Unit, Royal Free hospital, (UK); Dr Moscardó, Dr Valles, Hospital La Fe (Spain); Jose Rivera, Servicio de Hematología y Oncología Médica, Hospital Universitario Morales Meseguery Centro Regional de Hemodonación, IMIB-Arrixaca, Universidad de Murcia, Murcia 30003; Grupo de investigación CB15/00055 del Centro de

References 1. Podda G, Femia EA, Pugliano M, Cattaneo M. Congenital defects of platelet function. Platelets. 2012;23(7):552-563. 2. Gresele P, Bury L, Falcinelli E. Inherited platelet function disorders: algorithms for phenotypic and genetic investigation. Semin Thromb Hemost 2016;42(3):292305. 3. Gresele P, Harrison P, Bury L, et al. Diagnosis of suspected inherited platelet function disorders: results of a worldwide survey. J Thromb Haemost. 2014;12(9): 1562-1569. 4. Valera MC, Kemoun P, Cousty S, Sie P, Payrastre B. Inherited platelet disorders and oral health. J Oral Pathol Med. 2013;42(2):115-124. 5. Gresele P, Falcinelli E, Bury L. Diagnostic approach and management of inherited platelet function disorders. Hamostaseologie. 2016; 36(4):265-278. 6. Gresele P; Subcommittee on Platelet Physiology. Diagnosis of inherited platelet function disorders: guidance from the SSC of the ISTH. J Thromb Haemost. 2015; 13(2):314-322. 7. Kirchmaier CM, Pillitteri D. Diagnosis and Management of Inherited Platelet Disorders. Transfus Med Hemother. 2010;37(5):237-246. 8. Noris P, Schlegel N, Klersy C, et al. Analysis of 339 pregnancies in 181 women with 13 different forms of inherited thrombocytopenia. Haematologica. 2014;99(8):13871394. 9. Civaschi E, Klersy C, Melazzini F, et al. Analysis of 65 pregnancies in 34 women

1202

10.

11.

12.

13.

14.

15.

16.

Investigación Biomédica en Red de Enfermedades Raras (CIBERER), Instituto de Salud Carlos III (ISCIII), Madrid (Spain); Diego Mezzano, Department of Hematology-Oncology, School of Medicine, Pontificia Universidad Católica de Chile, Santiago, (Chile);Dr Stieltje, Dr Horellou, Dr Roussel-Robert, Cochin Hospital (France); Cécile Lavenu-Bombled, Bicetre (France); Marie Christine Alessi, Marseille (France); MF Hurtaud-Roux, Robert Debré Hospital Paris (France); Christian Gachet, Arnaud Dupuis, Hôpitaux Universitaires De Strasbourg (France); Adam Cuker, UPENN/Philadelphia (United States); Teresa Seara Sevivas, CHUC (Portugal); Paola Giordano, Giuseppe Lassandro, University Of Bari, Department Of Biomedical Science and Oncology - Pediatric Unit "F. Vecchio" (Italy); Elvira Grandone, I.R.C.C.S. Casa Sollievo Della Sofferenza (Italy); Lorenzo Alberio, Inselspital, Bern, Ch (Switzerland); Katrien Devreese, Ghent University Hospital (Belgium); Tantawy Azza, Iman Ragab, Ain Shams University (Egypt); Maha Othman, Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Ontario (Canada); Shinji Kunishima, Nagoya Medical Center (Japan). Funding This study was promoted by the Scientific Working Group on Thrombocytopenias and Platelet Function Disorders of the European Hematology Association (EHA). This study was supported in part by a grant to PG from Telethon (Protocol #GGP15063) and in part by a fellowship grant to EF from Fondazione Umberto Veronesi. NB was supported by FISFondos FEDER CP14/00024 and PI15/01457.

with 5 different forms of inherited platelet function disorders. Br J Haematol. 2015; 170(4):559-563. Tosetto A, Balduini CL, Cattaneo M, et al. Management of bleeding and of invasive procedures in patients with platelet disorders and/or thrombocytopenia: Guidelines of the Italian Society for Haemostasis and Thrombosis (SISET). Thromb Res. 2009; 124(5):e13-18. Bolton-Maggs PH, Chalmers EA, Collins PW, et al. A review of inherited platelet disorders with guidelines for their management on behalf of the UKHCDO. Br J Haematol. 2006;135(5):603-633. García-Matte R, María Constanza Beltrán M, Ximena Fonseca A, Pamela Zúñiga C. Management of children with inherited mild bleeding disorders undergoing adenotonsillar procedures. Int J Pediatr Otorhinolaryngol. 2012;76(2):291-294. Aryal KR, Wiseman D, Siriwardena AK, Bolton-Maggs PH, Hay CR, Hill J. General surgery in patients with a bleeding diathesis: how we do it. World J Surg. 2011;35(12):2603-2610. Kabashima A, Ueda N, Yonemura Y, et al. Surgical treatment of cecal cancer in a patient with Glanzmann's thrombasthenia: report of a case. Surg Today. 2009;39(11): 1002-1005. Sheikh AY, Hill CC, Goodnough LT, Leung LL, Fischbein MP. Open aortic valve replacement in a patient with Glanzmann's thrombasthenia: a multidisciplinary strategy to minimize perioperative bleeding. Transfusion. 2014;54(2):300-305. Gopalakrishnan A, Veeraraghavan R, Panicker P. Hematological and surgical management in Glanzmann's thrombas-

17.

18.

19.

20.

21.

22.

23.

24.

thenia: a case report. J Indian Soc Pedod Prev Dent. 2014;32(2):181-184. Erduran E, Aksoy A, Zaman D. The use of recombinant FVIIa in a patient with Glanzmann thrombasthenia with uncontrolled bleeding after tonsillectomy. Blood Coagul Fibrinolysis. 2009;20(3):215-217. d'Oiron R, Ménart C, Trzeciak MC, et al. Use of recombinant factor VIIa in 3 patients with inherited type I Glanzmann's thrombasthenia undergoing invasive procedures. Thromb Haemost. 2000;83(5):644-647. Hennewig U, Laws HJ, Eisert S, Göbel U. Bleeding and surgery in children with Glanzmann thrombasthenia with and without the use of recombinant factor VIIa. Klin Padiatr. 2005;217(6):365-370. Hartman MJ, Caccamese JF Jr, Bergman SA. Perioperative management of a patient with Bernard-Soulier syndrome for third molar surgery. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2007;103(5):626-629. Balci YI, Gözkeser E, Polat A, Gürses M, Kara CO, Herek Ö. Perioperative management of tonsilloadenoidectomy and circumcision of a patient with Bernard-Soulier syndrome: case report. Blood Coagul Fibrinolysis. 2014;25(8):907-908. Kostopanagiotou G, Siafaka I, Sikiotis C, Smyrniotis V. Anesthetic and perioperative management of a patient with BernardSoulier syndrome. J Clin Anesth. 2004;16 (6):458-460. Lederer DJ, Kawut SM, Sonett JR, et al. Successful bilateral lung transplantation for pulmonary fibrosis associated with the Hermansky-Pudlak syndrome. J Heart Lung Transplant. 2005;24(10):1697-1699. del Pozo Pozo AI, Jiménez-Yuste V, Villar A, Quintana M, Hernández-Navarro F.

haematologica | 2017; 102(7)

Surgery in inherited platelet disorders

Successful thyroidectomy in a patient with Hermansky-Pudlak syndrome treated with recombinant activated factor VII and platelet concentrates. Blood Coagul Fibrinolysis. 2002;13(6):551-553. 25. Poon MC, d'Oiron R, Zotz RB, Bindslev N, Di Minno MN, Di Minno G; Glanzmann Thrombasthenia Registry Investigators. The international, prospective Glanzmann Thrombasthenia Registry: treatment and outcomes in surgical intervention. Haematologica. 2015;100(8):1038-1044. 26. Pecci A, Gresele P, Klersy C, et al. Eltrombopag for the treatment of the inherited thrombocytopenia deriving from MYH9 mutations. Blood. 2010; 116(26): 5832-5837. 27. Pecci A, Barozzi S, d'Amico S, Balduini CL. Short-term eltrombopag for surgical preparation of a patient with inherited thrombocytopenia deriving from MYH9 mutation.

haematologica | 2017; 102(7)

Thromb Haemost. 2012;107(6):1188-1189. 28. Favier R, Feriel J, Favier M, Denoyelle F, Martignetti JA. First successful use of eltrombopag before surgery in a child with MYH9-related thrombocytopenia. Pediatrics. 2013;132(3):e793-795. 29. Miller AB, Hoogstraten B, Staquet M, Winkler A. Reporting results of cancer treatment. Cancer. 1981;47(1):207-214. 30. Rodeghiero F, Tosetto A, Abshire T, et al. ISTH/SSC bleeding assessment tool: a standardized questionnaire and a proposal for a new bleeding score for inherited bleeding disorders. J Thromb Haemost. 2010;8(9):2063-2065. 31. Mehran R, Rao SV, Bhatt DL, et al. White H. Standardized bleeding definitions for cardiovascular clinical trials: a consensus report from the Bleeding Academic Research Consortium. Circulation 2011; 123(23):2736-2747.

32. Federici AB, Bucciarelli P, Castaman G, et al. The bleeding score predicts clinical outcomes and replacement therapy in adults with von Willebrand disease. Blood. 2014; 123(26):4037-4044. 33. Estcourt LJ, Birchall J, Lowe D, GrantCasey J, Rowley M, Murphy MF. Platelet transfusions in haematology patients: are we using them appropriately? Vox Sang. 2012;103(4):284-293. 34. Charlton A, Wallis J, Robertson J, Watson D, Iqbal A, Tinegate H. Where did platelets go in 2012? A survey of platelet transfusion practice in the North of England. Transfus Med. 2014;24(4):213-218. 35. Levi M. Safety of prohemostatic interventions. Semin Thromb Hemost. 2012; 38(3):292-8. 36. Sadler JE. Von Willebrand disease type 1: a diagnosis in search of a disease. Blood. 2003;101(6):2089-2093.

1203

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Myeloid Leukemia

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1204-1214

Epigenetically induced ectopic expression of UNCX impairs the proliferation and differentiation of myeloid cells

Giulia Daniele,1 Giorgia Simonetti,2 Caterina Fusilli,3 Ilaria Iacobucci,2 Angelo Lonoce,1 Antonio Palazzo,1 Mariana Lomiento,4 Fabiana Mammoli,4 Renè Massimiliano Marsano,1 Elena Marasco,2 Vilma Mantovani,5,6 Hilmar Quentmeier,7 Hans G Drexler,7 Jie Ding,7 Orazio Palumbo,8 Massimo Carella,8 Niroshan Nadarajah,9 Margherita Perricone,2 Emanuela Ottaviani,2 Carmen Baldazzi,2 Nicoletta Testoni,2 Cristina Papayannidis,2 Sergio Ferrari,4 Tommaso Mazza,3 Giovanni Martinelli2 and Clelia Tiziana Storlazzi*1

Department of Biology, University of Bari “A. Moro”, Italy; 2Department of Experimental, Diagnostic and Specialty Medicine, University of Bologna, Italy; 3IRCCS Casa Sollievo della Sofferenza, Bioinformatics Unit, San Giovanni Rotondo, Italy; 4Department of Life Science, University of Modena and Reggio Emilia, Modena, Italy; 5Center for Applied Biomedical Research (CRBA), S. Orsola-Malpighi Hospital, Bologna, Italy; 6Unit of Medical Genetics, Department of Medical and Surgical Sciences, S. Orsola-Malpighi Hospital University of Bologna, Italy; 7Leibniz-Institute DSMZ-German Collection of Microorganisms and Cell Cultures, Department of Human and Animal Cell Lines, Braunschweig, Germany; 8Medical Genetics Unit, IRCCS “Casa Sollievo della Sofferenza (CSS)” Hospital, San Giovanni Rotondo, Italy; 9MLL Münchner Leukämielabor GmbH, München, Germany and 10Institute of Hematology “L. e A. Seràgnoli” S.Orsola-Malpighi Hospital, Bologna, Italy 1

ABSTRACT

Correspondence: cleliatiziana.storlazzi@uniba.it

Received: December 27, 2016. Accepted: April 12, 2017. Pre-published: April 14, 2017. doi:10.3324/haematol.2016.163022 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1204 ©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.

1204

e here describe a leukemogenic role of the homeobox gene UNCX, activated by epigenetic modifications in acute myeloid leukemia (AML). We found the ectopic activation of UNCX in a leukemia patient harboring a t(7;10)(p22;p14) translocation, in 22 of 61 of additional cases [a total of 23 positive patients out of 62 (37.1%)], and in 6 of 75 (8%) of AML cell lines. UNCX is embedded within a lowmethylation region (canyon) and encodes for a transcription factor involved in somitogenesis and neurogenesis, with specific expression in the eye, brain, and kidney. UNCX expression turned out to be associated, and significantly correlated, with DNA methylation increase at its canyon borders based on data in our patients and in archived data of patients from The Cancer Genome Atlas. UNCX-positive and -negative patients displayed significant differences in their gene expression profiles. An enrichment of genes involved in cell proliferation and differentiation, such as MAP2K1 and CCNA1, was revealed. Similar results were obtained in UNCX-transduced CD34+ cells, associated with low proliferation and differentiation arrest. Accordingly, we showed that UNCX expression characterizes leukemia cells at their early stage of differentiation, mainly M2 and M3 subtypes carrying wild-type NPM1. We also observed that UNCX expression significantly associates with an increased frequency of acute promyelocytic leukemia with PML-RARA and AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1 classes, according to the World Health Organization disease classification. In summary, our findings suggest a novel leukemogenic role of UNCX, associated with epigenetic modifications and with impaired cell proliferation and differentiation in AML.

haematologica | 2017; 102(7)

Epigenetic regulation of UNCX homeobox gene in AML

Introduction Homeobox (HB) genes encode transcription factors involved in key cellular processes such as body patterning, embryonic organogenesis, and cell-fate decisions.1,2 A growing body of literature has highlighted the crucial role of HB genes in normal hematopoiesis and leukemogenesis.3-6 Among HB genes, clustered HOX genes can promote the proliferation and inhibit the differentiation of hematopoietic progenitor cells and cause acute myeloid leukemia (AML)7 and acute lymphoid leukemia.8 Furthermore, several non-clustered HB genes, such as those belonging to the NKL subclass2 or to the Parahox (CDX)9 HB gene family, are critically involved in normal hematopoiesis and in leukemogenesis through their deregulation or ectopic expression. Notably, recent studies have highlighted a correlation between HB gene overexpression and mutations in epigenetic regulators.8,10 Alterations of DNA methylation are now widely considered a hallmark of cancer,11 although the precise leukemogenic mechanisms involving HB genes have still not been completely elucidated. Recently, Jeong et al. identified extended genomic regions exhibiting a low methylation level (≤10%), named “canyons”, spanning conserved domains that frequently harbor transcription factor genes (mainly HB genes) in murine hematopoietic stem cells (HSCs).10 Since several deregulated HB genes in AML fall inside canyons, this mechanism was proposed as a novel model for gene expression alteration in leukemogenesis.10 The starting point of this study was the identification of an M5 AML patient harboring a t(7;10)(p22;p14) translocation as the sole cytogenetic abnormality. This rearrangement was associated with the ectopic expression of the HB gene UNCX (NM_001080461, 7p22.3), likely as a result of a position effect. UNCX has tissue-specific expression in the eye, brain, and kidney, and it encodes a transcription factor involved in somitogenesis12,13 and neurogenesis.14 The murine Uncx gene was shown to map within a large canyon (23 kb) entirely covered by the repressive H3K27me3 histone mark in HSCs.10 Notably, UNCX expression has never been associated with cancer. We thus investigated the ectopic expression of UNCX in an independent and extensive AML cohort and performed genomic and functional studies to investigate its contribution to leukemogenesis.

Methylation analysis of the UNCX canyon DNA methylation ratios (MRs) of the UNCX canyon were determined through gene-specific amplification using in vitro transcription coupled with mass spectrometry (MassARRAY platform, Sequenom, San Diego, CA, USA).17 Statistical significance was obtained by comparing MRs of UNCX+ and UNCX– cases by using Student's t-test; the Satterthwaite correction was applied when unequal variance was present. To clearly visualize the methylation differences, the MR values were submitted to angular transformation.

Methylation and correlation analysis of the UNCX canyon in AML samples from The Cancer Genome Atlas (TCGA) We selected a total of 111 AML samples from the GDC Data Portal (https://gdc.cancer.gov), with both methylation and expression profile data. These subjects were divided according to the median value of UNCX expression (FPKM=0.0259) in UNCX-TCGA-positive (n=55) and UNCX-TCGA-negative (n=56). We compared the DNA methylation profile of UNCX as well as the whole genome (considering a minimum difference of 2-folds between groups) by the Mann-Whitney test. Spearman correlation was calculated between methylation and expression values within both sample sets. Correlation values were deemed significant at P<0.05.

DNMT3A mutational analysis A full description of the analytical methods used is provided in the Online Supplementary Methods.

Gene-expression profiling and pathway analyses Exon array was performed in two subgroups of our patient cohort [UNCX+ (ns. 1-Dx, 9, and 16) and UNCX– (ns. 13, 41, and 49)]. Similarly, 173 AML TCGA samples were divided according to the median value of UNCX expression in UNCX-TCGA-positive (n=47) and UNCX-TCGA-negative (n=126) cases. Pathway analysis was conducted on differentially expressed genes through QIAGEN’s Ingenuity Pathway Analysis software (IPA, QIAGEN, Redwood City, CA, USA).

Retrovirus-mediated UNCX expression in CB CD34+ cells Ectopic UNCX expression was achieved by retrovirus-mediated transduction of human cord blood (CB) CD34+ cells.18 Proliferation and differentiation rates were determined by colony forming cell (CFC) assays at 14 days after seeding. Flow cytometry analysis provided quantitative information regarding the maturation stage of infected cells.18 Cell morphology was assessed by MayGrunwald-Giemsa staining.

Methods

Correlation between UNCX expression and clinical/molecular features in TCGA patients

Patients, cell lines, and normal tissues

A total of 161 out of 173 TCGA AML samples were analyzed for potential associations between UNCX and clinical/molecular features. UNCX-TCGA-positive and -negative patients (Online Supplementary Table S2) were compared using the χ2 test or Fisher’s exact test. Continuous variables and medians of distributions were analyzed by Mann-Whitney U test or Kruskal-Wallis test for multiple comparisons.

We studied 62 AML patients (Table 1), including Case 1 with the t(7;10)(p22;p14) translocation, 75 AML and 14 additional cancer cell lines, and 6 normal tissues (Online Supplementary Table S1A and B). The use of samples was approved by the Policlinico S. Orsola-Malpighi Ethics Committee (ref. n. 253/2013/O/Tess of 29/10/2013).

Assessment of UNCX expression levels in AML UNCX expression was evaluated by RT-qPCR15,16 using a TaqMan UNCX Gene expression assay (Applied Biosystems, Milan, Italy). The TBP Endogenous Control (Applied Biosystems) was used as reference and Case 1 at onset (1-Dx) as calibrator. We classified patients on a median value of UNCX expression level (2-ΔΔCt=0.01300) as UNCX+ and UNCX–. haematologica | 2017; 102(7)

Results A novel t(7;10)(p22;p14) translocation in AML is associated with ectopic expression of the HB gene UNCX FISH results obtained in Case 1-Dx (Table 1) identified a breakpoint region on 7p22.3 within the fosmid clone 1205

G. Daniele et al. Table 1. Clinical, cytogenetic features, UNCX expression levels and DNMT3A, FLT3, and NPM1 mutational status of the 62 acute myeloid leukemia patients included in the study.

UNCX+

Case number 1-Dx 3 4 7 9 11 14 15 16 30 31 33 34 35 38 40 44 46 48 51 52

UNCX-

56 61 1-Rem 2 5 6 8 10 12 13 17 18

19 20 21 22 23 24 25 26 27 28 29 1206

Age/Sex

FAB subtype

Karyotype

DNMT3A mut V328I, R882C

FLT3 mut TKD

NPM1 mut WT

1.00000 0.13182 0.04966 0.01439 4.04648 0.01342 0.08304 0.74484 4.95883 0.10390 0.04932 0.03751 0.19981 0.02577 0.01774 0.19843 0.04565 0.01305 0.02259 0.02492 0.06999

/ / / WT / / / WT S708I / R882H L422L / WT / / / / / /

WT ITD WT WT WT ITD WT WT WT NA ITD WT TKD WT WT WT NA WT NA WT

WT WT WT WT WT NA NA WT MUT NA WT WT WT WT WT MUT WT WT WT WT

0.10236 3.72352 ND 0.00037 ND ND ND ND ND ND 0.00099

/ / V328I, R882C / / WT / / / WT /

WT WT WT WT NA WT TKD TKD TKD NA ITD

WT WT WT WT NA MUT NA WT MUT NA WT

0.00067

L422L

0.00011 ND ND 0.00065 0.00022 0.00038 ND 0.00031 ND ND 0.00512

WT WT WT L422L, R882C / F732del WT S663L / / WT

WT WT WT WT WT WT WT NA WT WT NA

WT WT WT WT NA WT WT NA WT WT NA

UNCX expression (2-ΔΔCt)

46/M M5 46,XY,t(7;10)(p22;p13)[18]/46,XY[2] 42/M M5 47,XY,+8[5]/46,XY,t(2;10)(q33;p13)[4]/46,XY[21] 20/F M5 46,XX[20] 44/F M5 NA 34/M M1 46,XY[13]* 43/F M5 46,XX [20] 54/M M0/M1 46,XY[20] 73/M M2 47,XY,+8[3]/46,XY[15] 54/F M3 46,XX,t(15,17)(q22,q12)[19]/46,XX[1] 81/F NI 46,XX[12] NA NI NA NA NI NA 83/M NI 47,XY,+mar[20] 43/M M4 46,XY,inv(16)[2] 52/F NI 46,XX,inv(16)[20] 74/F M3 46,XX,t(15;17)[9]/47,XX,+mar[13] 67/F M4 47,XX,+8[3]/48,XX,+8,+21[13] 49/M M1 Complex karyotype 64/M sAML to MDS 46,XY,+13[6] 57/M M1 NA 50/M M2 46,XY,t(3;?)(q21;?),add(7)(q34),der(7), del(14)(q23q32),add(16)(q22),-21,+1mar[4]/45, XY, t(3;?)(q21;?), add(7)(q34),-7, del(14)(q23q32), add(16)(q22),-21,+1mar[16] 77/F NI 46,XX[8] 53/F M2 46,XX[20] 46/M M5 46,XY[30] 24/M M5 46,XY[10] 40/M M5 46,XY[20] 50/F M5 46,XX,inv(9)(p11p13)[20] 71/F M5 46,XX,inv(9)(p11p13)[20] 46/M M5 46,XY[20] 61/M M5 47,XY,+8[3]/46,XY[5] NA/M M5 46,XX [20] 33/F M4 46,XX,t(11;21)(q13;q22)[4]/46, XX,add(21)(q22)[7]/46,XX[2] 76/M sAML to MDS 46,XY,del(5)(q13q31), del(11)(q21q23), add(15)(p13),-17,-18,+2mar[3]/46,XY, idem,del(4)(q27q33)[5]/46,XY[10] 77/M sAML to MDS 46,XY[20] 73/F M0 46,XX[20] 69/F sAML to MDS45,XX,t(1;3;13)(p34,q26,q14),-7[19]/46,XX[1] 77/F sAML to MDS 46,XX[20] 46/F NI 46,XX,del(7)(q22q35)[3] NA NI NA 50/F M0 46,XX[23]/45,XX,-7[5] 82/M M0/M1 46,XY[20] 63/F M4 46,XX,inv(16)(p13q22)[18] 61/F M1 46,XX[7]/47,XX,+11[13] NA NI NA

haematologica | 2017; 102(7)

Epigenetic regulation of UNCX homeobox gene in AML

62/M

36 37 39

32/F 50/F 69/M

41 42 43 45 47

NA 36/M 65/F 47/F 40/M

49 50 53 54 55 57 58

64/F NA NA NA 67/F NA 62/F

59 60 62

59/F 65/M 72/F

46,XX[2]/45,XX,t(3,21)(q26,q22), 0.01095 der(5q),-7,del(12)(p11,p13)[15]/45,XX,t(3,21) (q26,q22)der(5q),-7,del(11)(p13,p15), del(12)(p11,p13)[3] M1 46,XX[21] 0.00679 M1 46,XX[22] 0.01296 M5 43,XY,-7,hsr(11)(q13q23),-13, -17, 0.00528 del(20)(q11q13),-21-der(22)add(22)(p13), +1mar,1~3dmin[19]/44,XY,-7,hsr(11)(q13q23),-13,-17, del(20)(q11q13),-21-der(22)add(22)(p13), +2mar,1~3dmin[4] NI 46,XX [20] ND M0 NA 0.01141 M0/M1 46,XX[20] 0.00147 M2 46,XX[20] ND M5 46,XY,t(16;16)(p13;q22)[16]/46,XY, 0.00514 t(5;12)(q13;p13), t(16;16)(p13;q22)[4] M5 46,XX [20] ND NI NA 0.00383 NI NA 0.01296 NI NA 0.00372 M0 46,XX[20] 0.00737 NI NA 0.00034 M5 47,XX,+8,t(9;11)(p22;q23)[16]/48, ND XX,+8,+8,t(9;11)(p22;q23)[1]/50,XX,+8,+8, t(9;11)(p22;q23),+13,+19[3] sAML to MDS NA 0.00040 NI 46,XY,t(3;12)(p22;q24),+4,-15,+mar[19] 0.00039 M0/M1 46,XX,del(5)(q31q33)[2] ND

S714C

WT / WT

WT WT NA

MUT WT NA

/ / / / G550V, C554S

WT ITD WT ITD, TKD WT

WT MUT WT MUT WT

WT / / / / / /

WT ITD NA WT ITD TKD WT

WT NA NA WT MUT MUT WT

/ / /

WT WT WT

WT MUT WT

AML: acute myeloid leukemia; NA: not available; NI: not identified; sAML: secondary AML. *Karyotype on peripheral blood sample. UNCX+: > 0.01300; UNCX-: ≤ 0.01300. Values are rounded to the nearest ten thousandth. 1-Dx: case 1 at the time of diagnosis; 1-Rem: case 1 at the time of remission. ND: not detectable = UNCX Ct value >40; WT: wild-type; /: not tested due to lack of DNA material.

G248P85449H7 (WI2-1959P13) that contained UNCX as the only target gene (Online Supplementary Figure S1A and C). A breakpoint region on 10p14 was mapped to the G248P8034H6 (WI2-3164O12) fosmid clone within the coding sequence of the housekeeping gene UPF2, which is involved in nonsense-mediated decay of mRNAs19 and normally expressed in the bone marrow (BM) (http://biogps.org) (Online Supplementary Figure S1B and C). As a consequence of this rearrangement, UNCX was juxtaposed to the 3′ end of UPF2 in the derivative chromosome 7 [der(7)], as shown by FISH (Online Supplementary Figure S1), and ectopically expressed in Case 1-Dx, as detected by RT-qPCR (Table 1 and Figure 1). In addition, Case 1-Dx presented mutations in FLT3-TDK (Table 1) and WT1 (data not shown), which were not confirmed in Case 1 at complete remission (1-Rem), and wild-type NPM1 (Table 1).

UNCX is ectopically expressed in a subset of AML patients and cell lines To verify whether UNCX is expressed in AML independently of the t(7;10) translocation, UNCX transcript level was assessed by RT-qPCR in 61 additional AML cases. UNCX expression was detected in 37.1% (23 of 62) of our AML patient cohort (Table 1 and Figure 1A) and 8% (6 of 75) of the AML cell lines (Figure 1B and Online Supplementary Table S1A) at variable levels, regardless of haematologica | 2017; 102(7)

the French-American-British (FAB) subtype. Interestingly, UNCX expression was not detectable in Case 1-Rem (Table 1). None of the other investigated cases (ns. 9, 16, and 61) or cell lines harbored the t(7;10)(p22;p14) balanced translocation, as confirmed by FISH, or mutations in NPM1, FLT3 (Table 1), and WT1 (data not shown). UNCX expression was also detected in MEG-01 [chronic myeloid leukemia (CML)] and in brain cells. In normal cells, UNCX was not expressed in total BM, peripheral blood (PB), CB or BM CD34+ stem-progenitor cells (Online Supplementary Table S1B), and in hematopoietic lineages at different stage of maturation, according to Blueprint Epigenome data portal (http://blueprintdata.bsc.es/release_2016-08/#!/) (Online Supplementary Table S1C). In addition, we identified two alternative transcript isoforms for UNCX [UNCX-alternative 1 (UNCX-a1), GenBank KM587719 and UNCX-alternative 2 (UNCX-a2), GenBank KM587718], generated through the retention of distinct portions of intron II in both AML patients and cell lines (Online Supplementary Figure S2 and further details in the Online Supplementary Results). Moreover, to evaluate UNCX expression at protein level, we tested two different polyclonal antibodies specific for the N- and the C-terminus of the UNCX protein (ab105966, Abcam, Cambridge, UK, and AV47546, Sigma, respectively) by Western blot 1207

G. Daniele et al.

(WB). However, the lack of UNCX antibodies specificity (due to the detection of multiple non-specific bands) prevented the identification of the protein in HEK293T and Jurkat cell lines, indicated by the companies as WB positive controls, as well as in our AML cell lines. Notably, none of these antibodies has been referenced in any publication so far (further details in the Online Supplementary Methods).

UNCX ectopic expression is significantly associated with DNA methylation increase at UNCX canyon borders but is not correlated with DNMT3A mutations To assess whether epigenetic changes were responsible for the ectopic expression of UNCX in AML patients, we analyzed DNA methylation. UNCX exon 1 and also a 2.5kb region upstream of UNCX were significantly hypomethylated in all analyzed samples, regardless of UNCX expression level (Online Supplementary Figure S3A), as expected for genes within methylation canyons.10 Interestingly, we detected a significant increase of the MRs at UNCX 5′ and 3′ canyon borders in UNCX+ versus UNCX– samples. Specifically, three (Figure 2A, Online Supplementary Table S3A and Online Supplementary Figure S3B-D) and seven amplicons (Figure 2B, Online Supplementary Table S3B and Online Supplementary Figure S3E-M) at the 5′ and 3′ canyon borders, respectively, exhibited a significant increase in the average MR of CpGs. To corroborate this evidence, DNA methylation at UNCX canyon was found to be significantly higher (P<0.0001) in AML samples from TCGA expressing UNCX (UNCX-TCGA-positive) than in the UNCXTCGA-negative cohort (Figure 2C). Intriguingly, there was

a significant correlation between this DNA methylation increase and UNCX expression levels (rs=0.649, P<0.0001) (Online Supplementary Table S3C) in the UNCX-TCGApositive set. Moreover, a genome-wide methylation analysis yielded 270 genomic regions containing 289 differentially methylated genes, differing from UNCXTCGA-positive and UNCX-TCGA-negative samples by at least 2-fold (Online Supplementary Table S3D). We then evaluated whether mutations in DNMT3A could explain the altered methylation status of our UNCX+ cases. We detected approximately the same rate of DNMT3A mutation in UNCX+ and UNCX– patients from our internal cohort, with heterozygous lesions at position R882 and at other amino acid residues [3 of 8 (37.5%) in UNCX+, 6 of 19 (31.6%) in UNCX–] (Table 1). Notably, Case 1 showed the same mutations in both 1-Dx and 1Rem samples (Online Supplementary Table S3E), as expected for pre-leukemic lesions.20 In the TCGA cohort, we observed a significant association between mutations in DNMT3A and lack of UNCX expression (33.3% of mutant cases among UNCX-TCGA-negative vs. 9.1% among UNCX-TCGA-positive, P=0.0022 considering all patients’ subtypes; 33.6% of mutant cases among UNCX-TCGAnegative vs. 13.3% among UNCX-TCGA-positive, P=0.041 considering non-M3 AML). By analyzing the mutational status of genes involved in DNA methylation (DNMT1, 3A, 3B, IDH1/2, TET1/2, WT1), we observed a significant association with UNCX-TCGA-negative cases (58.9% of mutated cases among UNCX-TCGA-negative, vs. 25% among UNCX-TCGA-positive AML considering all cases, P=0.0002; 59.3% of mutated cases among UNCX-TCGA-negative vs. 36.7% among UNCX-TCGApositive AML, P=0.0384 considering non-M3 AML).

Figure 1. Expression levels of UNCX in acute myeloid leukemia (AML) patients and cell lines. RT-qPCR results showing UNCX expression in AML patients (A) and AML cell lines (B) in comparison to Case 1-Dx. Only positive samples exhibiting an expression level ≥0.10 are reported. The experiments were performed once and each sample was analyzed in triplicate.

1208

haematologica | 2017; 102(7)

Epigenetic regulation of UNCX homeobox gene in AML

Ectopic expression of UNCX in AML patients induces deregulation of genes involved in cell proliferation and differentiation To characterize the transcriptional program associated with UNCX expression in AML, we performed gene expression profiling of UNCX+ (ns. 1, 9, and 16) and UNCX– (ns. 13, 41, and 49) cases. A total of 596 genes were differentially expressed (414 up-regulated and 182 down-regulated in UNCX+ cases; Array Express accession n. E-MTAB-4098) (Online Supplementary Table S4A and Online Supplementary Figure S4A). TCGA AML samples were divided into 47 (27.2%) UNCX-TCGA-positive and 126 (72.8%) UNCX-TCGA-negative patients, according to the UNCX median expression value, which was zero, since more than half (i.e. 126) the individuals did not express UNCX at all. To check whether the partitioning strategy applied to both data sets matched, we considered the set of UNCX-TCGA-negative patients and sought for a gene fingerprint that better transcriptionally discriminated them from a subset of 12 “extremely positive” patients, namely expressing UNCX over the 75th percentile. A fingerprint of 199 on 20,530 genes was found by means of the sPLS-DA. This set of genes was fed to a PCA that, when applied on the samples of our internal data set, differentiated perfectly UNCX+ from UNCX– individuals (Online Supplementary Figure S4B and C). Analyses performed using Ingenuity Pathway Analysis software (IPA) (Online Supplementary Table S4B-D) identified a number of altered pathways, biological functions or diseases, and networks of genes involved in cell proliferation [mainly activation of mitogen-activated protein kinases (MAPKs), transforming growth factor-β (TGF-β), and phosphatidylinositol 3-kinase (PI3K) signaling pathways], cell-cycle regulation, hematopoiesis and hematologic disease, and cell death and survival. These results supported our hypothesis that UNCX could play an important role in differentiation (Figure 3A). Among the set of deregulated genes obtained from the analyses of both our cohort of samples and TCGA data, we tested the differential expression of 3 genes with an established role in leukemia (upregulation of CCNA1 and PIK3CB and downregulation of MAP2K1) by RT-qPCR in

all patients under study. We observed a statistically significant upregulation of CCNA1 and downregulation of MAP2K1 only in the 4 patients (ns. 1-Dx, 9, 16, and 61) showing the highest UNCX expression levels (UNCX≥1) (Online Supplementary Figure S4D), and not in the overall cohort of UNCX+ patients (Online Supplementary Figure S4E). Similarly, HOXA10 expression was significantly increased only in patients with UNCX ≥1 (Online Supplementary Figure S4D). Notably, HOXA10 was not identified as a deregulated gene in the GEP analysis but was found to be up-regulated in LUNCXIΔN cells, as described below.

UNCX expression strongly affects the proliferation and differentiation of normal myeloid cells in vitro To clarify the biological function of UNCX in hematopoietic cells, we ectopically expressed UNCX in CD34+ CB cells (LUNCXIΔN cells). CFC assays showed a significantly reduced number of colonies in LUNCXIΔN cells compared with LXIΔN (cells transduced with an empty vector) or non-transduced (NT) cells (Figure 4A), reflecting an UNCX-mediated effect on cell proliferation. To further assess the role of UNCX in the proliferation of myeloid cells, LUNCXIΔN, LXIΔN, and NT cells were seeded at the same density and counted every two days after infection for eight days. The proliferation rate was already reduced at day 5 post-infection (PI) in LUNCXIΔN cells (Figure 4B). Flow cytometry analysis at days 7, 10, and 14 PI revealed that CD34 tended to increase in LUNCXIΔN cells (mean value 29.1%, 15.9%, and 7.4%, respectively) compared with control LXIΔN cells (mean value 20.9%, 3.4%, and 0.6%, respectively), and this increase became statistically significant at day 14 PI (Figure 5A). RT-qPCR analysis confirmed the differential expression of CD34 at the transcript level; CD34 mRNA levels were significantly increased starting from day 5 PI (Online Supplementary Figure S5) to day 14 PI (Figure 5B) in LUNCXIΔN cells compared with LXIΔN cells. We also evaluated the expression level of HOXA10, KLF4, and MAFB, which are master regulators of mono-macrophage differentiation.21-24 We observed increasing levels of

Figure 2. DNA methylation levels at UNCX canyon in our acute myeloid leukemia (AML) patient cohort and TCGA samples. Mean values of DNA methylation ratios (MRs) of MassARRAY amplicons obtained for each of the 20 amplicons tested at both 5′ (A) and 3′ (B) UNCX canyon borders; P-values are reported above each amplicon showing a significant increase in the average MR in UNCX+ versus UNCX– patients. (C) Box and Whisker plot showing DNA methylation β values at the UNCX canyon in UNCX-TCGA-positive (n=55, median 0.3471, interquartile range 0.2898-0.3981) and UNCX-TCGA-negative (n=56, median 0.2771, interquartile range 0.2273-0.3348) patients. Significant differences were identified using the Mann-Whitney U-test (P<0.0001).

haematologica | 2017; 102(7)

1209

G. Daniele et al. A

Figure 3. Association analysis of TCGA acute myeloid leukemia (AML) cohort. (A) Treemap showing the activation state of the enriched processes for commonly differentially expressed genes between exon array and TCGA data; green: reduced activity, orange: enhanced activity. Higher significance levels for a process are reflected by a larger enclosing rectangle. (B and D) UNCX transcript levels were obtained by TCGA RNASeq data. (B) Mean value and standard deviation across FAB types are shown. UNCX expression was significantly higher in M3 cases compared with the other FAB subtypes (P<0.0001, except for M6/7). (C) Graph showing the distribution of UNCX-TCGA-positive and UNCX-TCGA-negative AML cases across FAB types (excluding M3). Percentages are reported in the graphs. (D) Mean value and standard deviation across World Health Organization (WHO) classes (here reported with abbreviations) are shown (AML with BCR-ABL1 and AML with t(9;11)(p21.3;q23.3); MLLT3-KMT2A are reported together due to low number of cases. UNCX expression was higher in acute promyelocytic leukemia (APL) with PLM-RARA and AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1.

1210

haematologica | 2017; 102(7)

Epigenetic regulation of UNCX homeobox gene in AML

Figure 4. Colony forming cell (CFC) assay and cell-proliferation analysis of UNCX-transduced cells. Number of colonies obtained in CFC assays (A) and proliferation curves (B) of non-transduced (NT), LXI∆N, and LUNCXI∆N CD34+ cells. All experimental data were verified in three independent experiments in triplicate.

HOXA10 and KLF4 transcripts from day 5 to 14 PI (Figure 5B and Online Supplementary Figure S5). In the same cells, MAFB expression was elevated at days 5 and 7, but decreased after day 10 PI in LUNCXIΔN cells compared with control cells (Figure 5B and Online Supplementary Figure S5). Notably, at day 14 PI, we observed a significant downregulation of MAP2K1 and upregulation of CCNA1 in LUNCXIΔN cells (confirming what was observed in AML patients; see above), which could indicate an effect of UNCX activation on cell proliferation and cell cycle control (Figure 5B). Moreover, morphological analysis with May-Grunwald-Giemsa staining demonstrated a persistence of immature myeloid cells at day 14 PI, which particularly affected the mono-macrophage lineage in the LUNCXIΔN sample (5% vs. 1% in the LXIΔN control sample) (Figure 5C). The persistence of undifferentiated cells of mono-macrophage lineage, even after several days PI, confirmed the flow cytometry and RT-qPCR results.

Ectopic expression of UNCX characterizes a specific subgroup of AML patients The TCGA data showed that UNCX expression was associated with lower age (median age of UNCX-TCGApositive patients 50 years; median age of UNCX-TCGAnegative patients 58.5 years; P=0.003) (Online Supplementary Table S2A), M3 FAB type (31.1% in UNCX-TCGA-positive patients; 0.9% in UNCX-TCGAnegative patients, with 93% of M3 cases expressing the gene; P<0.0001) (Figure 3B and Online Supplementary Table S2A), favorable prognosis according to karyotype-based classification (P=0.0001) (Online Supplementary Table S2A), and reduced fraction of bone marrow monocytes (median of UNCX-TCGA-positive patients 3.5%; median of UNCX-TCGA-negative patients 7.0%; P=0.0034) (Online Supplementary Table S2A). The same results were also confirmed by excluding M3 FAB cases from the analysis (Online Supplementary Table S2B), with a tendency towards increased UNCX expression in AML M2 cases compared to the other FAB types (P=0.0181). Accordingly, in the overall AML cohort M2 cases represented 45.1% of UNCX-TCGA-positive patients but only 20% of UNCXTCGA-negative patients (P=0.0233) (Figure 3C and Online Supplementary Table S2B). The analysis of UNCX expression, according to the WHO disease classification, showed the highest UNCX levels in APL with PML-RARA and haematologica | 2017; 102(7)

AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1 classes (P<0.0001) (Figure 3D and Online Supplementary Table S2A). On the other hand, we observed an increased frequency of AML with inv(16)(p13.1q22), AML with mutated RUNX1, and AML with mutated NPM1 among UNCX-TCGA-negative cases (Online Supplementary Table S2A). Indeed, UNCX expression showed an inverse correlation with NPM1c mutation, with 6.5% of UNCX-TCGA-positive and 32.1% of UNCX-TCGA-negative cases carrying the mutation (P=0.0029 excluding M3 cases). No association was observed with FLT3, RAS, and IDH1 mutational status (Online Supplementary Table S2C). The results suggest that UNCX expression characterized a specific subgroup of AML patients, according to their biological and molecular features.

Discussion We report for the first time the ectopic expression of UNCX in both 23 AML patients of our cohort (37.1%) and 8 AML cell lines (8%). Notably, the involvement of this HB gene and the identity of its downstream target genes in cancer have not previously been reported. Induced expression of UNCX in normal CD34+ CB cells in vitro markedly affected the proliferation and differentiation of the transduced cells. Indeed, UNCX expression resulted in both a reduction of the proliferation rate and an unprecedented persistence of CD34+ cells at days 10 (15.9%) and 14 (7.4%) PI. This result is noteworthy because in similar experiments,25 at day 10 PI, CD34+ HSCs were almost completely replaced by differentiated cells, and their contribution was reduced to approximately 7% of total cells. The observed reduction of CD34+ cell proliferation and their differentiation arrest at a very immature stage supported the hypothesis that UNCX exerts a potential leukemogenic effect over the myeloid lineage, which was corroborated by the transcriptional increase of HOXA10 and KLF4, along with the decrease of MAFB transcript levels. Notably, the expression of these genes, which are involved in mono-macrophage differentiation (HOXA10 and KLF4 in the early phase, MAFB in the late phase), was modulated along with UNCX activation in vitro. Upregulation of HOXA10 was also observed in a subgroup of UNCX+ patients of our cohort (UNCX expression level ≥1). This result is in line with the recent work of 1211

G. Daniele et al.

Yao et al. showing that Hoxa10 overexpression in murine bone marrow cells suppresses differentiation and induces expansion of a cell population derived from the common granulocyte/monocyte progenitor population and results in the AML phenotype.26 The identification of downstream target genes of UNCX will clarify its role in proliferation and differentiation. In parallel, the analysis of the TCGA AML cohort indicates that UNCX expression characterizes blast cells at early phases of differentiation, preferentially the M2 and M3 FAB types, which generally contain wild-type NPM1. Accordingly, UNCX positivity and NPM1c mutation are mutually exclusive both in our patient cohort and in the TCGA cohort. Moreover, the comparison between

TCGA RNA-Seq data and our exon array data suggests that UNCX expression is associated with a differentiation arrest in AML blasts, as confirmed by forcing the ectopic expression of UNCX in stem-progenitor cells, which induces an accumulation of immature cells of the monomacrophage lineage. The different maturation stage of UNCX-TCGA-positive AML blasts and UNCX-transduced normal stem-progenitor cells may explain the differences in the transcriptional signatures associated with UNCX expression in the two settings. Our GEP data and pathway analyses clearly show that the 3 UNCX+ patients (UNCX expression level ≥1) displayed remarkable deregulation of genes involved in hemopoiesis, proliferation, and cell cycle, such as those implicated in

1212

Figure 5. Delayed differentiation of CD34+ hematopoietic progenitor cells after retrovirus-mediated UNCX transfer. (A) Percentage values [including the calculated mean and the standard error of mean (SEM)] of CD34+ cells, measured in two independent UNCX transfer experiments (I and II), using flow cytometry at days 7, 10, and 14 post-infection (PI): non-transduced (NT) (blue line), LXI∆N (red line), and LUNCXI∆N (green line) CD34+ cells. *Two-tailed P≤0.05. (B) RTqPCR results of the genes encoding specific surface markers and transcription factors implicated in hematopoiesis, including MAP2K1 and CCNA1; the assay was performed on RNA extracted from cells at day 14 PI. *Two-tailed P≤0.05. (C) Morphological analysis after May-Grunwald-Giemsa staining at day 14 PI in NT, LXI∆N, and LUNCXI∆N CD34+ cells; the myeloid immature forms, characterized by a basophilic cytoplasm and a despiralized nucleus, are indicated with a red arrow.

haematologica | 2017; 102(7)

Epigenetic regulation of UNCX homeobox gene in AML MAPK, TGF-β, and PI3K signaling. All of these pathways play critical roles in proliferation, regulation of cell growth, differentiation, apoptosis, survival, and development in a wide range of cell types. Notably, although RTqPCR analysis did not validate these data in the overall cohort of UNCX+ patients, we observed MAP2K1 downregulation and CCNA1 upregulation in LUNCXIΔN cells, which showed UNCX expression level ≥1. Therefore, we speculate that abnormal levels of UNCX expression might be associated with specific transcriptional signatures and signaling pathway activation through a gene-dosage effect. However, further investigation is required to confirm this hypothesis. In AML patients (in both our and the TCGA cohorts), UNCX activation was associated and significantly correlated with a substantial increase in DNA methylation at both the 5ʹ and 3ʹ canyon borders. This methylation change might drastically affect UNCX transcription and result in its ectopic expression. However, here we found that DNMT3A mutations (both R882 and additional point mutations, including a few mutations not yet reported in the COSMIC database), did not associate with UNCX positivity, in contrast to what has been reported in mice where Dnmt3a mutations were directly correlated with shifts in DNA methylation at gene canyon borders.10 Our results are in line with previous reports of somatic mutations in DNMT3A in approximately 20% of de novo AML cases and 36% of cytogenetically normal AML cases.27,28 This result clearly suggests that DNMT3A mutations in humans are not directly correlated with UNCX canyon shrinkage and gene activation. In addition, analysis of the TCGA cohort showed the mutational status of alternative epigenetic modifier genes, classified according to the literature,29-31 and we did not observe any significant association with UNCX expression. Hence, future analyses are mandatory to clarify the epigenetic mechanism behind the regulation of UNCX. Since the Uncx canyon was found to be enriched in the repressive histone mark H3K27me3,10 we speculate that the increase in methylation might also induce a change in the histone mark configuration at the shrunk

References 8. 1. Shah N, Sukumar S. The Hox genes and their roles in oncogenesis. Nat Rev Cancer. 2010;10(5):361-371. 2. Homminga I, Pieters R, Meijerink JP. NKL homeobox genes in leukemia. Leukemia. 2012;26(4):572-581. 3. Abramovich C, Pineault N, Ohta H, Humphries RK. Hox genes: from leukemia to hematopoietic stem cell expansion. Ann N Y Acad Sci. 2005;1044:109-116. 4. Argiropoulos B, Humphries RK. Hox genes in hematopoiesis and leukemogenesis. Oncogene. 2007;26(47):6766-6776. 5. Jiang Q, Liu WJ. [Relationship between the HOX gene family and the acute myeloid leukemia-review]. Zhongguo Shi Yan Xue Ye Xue Za Zhi. 2013;21(5):1340-1344. 6. Alharbi RA, Pettengell R, Pandha HS, Morgan R. The role of HOX genes in normal hematopoiesis and acute leukemia. Leukemia. 2013;27(5):1000-1008. 7. Eklund EA. The role of HOX genes in

haematologica | 2017; 102(7)

10.

11. 12.

13.

canyon of UNCX, which would result in changes in chromatin condensation and in a remodeling of nuclear architecture. This hypothesis, together with that concerning a possible role of hydroxymethylation in UNCX activation, could not be verified here due to the lack of vital frozen cells for our patients. Notably, the future development of efficient anti-UNCX specific antibodies will help to understand the role of UNCX at the protein level in leukemogenesis. In fact, commercially available antibodies failed to identify UNCX in tumor cell lines, even in those with a high expression of the gene (Online Supplementary Table S1A and B). This finding could be explained by the documented poor expression of homeodomain-containing proteins, as reported in several studies32,33 and also for UNCX itself. Indeed, UNCX has been identified as a “missing protein” in the HEK293 cell line by means of bioinformatic predictions and high-throughput transcriptomic and proteomic technologies.34 In summary, our results suggest that ectopic activation of UNCX might represent an early epigenetic event in AML with a potential role in leukemogenesis because of its dramatic impact on the proliferation and differentiation of HSCs and progenitors. Our molecular and biological data indicate that UNCX expression may characterize a new subgroup of AML cases, which warrants further investigation. Moreover, although the molecular mechanisms underlying UNCX activation and its role in proliferation and differentiation remain to be clarified, this newly described alteration in AML, caused by an epigenetic modification affecting DNA methylation, introduces a previously undescribed scenario for the leukemogenic potential of epigenetic alterations. Funding This work was supported by the AIRC (Associazione Italiana per la Ricerca sul Cancro; AIRC IG n. 15413 for CTS), European LeukemiaNet, AIL (Associazione Italiana contro le Leucemie-Linfomi e Mieloma), AIL Modena, Fondazione Del Monte di Bologna e Ravenna, FIRB 2006, Ateneo RFO grants, and Progetto Regione-Università 2010-12 (L. Bolondi).

malignant myeloid disease. Curr Opin Hematol. 2007;14(2):85-89. Conway O'Brien E, Prideaux S, Chevassut T. The epigenetic landscape of acute myeloid leukemia. Adv Hematol. 2014;2014:103175. Rawat VP, Humphries RK, Buske C. Beyond Hox: the role of ParaHox genes in normal and malignant hematopoiesis. Blood. 2012;120(3):519-527. Jeong M, Sun D, Luo M, et al. Large conserved domains of low DNA methylation maintained by Dnmt3a. Nat Genet. 2014;46(1):17-23. Delpu Y, Cordelier P, Cho WC, Torrisani J. DNA methylation and cancer diagnosis. Int J Mol Sci. 2013;14(7):15029-15058. Sewell W, Sparrow DB, Smith AJ, et al. Cyclical expression of the Notch/Wnt regulator Nrarp requires modulation by Dll3 in somitogenesis. Dev Biol. 2009;329(2):400-409. Skuntz S, Mankoo B, Nguyen MT, et al. Lack of the mesodermal homeodomain protein MEOX1 disrupts sclerotome polar-

14.

15.

16.

17.

18.

ity and leads to a remodeling of the craniocervical joints of the axial skeleton. Dev Biol. 2009;332(2):383-395. Sammeta N, Hardin DL, McClintock TS. Uncx regulates proliferation of neural progenitor cells and neuronal survival in the olfactory epithelium. Mol Cell Neurosci. 2010;45(4):398-407. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25(4):402-408. Storlazzi CT, Lonoce A, Guastadisegni MC, et al. Gene amplification as double minutes or homogeneously staining regions in solid tumors: origin and structure. Genome Res. 2010;20(9):1198-1206. Ehrich M, Bocker S, van den Boom D. Multiplexed discovery of sequence polymorphisms using base-specific cleavage and MALDI-TOF MS. Nucleic Acids Res. 2005;33(4):e38. Grande A, Montanari M, Tagliafico E, et al. Physiological levels of 1alpha, 25 dihydroxyvitamin D3 induce the monocytic com-

1213

G. Daniele et al.

19.

20.

21.

22.

23.

1214

mitment of CD34+ hematopoietic progenitors. J Leukoc Biol. 2002;71(4):641-651. Weischenfeldt J, Waage J, Tian G, et al. Mammalian tissues defective in nonsensemediated mRNA decay display highly aberrant splicing patterns. Genome Biol. 2012;13(5):R35. Shlush LI, Zandi S, Mitchell A, et al. Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature. 2014;506(7488):328-333. Feinberg MW, Cao Z, Wara AK, Lebedeva MA, Senbanerjee S, Jain MK. Kruppel-like factor 4 is a mediator of proinflammatory signaling in macrophages. J Biol Chem. 2005;280(46):38247-38258. Gemelli C, Montanari M, Tenedini E, et al. Virally mediated MafB transduction induces the monocyte commitment of human CD34+ hematopoietic stem/progenitor cells. Cell Death Differ. 2006;13 (10):1686-1696. Gemelli C, Orlandi C, Zanocco Marani T, et al. The vitamin D3/Hox-A10 pathway supports MafB function during the monocyte differentiation of human CD34+ hemopoietic progenitors. J Immunol. 2008;

181(8):5660-5672. 24. Gemelli C, Zanocco Marani T, Bicciato S, et al. MafB is a downstream target of the IL-10/STAT3 signaling pathway, involved in the regulation of macrophage de-activation. Biochim Biophys Acta. 2014;1843 (5):955-964. 25. Salati S, Zini R, Bianchi E, et al. Role of CD34 antigen in myeloid differentiation of human hematopoietic progenitor cells. Stem Cells. 2008;26(4):950-959. 26. Yao J, Fang LC, Yang ZL, et al. Mixed lineage leukaemia histone methylases 1 collaborate with ERalpha to regulate HOXA10 expression in AML. Biosci Rep. 2014; 34(6):e00156. 27. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid leukemia. N Engl J Med. 2010;363(25):2424-2433. 28. Marcucci G, Metzeler KH, Schwind S, et al. Age-related prognostic impact of different types of DNMT3A mutations in adults with primary cytogenetically normal acute myeloid leukemia. J Clin Oncol. 2012; 30(7):742-750. 29. Cancer Genome Atlas Research N. Genomic and epigenomic landscapes of

30.

31.

32. 33.

34.

adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059-2074. Garg M, Nagata Y, Kanojia D, et al. Profiling of somatic mutations in acute myeloid leukemia with FLT3-ITD at diagnosis and relapse. Blood. 2015; 126(22): 2491-2501. Rampal R, Alkalin A, Madzo J, et al. DNA hydroxymethylation profiling reveals that WT1 mutations result in loss of TET2 function in acute myeloid leukemia. Cell Rep. 2014;9(5):1841-1855. Geng LN, Tyler AE, Tapscott SJ. Immunodetection of human double homeobox 4. Hybridoma. 2011;30(2):125-130. Kawagoe H, Humphries RK, Blair A, Sutherland HJ, Hogge DE. Expression of HOX genes, HOX cofactors, and MLL in phenotypically and functionally defined subpopulations of leukemic and normal human hematopoietic cells. Leukemia. 1999;13(5):687-698. Garin A, Odriozola L, Martinez-Val A, et al. Detection of missing proteins using the PRIDE database as a source of mass-spectrometry evidence. J Proteome Res. 2016; 15(11):4101-4115.

haematologica | 2017; 102(7)

ARTICLE

Acute Myeloid Leukemia

A three-dimensional ex vivo tri-culture model mimics cell-cell interactions between acute myeloid leukemia and the vascular niche

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Laura J. Bray,1,2 Marcus Binner,1 Yvonne Körner,1 Malte von Bonin,3-5 Martin Bornhäuser3-5 and Carsten Werner1

Max Bergmann Center of Biomaterials, Leibniz Institute of Polymer Research Dresden, Germany; 2Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia; 3Universitätsklinikum Carl-Gustav Carus, Faculty of Medicine, Technische Universität Dresden, Saxony, Germany; 4German Cancer Center (DKFZ), Heidelberg, Germany and 5German Cancer Consortium (DKTK), partner site Dresden, Germany 1

Haematologica 2017 Volume 102(7):1215-1226

ABSTRACT

x vivo studies of human disease, such as acute myeloid leukemia, are generally limited to the analysis of two-dimensional cultures which often misinterpret the effectiveness of chemotherapeutics and other treatments. Here we show that matrix metalloproteinase-sensitive hydrogels prepared from poly(ethylene glycol) and heparin functionalized with adhesion ligands and pro-angiogenic factors can be instrumental to produce robust three-dimensional culture models, allowing for the analysis of acute myeloid leukemia development and response to treatment. We evaluated the growth of four leukemia cell lines, KG1a, MOLM13, MV4-11 and OCI-AML3, as well as samples from patients with acute myeloid leukemia. Furthermore, endothelial cells and mesenchymal stromal cells were co-seeded to mimic the vascular niche for acute myeloid leukemia cells. Greater drug resistance to daunorubicin and cytarabine was demonstrated in three-dimensional cultures and in vascular co-cultures when compared with two-dimensional suspension cultures, opening the way for drug combination studies. Application of the C-X-C chemokine receptor type 4 (CXCR4) inhibitor, AMD3100, induced mobilization of the acute myeloid leukemia cells from the vascular networks. These findings indicate that the three-dimensional tri-culture model provides a specialized platform for the investigation of cell-cell interactions, addressing a key challenge of current testing models. This ex vivo system allows for personalized analysis of the responses of patients’ cells, providing new insights into the development of acute myeloid leukemia and therapies for this disease.

Introduction At the interface of in vitro culture models and complex animal models are sophisticated ex vivo models, which rely on our ability to replicate tissue microenvironments in order to sustain the growth of donor cells. Cell-cell and cell-matrix interactions, together with the signaling mechanisms between cells residing within spatially distinct niches, are important for the analysis of disease development and progression, and responses to drugs. Acute myeloid leukemia (AML) is a disease associated with 5-year survival rates of less than 40% in adults,1-3 although this figure decreases to less than 10% for adults aged over 65 years old.2,3 AML is characterized by an uncontrolled expansion of immature blasts resulting in a reduced normal blood cell production. Leukemic cell proliferation and resistance to chemotherapy have remained difficult to investigate ex vivo4,5 since conventional two-dimensional (2D) cell cultures cannot provide long-term maintenance of primary leukemic cells haematologica | 2017; 102(7)

Correspondence: Laura.bray@qut.edu.au

Received: October 8, 2016 Accepted: March 27, 2017 Pre-published: March 30, 2017. doi:10.3324/haematol.2016.157883 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1215 ©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.

1215

L.J. Bray et al.

without the support of additional growth factors or stromal cells and lack the microenvironmental stimuli of the bone marrow. Different bioengineered three-dimensional (3D) culture systems have, therefore, been developed to study AML cells in vitro more realistically using, however, stiff and porous materials as scaffolds and mono-cultures or co-cultures of AML with mesenchymal stromal cells (MSC).6-8 While these systems replicated important aspects of the stromal microenvironment, they did not allow for the exploration of leukemic-vascular cell-cell interactions which are critical for leukemia biology and progression.9 The vascular niche, so-called due to its density of blood vessels, is a location where endothelial cells and mural cells, such as pericytes, generate a microenvironment that influences the behavior of hematopoietic and leukemic stem and progenitor cells.10 In particular, angiogenesis is promoted by the bone marrow stroma and leukemic blasts and further increases in conditions such as AML and acute lymphoblastic leukemia.11-13 Activation by angiogenic growth factors and cytokines, such as vascular endothelial growth factor, stromal cell-derived factor 1 and fibroblast growth factor 2, modify the vascular niche to promote malignant growth.14 While a relationship between AML and vascular endothelium seems likely to contribute to the progression of AML, the mechanisms involved in these interactions are not yet understood.15-17 To recapitulate AML-vascular niche interactions in vitro, we used a set of thoroughly defined and widely tunable star-shaped poly(ethylene glycol) (starPEG)-heparin hydrogels18,19 to grow cells from four leukemia lines, KG1a, MOLM13, MV4-11 and OCI-AML3, as well as primary cells from AML patients, with human vascular endothelial cells as well as MSC. To support the specific requirements of the chosen cell types, the gel matrices were functionalized with precisely adjusted amounts of covalently attached adhesion receptor ligand peptide (RGD) motifs. Matrix metalloproteinase-responsive peptide sequences were incorporated as hydrogel crosslinkers to allow for localized cellular remodeling, thus supporting proliferation and migration of cells within the 3D gel cultures. A combination of growth factors (vascular endothelial growth factor, fibroblast growth factor 2, stromal cell-derived factor 1) known to associate tightly with the glycosaminoglycan heparin was applied to customize the gel matrices according to previously established protocols to afford sustainable administration resulting in the formation and maintenance of 3D endothelial cell capillary networks within the gels.20 To validate the culture model, we determined the impact of chemotherapeutics and signaling inhibitors on the tri-culture system and assessed the relevance of our findings. We applied a multi-disciplinary in vitro approach that integrates biological and physical techniques with human samples, ultimately extending our understanding of the impact of treatments on cell-cell interactions.

Methods Culture of cell lines KG1a, MOLM13, MV4-11 and OCI-AML3 cell lines were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ; Braunschweig, Germany) and used within 15 passages. KG1a, MOLM13 and MV4-11 cells were cultured in 1216

medium consisting of Roswell Park Memorial Institute (RPMI, Life Technologies, Darmstadt, Germany) medium supplemented with GlutaMax (Life Technologies), 10% fetal bovine serum (FBS; Hyclone Thermo Scientific, Schwerte, Germany) and 1% penicillin/streptomycin solution (PS; Life Technologies). OCI-AML3 cells were cultured in Dulbecco modified Eagle medium (DMEM, Life Technologies) supplemented with 10% FBS and 1% PS.

Culture of primary donor cells The studies were approved by the institutional review boards of all participating centers of the Study Alliance Leukemia in agreement with the Declaration of Helsinki and registered with National Clinical Trial numbers 00180115 (AML96 trial), 00180102 (AML2003 trial) and 00180167 (AML60+ trial). Written informed consent had been obtained from each patient. Three peripheral blood samples derived from patients with AML were obtained with ethical permission from the Uniklinikum Dresden, prepared as previously described,21 frozen, and thawed directly for experiments in hydrogels. Primary AML cells were cultured in medium consisting of StemSpan SFEM (Stem Cell Technologies, Grenoble, France) supplemented with 2% FBS for human myeloid long-term culture (Stem Cell Technologies), 1% L-glutamine, 1% PS solution (both from Life Technologies), 10 ng/mL FLT3L, 10 ng/mL stem cell factor, 10 ng/mL thrombopoietin, and 10 μg/mL interleukin 3 (all from R&D Systems, Minneapolis, USA). OCIAML3 cells were cultured in DMEM supplemented with 10% FBS and 1% PS. Human umbilical vein endothelial cells (HUVEC) were isolated as previously described22 and cultured in endothelial cell growth medium (Promocell, Heidelberg, Germany). MSC were derived from healthy volunteer donors after informed consent. The use of surplus bone marrow cells for MSC generation was approved by the ethics committee of the Technical University Dresden (Ethics approval ID: EK127042009). Bone marrowderived MSC were isolated as previously described23 and cultured in DMEM supplemented with 10% FBS and 1% PS. HUVEC and MSC were utilized for experiments between passages 1 - 6.

Statistics All statistical analyses were performed using GraphPad Prism version 6. Data were analyzed using either one-way or two-way analysis of variance (ANOVA) (depending on the experiment) with post-hoc Tukey multiple comparisons’ tests. Levels of statistical significance are reported (*P<0.05, **P<0.01, ***P<0.001 or ****P<0.0001).

Results Acute myeloid leukemia-vascular interactions within starPEG-heparin hydrogels A tri-culture model of AML-vascular interactions was established utilizing hydrogels of approximately 200-300 Pa stiffness (storage modulus) (Figure 1A), as this stiffness was reported optimal for the development of a robust endothelial network,20 and preliminary data showed that the endothelial cells would not form vascular networks in γ1.0 (1500 Pa) hydrogels (data not shown). Hydrogels were stable after casting and easily transferred into medium (Figure 1B). All AML cell lines, KG1a, MOLM13, MV4-11 and OCI-AML3, grew similarly within the 3D model. The colonization and proliferation of AML cells was visualized via light microscopy (Figure 1C) and confocal microscopy (Figure 1D-H) after 1 week of tri-culture. AML cell lines predominantly grew along the HUVEC-MSC vascular nethaematologica | 2017; 102(7)

An ex vivo model of acute myeloid leukemia

work. The AML cell lines were determined to proliferate in AML-HUVEC and AML-MSC co-cultures in a similar way as in the AML-HUVEC-MSC tri-cultures (Figure 2). AML cell proliferation resulted in a heterogeneous mixture of spheroids and loose cell clumps in contact with the vascular network (Figure 1E,F). Primary donor cells from patients with leukemia (pAML) grew more slowly than the AML cell lines, resulting in less dense areas of AMLvascular interactions (Figure 1I-K). However, pAML cells appeared as clumps within the vascular network branching (Figure 1I), rarely formed spheroids, and showed pref-

erence for single cell adherence and growth to the HUVEC-MSC network (Figure 1J,K). The HUVEC-MSC population displayed expression of CD31, M-CAM, ÎąSMA, and CD90 (Online Supplementary Figure S1).

Acute myeloid leukemia cells form spheroids within starPEG-heparin hydrogels AML mono-culture models were generated by culturing KG1a, MOLM13, MV4-11 and OCI-AML3 cells within starPEG-heparin hydrogels of various stiffness or in MatrigelTM over 2 weeks. To determine the optimal stiff-

Figure 1. A three-dimensional culture model of acute myeloid leukemia-vascular interactions. (A) A biohybrid starPEG-heparin hydrogel was utilized which allows for the culture of AML mono-cultures and tri-cultures with HUVEC and bone marrow-derived MSC. B) Macroscopic image of final cast hydrogel before culture. Scale bar = 5 mm. (C) Light microscope and (D,H) confocal images of the OCI-AML3 cell line (as a representative of all cell lines utilized in this study) in tri-culture with HUVEC and MSC after 7 days depicting (D, G, H) CD31 and (E, F) CD45 expression. Images display leukemia cell growth primarily along the vascular endothelial cells. (I) Light microscope and confocal images (J,K) of primary donor cells from a patient with AML in tri-culture with HUVEC and MSC after 7 days. Scale bar = 100 and 200 Îźm as indicated. Images display the preference of leukemia cells to attach to and grow along vascular structures or within vascular branching.

haematologica | 2017; 102(7)

1217

L.J. Bray et al.

ness for the culture of AML cells within starPEG-heparin hydrogels, three different cross-linking degrees were investigated (γ0.75, γ1, and γ1.5, where γ is the molar ratio of PEG to heparin). As previously described, the approximate stiffness of these cross-linking degrees equate to 500 Pa, 1500 Pa, and 3000 Pa (storage modulus), respectively.24 Hydrogels prepared at γ0.75 often disintegrated during the culture period. Proliferation was measured using PrestoBlue reagent. Hydrogels prepared at γ1 allowed the greatest proliferation when compared with 2D cultures and contained spheroids of the largest size amongst all cross-linking degrees (Figure 3A-D). MatrigelTM cultures consistently allowed the greatest proliferation compared with both 2D cultures and starPEG-heparin cultures. Human pAML cells derived from three different donors were only cultured in γ1 hydrogels. In all three donor samples, a heterogeneous population of cells was present, as evidenced by the mixture of spheroids, loose cell clumps, and single cells (Figure 3E). In some cases, elongated cells were also present.

Acute myeloid leukemia cells display similar functionality and phenotype between two- and three-dimensional cultures Migration assays and colony-forming assays were performed to determine whether the different matrices affected AML migratory abilities and clonogenicity, respectively. Migration of AML cells cultured in starPEGheparin hydrogels was decreased compared to that in 2D cultures (Online Supplementary Figure S2A; KG1a: 67.51% ± 32.66; MOLM13: 66.93% ± 38.17; MV4-11: 29.82% ± 15.95; OCI-AML3: 38.27% ± 23.97). This might be due to residual hydrogel fragments blocking the pores. In the colony-forming experiment, while the number of MV4-11 colonies derived from 3D cultures was decreased compared with that from 2D cultures (55.45% ± 17.65), KG1a, MOLM13 and OCI-AML3 colony formation was not or only marginally reduced in 3D cultures (Online Supplementary Figure S2B,C; 87.08% ± 2.103; 98.06% ± 0.02; 88.16% ± 42.73). Flow cytometry analysis of 3D and 2D mono-cultures showed similar marker expression with the following exceptions. MOLM13 3D mono-cultures displayed higher mean fluorescence intensities (MFI) of

CD33 and CD54, while MOLM13 2D cultures displayed higher MFI of CD29 (Online Supplementary Figure S3A). MV4-11 2D cultures showed higher MFI of HLA-DR and CD29 compared with 3D cultures. OCI-AML3 2D cultures demonstrated higher MFI of CD45 and CD54 than 3D cultures. Positive expression of target antigens in AML 3D and 2D cultures is detailed in Online Supplementary Table S1. Immunostaining for integrin α4, CD44 and CD54 was similar between all AML cells (Online Supplementary Figure S3B). Higher MFI of CD45 and HLADR was found in mono-cultures and higher MFI of CD49e and CD29 was found in pAML tri-cultures compared to mono-cultures. Immunostaining revealed similar expression of CD44 and CD34 in all three donors (Online Supplementary Figure S3C).

Acute myeloid leukemia three-dimensional tri-cultures demonstrate increased resistance to daunorubicin and cytarabine

The AML γ0.63 tri-cultures and γ1 mono-cultures were treated with seven different doses of daunorubicin and cytosine β-D-arabinofuranoside (cytarabine, AraC) and compared with 2D mono-cultures for 24 h. Compared with the 3D mono-cultures and tri-cultures, AML cells cultured in 2D were most sensitive to the drug doses applied (Figure 4B). Significant differences were found between KG1a 2D and 3D tri-culture cell viability after treatment with 0.20 μM daunorubicin (Figure 4B), and at all cytarabine concentrations tested except for 1,000 μM (Online Supplementary Figure S4B). MOLM13 tri-cultures were significantly more resistant than 2D cultures when treated with 0.10, 0.20 or 1.00 μM of daunorubicin (Figure 4B), and at most cytarabine concentrations (Online Supplementary Figure S4B). The number of viable cells was significantly higher in MV4-11 3D tri-cultures than 2D mono-cultures when treated with 0.01, 0.10, 0.20, 1.00, or 5.00 μM daunorubicin (Figure 4B), and at all cytarabine doses tested (Online Supplementary Figure S4B). Compared with 2D mono-cultures, OCI-AML3 3D tri-cultures displayed higher numbers of viable cells at 0.10 and 0.20 μM daunorubicin treatment (Figure 4B), and significantly increased resistance at all cytarabine concentrations tested

Figure 2. OCI-AML3 co-culture with either human umbilical vein endothelial cells or mesenchymal stem cells. Confocal images of the OCI-AML3 cell line (as a representative of all cell lines utilized in this study) in co-culture with MSC (left) or HUVEC (right) after 7 days of culture. Images display CD45+ leukemia cell growth along both network types. Scale bar = 100 μm.

1218

haematologica | 2017; 102(7)

An ex vivo model of acute myeloid leukemia

Figure 3. Three-dimensional mono-cultures for the culture of acute myeloid leukemia spheroids. Growth of (A) KG1a, (B) MOLM13, (C) MV4-11 and (D) OCI-AML3 cell lines was investigated in 3D. AML cells were cultured in starPEG-heparin hydrogels (upper left image) or in MatrigelTM (lower left image). Comparison of cell growth (upper right graph) and spheroid size (lower right graph) over 14 days was performed using three different cross-linking degrees of hydrogel (γ0.75, γ1, γ1.5). (E) Primary donor cells from two patients with AML were cultured in starPEG-heparin hydrogels as a mono-culture. Cell line experiments were performed at least twice with three technical replicates (n=2). Data are displayed as mean ± SD. pAML experiments were performed once in triplicate (n=1). The graph displays only the means. Cultures displayed heterogeneous distribution of cells, with each donor’s cell growing both as spheroids and single cells. Three photographs per technical replicate were utilized for ImageJ spheroid measurements as a minimum. *Indicates statistical significance: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

haematologica | 2017; 102(7)

1219

L.J. Bray et al.

Figure 4. Treatment of three-dimensional mono-cultures and tri-cultures with daunorubicin. (A) Light microscope images of AML cell lines untreated and treated with 5 μM (daunorubicin) DNR at day 7 after treatment. (B) AML cells lines were treated with daunorubicin for 24 h as 2D or 3D mono-cultured, or 3D tri-cultures. Graphs display mean ± SEM. (C) Mean spheroid diameter of 3D mono-cultures after treatment with daunorubicin (± SD). (D) pAML cells from a patient with AML were cultured as 3D mono-cultures (left confocal images) or 3D tri-cultures (right confocal images). Confocal cross-section (bottom confocal image) shows homogeneous distribution of AML mono-culture throughout hydrogel. Mono-cultures (right graph) and tri-cultures (left graph) were treated with two concentrations of daunorubicin. Graphs show only the means. Cell line experiments were performed at least three times in triplicate (n=3-4). pAML experiments were performed once in triplicate (n=1). Three photographs per technical replicate were utilized for ImageJ spheroid measurements as a minimum. Insert in (D) shows daunorubicin uptake into a pAML spheroid (lower right). *Indicates statistical significance between 2D cultures and 3D tri-cultures: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

1220

haematologica | 2017; 102(7)

An ex vivo model of acute myeloid leukemia

(Online Supplementary Figure S4B). AML spheroid size in mono-cultures was inversely correlated with daunorubicin concentration (Figure 4A,C). Generally, EC50 values for tri-cultures were increased in all cell lines when compared with 2D and 3D mono-cultures (Table 1). While pAML cells from donors 1 and 2 only demonstrated reduced viability at 5.00 μM daunorubicin, pAML cells from donor 3 had increased susceptibility to daunorubicin treatment with reduced viability at both 0.20 and 5.00 μM (Figure 4D). Cells from donor 1 demonstrated resistance to cytarabine therapy, while those from donor 2 showed a trend towards decreased viability upon treatment (Online Supplementary Figure S4D). The binding constants (Kb) for daunorubicin and cytarabine to the hydrogels reflect the hydrophilic character and number of positively charged groups of the molecules (Table 2). Cytarabine has a higher Kb than daunorubicin, indicating that cytarabine has a slightly higher affinity to the hydrogels (stronger electrostatic interaction with heparin) than daunorubicin. The lower the affinity of a drug to the hydrogels, the easier it is for the drug to diffuse through the hydrogels (Online Supplementary Figure S7).

The CXCR4 antagonist, AMD3100, induced mobilization of acute myeloid leukemia cells from the vascular network but failed to increase daunorubicin efficiency in three-dimensional tri-cultures The CXCR4/CXCR12 axis is thought to be the mechanism through which leukemia cells are protected from chemotherapeutics by the stromal microenvironment.25-27 Therefore, antagonists targeting this mechanism (e.g. AMD3100) have been developed for use in combination with chemotherapy such as cytarabine or daunorubicin.28,29 Sorafenib, a multi-tyrosine kinase inhibitor, has also been shown to have antileukemic efficacy when added to standard chemotherapy.30 The application of AMD3100 significantly reduced AML-vascular contact in KG1a, MOLM13 and MV4-11 cultures (Figure 5A,B). When OCI-AML3 cultures were treated with 2.5 μg/mL AMD3100, a contrasting effect occurred whereby AMLvascular contact increased significantly upon treatment (Figure 5B). Similar results were found for the pAML cultures (Figure 5C). Donors 1 and 3 both displayed decreased AML-vascular contact upon application of AMD3100, while donor 2 showed increased contact (Figure 5C,D). Numbers of viable cells remained unchanged in all three donors after AMD3100 treatment (Figure 5D). Using 3D analysis, no changes in vascular network volume were visualized (Online Supplementary Figure S8). However using 2D maximum projection analysis, increased percentage vessel area was found in KG1a and OCI-AML and increased total vessel length was found in OCI-AML3 after AMD3100 treatment (Online Supplementary Figure S8). In the pAML cultures, 3D analysis showed increased network volume in donors 1 and 2 and decreased network volume in donor 3 (Online Supplementary Figure S9). Using 2D analysis it was seen that donors 1 and 3, but not donor 2, had decreased vessel area and length upon treatment with AMD3100 (Online Supplementary Figure S9). CXCR4 expression was similar between MOLM-13, MV4-11 and OCI-AML3 cell lines, whereas some heterogeneity was seen between pAML donors 1 and 2 (Online Supplementary Figure S10). When AMD3100 or sorafenib was utilized as a pre-treatment for daunorubicin, no significant increase in daunorubicin toxhaematologica | 2017; 102(7)

icity was quantified for all cell lines tested (Online Supplementary Figure S5A,B). The binding constants for AMD3100 and sorafenib to the starPEG-heparin hydrogels are presented in Table 2. AMD3100 has a higher Kb than sorafenib, thus indicating that AMD3100 has a higher affinity to the hydrogels whereas the more hydrophobic sorafenib has a low affinity to the hydrogels (Online Supplementary Figure S7).

The combination of daunorubicin and cytarabine treatment resulted in complete ablation of acute myeloid leukemia three-dimensional tri-cultures To mimic the clinical scenario as closely as possible, the approximate values for standard induction therapy, with a combination of cytarabine and daunorubicin, were calculated for traditional human and murine treatments and applied to the 3D cultures based upon the dimensions of the hydrogels (Figure 6A). Five days after treatment, the percentage of live cells was approximately 10% and below (Figure 6B). Tri-cultures were maintained in normal medium for 14 days after treatment to examine any relapse or regrowth of AML cells. However, at day 14 after treatment, cell activity was completely obliterated, or in any case under the detection limit for the cell viability assay. Similar results were observed in the pAML cultures, with all three donors showing few or no metabolically active cells 5 days after treatment (Figure 6C).

Discussion Vascular endothelial cells are a critical component of the hematopoietic microenvironment that regulates blood cell

Table 1. EC50 half-maximal response values for daunorubicin in KG1a, MOLM13, MV4-11 and OCI-AML3 models.

Model

Daunorubicin EC50 (μM)

KG1a 2D KG1a 3D KG1a Tri-culture MOLM13 2D MOLM13 3D MOLM13 Tri-culture MV4-11 2D MV4-11 3D MV4-11 Tri-culture OCI-AML3 2D OCI-AML3 3D OCI-AML3 Tri-culture

0.105 0.092 0.239 0.020 0.038 0.303 0.006 0.012 0.178 0.057 0.132 0.312

Table 2. Observed binding constants of drugs to starPEG-heparin hydrogels.

Drug

Observed binding constant (Kb)

Daunorubicin Cytosine β-D-arabinofuranoside (cytarabine) AMD3100 Sorafenib

0.87 ± 0.07 M-1 3 ± 0.3 M-1 12.73 ± 0.17 M-1 0.79 ± 0.08 M-1 1221

L.J. Bray et al.

production. The stromal microenvironment promotes malignant progression through signaling factors, cellmatrix and cell-cell interactions or alterations in the surrounding matrix.31 Although the role of the stromal microenvironment in solid tumor development and progression is well established,32-34 the role of the stromal niche in AML progression is relatively unknown. Cell-cell interactions between the vascular niche and malignant cell types may play an important role in the pathophysiology of AML. Higher microvascular density is known to occur in the bone marrow microenvironment of patients with

hematologic malignancies,12,35 and may correlate with resistance to therapy.36 Signaling cross-talk between endothelial and AML cells has been demonstrated in in vitro cultures.37-39 Threfore, AML-vascular niche interactions in AML progression define a pressing area of research.9 The growth and capillary network formation of HUVEC within the applied platform of biohybrid glycosaminoglycan hydrogels has previously been demonstrated.20 Typically, hematopoietic stem cells are found in the vicinity of vascular endothelium.40 In fact, endothelial cells are

1222

Figure 5. Treatment of three-dimensional acute myeloid leukemia tricultures with the CXCR4 inhibitor, AMD3100. (A) Confocal images of AML cell lines untreated (upper row) or treated (lower row) with 2.5 µg/mL AMD3100. (B) Percentage of AML cell-cell contact with HUVEC and MSC compared with respective untreated control samples (left) and viability of AML tri-cultures after treatment with AMD3100 (right). (C) pAML from a patient with AML untreated (left) or treated (right) with 2.5 μg/mL AMD3100. Data displayed as mean ± SD. (D) Percentage of pAML contact with HUVEC and MSC after AMD3100 treatment compared with the untreated control sample (left) and viability of AML tri-cultures after treatment with AMD3100 (right). Data are displayed as mean ±SD (variability within experiment, n=1; left), and mean only (right). Cell line experiments were performed at least three times in triplicate (n=34). pAML experiments were performed once in triplicate (n=1). Three photographs per technical replicate were utilized for ImageJ AML-vascular contact measurements as a minimum. *Indicates statistical significance between treated samples and respective control: *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

haematologica | 2017; 102(7)

An ex vivo model of acute myeloid leukemia

haematologica | 2017; 102(7)

Figure 6. “Induction” treatment and “relapse” of acute myeloid leukemia tri-cultures. (A) Light microscope images of AML cell lines treated with human (hu) equivalent induction dose (left), murine (ms) equivalent induction dose (center) or daunorubicin (DNR) only (right). Doses were based upon standard clinical treatments (hu) or standard murine model chemotherapy regimes (ms) for AML. (B) Percentage of AML live cells after induction treatment and 14 days after treatment (relapse; R) compared with the respective untreated control sample. (C) Primary donor cells (pAML) from a patient with AML treated with hu induction (left). Percentage of pAML cell viability compared with that of the untreated control sample (right). Experiments were performed once in triplicate (n=1).

1223

L.J. Bray et al.

reported to increase the proliferation of AML cells when in co-culture.41 AML cells have a reduced cell cycle activity once attached to endothelial cells and are protected from standard chemotherapy. In this context, our results demonstrate the preference for AML cells to adhere to and proliferate on the endothelial network. While pAML cells did not proliferate as fast as the cell lines, they remained in close proximity to the endothelial network. Future studies may explore the role of specific integrins and adhesion molecules in the attachment of AML cells to the vascular endothelium and their localization during 3D mono- and tri-cultures, and integrin inhibition may reveal specific cascades promoting AML-vascular interactions. In this study, four different AML cell lines were tested to represent various subtypes of AML:42 KG1a, an M0 classified immature AML cell line, MOLM13, an M5a classified AML cell line known as acute monocytic leukemia, MV4-11, an M5 classified cell line derived from biphenotypic B-myelomonocytic leukemia, and OCIAML3, an M4 classified AML known as acute myelomonocytic leukemia. In this study, pAML cells derived from three different patients were utilized for the various experiments (Online Supplementary Table S2). Donor 1 was a 77-year old male with M2 classified AML (with myelodysplasia-related changes) whose cells were harvested from the bone marrow at initial diagnosis. The AML was negative for FLT3-ITD. Donor 2 was a 65-year old male with M1 classified AML (without maturation) whose cells were harvested from the peripheral blood at initial diagnosis. His AML was also negative for FLT3ITD. Donor 3 was a 75-year old male with M1 classified AML (with myelodysplasia-related changes): his cells were harvested from peripheral blood during a progressive/refractory stage of disease. Molecular genetics analysis was not available for this donor due to a lack of molecular testing. With regards to the growth of the four AML cell lines tested, it was seen that the cell viability was reduced in 3D cultures when compared with 2D cultures. AML cells grown within MatrigelTM developed loose clumps of cells, whereas in hydrogels the cells formed tight spheroids. When pAML cells were cultured within our customized biohybrid hydrogels, a heterogeneous population developed, which is typical of an unsorted peripheral blood sample. These results underline the softer mechanical properties and inherent growth factor presence within MatrigelTM in comparison with our glycosaminoglycan-based hydrogels. When cultured to specific time points, the established hydrogel mono-cultures effectively allow the production of AML spheroids of different sizes for subsequent studies. In our culture model, the AML tri-culture displayed increased resistance at some daunorubicin concentrations using MOLM-13 and MV4-11 cell lines when compared with 2D and 3D mono-cultures, and to a lesser extent in OCI-AML3 cell lines. However, this effect was not visualized in KG1a cultures. Moreover, when cytarabine was applied, our 3D model did not show decreased cell viability, although an increase in treatment concentration or duration of treatment may expose different effects. These findings are similar to those previously reported by other research groups, who suggested that AML cells that are adherent to endothelial cells become more resistant to cytarabine.17,43 Moreover, cell adhesion-mediated drug resistance to HUVEC and MSC, regulated by integrins, can lead to a decrease in response to chemotherapy.44 1224

Another important method for the clinical treatment of AML is the use of CXCR4 antagonists to block the CXCR4/CXCL12 axis and prevent AML cells from niche protection, and thus, in theory, make combination therapies more effective. In our hydrogel-based 3D tri-culture model, a significant increase in AML mobilization was observed when AMD3100 was applied to the cultures, except in the case of OCI-AML3. It could be that AMD3100 affects this subtype of AML in an opposing manner, and this method of treatment would not, therefore, be recommended clinically, although, given the high expression of CXCR4 on the surface of OCI-AML3 cells, also associated with poor prognosis,45,46 it could be that a more potent antagonist is required.47 It is important to note that differences in HUVEC/MSC network density were sometimes visualized between leukemia cell types in these experiments. It may, therefore, be possible that AMD3100 is in fact altering the network structures, thus making it more difficult for leukemia cells to come into contact with the network, rather than inducing mobilization of leukemia cells. More in-depth 3D quantification is needed to analyze microenvironmental changes in the future and further studies are required to determine the mechanistic effects of AMD3100 on different AML subtypes. Other than the application of AMD3100, there are several clinically relevant aspects of our study. Induction treatment is given to AML patients as a ‘first-line’ therapy. Usually, induction therapy results in ablation of the bulk of malignant cells in order to remove the leukemia from the patients’ hematopoietic system. To eradicate the residual leukemic cells, further consolidation therapy is applied, ideally to extinguish even minimal residual disease and to achieve a cure. Our data here show the effectiveness of equivalent induction therapy doses on our 3D microenvironment model, and complete ablation of leukemia cell lines without re-growth after 2 weeks. Future experiments should explore the development of long-term cultures (3-6 months) and the effects of induction therapy, relapse and consolidation therapy. As the minimal residual disease status after chemotherapy is of major prognostic significance,48 the biology and dynamics of residual leukemia cells after therapy could be an interesting application of the novel tri-culture model. Using specimens from AML patients as well as cell lines, we provide experimental evidence supporting the clinical relevance of our results. We have demonstrated not only that AML interactions with vascular endothelial cells are modulated by the CXCR4/CXCL12 axis within our engineered microenvironments, but also that not all subtypes of AML respond in the same way. We must also state the limitations of the study here, in which pAML experiments were performed with three patients’ samples in parallel triplicates; the conclusions drawn from our studies should, therefore, be considered exploratory. Although detailed risk classifications, based mainly on genetic abnormalities, have been shown to project the outcome of patient cohorts accurately, the outcome of individual patients still remains quite unpredictable. Additional information, for example on responsiveness towards a treatment approach, would therefore provide a new perspective to refine outcome predictions. Standard treatments are recommended for AML patients because of known improved outcomes, however with the absence of a personalized testing system, patients who do not respond to standard treatments haematologica | 2017; 102(7)

An ex vivo model of acute myeloid leukemia

are left without targeted disease management. Hence, our findings provide a new perspective on the impact of drug treatments, both standard and those in trials, on cell-cell interactions in a 3D microenvironment. In summary, our AML tri-culture model enables extensive analysis of leukemia-vascular interactions. Relying on a thoroughly tunable hydrogel matrix platform, AML cells have been, for the first time, successfully co-cultured with both HUVEC and MSC. Importantly, the approach offers exciting, unprecedented options for the visualization of AML-vascular interactions after the application of anticancer compounds, which is particularly important for investigating the impact of chemotherapeutics on postapplication cellular responses. The response of our model to the administration of anti-CXCR4 demonstrated mobilization of leukemia cells from the vascular niche, supporting the in vivo-like behavior of the organotypic scaffold. Moreover, our 3D AML co-cultures showed a significantly increased resistance to chemotherapeutics compared with 2D cultures, confirming the greater clinical relevance of the hydrogel-based model we presented. We anticipate

References 1. Appelbaum FR, Rowe JM, Radich J, Dick JE. Acute myeloid leukemia. Hematology Am Soc Hematol Educ Program. 2001;62-86. 2. NCI Network. Trends in incidence and outcome for haematological cancers in England: 2001-2010. London: Public Health England, 2014. 3. Visser O, Trama A, Maynadie M, et al. Incidence, survival and prevalence of myeloid malignancies in Europe. Eur J Cancer. 2012;48(17):3257-3266. 4. Pickup MW, Mouw JK, Weaver VM. The extracellular matrix modulates the hallmarks of cancer. EMBO Rep. 2014;15(12):1243-1253. 5. Radisky D, Muschler J, Bissell MJ. Order and disorder: the role of extracellular matrix in epithelial cancer. Cancer Invest. 2002;20(1):139-153. 6. Aljitawi OS, Li D, Xiao Y, et al. A novel three-dimensional stromal-based model for in vitro chemotherapy sensitivity testing of leukemia cells. Leuk Lymphoma. 2014;55(2):378-391. 7. Blanco TM, Mantalaris A, Bismarck A, Panoskaltsis N. The development of a three-dimensional scaffold for ex vivo biomimicry of human acute myeloid leukaemia. Biomaterials. 2010;31(8):22432251. 8. Velliou EG, Dos Santos SB, Papathanasiou MM, et al. Towards unravelling the kinetics of an acute myeloid leukaemia model system under oxidative and starvation stress: a comparison between two- and threedimensional cultures. Bioprocess Biosyst Eng. 2015;38(8):1589-1600. 9. Trujillo A, McGee C, Cogle CR. Angiogenesis in acute myeloid leukemia and opportunities for novel therapies. J Oncol. 2012;2012:128608. 10. Kopp HG, Avecilla ST, Hooper AT, Rafii S. The bone marrow vascular niche: home of HSC differentiation and mobilization. Physiology (Bethesda). 2005;20:349-356.

haematologica | 2017; 102(7)

that the approach will become instrumental in future extensive chemotherapeutic-signaling inhibitor combination treatment analyses as well as in the definition of individualized treatment modalities for AML patients. Beyond that, identification of the signaling mechanisms of endothelial cells within the vascular niche which encourage AML growth or provide resistance to conventional chemotherapy might provide new insights for the development of novel therapeutic strategies for AML patients. Acknowledgments LJB was supported by the Endeavour Awards as part of the Prime Ministerâ&#x20AC;&#x2122;s Australia Awards. Financial support was provided by the German Research Foundation (Deutsche Forschungsgemeinschaft) through grant numbers: SFB-TR 67, WE 2539-7 and FOR/EXC999, by the Leibniz Association (SAW2011-IPF-2 68) and by the European Union through the Integrated Project ANGIOSCAFF (Seventh Framework Program). The authors gratefully thank Dr Manja Wobus, Dr Mikhail Tsurkan, Mrs Juliane Drichel, Mrs. Milauscha Grimmer and Mr Christoph Hentschel for their advice and assistance.

11. Dias S, Hattori K, Heissig B, et al. Inhibition of both paracrine and autocrine VEGF/ VEGFR-2 signaling pathways is essential to induce long-term remission of xenotransplanted human leukemias. Proc Natl Acad Sci USA. 2001;98(19):10857-10862. 12. Hussong JW, Rodgers GM, Shami PJ. Evidence of increased angiogenesis in patients with acute myeloid leukemia. Blood. 2000;95(1):309-313. 13. Padro T, Ruiz S, Bieker R, et al. Increased angiogenesis in the bone marrow of patients with acute myeloid leukemia. Blood. 2000;95(8):2637-2644. 14. Butler JM, Kobayashi H, Rafii S. Instructive role of the vascular niche in promoting tumour growth and tissue repair by angiocrine factors. Nature Rev Cancer. 2010;10(2):138-146. 15. Doan PL, Chute JP. The vascular niche: home for normal and malignant hematopoietic stem cells. Leukemia. 2012;26(1):54-62. 16. Ribatti D. Bone marrow vascular niche and the control of tumor growth in hematological malignancies. Leukemia. 2010;24(7): 1247-1248. 17. Pezeshkian B, Donnelly C, Tamburo K, Geddes T, Madlambayan GJ. Leukemia mediated endothelial cell activation modulates leukemia cell susceptibility to chemotherapy through a positive feedback loop mechanism. PLoS One. 2013;8(4): e60823. 18. Freudenberg U, Hermann A, Welzel PB, et al. A star-PEG-heparin hydrogel platform to aid cell replacement therapies for neurodegenerative diseases. Biomaterials. 2009;30(28):5049-5060. 19. Tsurkan MV, Chwalek K, Prokoph S, et al. Defined polymer-peptide conjugates to form cell-instructive starPEG-heparin matrices in situ. Adv Mater. 2013;25(18): 2606-2610. 20. Chwalek K, Tsurkan MV, Freudenberg U, Werner C. Glycosaminoglycan-based hydrogels to modulate heterocellular communication in in vitro angiogenesis models.

Sci Rep. 2014;4:4414. 21. von Bonin M, Wermke M, Cosgun KN, et al. In vivo expansion of co-transplanted T cells impacts on tumor re-initiating activity of human acute myeloid leukemia in NSG mice. PLoS One. 2013;8(4):e60680. 22. Chwalek K, Levental KR, Tsurkan MV, Zieris A, Freudenberg U, Werner C. Twotier hydrogel degradation to boost endothelial cell morphogenesis. Biomaterials. 2011;32(36):9649-9657. 23. Oswald J, Boxberger S, Jorgensen B, et al. Mesenchymal stem cells can be differentiated into endothelial cells in vitro. Stem Cells. 2004;22(3):377-384. 24. Bray LJ, Binner M, Holzheu A, et al. Multiparametric hydrogels support 3D in vitro bioengineered microenvironment models of tumour angiogenesis. Biomaterials. 2015;53:609-620. 25. Burger JA, Kipps TJ. CXCR4: a key receptor in the crosstalk between tumor cells and their microenvironment. Blood. 2006;107(5):1761-1767. 26. 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. 27. Tavor S, Petit I. Can inhibition of the SDF1/CXCR4 axis eradicate acute leukemia? Semin Cancer Biol. 2010;20(3):178-185. 28. Burger JA, Peled A. CXCR4 antagonists: targeting the microenvironment in leukemia and other cancers. Leukemia. 2009;23(1):43-52. 29. Juarez J, Bradstock KF, Gottlieb DJ, Bendall LJ. Effects of inhibitors of the chemokine receptor CXCR4 on acute lymphoblastic leukemia cells in vitro. Leukemia. 2003;17(7):1294-1300. 30. Rollig C, Serve H, Huttmann A, et al. Addition of sorafenib versus placebo to standard therapy in patients aged 60 years or younger with newly diagnosed acute myeloid leukaemia (SORAML): a multicentre, phase 2, randomised controlled trial.

1225

L.J. Bray et al. Lancet Oncol. 2015;16(16):1691-1699. 31. Micke P, Ostman A. Tumour-stroma interaction: cancer-associated fibroblasts as novel targets in anti-cancer therapy? Lung Cancer. 2004;45 (Suppl 2):S163-175. 32. Dvorak HF, Weaver VM, Tlsty TD, Bergers G. Tumor microenvironment and progression. J Surg Oncol. 2011;103(6):468-474. 33. Hanahan D, Coussens LM. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell. 2012;21(3):309-322. 34. Lathia JD, Heddleston JM, Venere M, Rich JN. Deadly teamwork: neural cancer stem cells and the tumor microenvironment. Cell Stem Cell. 2011;8(5):482-485. 35. Ribatti D. Is angiogenesis essential for the progression of hematological malignancies or is it an epiphenomenon? Leukemia. 2009;23(3):433-434. 36. Ayala F, Dewar R, Kieran M, Kalluri R. Contribution of bone microenvironment to leukemogenesis and leukemia progression. Leukemia. 2009;23(12):2233-2241. 37. Hatfield K, Oyan AM, Ersvaer E, et al. Primary human acute myeloid leukaemia cells increase the proliferation of microvascular endothelial cells through the release of soluble mediators. Br J Haematol.

1226

2009;144(1):53-68. 38. Hatfield KJ, Evensen L, Reikvam H, Lorens JB, Bruserud O. Soluble mediators released by acute myeloid leukemia cells increase capillary-like networks. Eur J Haematol. 2012;89(6):478-490. 39. Liesveld JL, Rosell KE, Lu C, et al. Acute myelogenous leukemia--microenvironment interactions: role of endothelial cells and proteasome inhibition. Hematology. 2005;10(6):483-494. 40. Kiel MJ, Yilmaz OH, Iwashita T, Yilmaz OH, Terhorst C, Morrison SJ. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell. 2005;121(7): 1109-1121. 41. Hatfield K, Ryningen A, Corbascio M, Bruserud O. Microvascular endothelial cells increase proliferation and inhibit apoptosis of native human acute myelogenous leukemia blasts. Int J Cancer. 2006;119 (10):2313-2321. 42. Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the acute leukaemias. French-American-British (FAB) co-operative group. Br J Haematol. 1976;33(4):451-458. 43. Cogle CR, Goldman DC, Madlambayan

44.

45.

46.

47.

48.

GJ, et al. Functional integration of acute myeloid leukemia into the vascular niche. Leukemia. 2014;28(10):1978-1987. Damiano JS, Cress AE, Hazlehurst LA, Shtil AA, Dalton WS. Cell adhesion mediated drug resistance (CAM-DR): role of integrins and resistance to apoptosis in human myeloma cell lines. Blood. 1999;93(5):1658-1667. Konoplev S, Rassidakis GZ, Estey E, et al. Overexpression of CXCR4 predicts adverse overall and event-free survival in patients with unmutated FLT3 acute myeloid leukemia with normal karyotype. Cancer. 2007;109(6):1152-1156. Rombouts EJ, Pavic B, Lowenberg B, Ploemacher RE. Relation between CXCR-4 expression, Flt3 mutations, and unfavorable prognosis of adult acute myeloid leukemia. Blood. 2004;104(2):550-557. Cho BS, Zeng Z, Mu H, et al. Antileukemia activity of the novel peptidic CXCR4 antagonist LY2510924 as monotherapy and in combination with chemotherapy. Blood. 2015;126(2):222-232. Shayegi N, Kramer M, Bornhauser M, et al. The level of residual disease based on mutant NPM1 is an independent prognostic factor for relapse and survival in AML. Blood. 2013;122(1):83-92.

haematologica | 2017; 102(7)

ARTICLE

Acute Myeloid Leukemia

Precision and prognostic value of clone-specific minimal residual disease in acute myeloid leukemia Pierre Hirsch,1,2,3 Ruoping Tang,1,2,3 Nassera Abermil,4 Pascale Flandrin,1,2,5 Hannah Moatti,1,2 Fabrizia Favale,1,2,4 Ludovic Suner,1,2,4 Florence Lorre,6 Christophe Marzac,4 Fanny Fava,1,2,3 Anne-Claire Mamez,3 Simona Lapusan,3 Françoise Isnard,3 Mohamad Mohty,1,3 Ollivier Legrand,1,2,3 Luc Douay,1,2,4 Chrystele Bilhou-Nabera1,2,4 and François Delhommeau1,2,4

Sorbonne Universités, UPMC Univ Paris 06, INSERM, APHP Hôpital Saint-Antoine, Centre de Recherche Saint-Antoine (CRSA), Paris; 2Sorbonne Universités, UPMC Univ Paris 06, GRC n°7, Groupe de Recherche Clinique sur les Myéloproliférations Aiguës et Chroniques MYPAC, Paris; 3AP-HP, Hôpital Saint-Antoine, Service d’Hématologie Clinique et de Thérapie Cellulaire, Paris; 4AP-HP, Hôpital Saint-Antoine, Service d’Hématologie Biologique, Paris; 5Université de Saint Etienne, Laboratoire d'Hématologie, CHU de Saint-Etienne and 6AP-HP, Hôpital Saint-Antoine, Laboratoire Commun de Biologie et Génétique Moléculaires, Paris, France 1

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1227-1237

ABSTRACT

he genetic landscape of adult acute myeloid leukemias (AML) has been recently unraveled. However, due to their genetic heterogeneity, only a handful of markers are currently used for the evaluation of minimal residual disease (MRD). Recent studies using multitarget strategies indicate that detection of residual mutations in less than 5% of cells in complete remission is associated with a better survival. Here, in a series of 69 AMLs with known clonal architecture, we design a clone-specific strategy based on fluorescent in situ hybridization and high-sensitivity next generation sequencing to detect chromosomal aberrations and mutations, respectively, in follow-up samples. The combination of these techniques allows tracking chromosomal and genomic lesions down to 0.5-0.4% of the cell population in remission samples. By testing all lesions in follow-up samples from 65 of 69 evaluable patients, we find that initiating events often persist and appear to be, on their own, inappropriate markers to predict short-term relapse. In contrast, the persistence of two or more lesions in more than 0.4% of the cells from remission samples is strongly associated with lower leukemia-free and overall survivals in univariate and multivariate analyses. Although larger prospective studies are needed to extend these results, our data show that a personalized, clone-specific, MRD followup strategy is feasible in the vast majority of AML cases. Introduction Acute myeloid leukemias (AMLs) are heterogeneous diseases which occur after accumulation of various chromosomal and genomic lesions in hematopoietic stem or progenitor cells.1-5 Some of these events, including DNMT3A, TET2 and ASXL1 mutations, also occur with aging, leading to clonal hematopoiesis of indeterminate potential (CHIP).6-9 High throughput and single cell derived analyses of AML clonal architecture have shown that CHIP lesions are founding events in the leukemic clone.2-5 Moreover, some mutations, such as those affecting DNMT3A or TET2, can still be detected after treatment, whereas others, such as NPM1 mutations, are mostly not detected.2,3,10 This suggests that most relapses emerge from a resistant pre-leukemic clone that behaves as a disease reservoir. Despite this greater understanding of leukemogenesis mechanisms, AML is still associated with poor prognosis.11 The initial level of response to treatment has been haematologica | 2017; 102(7)

Correspondence: francois.delhommeau@aphp.fr

Received: November 9, 2016. Accepted: March 13, 2017. Pre-published: March 16, 2017. doi:10.3324/haematol.2016.159681 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/6/xxx ©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.

1227

P. Hirsch et al. Table 1. Comparison of patients' characteristics according to clonal clearance.

All patients‡ (n=68) Variable

Sex M/F: n (%) 37 (55) / 31 (45) Age (years): median (range) 58 (18-84) FAB: n (%) M0-M2 39 (57) M4-M5 26 (38) M6-M7 3 (4) Cytogenetic group* Good: n (%) 7 (11) Intermediate: n (%) 43 (65) Poor: n (%) 16 (24) FLT3-ITD : n (%) 15 (22) NPM1 mut: n (%) 23 (34) Initial WBC (x109/L): 14.2 (0.5-350) Median (range) Day of CR1 evaluation: 47 (28-194) median (range)

Analysis of all target clearance (n=59) Clearance No clearance P (VCF<3.33%) (VCF≥3.33%) (n=33) (n=26) 18 (55) / 15 (45) 53 (18-84)

15 (58) / 11 (42) 63 (37-80)

18 (55) 15 (45) 0 (0)

13 (50) 11 (42) 2 (8)

NS 0.0176 NS

Analysis of persisting markers in CR (n=58) 0 or 1 marker Two markers P (n=31) or more (n=27) 19 (61) / 12 (39) 13 (48) / 14 (52) 54 (19-84) 62 (18-80) 20 (65) 11 (35) 0 (0)

11 (41) 15 (55) 1 (4)

0.0532 5 (16) 18 (56) 9 (28) 7 (21) 11 (33) 12.2 (0.5-350)

0 (0) 22 (88) 3 (12) 7 (26) 10 (38) 15.1 (0.82-117)

45 (29-194)

47 (28-139)

NS NS NS

5 (16) 19 (63) 6 (20) 8 (25) 11 (35) 14.6 (0.5-350)

0 (0) 21 (81) 5 (19) 6 (22) 10 (37) 13.9 (0.82-117)

NS NS NS

47 (29-194)

43 (28-139)

One patient had no target (normal karyotype and no mutation) and was excluded from the analyses. *Not available (NA) in 2 patients (failure). CR: complete remission; VCF: variant cell fractions; M: male; F: female; NS: not significant; mutated; CR1: first complete remission.FAB: French American and British classification; WBC: white blood cell count; mut: mutated. ‡

identified as a major prognostic factor in adults. Many tools have been developed to evaluate the minimal residual disease (MRD) in complete remission (CR). Cytogenetic12-14 and flow cytometry15-20 follow-up strategies can provide prognostic information, but their use is limited by either a poor sensitivity or a lack of specific assessable markers. Molecular MRD tools have been well validated in AMLs with recurrent gene fusions, mostly those with core-binding factor translocations18,21-23 or with NPM1 mutations.24-26 These markers cover up to 50% of all AMLs.11 However, with the recent understanding of comprehensive genetic landscapes of AMLs,27,28 nearly all patients could, in theory, be assessed for MRD by specific molecular markers. A few studies have investigated alternative MRD markers. Most of these studies used strategies focused on one or two genes. Among them, RUNX1 mutation evaluation seems to be of particular interest29 due to its strong prognostic value and a mutual exclusion with NPM1 mutated cases. IDH1 and IDH2 mutations and FLT3-ITD detection could also be useful tools,30-32 but these are frequently either associated with NPM1 mutations or lost at relapse.31,33 Finally, DNMT3A mutations seem to be of little interest because they frequently persist at a high level in CR,2,3,10,30 even in long-term responders.34 One recent study analyzed the clearance of all variants found in 50 AMLs.10 Such clearance after one course, defined as variant allele frequency (VAF) below 2.5%, was associated with both better event-free survival and overall survival (OS) probabilities. In the present study, we asked whether an architecture-based clone-specific MRD strategy could provide more powerful prognostic information than evaluating the clearance of all variants. A combination of highly sensitive NGS (HS-NGS) assay and chromosomal analyses was used to monitor MRD. We found only a trend towards association of a residual clonal disease below 3.33% with better leukemia-free survival (LFS). Strikingly, the earliest 1228

event of the clonal architecture was frequently detected, even in long term-responders. In contrast, our data revealed that persistence of the two or three first lesions of the clone was strongly associated with a poor prognosis, and was predictive for relapse at one year.

Methods Patients’ samples Bone marrow (BM) and blood samples were collected from AML patients after informed consent. CD3+ cells were isolated from CR blood samples.3 The study was conducted with the approval of the MyPAC clinical research group according to French law and the Declaration of Helsinki. Forty-five AMLs were prospectively included in the study between 2013 and 2015. Thirteen patients who experienced relapse between 2013 and 2015, and 11 other patients diagnosed before 2013, were retrospectively included (Table 1 and Figure 1). Of these 69 patients, 46 have already been reported.3 Patients received an initial anthracycline- and cytarabine-based induction treatment, and a cytarabinebased post-induction treatment. Twenty-two patients received hematopoietic stem cell transplantation (HSCT) in first CR (Online Supplementary Table S1).

Targeted resequencing Sequencing was performed using a 122 gene panel (HaloPlex Target Enrichment System®, Agilent Technologies) on a MiSeq® sequencer (IlluminaINC). Whole exome sequencing was performed in 4 patients with KMT2A translocations. Detailed protocols have been reported previously3 (Online Supplementary Tables S1 and S2). After alignment, described variants were called using an Ensembl database. Non-described variants of potential interest were also sequenced using the Sanger method in the CD3+ fraction or in the CR samples. haematologica | 2017; 102(7)

Residual clonal architecture in myeloid leukemia

Figure 1. Flow-chart of inclusion in survival analyses.

FLT3-ITD, NPM1 and CEBPA mutation detection Mutation detection was performed with standard routine techniques.3

Cytogenetic and fluorescent in situ hybridization analyses Cytogenetic analysis was performed on R-banding metaphases after 24-hour culture using standard procedures. Chromosomal rearrangements were confirmed by fluorescent in situ hybridization (FISH) analysis on 200 interphase nuclei, as described.3 FISH was performed in CR samples (200-400 nuclei) to evaluate the clearance of chromosomal abnormalities with a theoretical detection threshold of 0.5%.

Variant cell fraction determination Variant cell fractions (VCFs) were determined from frequencies of nuclei harboring chromosomal aberrations, variant allele frequency (VAFs) for single nucleotide variants (SNVs), indels, and single nucleotide polymorphisms (SNPs) in sequenced regions with loss of heterozygosity, peak height ratios of high-resolution sizing of fluorescent dye-labeled PCR amplification for FLT3-ITDs. VCF and VCF confidence intervals were used to determine the clonal architecture, as already described.3

High-sensitivity targeted sequencing Four different targeted-resequencing panels using HaloPlex HS Target Enrichment System® (Agilent Technologies) were designed, each covering lesions from 10-25 distinct patients. CR samples from each patient were analyzed with appropriate panels as described.3 With this assay, single DNA fragments are tagged with unique random 10-nucleotide indexes before the first PCR haematologica | 2017; 102(7)

amplification step of library preparation. After the PCR step, each tagged DNA fragment is amplified, generating an amplicon family. When libraries are sequenced, raw reads are generated from amplicons, and a family read corresponds to the group of reads harboring the same random index generated from a unique amplicon family. This improves discrimination of variant nucleotides from background sequencing errors, and allows a more confident detection of low frequency variants.

Digital droplet PCR Digital droplet PCR (ddPCR) experiments were performed using a ddPCR droplet generator system (Biorad), according to the manufacturer’s protocol, and 40 PCR cycles (Biorad iCycler). Droplet reading was performed with QX200 droplet reader (Biorad). Quantasoft software v.1.7.4 (Biorad) was used for result interpretation. Primers and probes are described in Online Supplementary Table S3.

Statistical analyses Associations between patients’ characteristics were analyzed using Fisher, Mann-Whitney or χ2 tests. The Spearman test was used to assess the correlation between ddPCR and HS-NGS results. Standard definitions were used for LFS, OS and CR. Probabilities of survival were calculated using the Kaplan-Meier method. Differences between survival distributions were evaluated by the log-rank test. Cox models were constructed for multivariate analyses, including all variables of interest. Survival analyses were performed with and without data censoring at allo-HSCT. The median follow-up time for surviving patients was 24.2 months. P<0.05 was considered significant. We used StatView software (v.5.0) for statistical analyses (SAS Institute Inc., San Diego, CA, USA). 1229

P. Hirsch et al.

Results Definition of MRD target panels using AML clonal architecture at diagnosis To set up a multitarget MRD follow-up strategy, we first established the clonal composition of AMLs by combining cytogenetic, standard molecular and NGS data, which led to the detection of 63 chromosomal abnormalities and 235 mutations. In total, a median of 4 (range 0-10) chromosomal or genetic events were identified per patient, with a median of 3 (0-10) mutations (Online Supplementary Tables S1 and S2). The clonal architecture inferred from VCFs (Figure 2A) indicated that most patients had a mixture of founding lesions (i.e. VCFs >0.5) and subclonal lesions (i.e.

VCFs <0.5) (Figure 2B). Overall, in 1 of 69 patients no target was found (normal karyotype and no mutation). In 3 additional cases, no material was available at the time of first CR (CR1). These 4 patients were excluded from further analyses.

Detection of low variant allele fractions with the high-sensitivity NGS assay To determine the threshold of detection of the high-sensitivity NGS (HS-NGS) assay, we performed serial dilutions (1/20, 1/100, 1/400) of 4 samples with 29 known variants (VAFs ranging from 1.6% to 48.1%) in non-mutated control genomic DNA. With the HS-NGS assay, confident detection of variants relies on the presence of mutant family

1230

Figure 2. Clonal composition of acute myeloid leukemias (AMLs) at diagnosis. (A) Distribution of variant cell fractions (VCFs) of individual diagnosis samples: colored dots indicating the fraction of cells bearing genetic or chromosomal lesion are distributed along one line per patient. Error bars represent VCF confidence intervals. Distinct categories of clonal composition are listed in the left legend and indicated by individual colored arrows. (B) VCFs of CHIP mutations, signaling lesions, and all other lesions. Error bars indicate VCF confidence intervals. Vertical bars indicate median VCFs.

haematologica | 2017; 102(7)

Residual clonal architecture in myeloid leukemia

Residual genomic landscape in post-treatment samples

reads with unique molecular indexes. When paired end sequencing is performed, a unique molecular index yields two family reads. We therefore considered as positive specimen samples with more than two mutant and distinct family reads, meaning that more than one variant DNA molecule was tagged by unique indexes in the sample. By this approach, we detected all diluted variants with theoretical VAFs above 0.12%. Moreover, we found that all variants with VAFs measured over 0.18% were supported by 6 family reads or more (Figure 3A). To address the issue of false positive detection, we screened all samples expected to be wild-type for 31 given SNVs. For each variant, we calculated the threshold of detection as the median percentage of positive family reads in expected negative samples + 2 standard deviations (SD) (Figure 3B). For example, 45 expected negative samples were screened for the DNMT3A p.R882H mutation. In these samples, with a median of 5106 (178245974) family reads per sample, variants were detected in 0.06±0.059% of family reads (median±SD). The detection threshold for this variant was consequently calculated at 0.18%. Detection thresholds of 30 other SNV targets ranged from 0 to 0.19%. We next evaluated the threshold of detection of indel variants. We first focused on a CUX1 indel (p.E221fs. chr7:101758539 AG>A). Out of 34 expected negative samples, we only found one mutant read family in one sample. We then analyzed NPM1 type A mutation. In 83 expected negative samples, we detected mutants in 4 cases, with a maximum level of 0.17% of read families. In these 4 cases, bulk AML samples with high NPM1 type A mutation ratios (VAF >20%) were processed in the same library preparation experiment. This suggests inter-sample cross contamination during the process that could probably be avoided. For further analyses, we consequently set a consensus SNV and indel detection threshold at 0.2% VAF (i.e. VCF of 0.4% for heterozygous lesions). In further experiments, all variants with VAF less than 0.2% were considered as not detected. DdPCR experiments were performed to test 17 remission samples for targets detected with VAF less than 2% using HS-NGS. VAFs obtained using ddPCR were correlated to those obtained using HS-NGS (r²=0.92, P=0.001) (Figure 3C and D).

To evaluate the persistence of both chromosomal and genomic lesions at the time of CR1, we performed a combination of karyotypic, FISH, and HS-NGS analyses in the 65 patients with available material and with at least one lesion identified at diagnosis (Figures 4 and 5 and Online Supplementary Table S1). Among the 281 lesions found in these patients, three chromosomal abnormalities were not evaluated due to karyotype failure in one patient (2015035), and five subclonal lesions were not included in the follow-up NGS panels (Figure 4B). Of the 273 remaining lesions, 83 were still detected with VCFs over 2%, which we arbitrarily considered at high level. Those events included mutations in DNMT3A (19 of 21 mutations at diagnosis), TET2 (14 of 23), ASXL1 (4 of 5), EZH2 (3 of 3), IDH1 (4 of 7), TP53 (5 of 6), SRSF2 (4 of 7), and U2AF1 (2 of 3). The earliest event of the clonal architecture (Figure 2) was detected at high level in CR1 samples from 30 of 65 patients. Two lesions were still detectable at high level (>2%) in CR1 in 20 of 65 patients. Detection of one or more events at high level in CR did not correlate with blast count evaluated by morphological examination in CR BM samples. Persistence of a high DNMT3A, U2AF1, TET2, or SRSF2 mutation burden, despite multiple chemotherapy courses, was observed in 8 patients who did not experience relapse (Figure 5B and E). Clearance of these events was observed after BM transplantation in 6 patients (UPN 2013-003, 2014-008, 2014-010, 2014-016, 2014-020, and 2015-036), including one who finally relapsed with a mutational profile similar to that at diagnosis (2014-016) (Figure 5C). One hundred and fifty-eight out of 273 evaluable events were not detected in CR. These cleared events included most mutations in FLT3 (23 of 25), NRAS (11 of 12), KIT (3 of 3), NPM1 (18 of 22), CEBPA (5 of 6), WT1 (6 of 8), IDH2 (2 of 3) and BCOR (4 of 5). Among these mutations, some were missing at relapse, indicating clonal evolution, including mutations in FLT3, NRAS, and NPM1 (Online Supplementary Table S1). In 3 patients, we detected the rise of minor subclones independent of the initial leukemic clones. This was observed after the first induction course in 2 patients and at relapse in the third (Figure 5D). In all 3 patients, the

Table 2. Results from multivariate analyses for leukemia-free survival and overall survival in the 58 evaluable patients with two or more minimal residual disease targets.

Variable Age (continuous) Cytogenetic (poor vs. other) 0-1 marker vs. 2 or more markers in CR1 NPM1 status (WT vs. MUT) FLT3 (no ITD vs. ITD)

LFS 95% CI

All patients (n=58) OS P HR 95% CI

Intermediate cytogenetics (n=40) LFS OS 95% CI P HR 95% CI

0.986 0.951-1.023 0.46 1.553 0.370-6.528 0.54

0.9860.935-1.04 0.60 0.4210.09-1.962 0.27

0.962 0.915-1.012 0.13 NOT INCLUDED

0.936

0.868-1.009 0.084 NOT INCLUDED

0.109 0.031-0.390 0.0006

0.0710.01-0.480 0.006

0.075 0.016-0.342 0.0008

0.028

0.002-0.434 0.010

NOT INCLUDED

1.316 0.424-4.087 0.63

2.088

0.337-12.9 0.42

NOT INCLUDED

0.336 0.094-1.195 0.092

0.118

0.14-0.966 0.046

LFS: leukemia-free survival; OS: overall survival; HR: Hazard Ratio; CI: Confidence Interval; CR1: first complete remission; WT: wild-type; MUT: mutant; ITD: internal tandem duplication.

haematologica | 2017; 102(7)

1231

P. Hirsch et al.

Figure 3. Determination of the threshold of detection of the high sensitivity next-generation sequencing (NGS) assay. (A) Sensitivity assay. Four acute myeloid leukemia (AML) DNA samples were diluted into control DNA (20x, 100x, 400x dilutions). Circles indicate variant allele frequencies (VAFs) determined using the HSNGS assay. Circle size is proportional to the number of reads supporting each variant as indicated in the right key. Arrows indicate variants which were considered as not confidently detected because they were supported by less than 3 family reads. (B) Threshold of detection of two indel variants and 31 single nucleotide variants (SNVs). Histograms represent the median VAF of each variant in multiple expected negative samples. Error bars indicate two standard deviations. The subsequent consensus 0.2% detection threshold of the highly sensitive NGS (HS-NGS) assay is represented by the red dotted line. (C) Representative ddPCR dotplot analyses of three mutations. Positive controls, negative controls, and complete remission samples are shown. Squares indicate the areas of positive droplets (+/+: mutant, Âą: mutant + wild-type, wt: wild-type). VAFs are noted. (D) Analysis of 17 mutations by ddPCR and HS-NGS. Colored circles correspond to the minimal residual disease samples shown in (B).

1232

haematologica | 2017; 102(7)

Residual clonal architecture in myeloid leukemia

clone involved DNMT3A variants (p.R882 and p.R729 mutations) suggesting a selection of a pre-existent CHIP by treatment pressure.

Prognostic value of clonal response in CR We first tried to determine if the response level of all tar-

gets after one course was associated with a prognostic value in patients who reached cytological CR. Nine patients were excluded from these analyses (6 needed two courses or more to reach CR and in 3 no material was available) (Figure 1). The 59 remaining patients were then distributed into two groups. To discriminate good from

Figure 4. Residual genomic landscape of 65 acute myeloid leukemia (AML) patients after treatment. (A) Percentages of lesions above or below the threshold of detection, i.e. 0.4% variant cell fraction (VCF) for high-sensitivity next-generation sequencing (HS-NGS) assay, 0.5% VCF for fluorescent in situ hybridization (FISH) assay, or 5% VCF for karyotype evaluation, in first complete remission (CR1) samples according to the occurrence of further relapse or not. The shaded area represents lesions from patients who went into relapse. (B) Co-mutation table showing the detection, as indicated in the bottom key, of initial lesions at the time of CR1 in the 65 patients. Data from patients who further experienced relapse (left) and data from those who did not (right). Numbers in boxes indicate multiple lesions in a gene or pathway. Letters below the table indicate the method used for the detection of cytogenetic aberrations. K: karyotype; N: not done.

haematologica | 2017; 102(7)

1233

P. Hirsch et al.

poor responders, we used a VCF of 3.33% as threshold. This value corresponds to the median VCF of the first lesion, determined by HS-NGS quantification in the CR1 samples. Patients with VCF less than 3.33% were considered as good responders whereas patients with VCF 3.33% or more were considered as poor responders. The latter were significantly older than good responders (63 vs. 53 years; P=0.01), and tended to more frequently harbor intermediate risk cytogenetics (P=0.0532) (Table 1). There was a trend to lower LFS probability in poor responders (31.7±9.9% vs. 51.7±9.8% at 2 years; P=0.08) with no translation in OS (Figure 6A). The difference in LFS became significant when censoring data after allo-HSCT (20.1+11.3% vs. 63.6+11%; P=0.01) (Online Supplementary Figure S1). When focusing on the 40 patients with intermediate cytogenetics, good and poor responders had similar LFS (Figure 6A) and OS probabilities. Similar results were observed when using 2% and 5% VCF thresholds to separate the two groups (data not shown). The earliest events of the clonal architecture retrieved in poor responder patients (i.e. VCF>3.33%) were mutations in DNMT3A (8 of 10 patients with DNMT3A mutation as earliest event), TET2 (n=8 of 10), components of the splice machinery (n=3 of 3), TP53 (n=2 of 2) or ASXL1 (n=1 of 1).

Conversely, all 10 patients with t(8;21), inv(16) or KMT2A translocations as earliest events were good responders (Figures 4 and 5 and Online Supplementary Table S1). This suggests a frequent resistance of the pre-leukemic clone when the initial lesion is one of the major CHIP lesions.

Prognostic value of the persistence of multiple markers in CR The earliest lesion of the clone was frequently detected in CR samples, which may blur the prognostic value of persisting events. We therefore performed a second analysis to evaluate if the detection of two or more markers had a prognostic impact. The 58 patients with more than one evaluable event and who reached CR in one course were separated into two groups (Table 1): 1) patients with 0 or 1 marker (responders) above the detection threshold after treatment (i.e. VCF ≥0.4%, n=31); and 2) patients with 2 or more detectable lesions (nonresponders, n=27). LFS at two years was 64.9±9.3% in responders and 19.8±8.7% in non-responders (P=0.001). The OS probability at two years was higher in responders (84±6.6% vs. 57.1±10.5%; P=0.023) (Figure 6B). When focusing on the 40 patients with intermediate cytogenetics, non-responders had lower LFS at two years

Figure 5. Multi-target monitoring of minimal residual disease (MRD) in 16 representative acute myeloid leukemia (AML) patients. (A-C) Histograms represent sequential analyses of variant cell fractions (VCF) in patients with early clearance of all targets (A), with one or more targets persisting at a high level in complete remission (CR) and who did not experience relapse (B), with one or more targets persisting at high level in CR and who experienced relapse (C). Red dotted lines indicate the 0.4% threshold of detection of the next-generation sequencing (NGS) assay. Black dotted lines indicate hematopoietic stem cell transplantation (HSCT). D: day of sampling or HSCT; DG: diagnosis sample; RL: relapse sample. (D) Emergence of DNMT3A mutant clones in 3 treated AML patients. Histograms are as in (AC). Shaded areas show the global evolution of the initial AML clone. Arrows indicate the progression of independent DNMT3A mutant clones. (E) Persistence of an isolated major DNMT3A mutant clone in a patient with long-term (46 months) CR.

1234

haematologica | 2017; 102(7)

Residual clonal architecture in myeloid leukemia

(57±11.8% vs. 19.4±10.5; P=0.0048) and a trend to lower OS (85±8% vs. 61±11.9%; P=0.07) (Figure 6B). Similar results were observed when restricting the analyses to the 42 prospectively included patients with two or more MRD targets (LFS at 2 years 73±10% vs. 24±10%, P=0.0026, and OS at 2 years 90.2±6.6% vs. 62.8±11.5%, P=0.036) (data not shown). Results were the same when censoring observations after allo-HSCT (Online Supplementary Figure S1). Multivariate Cox models were constructed including variables of clinical interest and censoring data after alloHSCT. In all analyses, persistence of 2 or more lesions was

an independent variable for LFS and OS (Table 2). Finally, we analyzed the 50 patients with 3 or more lesions. The persistence of 3 or more markers after one course was associated with a very high risk of relapse (LFS 23.5±10.3% vs. 75.8±7.5% at one year, P<0.0001; median LFS at 7 months in the non-responder group and not reached after 2 years in the responder group), and a lower OS probability (45.2±13.5% vs. 84.8±6.2 at 2 years, P=0.026) with similar results when focusing on the 36 analyzable intermediate-risk cytogenetics patients (Figure 6C) and when censoring data after allo-HSCT (Online Supplementary Figure S1).

B Figure 6. Persistence of multiple genetic lesions in first complete remission (CR1) is associated with poor prognosis. (A) Leukemia-free survival and overall survival according to the clonal response. Variant cell fraction (VCF) less than 3.33% indicates good responders (i.e. all lesions were found below the 3.33% VCF value). VCF 3.33% or more indicates poor responders (i.e. at least one lesion was found above the threshold VCF of 3.33%). Data from the 59 evaluable patients (top) and from patients with intermediate-risk cytogenetics (bottom) are shown. (B) Leukemia-free survival and overall survival according to the detection of 0-1 lesions or 2 or more lesions above the 0.4% VCF detection threshold. Data from the 58 evaluable patients (top) and from patients with intermediate-risk cytogenetics (bottom) are shown. (C) Leukemia-free survival according to the detection of 0-2 lesions or 3 or more lesions above the 0.4% VCF detection threshold. Data from the 50 evaluable patients (left) and from patients with intermediate-risk cytogenetics (right) are shown.

haematologica | 2017; 102(7)

1235

P. Hirsch et al.

Discussion In the present work, we used a multi-target strategy to monitor MRD in a series of 69 patients. In most cases, the clonal compositions, as determined by cytogenetic and NGS techniques, allowed personalized follow-up panels to be set up using a combination of FISH and HS-NGS. The persistence of two or more lesions in CR1 is associated with a high risk of relapse in patients with all cytogenetic profiles, and in patients with intermediate-risk cytogenetics. Our results highlight the benefits of a high-sensitivity multi-target evaluation of MRD in AML patients, based on the follow up of the two or three first lesions of the clonal architecture. In our series, the clearance of all genetic events only tended to be associated with LFS, in contrast to a previous retrospective study.10 This difference could be explained in part by patient selection, treatment diversity, and the small numbers of cases included. However, we found similar results when restricting the analysis to prospectively included patients. Moreover, setting the threshold of 5% used in the Klco study10 to discriminate good from poor responders did not alter our conclusions. The difference became significant when censoring data after allo-HSCT, with LFS probabilities lower in poor responders but showing an increase in good responders. This suggests that allo-HSCT could, in part, minimize the prognostic impact of clearing all targets after one course. Patients with favorable cytogenetics were always good responders in our study, and when we focused on patients with an intermediate-risk karyotype, clearance of all events lost all prognostic impact. The analysis of the genetic landscape in CR indicated that this was mainly due to the frequent persistence of pre-leukemic hematopoiesis after treatment (persistence of the main CHIP lesions), which had no prognostic impact at two years of follow up. CHIP seems to be resistant to AML treatment, especially in the case of TET2 or DNMT3A mutations, which account for longterm remissions with persistent clonal hematopoiesis.3,30,34 Furthermore, we identified 3 cases with the emergence of DNMT3A mutant cells distinct from the initial AML clone after treatment pressure, as previously described.35 In several patients, repeated chemotherapy courses did not impact CHIP, and only HSCT led to the exhaustion of the mutant clone. Thus, MRD evaluation focused on CHIP lesions may be of potential interest to monitor post-graft response and the risk of long-term recurrence, as late relapse can occur after the re-evolution of persistent preleukemic clones.3 The persistence of the two or three first events of the clone above the threshold of 0.4% after one course was predictive of poor outcome in our patients. This result had

References 1. Welch JS, Ley TJ, Link DC, et al. The origin and evolution of mutations in acute myeloid leukemia. Cell. 2012;150(2):264-278. 2. Shlush LI, Zandi S, Mitchell A, et al. Identification of pre-leukaemic haematopoietic stem cells in acute leukaemia. Nature. 2014; 506(7488):328-333.

1236

not been reported in previous studies using thresholds around 2%.10 Larger prospective cohorts should, however, be investigated to validate these results. Indeed, in our series, the number of patients included is a limitation to the multivariate analysis conclusions. The impact of treatment strategies according to MRD levels and, in particular, the value of HSCT in good responders, according to the different genotype/karyotype subgroups could also not be properly evaluated. Interestingly, in 5 patients with available material, molecular re-evolution preceded cytological relapse by 1-9 months. This suggests that multi-target MRD monitoring could also be a useful tool for early therapeutic intervention. Our study indicates that a personalized MRD strategy could easily be used in daily practice. The 122 gene panel represents a suitable and affordable diagnostic tool, and could probably be refined by reducing the number of targets to the 50-60 most frequently mutated genes in AML.20,28,29 With the combination of this panel and simple cytogenetic analyses, nearly all AML patients may have two or more evaluable lesions, and could be eligible for MRD monitoring. However, in our study, CBF patients were all good responders. Several studies have shown that CBF translocations are the earliest events in the clone,3,36 and CBF transcripts monitoring has been proven to be associated with prognosis. Consequently, our multi-target strategy appears not to be useful in these patients. In the current study, MRD analysis was performed using patient-specific HS-NGS assays. To reduce the delays involved in MRD analysis, ddPCR could be an alternative strategy. Indeed, personalized ddPCR probes can be designed soon after the diagnostic NGS screening, and can be made available by the end of the first treatment course. As results from ddPCR and HS-NGS assays are correlated, this clone-specific ddPCR-based strategy should be evaluated in large prospective studies. In conclusion, our study shows the prognostic value of a personalized clone-specific MRD evaluation that can be used in most AML patients. Detection of two or more events in more than 0.4% of the cells after one course is strongly associated with lower survival, in particular in patients with intermediate-risk cytogenetics. Forty-five consecutive patients were prospectively investigated, but larger studies are needed to confirm the results and to evaluate whether similar ddPCR and FISH based strategies could be useful to guide treatment decisions. Funding This project was funded by ARC foundation (N_EML20110602421), RĂŠgion Ile-de-France (N_2012-2eml-06-UPMC_12016710), Association Laurette Fugain (N_J15I409 to FD), Institut National du Cancer (PH).

3. Hirsch P, Zhang Y, Tang R, et al. Genetic hierarchy and temporal variegation in the clonal history of acute myeloid leukaemia. Nat Commun. 2016;7:12475. 4. Corces-Zimmerman MR, Hong W-J, Weissman IL, Medeiros BC, Majeti R. Preleukemic mutations in human acute myeloid leukemia affect epigenetic regulators and persist in remission. Proc Natl Acad Sci USA. 2014;111(7):2548-2553.

5. Jan M, Snyder TM, Corces-Zimmerman MR, et al. Clonal evolution of preleukemic hematopoietic stem cells precedes human acute myeloid leukemia. Sci Transl Med. 2012;4(149):149ra118. 6. Genovese G, KĂ¤hler AK, Handsaker RE, et al. Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med. 2014;371(26):2477-2487. 7. Busque L, Patel JP, Figueroa ME, et al.

haematologica | 2017; 102(7)

Residual clonal architecture in myeloid leukemia

10.

11. 12.

13.

14.

15.

16.

17.

Recurrent somatic TET2 mutations in normal elderly individuals with clonal hematopoiesis. Nat Genet. 2012; 44(11): 1179-1181. Jaiswal S, Fontanillas P, Flannick J, et al. Agerelated clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371(26):2488-2498. Steensma DP, Bejar R, Jaiswal S, et al. Clonal hematopoiesis of indeterminate potential and its distinction from myelodysplastic syndromes. Blood. 2015; 126(1):9-16. Klco JM, Miller CA, Griffith M, et al. Association Between Mutation Clearance After Induction Therapy and Outcomes in Acute Myeloid Leukemia. JAMA. 2015; 314(8):811-822. Döhner H, Weisdorf DJ, Bloomfield CD. Acute Myeloid Leukemia. N Engl J Med. 2015;373(12):1136-1152. Hirsch P, Labopin M, Viguié F, et al. Interest of cytogenetic and FISH evaluation for prognosis evaluation in 198 patients with acute myeloid leukemia in first complete remission in a single institution. Leuk Res. 2014;38(8):907-912. Chen Y, Cortes J, Estrov Z, et al. Persistence of cytogenetic abnormalities at complete remission after induction in patients with acute myeloid leukemia: prognostic significance and the potential role of allogeneic stem-cell transplantation. J Clin Oncol. 2011;29(18):2507-2513. Marcucci G, Mrózek K, Ruppert AS, et al. Abnormal cytogenetics at date of morphologic complete remission predicts short overall and disease-free survival, and higher relapse rate in adult acute myeloid leukemia: results from cancer and leukemia group B study 8461. J Clin Oncol. 2004; 22(12):2410-2418. Terwijn M, van Putten WLJ, Kelder A, et al. High prognostic impact of flow cytometric minimal residual disease detection in acute myeloid leukemia: data from the HOVON/SAKK AML 42A study. J Clin Oncol. 2013;31(31):3889-3897. Freeman SD, Virgo P, Couzens S, et al. Prognostic relevance of treatment response measured by flow cytometric residual disease detection in older patients with acute myeloid leukemia. J Clin Oncol. 2013;31(32):4123-4131. Kern W, Voskova D, Schoch C, Hiddemann W, Schnittger S, Haferlach T. Determination of relapse risk based on assessment of mini-

haematologica | 2017; 102(7)

18.

19.

20.

21.

22.

23.

24.

25.

26.

mal residual disease during complete remission by multiparameter flow cytometry in unselected patients with acute myeloid leukemia. Blood. 2004;104(10):3078-3085. Perea G, Lasa A, Aventín A, et al. Prognostic value of minimal residual disease (MRD) in acute myeloid leukemia (AML) with favorable cytogenetics [t(8;21) and inv(16)]. Leukemia. 2006;20(1):87-94. Buccisano F, Maurillo L, Spagnoli A, et al. Cytogenetic and molecular diagnostic characterization combined to postconsolidation minimal residual disease assessment by flow cytometry improves risk stratification in adult acute myeloid leukemia. Blood. 2010;116(13):2295-2303. Walter RB, Gooley TA, Wood BL, et al. Impact of pretransplantation minimal residual disease, as detected by multiparametric flow cytometry, on outcome of myeloablative hematopoietic cell transplantation for acute myeloid leukemia. J Clin Oncol. 2011;29(9):1190-1197. Jourdan E, Boissel N, Chevret S, et al. Prospective evaluation of gene mutations and minimal residual disease in patients with core binding factor acute myeloid leukemia. Blood. 2013;121(12):2213-2223. Yin JAL, O’Brien MA, Hills RK, Daly SB, Wheatley K, Burnett AK. Minimal residual disease monitoring by quantitative RT-PCR in core binding factor AML allows risk stratification and predicts relapse: results of the United Kingdom MRC AML-15 trial. Blood. 2012;120(14):2826-2835. Zhu H-H, Zhang X-H, Qin Y-Z, et al. MRDdirected risk stratification treatment may improve outcomes of t(8;21) AML in the first complete remission: results from the AML05 multicenter trial. Blood. 2013; 121(20):4056-4062. Krönke J, Schlenk RF, Jensen K-O, et al. Monitoring of minimal residual disease in NPM1-mutated acute myeloid leukemia: a study from the German-Austrian acute myeloid leukemia study group. J Clin Oncol. 2011;29(19):2709-2716. Gorello P, Cazzaniga G, Alberti F, et al. Quantitative assessment of minimal residual disease in acute myeloid leukemia carrying nucleophosmin (NPM1) gene mutations. Leukemia. 2006;20(6):1103-1108. Lambert J, Lambert J, Nibourel O, et al. MRD assessed by WT1 and NPM1 transcript levels identifies distinct outcomes in

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

AML patients and is influenced by gemtuzumab ozogamicin. Oncotarget. 2014; 5(15):6280-6288. Cancer Genome Atlas Research Network. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013;368(22):2059-2074. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic Classification and Prognosis in Acute Myeloid Leukemia. N Engl J Med. 2016;374(23):2209-2221. Kohlmann A, Nadarajah N, Alpermann T, et al. Monitoring of residual disease by nextgeneration deep-sequencing of RUNX1 mutations can identify acute myeloid leukemia patients with resistant disease. Leukemia. 2014;28(1):129-137. Debarri H, Lebon D, Roumier C, et al. IDH1/2 but not DNMT3A mutations are suitable targets for minimal residual disease monitoring in acute myeloid leukemia patients: a study by the acute leukemia french association. Oncotarget. 2015;6 (39):42345-42353. Abdelhamid E, Preudhomme C, Helevaut N, et al. Minimal residual disease monitoring based on FLT3 internal tandem duplication in adult acute myeloid leukemia. Leuk Res. 2012;36(3):316-323. Bibault J-E, Figeac M, Hélevaut N, et al. Next-generation sequencing of FLT3 internal tandem duplications for minimal residual disease monitoring in acute myeloid leukemia. Oncotarget. 2015;6(26):2281222821. Krönke J, Bullinger L, Teleanu V, et al. Clonal evolution in relapsed NPM1-mutated acute myeloid leukemia. Blood. 2013; 122(1):100108. Bhatnagar B, Eisfeld A-K, Nicolet D, et al. Persistence of DNMT3A R882 mutations during remission does not adversely affect outcomes of patients with acute myeloid leukaemia. Br J Haematol. 2016;175(2):226236. Wong TN, Miller CA, Klco JM, et al. Rapid expansion of pre-existing non-leukemic hematopoietic clones frequently follows induction therapy for de novo AML. Blood. 2016;127(7):893-897. Wang Y-Y, Zhao L-J, Wu C-F, et al. C-KIT mutation cooperates with full-length AML1ETO to induce acute myeloid leukemia in mice. Proc Natl Acad Sci USA. 2011;108 (6):2450-2455.

1237

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1238-1246

Host virus and pneumococcus-specific immune responses in high-count monoclonal B-cell lymphocytosis and chronic lymphocytic leukemia: implications for disease progression Ignacio Criado,1 Santiago Muñoz-Criado,2 Arancha Rodríguez-Caballero,1 Wendy G. Nieto,1 Alfonso Romero,3 Paulino Fernández-Navarro,4 Miguel Alcoceba,5 Teresa Contreras,6 Marcos González,5 Alberto Orfao,1 Julia Almeida1 and The Primary Health Care Group of Salamanca for the Study of MBL

Cancer Research Centre (IBMCC, USAL-CSIC), Department of Medicine and Cytometry Service (NUCLEUS), University of Salamanca and IBSAL, Salamanca; 2Microbiology Service, University Hospital of Salamanca; 3Gerencia de Atención Primaria de Salud, Centro de Atención Primaria de Salud Miguel Armijo, Salamanca, Sanidad de Castilla y León (SACYL); 4Centro de Atención Primaria de Salud de Ledesma, Salamanca, Sanidad de Castilla y León (SACYL); 5Hematology Service, University Hospital of Salamanca, IBMCC, IBSAL and Department of Medicine, University of Salamanca and 6Biochemistry Service, University Hospital of Salamanca, Spain. 1

*AO and JA contributed equally to this work and both should be considered as senior authors.

ABSTRACT

Correspondence: orfao@usal.es

Received: October 26, 2016. Accepted: April 5, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.159012 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1238 ©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.

1238

atients diagnosed with chronic lymphocytic leukemia (CLL) display a high incidence of infections due to an associated immunodeficiency that includes hypogammaglobulinemia. A higher risk of infections has also been recently reported for high-count monoclonal B-cell lymphocytosis, while no information is available in low-count monoclonal B-cell lymphocytosis. Here, we evaluated the status of the humoral immune system in patients with chronic lymphocytic leukemia (n=58), as well as in low- (n=71) and high- (n=29) count monoclonal B-cell lymphocytosis versus healthy donors (n=91). Total free plasma immunoglobulin titers and specific levels of antibodies against cytomegalovirus, Epstein-Barr virus, influenza and S.pneumoniae were measured by nephelometry and ELISA-based techniques, respectively. Overall, our results show that both CLL and high-count monoclonal Bcell lymphocytosis patients, but not low-count monoclonal B-cell lymphocytosis subjects, present with relatively high levels of antibodies specific for the latent viruses investigated, associated with progressively lower levels of S.pneumoniae-specific immunoglobulins. These findings probably reflect asymptomatic chronic reactivation of humoral immune responses against host viruses associated with expanded virus-specific antibody levels and progressively decreased protection against other micro-organisms, denoting a severe humoral immunodeficiency state not reflected by the overall plasma immunoglobulin levels. Alternatively, these results could reflect a potential role of ubiquitous viruses in the pathogenesis of the disease. Further analyses are necessary to establish the relevance of such asymptomatic humoral immune responses against host viruses in the expansion of the tumor B-cell clone and progression from monoclonal B-cell lymphocytosis to CLL. haematologica | 2017; 102(7)

Pathogen-specific antibodies in MBL and CLL

Introduction Chronic lymphocytic leukemia (CLL) is the most common leukemia in adults in Western countries. It is characterized by an expansion of 5×109/L or more clonal B lymphocytes in peripheral blood (PB) that co-express CD5, CD19, CD23 and CD200, together with abnormally low levels of CD20, CD22, CD79b and surface immunoglobulins (sIg).1-4 CLL typically occurs in elderly patients and has a highly variable clinical course.5 Despite the heterogeneous clinical outcome, the majority of CLL patients share a profound immune dysregulation which is already detected at the earliest stages of the disease, and that progressively becomes more severe during clinical observation, leading to patient death even in the absence of disease progression.6 The precise mechanisms underlying such immune dysregulation in CLL are not fully understood; however, hypogammaglobulinemia has been identified as one of the major factors involved,6-8 both in the immunodeficiency status and death of CLL patients.9,10 Thus, hypogammaglobulinemia is present in up to 85% of patients. During the course of disease, a direct association has been reported between the stage and duration of disease and the severity of hypogammaglobulinemia.11,12 As a result, infection is one of the most prevalent causes of morbidity and mortality in CLL.13 Approximately 80% of CLL patients have infections during the course of the disease; such infections particularly involve the respiratory tract, pneumonia accounting for approximately 75% of all pulmonary complications in CLL.14 Recent studies have reported that subjects at earlier stages of the disease [e.g. high-count monoclonal B-cell lymphocytosis (MBLhi)] also have an increased risk of infections and a greater rate of infection-related deaths.15 Thus, hospitalization due to infection is significantly more common among MBLhi cases than in the general population (16% vs. 2.6% after a median follow-up period of 10 years, respectively), the overall frequency of infection in MBLhi individuals being similar to that of newly-diagnosed CLL patients (18%).15 Since vaccination represents an effective strategy to decrease the risk of infection in immunocompromised patients, the potential definition of optimal vaccination strategies in MBLhi and CLL requires a more in depth and comprehensive understanding of the dysregulated immunological mechanisms in these patients. In order to gain further insight into the nature, relevance and clinical significance of hypogammaglobulinemia in CLL and MBL patients, we evaluated the soluble levels of plasma antibodies specific for ubiquitous pulmonary infection-associated pathogens (i.e. influenza A and B viruses and S.pneumoniae) as well as other ubiquitous host pathogens, such as cytomegalovirus (CMV) and EpsteinBarr virus (EBV), in newly-diagnosed untreated CLL patients at different stages of the disease (Binet A vs. Binet B/C), pre-leukemic MBLhi, and low-count MBL (MBLlo) subjects versus a large group of age- and sex-matched healthy individuals from the same geographical area.

into four subgroups: healthy donors (controls; n=91), CLL-like MBLlo (n=71), CLL-like MBLhi (n=29), and newly-diagnosed previously untreated CLL patients (n=58). According to the World Health Organization (WHO) 2016 criteria,16 MBL was diagnosed whenever less than 5x109/L clonal B cells with a CLL phenotype were present in PB, in the absence of other signs of disease; otherwise, diagnosis of CLL was established. Within CLL, 32 patients were classified as early stage CLL (Binet A), while the remaining 26 corresponded to advanced-stage CLL (Binet B/C).4 In turn, MBLlo and MBLhi cases were discriminated based on a cut-off value of less than 0.5x109/L circulating clonal B cells with CLL-like phenotype, as described elsewhere.17 Additional information about the inclusion and exclusion criteria for selection of controls and patients, as well as procedures for sample collection and storage are detailed in the Online Supplementary Methods. The study was approved by the local Ethics Committee of the University Hospital of Salamanca, and conducted in accordance with the Declaration of Helsinki.

Immunophenotypic studies Immunophenotypic studies were performed on erythrocytelyzed PB samples, using a high-sensitive multicolor flow cytometry approach, previously described in detail.18 For this purpose, PB white blood cells (WBC) were systematically stained with the monoclonal antibody (MAb) combinations detailed in Online Supplementary Table S1. For flow cytometry data analysis, the INFINICYTTM software (Cytognos S.L., Salamanca, Spain) was used. All cases showed a clonal-imbalanced surface membrane (Sm) immunoglobulin (Ig)-kappa : SmIg-lambda ratio of >3:1 or <1:319and/or an aberrant CD5+ CLL(-like) B-cell population. The minimum number of clustered events required to define an abnormal B-cell population was 50 cells or more.

Measurement of soluble plasma levels of anti-viral and streptococcus pneumoniae (pneumococcus)-specific antibodies Exposure to CMV, EBV, influenza A and B viruses, and pneumococcus were measured by immunoenzymatic-based approaches, including either enzyme-linked immunosorbent (ELISA) or chemiluminescent immune assays, using commercially available kits, as detailed in the Online Supplementary Methods and Online Supplementary Table S2. Of note, analysis of influenza A- and influenza B-specific IgM and IgG and S.pneumoniae-specific IgG plasma levels was restricted to those subjects who had not been vaccinated against influenza and S.pneumoniae, respectively, during the 9-year period prior to the study (Online Supplementary Methods). In each patient, total plasma levels of IgM, IgG and IgA were systematically measured in parallel by nephelometry.

Quantitation of CMV and EBV viral copy number in plasma Detection and quantitation of CMV and EBV viral load in plasma was determined in a subset of 177 and 191 subjects, respectively, using commercially available kits: COBAS®AMPLiPrep/COBAS®TaqMan (Roche Diagnostics, Basel, Switzerland) and EBV R-gene® (BioMerieux, Verniolle, France), with strict adherence to the manufacturers’ instructions.

Results Methods Controls and patients

Clinical and laboratory features of MBL versus CLL patients

A total of 249 individuals were prospectively studied between November 2007 to November 2012. These subjects were classified

Overall, 249 individuals, including 119 males (48%) and 130 females (52%), with a mean age of 68±11 years were

haematologica | 2017; 102(7)

1239

I. Criado et al.

studied; there was a similar distribution according to age across the different patient groups and controls. Interestingly, while females predominated among MBLlo cases (male/female ratio 1:2), MBLhi and CLL showed a significantly (P<0.01) higher male/female ratio (5:1 and 1.2:1, respectively) (Table 1). As expected, abnormal blood cell counts were found only in MBLhi and CLL patients (but not in MBLlo), including lower platelet counts and hemoglobin levels among stage B/C CLL. Likewise, the absolute number of PB clonal B cells/μl progressively increased from MBLlo subjects to advanced-stage CLL patients (P<0.05). CLL patients also showed a greater frequency of IGHV unmutated cases (from 20% in MBLlo to 26% in MBLhi, 41% in CLL stage A and 64% in CLL stage B/C; P=0.04), whereas MBLlo cases showed a significantly lower frequency of cytogenetically altered CLL-like clones compared to both MBLhi and CLL (30% vs. 68% and 70%, respectively; P=0.002) (Table 1). Of note, all subjects were from the same geographical area (Province of Salamanca, Northwest-Central Spain) and, therefore, shared a similar antigen environment.

Soluble Ig plasma levels in MBL and CLL versus healthy controls Whereas total Ig plasma levels were within the normal range in MBLlo cases, they were significantly decreased in MBLhi and CLL patients versus both controls and MBLlo cases (Figure 1A). Interestingly, progressively lower levels

of total plasma Igs were found from MBLhi to stage A and stage B/C CLL cases, the latter two groups showing significantly lower amounts of plasma Igs versus MBLhi cases (P=0.04 for stage A and P=0.02 for stage B/C) (Figure 1A). In more detail, none of the MBLlo subjects presented with decreased total Ig plasma levels below the normal range, while the frequency of hypogammaglobulinemia increased from MBLhi to early- and advanced-stage CLL patients: 7% versus 16% and 19%, respectively (P<0.001) (Figure 1A). Of note, progressively lower levels of plasma Igs were observed from MBLhi to advanced stage CLL also for each Ig isotype (Figure 1B, C and D), particularly for IgM and IgA. Thus, all except one MBLlo case showed normal amounts of IgM, IgG and IgA plasma levels; in contrast, 17% of MBLhi subjects, 38% stage A CLL and 46% stage B/C CLL patients had decreased amounts of plasma IgM (P<0.0001). In addition, 10% of MBLhi cases had decreased IgA plasma levels versus 16% of stage A CLL and 46% of stage B/C CLL patients (P<0.0001); in turn, IgG plasma levels were decreased in 14% of MBLhi cases and 24% of CLL patients (P>0.05) versus none of the MBLlo subjects.

CMV-, EBV- and influenza-specific IgM and IgG plasma levels in MBL and CLL patients versus healthy controls As expected for a Mediterranean country, more than 90% of adults analyzed here had been exposed to both CMV and EBV before sample collection, regardless of the

Table 1. Clinical and laboratory characteristics of controls versus monoclonal B-cell lymphocytosis subjects and chronic lymphocytic leukemia patients.

Age (years) Sex (M/F) Hemoglobin (g/L) N. of platelets x109/L N. of leukocytes/μl N. of lymphocytes/μl N. of B lymphocytes/μl

Healthy donors (n=91)

MBLlo (n=79)

MBLhi (n=29)

CLL Stage A (n=32)

CLL Stage B/C (n=26)

CLL (n=58)

70 (43-87) 42% / 58% 147 (106-181) 226 (90-388) 6,200 (3550-11,240) 1678 (766-4124) 138 (31-776) NA

68 (52-85) 83% / 17% 144 (130-190) 198 (85-386) 11,550 (7154-15,660) 5250 (2291-9333) 3.097 (978-4773) 3.035 (921-4844) 74%/26%

67 (45-85) 56% / 44% 145 (120-174) 211 (112-408) 27,310 (13,520-393,530) 18,591 (6469-381,409) 17,727 (5134-369,288) 17.686 (5065-369,288) 59%/41%

70 (41-85) 54% / 46% 118 (88-164) 137 (67-271) 53,880 (16,630-289,420) 50,346 (12,779-282,098) 41.493 (8176-276,367) 41.442 (8019-276,367) 36%/64%

68 (41-85) 55% / 45% 136 (88-174) 174 (67-408) 34,920 (13,520-393,530) 25.939 (6469-381,409) 21.352 (5134-369,288) 21.130 (5065-369,288) 49%/51%

72 (43-95) 35% / 65% 144 (99-177) 222 (119-262) 6,090 (3650-9400) 1774 (317-3749) 139 (31-478) 0.731 (0024-82.24) *80%/20%

NA NA NA NA NA

*30% *30% *6% *0% *0%

68% 39% 25% 7% 0%

69% 50% 6% 13% 0%

71% 39% 33% 13% 0%

70% 45% 18% 13% 0%

N. of clonal B lymphocytes/μl Mutational status (Mut/UMut) Cytogenetic alterations % cases altered % cases del13q14(D13S25) % cases trisomy cr. 12 % cases del11q22(ATM) % cases del17p13(TP53)

P<0.01b P<0.01a P<0.01a P<0.01ab P<0.01ab P<0.01ab P<0.01ab P<0.05c P<0.01c NS NS NS NS

Results expressed either as median (range) or as percentage of cases for continuous and categorical variables, respectively. The CLL group includes both CLL Binet stage A and CLL Stage B/C cases. aCLL versus all other groups. bMBLhi versus all other groups. cCLL versus healthy individuals. CLL: chronic lymphocytic leukemia; F: female; M: male; MBLhi: high-count monoclonal B lymphocytosis; MBLlo: low-count monoclonal B lymphocytosis; Mut: mutated; NA: not applicable; ND: not determined; NS: no statistically significant differences detected (P>0.05); Umut: unmutated. *Sample size restricted to 23 subjects in which molecular and cytogenetic determinations were performed.

1240

haematologica | 2017; 102(7)

Pathogen-specific antibodies in MBL and CLL

diagnostic subgroup (Online Supplementary Table S3). In virtually every case the pattern of plasma antibodies specific for both viruses was consistent with past infection (i.e. CMV- or EBV-specific IgG-positive and IgM-negative plasma antibodies). In contrast, variable percentages of cases from the different study groups showed influenza virusspecific plasma Igs for the strains evaluated (Online Supplementary Table S3); in most of these cases, the pattern observed also corresponded to past exposure to the viruses. From the whole series of subjects who showed influenza virus-specific plasma Igs (n=127), 36 reported that they had been vaccinated against influenza virus before their recruitment; no statistically significant differences were observed in the distribution of these subjects in the distinct groups of individuals under study (Online Supplementary Table S3). Those patients found to have been previously exposed to any of the viruses investigated (i.e. those who showed increased plasma levels of at least one of the virus-specific Ig tested) were further evaluated for the corresponding pathogen-specific Ig plasma levels. Overall, plasma levels of pathogen-specific IgM and IgG

antibodies did not follow the pattern observed for total IgM and IgG plasma levels in the different groups of subjects analyzed (Figure 1). Thus, there was no reduction of specific IgM and IgG against CMV, EBV viral capside antigen (VCA) and influenza A and B in MBLhi and even in CLL patients versus both controls and MBLlo (Online Supplementary Figure S1). Regarding CMV-specific IgM and IgG titers and the amount of plasma IgM antibodies against VCA-EBV and the influenza virus, no significant differences were actually observed among individuals of the different groups studied (e.g. controls, MBLhi and CLL) (Online Supplementary Figure S1A, C and F). In contrast, VCA-EBV-specific IgG plasma levels were higher (P=0.01) in CLL patients versus both controls and MBLlo cases (Online Supplementary Figure S1D). However, clear differences emerged (or they increased) when the ratio between the plasma levels of each of these pathogen-specific IgG antibodies (CMV-, VCA-EBV- and influenza-specific IgG) plasma levels and the overall amount of plasma IgG per subject/patient was considered (Figure 2). Thereby, the CMV-specific IgM/total IgM and CMV-, VCA-EBV-specif-

Figure 1. Soluble immunoglobulin immunoglobulin (Ig) plasma levels in monoclonal B lymphocytosis (MBL) and chronic lymphocytic leukemia (CLL) versus healthy donors. (A) The overall amount of plasma immunoglobulins (md/dl) determined by conventional nephelometry is shown for the different groups of individuals analyzed. (B-D) IgM, IgG and IgA plasma levels within the different groups of individuals studied, respectively. Notched boxes represent 25th and 75th percentile values; the lines in the middle correspond to median values (50th percentile) and vertical lines represent the highest and lowest values that are neither outliers nor extreme values. Vertical dotted lines represent the inferior limit value of normality for each immunoglobulin. Dotted lines represent the lower limit of normality for each immunoglobulin (40 mg/dl; 700 mg/dl; and 70 mg/dl). Numbers under dotted lines represent the percentage of cases with plasma levels of the corresponding immunoglobulin found to be below normal values. *P<0.05 versus healthy donors and MBLlo; **P<0.05 versus healthy donors, MBLlo and MBLhi. MBLhi: high-count monoclonal B lymphocytosis; MBLlo: low-count monoclonal B lymphocytosis.

haematologica | 2017; 102(7)

1241

I. Criado et al.

ic IgG/total IgG ratios were significantly higher in CLL (Pâ&#x2030;¤0.001), particularly in stage B/C CLL cases (Pâ&#x2030;¤0.02) versus healthy donors and MBLlo subjects. Likewise, the influenza-specific IgG/total IgG ratio tended to be higher (P=0.056) for CLL patients compared to healthy donors and MBLlo cases (Figure 2). Of note, MBLhi also showed significantly higher anti-VCA-EBV-specific IgG/total IgG plasma levels than controls and MBLlo cases (Figure 2D). An exception to this general pattern was the EBNA-specific IgG plasma levels, which were found to be significantly reduced (vs. healthy donors) in both MBLhi (P=0.01) and CLL patients (P=0.002), particularly in stage B/C CLL (P=0.002) (Online Supplementary Figure S1E).

S.Pneumoniae-specific IgG plasma levels in MBL and CLL versus healthy controls As mentioned above, S.pneumoniae-specific IgG plasma levels were quantified in those subjects who reported no previous administration of anti-PCP (Pneumococcal Capsular Polysaccharide) vaccination (Figure 3). Their amount, as well as the pathogen specific IgG/total IgG ratio were within the normal range in all MBLlo subjects

and healthy controls analyzed (Figure 3A and 3). However, S.pneumoniae-specific IgG plasma levels were significantly reduced in MBLhi and CLL patients versus both controls and MBLlo (Figure 3A), in contrast to what was observed for virtually all viral pathogens described above, except the EBNA-specific IgG antibodies. Interestingly, no statistically significant differences were observed between MBLhi and CLL as regards the overall amount of anti-S.pneumoniae-specific IgG plasma levels. Of note, the ratio between the anti-S.pneumoniae-specific IgG and total IgG plasma levels was similar among the distinct groups of subjects analyzed (Figure 3B), as both the S.pneumoniae-specific IgG and the overall IgG plasma levels directly correlated within each group of subjects analyzed.

CMV and EBV viral load and virus-specific Ig titers Overall, viral load for CMV was studied in plasma from 177 subjects (53 controls, 56 MBLlo, 22 MBLhi and 46 CLL patients). No viral DNA was detected in any sample except 3 cases (1 MBLhiand 2 CLL Binet A subjects), in which the viral load could not be precisely quantified, as

Figure 2. Ratio between pathogen-specific immunoglobulin (Ig) plasma levels and total immunoglobulin plasma levels per Ig isotype in monoclonal B lymphocytosis (MBL) and chronic lymphocytic leukemia (CLL) patients versus healthy subjects. (A and B) Ratio between cytomegalovirus (CMV)-specific IgM and IgG plasma titers and the overall plasma IgM and IgG levels, respectively. (C and D) Ratio between viral capside antigen (VCA)-Epstein-Barr virus (EBV)-specific IgM and IgG titers in plasma and the overall amount of IgM and IgG in plasma, respectively. (E) Anti Epstein-Barr nuclear antigen (EBNA)-EBV-specific IgG/total IgG plasma level ratio. (F and G) Influenza (strains A + B)-specific/total IgM and IgG ratios, respectively. Only data on seropositive subjects for each pathogen are included in this figure. (F and G) Data presented correspond only to subjects who referred no previous vaccination against influenza. Notched boxes represent 25th and 75th percentile values; the lines in the middle correspond to median values (50th percentile), whereas vertical lines represent the highest and lowest values that are neither outliers nor extreme values. *P<0.05 versus healthy donors and MBLlo; **P<0.05 versus healthy donors, MBLlo and MBLhi.

1242

haematologica | 2017; 102(7)

Pathogen-specific antibodies in MBL and CLL

linemia of 29%20 vs. 14% in our series). This apparent discrepancy might be due to the fact that our series mostly comprised MBLhi cases with lower numbers of clonal B cells studied at diagnosis, while 4 of 7 MBLhi cases reported by Glancy et al. to have low IgG titers, had absolute lymphocyte counts more than 4x109/L. Nevertheless, it should be noted that we did not find any correlation between soluble Ig plasma levels and the number of clonal B cells in PB, within each group of subjects analyzed (data not shown). In contrast, a statistically significant direct correlation was found between total Ig plasma levels and the number of normal residual B cells among CLL patients (r=0.29, P=0.04). Therefore, presence of hypogammaglobulinemia in MBLhi cases could also reflect a defective normal residual B-cell function and it might contribute to explain the near 3-fold increased frequency of infection observed among these subjects versus the general population of the same age and having the same comorbidities, to that of newly-diagnosed CLL.15 Altogether, these findings suggest that antibody-related immunodeficiency might emerge before the onset of CLL, already at an MBLhi state, preceding (or potentially favoring) malignant transformation and progression of the disease. Despite a progressive reduction of (total) soluble Ig plasma levels from MBL to advanced CLL, similar levels of CMV-specific IgM and IgG, VCA-EBV-specific IgM and influenza-specific IgM and IgG were found among the five groups analyzed (i.e. healthy donors, MBLlo, MBLhi, early CLL and advanced stage CLL). Indeed, VCA-EBV-specific IgG levels were even increased in CLL patients versus healthy subjects. Furthermore, when the ratio between each of these Ab levels and the total plasma levels of the corresponding Ig isotype (IgM or IgG) were considered, progressively greater fractions of the above referred antigen-specific antibodies per-isotype were found from MBLhi to stage A and stage B/C CLL patients, respectively. Our findings on the antibody levels against CMV confirm previous results on CLL reported by Vanura et al. who also showed that, despite progressive decay of total IgM and

it was below the limit of quantification of the method (<137 IU/ml). In turn, EBV DNA load (measured in 191 samples: 57 controls, 59 MBLlo, 23 MBLhi, 53 CLL patients) was detected in plasma from 7 of 53 Binet A CLL patients (13%), while systematically undetectable in the other three groups (P<0.0001). No statistically significant differences in gender distribution, age, number of clonal B cells and EBV (VCA)-specific IgM and IgG titers were found between CLL cases with quantifiable EBV DNA in plasma versus negative CLL cases. Also, no statistical correlation was found between the number of EBV DNA copies (median of 3.6 DNA copies/Îźl; range 1.4-22.8 DNA copies/Îźl) and EBV-specifc immunoglobulin titers in plasma for those 7 EBV-viral load-positive CLL cases.

Discussion Infection is one of the most frequent causes of death in CLL (approx. 30-50% of CLL patients).8 Although the specific mechanisms underlying immune dysregulation in CLL are not fully understood8, hypogammaglobulinemia, together with T-cell abnormalities, are common features of the CLL-associated immunodeficiency status, the former affecting up to 85% of the patients already at diagnosis or during the course of their disease.9,10 The frequency and severity of hypogammaglobulinemia (at the expense of all major Ig isotypes) increase from MBL subjects to early and advanced stage CLL patients. Here, we confirm and extend on these observations. Thus, we show for the first time that total soluble Ig plasma levels are within normal values in MBLlo subjects, regardless of the Ig isotype evaluated; in contrast, hypogammaglobulinemia was a relatively frequent feature of MBLhi cases. Of note, the degree of decreased IgM and IgG plasma levels in MBLhi was similar to that observed in stage A CLL. In a recent study, Glancy et al. have even reported a higher frequency of decreased IgG levels in MBLhi (i.e. 7 of 24 MBLhi cases, which represents a frequency of IgG hypogammaglobu-

Figure 3. Streptococcus pneumoniae-specific IgG plasma levels in monoclonal B lymphocytosis (MBL) and chronic lymphocytic leukemia (CLL) patients versus healthy controls. (A) Titers of antibody-specific plasma levels against the pneumococcal polysaccharide antigen for each group of individuals analyzed. (B) Ratio between anti-pneumococcus-specific IgG and total IgG plasma levels for each group of subjects investigated. Only data from those subjects that did not receive vaccination against S.pneumoniae are displayed. Notched boxes represent 25th and 75th percentile values; the lines in the middle correspond to median values (50th percentile), while vertical lines represent the highest and lowest values that are neither outliers nor extreme values. *P<0.05 versus healthy donors and MBLlo. MBLhi: high-count monoclonal B lymphocytosis; MBLlo: low-count monoclonal B lymphocytosis.

haematologica | 2017; 102(7)

1243

I. Criado et al.

IgG subclasses, the CMV-specific immune response may be preserved even in CLL cases with advanced disease.21 Here, we confirm and extend on these findings by showing for the first time that: i) this behavior is already detectable at the MBLhi stage; and ii) it is also common to other antibody responses against EBV and the influenza virus in non-vaccinated individuals, despite the mechanisms by which influenza infects cells are completely different from those of CMV and EBV.22-24 As mentioned above, we did observe decreased titers of EBV-specific IgG levels in both MBLhi and CLL; interestingly, this was restricted to the antibody response against the EBNA-EBV antigen, but not the VCA-EBV antigen. The EBNA-EBV protein is located in the nucleus of infected host cells and it acts as a transcription factor for the virus, allowing for its replication inside the cell;23 in contrast, the VCA-EBV protein is a structural component of the capside of the virus.25 Therefore, the (humoral) immune response against the VCA-antigen might only occur if infected cells are lysed and active viral replications occurs. Therefore, our results suggest that like CMV, EBV probably undergoes a mild (undetectable) reactivation, whenever an immunodeficiency state has been acquired, but fully bloomed EBV and CMV infections can still be controlled, as reflected by the preserved production of specific antibodies against both viruses in MBLhi and CLL patients21 and the detection of quantifiable EBV DNA in plasma of CLL cases but not MBL. Long-term monitoring of virus-specific Ig plasma levels in CLL versus MBL versus healthy donors is required to validate this hypothesis. In contrast to the general pattern found for the plasma levels of antibodies against the ubiquitous viruses here investigated, a significant reduction was observed in the plasma levels of pneumococcus-specific IgG from MBLhi to stage B/C CLL, in parallel to the overall decrease in total IgG plasma levels. These findings further suggest that, while the antibody-mediated immune response against ubiquitous pathogens (e.g. human herpesviruses and influenza virus) is still preserved and the virus is actively controlled in immunocompromised MBLhi and CLL patients, protection against other pathogens (i.e. pneumococcus) is progressively lost, placing these patients at risk of severe infection and death. In line with this hypothesis, CMV disease is infrequent among untreated CLL patients compared to other immunocompromised patients.13,26 In contrast, CLL patients frequently present respiratory tract infections caused by encapsulated bacteria, particularly Streptococcus pneumoniae and Haemophilus influenza,27 further supporting a unique dysregulation of immunesurveillance against infectious agents in MBLhi and CLL. To the best of our knowledge, no studies have been reported so far about the immune response profile against different pathogens in MBLlo subjects. As no differences were detected in both total and pathogen specific Ig plasma titers in MBLlo versus age-matched healthy subjects of the same geographical area, it might be expected that the antibody response of these subjects remains normal or at most little altered. Altogether, these findings suggest that the onset of dysregulated antibody-based immune responses might occur in the transition from MBLlo to MBLhi and CLL, being associated with a clinically silent reactivation of preserved T-cell dependent antibody responses against host viruses. If this holds true, and chronic baseline activation of antibody responses against host viruses occurs in MBLhi and CLL patients, such a 1244

response could also potentially affect the tumor clone and contribute to its expansion and progression of the disease. In line with this hypothesis, it has been shown that most MBLlo subjects show (oligoclonal) expansions of CD4+/CD8+ double-positive T cells28 which have a limited TCRvÎ˛ repertoire and participate in immune responses against chronic viral infections, particularly against CMV.29 There is even stronger evidence to suggest that CLL evolves from repetitive activation of particular B-cell clones through B-cell receptor (BCR) triggering by conventional antigens,30 which, in the light of the results reported here, increase in the CMV- and EBV-specific IgG/total Ig ratio in both MBLhi and CLL patients. This might further suggest a potential role for ubiquitous viruses in the pathogenesis of the disease. Previous findings showing an association between the presence of CMV- and EBV-DNA in blood of CLL patients who express stereotyped IGHV4-34 BCRs31 would further support this hypothesis. However, here we only analyzed a relatively limited number of cases within each study group, particularly within the MBLhi group, and, therefore, further long-term longitudinal studies in MBL and CLL in larger series of subjects are necessary in order to elucidate the value of (total and pathogenspecific) Ig plasma levels, as a surrogate marker for a normal versus abnormal B-cell function, and to determine both the risk of progression from MBL to CLL and the potential need for adoption of specific active immunotherapy measures for patients at risk of life-threatening infections. In this regard, extensive research on the effectiveness of vaccines, particularly against influenza and S.pneumoniae, has been carried out in CLL, while there is limited information on MBLhi.15 Thus, response to vaccination against both polysaccharide (e.g. classical multivalent pneumococcal vaccines32,33) and protein antigens (e.g. tetanus toxoid and influenza virus34,35) has been shown to be associated with poor seroprotective responses in CLL, even after various doses. Such defective antibody responses have been related to a broad variety of immune defects including complement dysregulation, T-cell impaired function and altered antigen presentation, in addition to B-cell deficiency.8,9,27,36,37 Because of this, vaccination of CLL patients early after diagnosis, and particularly even at the MBL stage when better responses might be expected,8,33 has been proposed as a potentially effective strategy to improve serological immune protection of CLL patients against the most common pathogens. Parallel analyses focused on the humoral immunity and immune responses other than just the evaluation of plasma antibody levels are required to fully understand the uniqueness of the immunodeficiency status of MBLhi and CLL patients. In summary, we report on the existence of a significant and selective, defective antibody protection against S.pneumoniae in CLL which emerges already among MBLhi to early stage CLL and worsens through progression of the disease. Such an immune defect might be associated with an active, but silent, response against host viruses such as CMV, EBV and influenza, for which preserved antibody serum levels are detected, even in advanced CLL. These results suggest that chronic viral re-activation might contribute to the preserved host virus-specific antibody titers through sustained immune responses, which might also favor parallel expansion of the tumor B-cell clone and progression from MBLhi to CLL. Further studies in larger MBL and CLL patient cohorts with long-term follow up and haematologica | 2017; 102(7)

Pathogen-specific antibodies in MBL and CLL

sequential serological analyses are necessary to confirm this hypothesis. Primary Health Care Group of Salamanca for the study of MBL: list of members (alphabetical order): Alonso Martín, María Monserrat (C.S. Fuentes de Oñoro); Asensio Oliva, María Carmen (C.S. Santa Marta de Tormes), Bárez Hernández, Pilar (C.S. Garrido Sur); Cabo Sastre, Luis (C.S. Ledesma); Carreño Luengo, María Teresa (C.S. Ledesma); Casado Romo, José María (C.S. Alba de Tormes); Cubino Luis, Rocio (C.S. Sancti Spiritus); De Vega Parra, José (C.S. Peñaranda); Franco Esteban, Eloy (C.S. Pizarrales-Vidal); García García, María Concepción (C.S. Guijuelo); García Rodríguez, Bernardo Lucio (C.S. La Alberca); Garzón Martín, Agustín (C.S. Peñaranda); Goenaga Andrés, Rosario (C.S. Ledesma); Gómez Cabrera, Rosalia (C.S. Garrido Sur); Gómez Sánchez, Francisco (C.S. Periurbana Norte); González Moreno, Josefa (C.S. Guijuelo); González Vicente, Ángel Carlos (C.S. Aldeadávila de la Ribera); Guarido Mateos, José Manuel (C.S. Vitigudino); Hernández Sánchez, María Jesús (C.S. Vitigudino); Herraes Martín, Ricardo (C.S. La Alberca); Herrero Sánchez, Amparo (C.S. Fuentes de Oñoro); Jiménez Ruano, María Josefa (C.S. Garrido Norte); Jimeno Cascón, Teresa Basa (C.S. Elena Ginel Díez); Macías Kuhn, Francisco (C.S. Ledesma); Mateos Rubio, Pablo (C.S. Ledesma); Márquez Velasco, María Salud (C.S. Sancti Spiritus); Merino Palazuelo, Miguel (C.S. Fuentes de Oñoro); Miguel Lozano, Rubén (C.S. Garrido Norte); Montero Luengo, Juan (C.S. San Juan); Muriel Díaz, María Paz (C.S. Miguel Armijo); Pablos Regueiro, Araceli (C.S. Vitigudino); Pascual Martín, J. Antonio (C.S. Fuentes de Oñoro); Pastor Alcalá, Luis (C.S. Vitigudino); Pedraza García, Jesús (C.S. Vitigudino); Pérez Díaz, Manuel (C.S. Pizarrales-Vidal); Pérez García, Manuel (C.S. Alba de Tormes); Prieto Gutiérrez, María Teresa (C.S. Peñaranda); Ramos Arranz, Manuel (C.S. Ledesma); Ramos Mongue, Aurora Esther (C.S. Ledesma);

References 1. Strati P, Shanafelt TD. Monoclonal B-cell lymphocytosis and early-stage chronic lymphocytic leukemia: diagnosis, natural history, and risk stratification. Blood. 2015; 126(4):454-462. 2. Hallek M, Cheson BD, Catovsky D, Caligaris-cappio F, Dighiero G, Do H. 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. 3. Chiorazzi N, Rai KR, Ferrarini M. Chronic lymphocytic leukemia. N Engl J Med. 2005;352(8):804-815. 4. Hallek M. Chronic lymphocytic leukemia: 2015 Update on diagnosis, risk stratification, and treatment. Am J Hematol. 2015; 90(5):446-460. 5. Montillo M, Hamblin T, Hallek M, Montserrat E, Morra E. Chronic lymphocytic leukemia: novel prognostic factors and their relevance for risk-adapted therapeutic strategies. Haematologica. 2005; 90(3):391-399.

haematologica | 2017; 102(7)

Rodríguez Medina, Ana María (C.S. Alba de Tormes); Rodríguez Vegas, Margarita (C.S. Ledesma); Romo Cortina, Javier (C.S. Elena Ginel Díez); Roselló Carmen, Elena (C.S. Vitigudino); Sánchez Alonso, Begoña (C.S. Aldeadávila de la Ribera); Sánchez Bazo, Begoña (C.S. Aldeadávila de la Ribera), Sánchez White, Nicolás (C.S. Sancti Spiritus); Sandín Pérez, Rafael (C.S. San José); Sanz Santa-Cruz; Fernando (C.S. Capuchinos); Soto Jiménez, Francisco (C.S. Santa Marta de Tormes); Velasco Marcos, María Auxiliadora (C.S. Elena Ginel Díez); Vicente López, Horacio Marcos (C.S. Aldeadávila de la Ribera); Vicente Santos, M. Sebastián (C.S. Aldeadávila de la Ribera). Acknowledgments The authors thank María Teresa Blázquez Martín and María del Mar Clemente Aguilar for their technical support in both serological assays and quantitation of viral load in plasma. Funding This work was supported by the: RD06/0020/0035 and RD12/0036/0048 grants from Red Temática de Investigación Cooperativa en Cáncer (RTICC), Instituto de Salud Carlos III, Ministerio de Economía y Competitividad, (Madrid, Spain and FONDOS FEDER); CB16/12/00400 grant (CIBER-ONC, Instituto de Salud Carlos III, Ministerio de Economía y Competitividad, Madrid, Spain and FONDOS FEDER); the FIS PI06/0824-FEDER, PS09/02430-FEDER, PI12/00905FEDER and DTS15/00119-FEDER grants, from the Fondo de Investigación Sanitaria of Instituto de Salud Carlos III; the GRS206/A/08 grant, (Ayuda al Grupo GR37 de Excelencia, SAN/1778/2009) from the Gerencia Regional de Salud, (Consejería de Educación and Consejería de Sanidad of Castilla y León, Valladolid, Spain); FS/1-2010 and FS/19-2013 grants, from the Fundación Memoria D. Samuel Solórzano, (University of Salamanca, Salamanca, Spain).

6. Forconi F, Moss P. Perturbation of the normal immune system in patients with CLL. Blood. 2015;126(5):573-581. 7. Lanasa MC, Weinberg JB. Immunologic aspects of monoclonal B-cell lymphocytosis. Immunol Res. 2011;49(1-3):269-280. 8. Whitaker JA, Shanafelt TD, Poland GA, Kay NE. Room for improvement: Immunizations for patients with monoclonal B-cell lymphocytosis or chronic lymphocytic leukemia. Clin Adv Hematol Oncol. 2014;12(7):440-450. 9. Hamblin AD, Hamblin TJ. The immunodeficiency of chronic lymphocytic leukaemia. Br Med Bull. 2008;87(1):49-62. 10. Freeman JA, Crassini KR, Best OG, et al. Immunoglobulin G subclass deficiency and infection risk in 150 patients with chronic lymphocytic leukemia. Leuk Lymphoma. 2013;54(1):99-104. 11. Orfao A, Gonzalez M, San Miguel JF, et al. B-cell chronic lymphocytic leukaemia: prognostic value of the immunophenotype and the clinico-haematological features. Am J Hematol. 1989;31(1):26-31. 12. Parikh SA, Leis JF, Chaffee KG, et al. Hypogammaglobulinemia in newly diagnosed chronic lymphocytic leukemia: Natural history, clinical correlates, and outcomes. Cancer. 2015;121(17):2883-2891.

13. Morrison VA. Infectious complications in patients with chronic lymphocytic leukemia: pathogenesis, spectrum of infection, and approaches to prophylaxis. Clin Lymphoma Myeloma. 2009;9(5):365-370. 14. Ahmed S, Siddiqui AK, Rossoff L, Sison CP, Rai KR. Pulmonary Complications in Chronic Lymphocytic Leukemia. Cancer. 2003;98(9):1912-1917. 15. Moreira J, Rabe KG, Cerhan JR, et al. Infectious complications among individuals with clinical monoclonal B-cell lymphocytosis (MBL): a cohort study of newly diagnosed cases compared to controls. Leukemia. 2013;27(1):136-141. 16. Swerdlow SH, Campo E, Pileri SA, et al. The 2016 revision of the World Health Organization (WHO) classification of lymphoid neoplasms. Blood. 2016;127(20): 2375-2390. 17. Shanafelt TD, Ghia P, Lanasa MC, Landgren O, Rawstron AC. Monoclonal Bcell lymphocytosis (MBL): biology, natural history and clinical management. Leukemia. 2010;24(3):512-520. 18. Nieto WG, Almeida J, Romero A, et al. Increased frequency (12%) of circulating chronic lymphocytic leukemia-like B-cell clones in healthy subjects using a highly sensitive multicolor flow cytometry

1245

I. Criado et al. approach. Blood. 2009;114(1):33-37. 19. Nieto WG, Almeida J, Teodosio C, et al. Commentary: Comparison of current flow cytometry methods for monoclonal B cell lymphocytosis detection. Cytometry B Clin Cytom. 2010;78 Suppl 1:S4-9. 20. Glancy E, Siles R. Monoclonal B-cell lymphocytosis and hypogammaglobulinaemia. Br J Haematol. 2016;173(2):316-317. 21. Vanura K, Rieder F, Kastner M-T, et al. Chronic lymphocytic leukemia patients have a preserved cytomegalovirus-specific antibody response despite progressive hypogammaglobulinemia. PLoS One. 2013;8(10):e78925. 22. Sun X, Whittaker GR. Entry of influenza virus. Adv Exp Med Biol. 2013;790:72-82. 23. Cohen JI. Epstein-Barr virus infection. N Engl J Med. 2000;343(7):481-492. 24. Spector DH. Human cytomegalovirus riding the cell cycle. Med Microbiol Immunol. 2015;204(3):409-419. 25. Moss DJ, Burrows SR, Khanna R. EBV: immunobiology and host response. In: Arvin A, Campadelli-Fiume G, Mocarski E, Moore PS, Roizman B, Whitley R, et al. Human Herpesviruses: Biology, Therapy, and Immunoprophylaxis. Cambridge University Press; 2007;Chapter 51. 26. Laurenti L, Piccioni P, Cattani P, et al. Cytomegalovirus reactivation during alemtuzumab therapy for chronic lymphocytic

1246

27.

28.

29.

30.

31.

leukemia: incidence and treatment with oral ganciclovir. Haematologica. 2004; 89(10):1248-1252. Pasiarski M, Rolinski J, Grywalska E, et al. Antibody and plasmablast response to 13valent pneumococcal conjugate vaccine in chronic lymphocytic leukemia patients Preliminary report. PLoS One. 2014; 9(12):1-14. Fazi C, Scarfó L, Pecciarini L, et al. General population low-count CLL-like MBL persists over time without clinical progression, although carrying the same cytogenetic abnormalities of CLL. Blood. 2011; 118(25):6618-6625. Suni MA, Ghanekar SA, Houck DW, et al. CD4+CD8dim T lymphocytes exhibit enhanced cytokine expression, proliferation and cytotoxic activity in response to HCMV and HIV-1 antigens. Eur J Immunol. 2001;31(8):2512-2520. Widhopf GF 2nd, Goldberg CJ, Toy TL, et al. Nonstochastic pairing of immunoglobulin heavy and light chains expressed by chronic lymphocytic leukemia B cells is predicated on the heavy chain CDR3. Blood. 2008;111(6):3137-3144. Kostareli E, Hadzidimitriou A, Stavroyianni N, et al. Molecular evidence for EBV and CMV persistence in a subset of patients with chronic lymphocytic leukemia expressing stereotyped IGHV4-

32.

33.

34.

35.

36.

37.

34 B-cell receptors. Leukemia. 2009;23(5): 919-924. Sinisalo M, Aittoniemi J, Oivanen P, Käyhty H, Olander RM, Vilpo J. Response to vaccination against different types of antigens in patients with chronic lymphocytic leukaemia. Br J Haematol. 2001; 114(1): 107-110. Sánchez-Ramón S, Dhalla F, Chapel H. Challenges in the Role of Gammaglobulin Replacement Therapy and Vaccination Strategies for Hematological Malignancy. Front Immunol. 2016;7:317. Sinisalo M, Aittoniemi J, Kayhty H, Vilpo J. Vaccination against infections in chronic lymphocytic leukemia. Leuk Lymphoma. 2003;44(4):649-652. Pollyea DA, Brown JMY, Horning SJ. Utility of influenza vaccination for oncology patients. J Clin Oncol. 2010;28(14):24812490. Sinisalo M, Vilpo J, Itälä M, Väkeväinen M, Taurio J, Aittoniemi J. Antibody response to 7-valent conjugated pneumococcal vaccine in patients with chronic lymphocytic leukaemia. Vaccine. 2007;26(1):82-87. Van Der Velden AMT, Van Velzen-Blad H, Claessen AME, et al. The effect of ranitidine on antibody responses to polysaccharide vaccines in patients with B-cell chronic lymphocytic leukaemia. Eur J Haematol. 2007;79(1):47-52.

haematologica | 2017; 102(7)

ARTICLE

Non-Hodgkin Lymphoma

Inhibition of monocarboxyate transporter 1 by AZD3965 as a novel therapeutic approach for diffuse large B-cell lymphoma and Burkitt lymphoma

Richard A. Noble,1 Natalie Bell,1 Helen Blair,1 Arti Sikka,2 Huw Thomas,1 Nicole Phillips,1 Sirintra Nakjang,1 Satomi Miwa,3 Rachel Crossland,1 Vikki Rand,1 Despina Televantou,4 Anna Long,4,5 Hector C. Keun,2 Chris M. Bacon,1,4,5 Simon Bomken,1,6 Susan E. Critchlow7 and Stephen R. Wedge1

Northern Institute for Cancer Research, Newcastle University, Newcastle upon Tyne; Division of Cancer, Imperial College London; 3Institute for Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne; 4Cellular Pathology, Newcastle upon Tyne Hospitals NHS Foundation Trust; 5MRC/EPSRC Newcastle Molecular Pathology Node, Newcastle upon Tyne; 6Department of Pediatric and Adolescent Hematology and Oncology, Newcastle upon Tyne Hospitals NHS Foundation Trust and 7 AstraZeneca, Cambridge, UK 1

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1247-1257

ABSTRACT

nhibition of monocarboxylate transporter 1 has been proposed as a therapeutic approach to perturb lactate shuttling in tumor cells that lack monocarboxylate transporter 4. We examined the monocarboxylate transporter 1 inhibitor AZD3965, currently in phase I clinical studies, as a potential therapy for diffuse large B-cell lymphoma and Burkitt lymphoma. Whilst extensive monocarboxylate transporter 1 protein was found in 120 diffuse large B-cell lymphoma and 10 Burkitt lymphoma patients’ tumors, monocarboxylate transporter 4 protein expression was undetectable in 73% of the diffuse large B-cell lymphoma samples and undetectable or negligible in each Burkitt lymphoma sample. AZD3965 treatment led to a rapid accumulation of intracellular lactate in a panel of lymphoma cell lines with low monocarboxylate transporter 4 protein expression and potently inhibited their proliferation. Metabolic changes induced by AZD3965 in lymphoma cells were consistent with a feedback inhibition of glycolysis. A profound cytostatic response was also observed in vivo: daily oral AZD3965 treatment for 24 days inhibited CA46 Burkitt lymphoma growth by 99%. Continuous exposure of CA46 cells to AZD3965 for 7 weeks in vitro resulted in a greater dependency upon oxidative phosphorylation. Combining AZD3965 with an inhibitor of mitochondrial complex I (central to oxidative phosphorylation) induced significant lymphoma cell death in vitro and reduced CA46 disease burden in vivo. These data support clinical examination of AZD3965 in Burkitt lymphoma and diffuse large B-cell lymphoma patients with low tumor monocarboxylate transporter 4 expression and highlight the potential of combination strategies to optimally target the metabolic phenotype of tumors.

Introduction The increased reliance on glycolytic metabolism under aerobic conditions, termed the “Warburg effect”, is adopted by many tumor types and is characterized by an increased utilization of glucose and a corresponding greater efflux of lactate.1,2 Consequently, there has been much interest in targeting this recognized ‘hallmark of cancer’ for therapeutic benefit.3,4 One such approach has been to interfere with lactate transport via inhibition of monocarboxylate transporter (MCT) 1. MCT1 and MCT4 are cell membrane-localized, proton-coupled transporters of haematologica | 2017; 102(7)

Correspondence: steve.wedge@ncl.ac.uk

Received: December 22, 2016. Accepted: March 31, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.163030 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1247 ©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.

1247

R.A. Noble et al.

monocarboxylates such as lactate and pyruvate.5 MCT1 is expressed widely and possesses a comparatively high affinity for lactate, allowing it to function as an influx or efflux transporter depending upon the local lactate concentration gradient. In contrast, MCT4 predominantly fulfils an efflux transport role in highly glycolytic tissues. The function of both transporters is dependent upon an association with the transmembrane accessory protein CD147 (basigin; BSG) which ensures their correct orientation at the cell surface.6 MCT1 and MCT4 can be differentially over-expressed in cancer,7-11 and a subset of tumors express MCT1 in the absence of appreciable MCT4 protein. In such cells MCT1 inhibition can have significant consequences: preventing lactate efflux in highly glycolytic tumor types, and restricting access to lactate in more oxidative cancer types in which it may be utilized as a respiratory fuel.12,13 AZD3965 is an orally bioavailable MCT1 inhibitor, which is currently under phase I clinical investigation (NCT01791595).14 Recent studies have demonstrated that AZD3965 or structurally related MCT1 inhibitors can inhibit the bidirectional transport of lactate in cancer cells which lack MCT4 protein and this may inhibit the cells’ growth.7,12,14 We evaluated the metabolic and therapeutic effects of AZD3965 in aggressive forms of non-Hodgkin lymphoma, namely, diffuse large B-cell lymphoma (DLBCL) and Burkitt lymphoma (BL). We demonstrate in patients’ samples that these diseases often have high MCT1 and undetectable or negligible MCT4 protein expression. We also show that AZD3965 can inhibit lactate efflux sufficiently in DLBCL and BL cell lines to alter cellular metabolism and exert a profound cytostatic effect on lymphoma cell growth in vitro and in vivo. Finally, we demonstrate that combining AZD3965 with an inhibitor of oxidative phosphorylation (OXPHOS) can induce significant tumor cell death and reduce lymphoma disease burden in vivo. Collectively these studies define a clear opportunity for the use of AZD3965 in the clinical management of DLBCL and BL.

Methods Information concerning cell origin, authentication, culture conditions, western blotting and antibody and drug use are detailed in the Online Supplementary Information.

Determination of intracellular lactate, protein, cell growth and viability Lactate concentration was determined by colorimetric assay (Trinity Biotech, Co Wicklow, Ireland) and normalized to protein content. For growth inhibition assays, cells were plated overnight before treatment for 72 h and assessed using an XTT assay (Sigma, Saint-Louis, MO, USA). GI50 values were determined using GraphPad Prism software (version 6). Cell number and viability were determined concurrently after 72 h and 120 h of AZD3965 treatment using a hemocytometer and trypan blue exclusion, respectively.

Immunohistochemistry Formalin-fixed, paraffin-embedded, pre-treatment, diagnostic DLBCL and BL tissue samples were obtained from the Newcastle Hematology Biobank (National Research Ethics Service Committee Reference 07/H0906/109+5) and the 1248

Children’s Cancer and Leukaemia Tissue Bank (Reference 08/H0405/22+5), respectively. Immunohistochemistry for MCT1 and MCT4 was performed on the Ventana Benchmark (Tucson, AZ, USA) automated immunostaining platform using Optiview detection. Staining extent and intensity on tumor cells were evaluated by two hematopathologists (CMB and DT) and a summary H-score (0-300) calculated as previously described.15 In selected cases, double immunohistochemical staining for PAX5 (SP34 rabbit monoclonal antibody, Ventana) was used to distinguish tumor cell versus non-tumor MCT4 expression. DLBCL cell-of-origin classification was determined by immunostaining, as described in Culpin et al.,16 using the Hans algorithm.17

Metabolic assays Oxygen consumption rate and extracellular acidification rate were measured using a Seahorse XF24 analyzer (Agilent, Santa Clara, CA, USA). CA46 or CA46-R cells (2x105) were made to adhere to Seahorse 24-well plates using Cell-Tak tissue adhesive (Corning, Wiesbaden, Germany) 1 to 2 h prior to analysis. To examine intracellular metabolite concentrations, tumor cells were seeded in media containing dialyzed fetal calf serum (10%) and treated with dimethlysulfoxide vehicle or AZD3965 (100 nM) for 2 h under normoxic conditions. Following metabolite extraction, liquid chromatography-mass spectrometry analysis was performed as previously described.14 Extracellular metabolites in RPMI supplemented with dialyzed fetal calf serum (10%) were determined using nuclear magnetic resonance following incubation with AZD3965 (100 nM) for 24 h. The intracellular metabolite composition of tumor xenograft samples was assessed by gas chromatography-mass spectrometry. Additional details on metabolic assays are detailed in the Online Supplementary Information.

In vivo efficacy of AZD3965 For in vivo studies, luciferase-expressing CA46 cells18,19 were injected intravenously, via the tail vein, into NOD/LtSz-scid IL2Rγ null (NSG) mice within a laminar flow hood. Mice were imaged using an IVIS Spectrum pre-clinical imaging system (Perkin Elmer, Waltham, MA, USA) as previously described.20 IVIS spectrum operators were blinded to treatment assignments. Both AZD3965 (100 mg/kg, BID) and BAY 87-2243 (9 mg/kg, QD) or relevant vehicle controls were administered by oral gavage. Animal experiments were approved by Institutional Ethical Review Process Committees and performed under UK Home Office licenses.

Statistical tests Statistical significance was examined using a two-tailed Student t-test, with the exception of group comparisons within in vivo experiments which were performed using a two-way ANOVA with a Tukey test, or a Pearson χ2 test to examine whether post-treatment tumor volumes had decreased relative to pre-treated volumes. Data comparisons with P-values <0.05 were considered statistically significant.

Results Diffuse large B-cell lymphoma and Burkitt lymphoma are appropriate diseases for AZD3965 treatment We re-analyzed MCT1 and MCT4 published gene expression data from tumor cell lines and found DLBCL and BL cell lines to be among the lowest expressers of haematologica | 2017; 102(7)

AZD3965 treatment of lymphoma

MCT4 (SLC16A3) mRNA, particularly in comparison to those originating from diverse solid tumor types (Online Supplementary Figure S1).21 In contrast, MCT1 (SLC16A1) expression was less variable across cancer types. To determine protein expression in clinical lymphoma samples, we stained a cohort of 120 DLBCL patientsâ&#x20AC;&#x2122; samples for both MCT1 and MCT4 protein and categorized samples using an immunohistochemical H-score (Online Supplementary Figure S2). DLBCL samples were found to be negative for

MCT4 (H-Score = 0) in 73% of cases (Figure 1A), despite variable numbers of MCT4-positive stromal cells identified by morphology and the absence of a B-cell marker (PAX5) (data not shown). MCT4 protein staining was absent in both activated B-cell and germinal center B-cell cell-oforigin classifications (Figure 1B,C). The majority of samples had significant tumor cell MCT1 protein expression which was not significantly associated with MYC translocation status (data not shown). Specimens from ten patients

Figure 1. Diffuse large B-cell lymphoma and Burkitt lymphoma are appropriate diseases for AZD3965 treatment. (A) H-score analysis of MCT1 and MCT4 tumor cell protein expression in 120 samples from DLBCL and 10 BL patient samples. A sample was considered negative (H-score of 0) when no staining was evident on tumor cells, staining on stromal cells or inflammatory infiltrate being excluded from the analysis. Representative MCT1 and MCT4 immunohistochemical staining from two DLBCL samples (i and ii) and one BL sample (iii) are shown. (B) Pie charts indicate that the majority of DLBCL samples are MCT4 negative and the relative proportion of MCT4 negative samples is similar in both activated B-cell (ABC) and germinal center B-cell (GCB) subsets. (C) MCT4 vs. MCT1 H-score plot for the DLBCL samples in relation to ABC/GCB classification.

haematologica | 2017; 102(7)

1249

R.A. Noble et al.

were also stained and found to have uniformly strong MCT1 and undetectable MCT4 protein, with the exception of weak tumor cell staining (H-score ≤ 10) in inflamed, ulcerated areas within two intestinal BL tumor samples (Figure 1A). That the majority of DLBCL and BL patients’ specimens examined had little or no evidence of MCT4 protein expression but appreciable MCT1 suggests that these are potentially appropriate malignancies in which to examine MCT1 inhibitor treatment.

cell line BJAB was not markedly reduced (Figure 2D), <10% growth inhibition being evident following exposure to 10 μM AZD3965 (n=3; data not shown). Although AZD3965 induced a profound cytostatic effect in DLBCL and BL cells without MCT4, limited effects on cell viability were detected over a 72 h period (Figure 2G,H). Furthermore, when incubation with a >GI95 (72 h data) concentration of AZD3965 was prolonged to 120 h, only a modest loss of cell viability in Raji (23%) and a <5% change in CA46 viability (Figure 2I) were detectable.

AZD3965 induces rapid accumulation of lactate in human lymphoma cell lines and significantly inhibits their growth in vitro

Consequences of AZD3965 treatment on tumor cell metabolism and efficacy in an in vivo Burkitt lymphoma model

We assembled a panel of DLBCL and BL cell lines and determined their expression of MCT1 and MCT4 protein by western blotting (Figure 2A), with confirmatory immunohistochemistry on a subset (Online Supplementary Figure S3). The DLBCL cell lines selected included Farage, OCILY18, Pfeiffer and Toledo, which are representatives of the germinal center B-cell subtype, and RIVA, an example of the activated B-cell subtype.22,23 A selection of these cells have also been characterized according to the Consensus Cluster Classification, with Farage and OCILY18 being of the B-cell receptor subtype, and Pfeiffer and Toledo the OXPHOS subtype.24,25 The BL cell lines examined comprised the Epstein-Barr virus-positive Raji and Daudi cells and Epstein-Barr virus-negative BJAB, BL41, Ramos and CA46 lines. MCT1 protein expression was detectable in each lymphoma cell line, all BL cell lines being high expressers, but more variation being evident in the DLBCL cell lines, with RIVA being a high expresser, Toledo, Pfeiffer and OCILY18 being intermediate expressers and Farage a comparatively low expresser (Figure 2A). With the exception of BJAB, all cell lines were found to be negative for MCT4 and, consistent with such cells being reliant on MCT1 for lactate transport, they accumulated lactate intracellularly (>25 μg/mg protein, P<0.05) following AZD3965 treatment (Figure 2B). Notably, neither the absolute concentration of lactate attained following AZD3965 treatment nor the magnitude of accumulation relative to basal control conditions (i.e. fold change) correlated with MCT1 protein expression. In contrast, the MCT4-positive BJAB cell line did not show a significant change in intracellular lactate following MCT1 inhibition (P=0.16). An examination of the time-dependency of lactate accumulation in CA46 cells, revealed a rapid increase in intracellular lactate within the first 30 min following treatment with 100 or 1000 nM of AZD3965 (Figure 2C). While a reduction (of 40-50%) from peak lactate levels was observed by 24 h, a concentration of approximately 50 μg of lactate per mg of protein was largely maintained between 24 and 72 h (Figure 2C) suggesting attainment of an intracellular lactate equilibrium. Although maximal lactate accumulation was not evident until 6 h following incubation with 10 nM of AZD3965, the lactate concentration at 24 and 72 h was comparable to that achieved with 100 nM and 1000 nM AZD3965 treatment (Figure 2C). However, a concentration of 100 nM AZD3965 was chosen for all further single concentration experiments, given the more rapid onset of lactate accumulation. AZD3965 potently inhibited the proliferation of DLBCL and BL cell lines in which lactate accumulation was observed (Figure 2D-F; 72 h GI50 values ranged from 3 to 39 nM). However, proliferation of the MCT4-expressing

We examined the consequences of AZD3965 treatment (2 h incubation) on cellular metabolism in three DLBCL and two BL cell lines in vitro. MCT1 inhibition increased the intracellular levels of tricarboxylic acid cycle (TCA) intermediates across a number of the AZD3965-sensitive cell lines (Figure 3A; Online Supplementary Figure S4), potentially reflecting increased activity of the TCA cycle. We also observed changes in the glycolytic pathway, including lactate accumulation and increased levels of early glycolytic intermediates, in particular glucose-6phosphate, consistent with lactate inducing feedback inhibition of phosphofructokinase.26 The reduction in fructose-bisphosphate observed in CA46 and Daudi cells, would also be predicted to reduce pyruvate kinase activity and contribute to reduced glycolytic flux. To determine whether these effects could be reproduced in vivo, we grew CA46 BL cells subcutaneously in NSG mice and harvested tumors 2 h after mice had been given a single oral dose of AZD3965 (100 mg/kg). Tumor lactate accumulation was evidenced by both biochemical assay (Figure 3B) and gas chromatography – mass spectrometry analysis (Figure 3C). Reductions in glutamate and succinate were also observed in tumors (Figure 3C). Given that intracellular tumor lactate was elevated by AZD3965 treatment, we next sought to determine the consequences of this on CA46 growth in vivo. CA46 tumor cells, engineered to ectopically express firefly luciferase, were inoculated intravenously and their growth determined by bioluminescence in vivo imaging. Cell engraftment was confirmed 6 days after inoculation, prior to commencing oral treatment with AZD3965 or vehicle. AZD3965 treatment for 24 days inhibited tumor growth by 99% (Figure 3D,E). Reduced CA46 cell engraftment in AZD3965-treated animals was also evident from a lack of human CD20 staining in spleen (Figure 3F,G) and preservation of normal spleen weight. Evidence of CD20 staining was found in only 8% (1/13) of femora recovered from AZD3965-treated mice, whereas engraftment was observed in 86% (12/14) of vehicle-treated mice (Figure 3G and data not shown).

1250

Adaptive resistance to AZD3965 in vitro involves a greater dependency on oxidative phosphorylation To determine whether an adaptive resistance to AZD3965 could be induced in vitro, CA46 cells were cultured continuously in the presence of 10 nM of the compound for 4 weeks followed by 100 nM for 3 weeks. This resulted in cells with reduced sensitivity to the anti-proliferative effects of the compound (termed CA46-R cells) (Figure 4A). Significant AZD3965-induced intracellular lactate accumulation was observed in both CA46 and CA46haematologica | 2017; 102(7)

AZD3965 treatment of lymphoma

Figure 2. MCT1 inhibition induces rapid accumulation of lactate and significant anti-proliferative activity in diffuse large B-cell lymphoma and Burkitt lymphoma cell lines. (A) MCT1 and MCT4 protein expression in cell lines using GAPDH as a loading control. (B) Intracellular lactate in cell lines following 24 h incubation with AZD3965 (1 μM) or vehicle. (C) Concentration and time dependency of intracellular lactate accumulation in CA46 cells following treatment with AZD3965 or vehicle. (D-F) Sensitivity of BL or DLBCL cells treated with AZD3965 for 72 h assessed by XTT assay. (G and H) Cell number and viability following AZD3965 (100 nM) treatment for 72 h. (I) Cell viability following an extended 120 h exposure to AZD3965 (100 nM). Graphs show the means of ≥3 independent experiments ± SEM. *P<0.05, **P<0.01, ***P<0.001 by unpaired two-tailed t-test.

haematologica | 2017; 102(7)

1251

R.A. Noble et al.

Figure 3. AZD3965 alters cellular metabolism in vitro and in vivo causing growth inhibition. (A) Levels of tricarboxylic acid (TCA) cycle and glycolytic intermediates in cell lines following 2 h exposure to AZD3965 (100 nM) determined by liquid chromatography-mass spectrometry. Significantly altered metabolites (P<0.05) are expressed as log2 fold-change relative to vehicle-treated control. ÎąKG: alpha-ketoglutarate; FBP: fructose-bisphosphate; F1P: fructose-1-phosphate; F6P: fructose-6phosphate; GAP: glyceraldehyde-3-phosphate; G1P: glucose-1-phosphate; G6P: glucose-6-phosphate. (B) NSG mice with subcutaneous CA46 xenografts were treated with AZD3965 (100 mg/kg) or vehicle and tumors collected after 2 h. Lactate concentrations were normalized to protein. (C) Significantly altered (unpaired twotailed t-test) intra-tumoral metabolite levels determined by gas chromatography-mass spectrometry. (D) NSG mice were inoculated intravenously with luciferaseexpressing CA46 cells and 6 days later (treatment day 0) treated with AZD3965 (100 mg/kg, BID) or vehicle for 24 days. Representative images from two mice in the AZD3965 and vehicle-treated groups using different radiance scales (p/sec/cm2/sr) for mice prior to treatment and during treatment to avoid image saturation. (E) Mean total flux from AZD3965 and vehicle-treated mice (n=8 per group). (F) Spleen weights from AZD3965 and vehicle-treated mice. Reference historical spleen weights from NSG mice were 0.02â&#x20AC;&#x201C;0.05 g.19 (G) Immunohistochemical analysis of CA46 infiltration via anti-CD20 staining of bone marrow and spleen sections from mice treated with AZD3965 or vehicle. Statistical significance was assessed by an unpaired two-tailed t-test *P<0.05, ***P<0.001.

1252

haematologica | 2017; 102(7)

AZD3965 treatment of lymphoma

R cells and, although the concentration of lactate in CA46R was 28% less than in the parental cell line, the level attained was comparable to or greater than that achieved in other AZD3965-sensitive lines following drug treatment, including Daudi, Toledo, and Pfeiffer cells. There was no evidence of MCT4 being expressed as a potential

compensatory mechanism to mediate lactate efflux (Figure 4B). In contrast, the levels of MCT1 and its co-chaperone CD147 were lower in CA46-R cells, suggesting that the level of functional MCT1 may be reduced (Figure 4B). The doubling times of CA46 and CA46-R were comparable (Online Supplementary Figure S5A) and resistance was main-

haematologica | 2017; 102(7)

Figure 4. Acquired resistance to AZD3965 in vitro is associated with increased oxidative metabolism. (A) The sensitivity of CA46 and CA46-R cells to AZD3965 (72 h treatment) determined by an XTT assay and cell counting. (B) Intracellular accumulation of lactate determined after 24 h exposure to AZD3965 (1 ÎźM). MCT1, MCT4 and CD147 protein levels assessed by western blotting. (C) Extracellular acidification rate (ECAR) in CA46 and CA46R with and without treatment with AZD3965 (100 nM) or vehicle. Oxygen consumption rate (OCR) in CA46 and CA46-R cells, indicating the effects following addition of oligomycin, FCCP and antimycin. ECAR and OCR values (mean Âą SEM) are normalized to protein expression and representative of three independent experiments.

1253

R.A. Noble et al.

tained following culture in drug-free medium for 2 weeks (data not shown). To explore an altered metabolic phenotype, the relative consumption and release of metabolic substrates was assessed following 24 h of AZD3965 treatment. CA46-R showed increased glutamine uptake, decreased lactate release and increased pyruvate export (Online Supplementary Figure S5B), changes consistent with reduced glycolytic lactate production and increased glucose and glutamine oxidation with which to fuel TCA cycle activity. We also examined the respective contributions of glycolysis and OXPHOS in CA46 and CA46-R cells. Acute exposure to AZD3965 triggered a rapid decrease in extracellular acidification rate in CA46 cells but not in CA46-R cells which demonstrated a lower basal extracellular acidification rate (Figure 4C). CA46 and CA46-R differed markedly in their basal oxygen consumption rate, with CA46-R utilizing more oxygen (Figure 4C). Collectively, these measurements are indicative of CA46-R cells having a more oxidative metabolic phenotype (additional details are available in the Online Supplementary Information).

Combining AZD3965 with inhibitors of complex I triggers cell death Since inhibition of glycolysis will generate greater reliance on OXPHOS for ATP generation, simultaneous inhibition of mitochondrial complex I may trigger cell death,12 a phenomenon demonstrated in Raji cells by combining AZD3965 with metformin (Figure 5A). However, the concentrations of metformin required to demonstrate this effect were significantly in excess of those that can be achieved in mice following oral dosing.27 We therefore examined the potent complex I inhibitor BAY 87-2243, which would subsequently permit inhibition of OXPHOS to be studied in mice.28 While the BL cell lines, Raji and CA46 were insensitive to BAY 87-2243 monotherapy in vitro, the combination of AZD3965 with BAY 87-2243 induced profound cell death in both cell lines (Figure 5B,C). CA46-R cells were more sensitive to the growth inhibitory effect of BAY 87-2243 than parental CA46 cells, but a combination of BAY 87-2243 and AZD3965 was similarly required to induce cell death (Figure 5D). In contrast, cell death was not evident when AZD3965 was combined with BAY 87-2243 in MCT4 protein-expressing BJAB cells (data not shown).

Combining AZD3965 with BAY 87-2243 in vivo To examine MCT1 inhibition combined with complex I inhibition in vivo, mice were inoculated intravenously with luciferase-expressing CA46 cells (Figure 6A). Mean tumor engraftment was equivalent 12 days after inoculation (P>0.05 for all group comparisons by two-way ANOVA; Figure 6B), prior to dosing. Following treatment (72 h after the last dose) mice were re-imaged and AZD3965 monotherapy treatment again resulted in significant control of tumor growth, mean tumor volume being not significantly different from pre-treatment values (Figure 6B; P>0.05 by the two-tailed t-test). While tumor burden in control or BAY 87-2243-treated animals had increased markedly from each pre-treatment value, the combination of AZD3965 with BAY 87-2243 led to a reduction in mean tumor burden compared with the pretreated value in four of five mice, with signal intensities being reduced by between 10- to 267-fold (Figure 6B; P=0.01 by the Pearson Ď&#x2021;2 test). This is consistent with the 1254

induction of lymphoma cell death in vivo. The combination treatment regimen was well tolerated with no significant differences in body weight when compared to that in the control vehicle-treated group throughout treatment (data not shown). All tumor inhibitory effects were lost upon cessation of treatment as engraftment progressed in each drug-treated group, indicating that targeting of metabolism is likely to require continuous therapy (Figure 6C).

Discussion This study aimed to evaluate DLBCL and BL as potential tumor types appropriate for the clinical development of AZD3965, a novel MCT1 inhibitor currently undergoing phase I evaluation. Although a glycolytic phenotype and increased generation of lactate are implicated in the pathogenesis of both DLBCL and BL, the relative expression of MCT1 versus MCT4 in DLBCL has been less clear. A previous study examining clinical gene expression data confirmed high expression of MCT1 mRNA and low expression of MCT4 mRNA in BL but suggested that the converse was true in a cohort of non-Hodgkin lymphomas that would have contained predominantly DLBCL samples.12 Our examination of MCT1 and MCT4 protein using immunohistochemistry showed uniformly strong MCT1 staining in BL with a corresponding lack of MCT4. However, our analysis also indicated that the majority of DLBCL does not stain positive for MCT4 protein. DLBCL samples without MCT4 protein expression were observed among both cellof-origin subgroups, and in groups with and without any MYC aberration. This suggests that all major DLBCL subgroups contain patients with an MCT1-positive/MCT4negative protein expression profile, who may be appropriate candidates to receive AZD3965 treatment. Additional experiments examining neuroblastoma cell lines (IMR-32 and SH-SY5Y; Online Supplementary Figure S6) and a previous study on small cell lung cancer cells7 demonstrated that only partial sensitivity to AZD3965 can be observed in solid tumor cell lines lacking MCT4 expression, despite significant lactate accumulation. Solid tumor cell lines also have a much broader spectrum of MCT4 mRNA expression in comparison to those of hematologic origin (Online Supplementary Figure S1). Although these data do not exclude AZD3965 as being suitable for the treatment of a subset of solid tumors, the more potent GI50 and uniform response to AZD3965 treatment observed in DLBCL and BL cell lines suggest that these B-cell malignancies are better disease indications in which to initially examine the clinical activity of AZD3965. Importantly, MCT1 protein expression per se does not clearly correlate with the extent of lactate accumulation or growth inhibition observed in vitro following AZD3965 treatment. Prospective stratification of patients for AZD3965 treatment should prioritize the treatment of patients whose tumors do not stain positive for MCT4 protein, rather than treatment based upon the magnitude of MCT1 protein expression. The lack of effect of AZD3965 on both intracellular lactate concentration and growth of the MCT4-expressing BJAB lymphoma cell line is consistent with previous data showing that MCT4 overexpression in a breast cancer cell line or RAS-transformed fibroblasts confers resistance to MCT1 inhibitor treatment.12,29 haematologica | 2017; 102(7)

AZD3965 treatment of lymphoma

The effect of inhibiting MCT1 in DLBCL and BL cell lines in vitro was predominantly cytostatic. Encouragingly, however, we also observed a striking cytostatic response in vivo with AZD3965 treatment, which halted progressive splenic engraftment of the lymphoma. The incidence of BL is increased in elderly patients and the median age of

diagnosis for DLBCL is around 70 years.30,31 Given that not all of these patients will be fit enough to tolerate multiagent chemo-immunotherapy, a well-tolerated oral cytostatic therapy could have significant clinical utility in this group of patients or in individuals who have relapsed following current standard-of-care treatment.

Figure 5. Combining AZD3965 with inhibitors of mitochondrial complex I induces death of Burkitt lymphoma cells. Viable cell numbers were determined by cell counting with trypan blue exclusion over a 72 h period, following treatment with AZD3965, a complex I inhibitor, or the combination. (A) Raji cells treated with vehicle, AZD3965 (100 nM), metformin (1 mM) or the combination. (B) Raji cells treated with vehicle, AZD3965 (5 nM), BAY 872243 (100 nM) or the combination. (C) CA46 and (D) CA46-R cells treated with vehicle, AZD3965 (10 nM), BAY 87-2243 (10 nM) or the combination. All graphs show the means of â&#x2030;Ľ3 independent experiments Âą SEM.

haematologica | 2017; 102(7)

1255

R.A. Noble et al. A

Figure 6. Combining AZD3965 with an inhibitor of mitochondrial complex I in vivo. (A) Schema indicating treatment duration and scan intervals. (B) Pre- and post-treatment bioluminescent signals for mice within each group with a representative image from one of the mice that received the combination (inset). (C) Change in signal intensity subsequent to treatment. Graph shows the mean + SD total flux (n â&#x2030;Ľ5 per group).

The compensatory metabolic alterations observed after AZD3965 treatment in vitro, following either acute or chronic exposure, support an adaptive metabolic response that causes a greater reliance on OXPHOS and increased TCA activity. In order to elicit tumor cell death, when targeting the glycolytic tumor phenotype it may be necessary to inhibit multiple metabolic pathways, or nodes in a given pathway, ensuring that there is a basis for tumor selectivity with at least one of the approaches used. Our data confirm that a combination of AZD3965 with different complex I inhibitors induces rapid cell death in BL (Figure 5A-D) and DLBCL (data not shown) cell lines in vitro. We also verified that such a combination can reduce CA46 disease burden in vivo, in contrast to the cytostatic effect induced by administration of AZD3965 alone. Approaches to induce tumor cell death and impart curative activity would be particularly desirable in the treat1256

ment of children, in whom BL accounts for the majority of non-Hodgkin lymphoma.32 Collectively, the striking activity of AZD3965 monotherapy observed in DLBCL and BL cell lines, and its potential for use in combination, provides a rationale for examining the efficacy of this agent against these malignancies. Acknowledgments This work was supported by the Biotechnology and Biological Sciences Research Council [BB/K501700/1; 1247227]. Additional funding was provided by AstraZeneca and Cancer Research UK. Lindsay Rai-Rowcroft and Hilary Lewis (AstraZeneca) provided technical assistance with metabolomic studies. David Thwaites (Newcastle University) provided technical advice, Ross Maxwell and Ian Wilson (Newcastle University) further technical support and Elizabeth Want (Imperial College London) additional supervision. The IVIS spectrum was funded by the Wellcome Trust [087961]. haematologica | 2017; 102(7)

AZD3965 treatment of lymphoma

References 1. Warburg O. On the origin of cancer cells. Science. 1956;123(3191):309-314. 2. Lunt SY, Vander Heiden MG. Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu Rev Cell Dev Biol. 2011;27:441-464. 3. Martinez-Outschoorn UE, Peiris-Pages M, Pestell RG, Sotgia F, Lisanti MP. Cancer metabolism: a therapeutic perspective. Nat Rev Clin Oncol. 2017;14(1):11-31. 4. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144(5):646-674. 5. Halestrap AP. The monocarboxylate transporter family-structure and functional characterization. IUBMB Life. 2012;64(1):1-9. 6. Kirk P, Wilson MC, Heddle C, et al. CD147 is tightly associated with lactate transporters MCT1 and MCT4 and facilitates their cell surface expression. EMBO J. 2000;19(15): 3896-3904. 7. Polanski R, Hodgkinson CL, Fusi A, et al. Activity of the monocarboxylate transporter 1 inhibitor AZD3965 in small cell lung cancer. Clin Cancer Res. 2014;20(4):926-937. 8. Kim Y, Choi J-W, Lee J-H, Kim Y-S. Expression of lactate/H+ symporters MCT1 and MCT4 and their chaperone CD147 predicts tumor progression in clear cell renal cell carcinoma: immunohistochemical and The Cancer Genome Atlas data analyses. Hum Pathol. 2015;46(1):104-112. 9. Pinheiro C, Albergaria A, Paredes J, et al. Monocarboxylate transporter 1 is up-regulated in basal-like breast carcinoma. Histopathology. 2010;56(7):860-867. 10. Baek G, Tse YF, Hu Z, et al. MCT4 defines a glycolytic subtype of pancreatic cancer with poor prognosis and unique metabolic dependencies. Cell Rep. 2014;9(6):22332249. 11. Pertega-Gomes N, Vizcaino JR, MirandaGoncalves V, et al. Monocarboxylate transporter 4 (MCT4) and CD147 overexpression is associated with poor prognosis in prostate cancer. BMC Cancer. 2011;11:312. 12. Doherty JR, Yang C, Scott KE, et al. Blocking lactate export by inhibiting the Myc target MCT1 disables glycolysis and glutathione

haematologica | 2017; 102(7)

synthesis. Cancer Res. 2014;74(3):908-920. 13. Sonveaux P, Copetti T, De Saedeleer CJ, et al. Targeting the lactate transporter MCT1 in endothelial cells inhibits lactate-induced HIF-1 activation and tumor angiogenesis. PLoS One. 2012;7(3):e33418. 14. Bola BM, Chadwick AL, Michopoulos F, et al. Inhibition of monocarboxylate transporter-1 (MCT1) by AZD3965 enhances radiosensitivity by reducing lactate transport. Mol Cancer Ther. 2014;13(12):28052816. 15. Pfeifer M, Zheng B, Erdmann T, et al. AntiCD22 and anti-CD79B antibody drug conjugates are active in different molecular diffuse large B-cell lymphoma subtypes. Leukemia. 2015;29(7):1578-1586. 16. Culpin RE, Sieniawski M, Angus B, et al. Prognostic significance of immunohistochemistry-based markers and algorithms in immunochemotherapy-treated diffuse large B cell lymphoma patients. Histopathology. 2013;63(6):788-801. 17. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of the molecular classification of diffuse large B-cell lymphoma by immunohistochemistry using a tissue microarray. Blood. 2004;103(1):275-282. 18. Scherr M, Elder A, Battmer K, et al. Differential expression of miR-17~92 identifies BCL2 as a therapeutic target in BCRABL-positive B-lineage acute lymphoblastic leukemia. Leukemia. 2014;28(3):554-565. 19. Bomken S, Buechler L, Rehe K, et al. Lentiviral marking of patient-derived acute lymphoblastic leukaemic cells allows in vivo tracking of disease progression. Leukemia. 2013;27(3):718-721. 20. Pal D, Blair HJ, Elder A, et al. Long-term in vitro maintenance of clonal abundance and leukaemia-initiating potential in acute lymphoblastic leukaemia. Leukemia. 2016;30(8): 1691-1700. 21. Barretina J, Caponigro G, Stransky N, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483(7391):603-607. 22. Bradley WD, Arora S, Busby J, et al. EZH2 inhibitor efficacy in non-Hodgkin's lymphoma does not require suppression of H3K27 monomethylation. Chem Biol. 2014;21(11):1463-1475.

23. Cheng S, Coffey G, Zhang XH, et al. SYK inhibition and response prediction in diffuse large B-cell lymphoma. Blood. 2011;118(24): 6342-6352. 24. Polo JM, Juszczynski P, Monti S, et al. Transcriptional signature with differential expression of BCL6 target genes accurately identifies BCL6-dependent diffuse large B cell lymphomas. Proc Natl Acad Sci USA. 2007;104(9):3207-3212. 25. Chen L, Monti S, Juszczynski P, et al. SYKdependent tonic B-cell receptor signaling is a rational treatment target in diffuse large Bcell lymphoma. Blood. 2008;111(4):22302237. 26. Costa Leite T, Da Silva D, Guimaraes Coelho R, Zancan P, Sola-Penna M. Lactate favours the dissociation of skeletal muscle 6phosphofructo-1-kinase tetramers downregulating the enzyme and muscle glycolysis. Biochem J. 2007;408(1):123-130. 27. Dowling RJ, Lam S, Bassi C, et al. Metformin pharmacokinetics in mouse tumors: implications for human therapy. Cell Metab. 2016;23(4):567-568. 28. Ellinghaus P, Heisler I, Unterschemmann K, et al. BAY 87-2243, a highly potent and selective inhibitor of hypoxia-induced gene activation has antitumor activities by inhibition of mitochondrial complex I. Cancer Med. 2013;2(5):611-624. 29. Le Floch R, Chiche J, Marchiq I, et al. CD147 subunit of lactate/H+ symporters MCT1 and hypoxia-inducible MCT4 is critical for energetics and growth of glycolytic tumors. Proc Natl Acad Sci USA. 2011;108(40): 16663-16668. 30. Winkelmann N, Wedding U. Diffuse large Bcell non-Hodgkinâ&#x20AC;&#x2122;s lymphoma (DLBCLNHL). In: Wedding U, Audisio RA, eds. Management of Hematological Cancer in Older People. London: Springer; 2015; p.185-202. 31. Mbulaiteye SM, Anderson WF, Ferlay J, et al. Pediatric, elderly, and emerging adult-onset peaks in Burkitt's lymphoma incidence diagnosed in four continents, excluding Africa. Am J Hematol. 2012;87(6):573-578. 32. Hochberg J, Waxman IM, Kelly KM, Morris E, Cairo MS. Adolescent non-Hodgkin lymphoma and Hodgkin lymphoma: state of the science. Br J Haematol. 2009;144(1):24-40.

1257

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Non-Hodgkin Lymphoma

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1258-1265

Development of a modified prognostic index for patients with aggressive adult T-cell leukemia-lymphoma aged 70 years or younger: possible risk-adapted management strategies including allogeneic transplantation Shigeo Fuji,1,2 Takuhiro Yamaguchi,3 Yoshitaka Inoue,1,2,4 Atae Utsunomiya,5 Yukiyoshi Moriuchi,6 Kaoru Uchimaru,7 Satsuki Owatari,8 Takashi Miyagi,9 Jun Taguchi,10 Ilseung Choi,11 Eiichi Otsuka,12 Sawako Nakachi,13 Hisashi Yamamoto,14 Saiko Kurosawa,1 Kensei Tobinai2,15 and Takahiro Fukuda10

Department of Hematopoietic Stem Cell Transplantation, National Cancer Center Hospital, Tokyo; 2Juntendo University Graduate School of Medicine, Tokyo; 3Division of Biostatistics, Tohoku University Graduate School of Medicine, Sendai; 4Department of Hematology, Kumamoto University Hospital; 5Department of Hematology, Imamura Bunin Hospital, Kagoshima; 6Department of Hematology, Sasebo City General Hospital; 7 Department of Hematology/Oncology, Institute of Medical Science, The University of Tokyo; 8Department of Hematology, National Hospital Organization Kagoshima Medical Center; 9Department of Hematology, Heart-Life Hospital, Okinawa; 10Department of Hematology, Nagasaki University Hospital; 11Department of Hematology, National Hospital Organization Kyushu Cancer Center, Fukuoka; 12Department of Hematology, Oita Prefectural Hospital; 13Second Department of Internal Medicine, University of the Ryukyus, Okinawa; 14Department of Hematology, Toranomon Hospital, Tokyo and 15 Department of Hematology, National Cancer Center Hospital, Tokyo, Japan 1

ABSTRACT

Correspondence: s-fuji@pj8.so-net.ne.jp

Received: January 23, 2017. Accepted: March 17, 2017. Pre-published: March 24, 2017. doi:10.3324/haematol.2017.164996 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1258 ©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.

1258

dult T-cell leukemia-lymphoma is a distinct type of peripheral Tcell lymphoma caused by human T-cell lymphotropic virus type I. Although allogeneic stem cell transplantation after chemotherapy is a recommended treatment option for patients with aggressive adult T-cell leukemia-lymphoma, there is no consensus about indications for allogeneic stem cell transplantation because there is no established risk stratification system for transplant eligible patients. We conducted a nationwide survey of patients with aggressive adult T-cell leukemia-lymphoma in order to construct a new, large database that includes 1,792 patients aged 70 years or younger with aggressive adult T-cell leukemialymphoma who were diagnosed between 2000 and 2013 and received intensive first-line chemotherapy. We randomly divided patients into two groups (training and validation sets). Acute type, poor performance status, high soluble interleukin-2 receptor levels (> 5,000 U/mL), high adjusted calcium levels (≥ 12 mg/dL), and high C-reactive protein levels (≥ 2.5 mg/dL) were independent adverse prognostic factors used in the training set. We used these five variables to divide patients into three risk groups. In the validation set, median overall survival for the low-, intermediate-, and high-risk groups was 626 days, 322 days, and 197 days, respectively. In the intermediate- and high-risk groups, transplanted recipients had significantly better overall survival than non-transplanted patients. We developed a promising new risk stratification system to identify patients aged 70 years or younger with aggressive adult T-cell leukemia-lymphoma who may benefit from upfront allogeneic stem cell transplantation. Prospective studies are warranted to confirm the benefit of this treatment strategy. haematologica | 2017; 102(7)

Modified prognostic index for aggressive ATL

Introduction Adult T-cell leukemia-lymphoma (ATL) is a distinct type of peripheral T-cell lymphoma caused by human T-cell lymphotropic virus type I (HTLV-1).1 Patients with aggressive ATL such as the acute or lymphoma subtype have dismal outcomes, even with intensive chemotherapy.2-6 Allogeneic hematopoietic stem cell transplantation (alloHSCT) is a promising treatment option for patients with aggressive ATL.7-9 However, there is still no consensus on whether all patients with aggressive ATL should undergo upfront allo-HSCT, because there is no direct comparison of clinical outcomes among non-transplanted and transplanted patients using one large database. In addition, risk stratification of aggressive ATL in transplant eligible patients is not yet well established, mostly due to a lack of prospective randomized studies. A retrospective study of 807 patients in Japan described a prognostic index for acute and lymphoma type ATL (ATL-PI) that included stage, Eastern Cooperative Oncology Group performance status (ECOG PS), age, albumin, and soluble interleukin-2 receptor (sIL-2R) level.3 However, in that study, allo-HSCT recipients were excluded to establish ATL-PI, and a large proportion of patients older than 70 years, who are usually not candidates for allo-HSCT, were included. In patients with acute myeloid leukemia (AML), analysis of cytogenetic abnormalities and specific genes is widely used to improve risk stratification and identify patients who can benefit from upfront allo-HSCT. Such a prognostication system is needed for patients with ATL who are transplant eligible in order to reasonably consider the use of upfront allo-HSCT. Herein, we aim to develop a new prognostic index in patients with aggressive ATL aged 70 years or younger using a large database of 1,191 non-transplanted patients and 601 allo-HSCT recipients, thus making it possible to assess the impact of allo-HSCT in each risk group within this single database.

Methods Data Source We conducted a nationwide survey of patients with aggressive ATL to construct a new large database. This study was approved by the institutional review board of the National Cancer Center in Tokyo, Japan (No. 2014-179). First, we invited 232 hospitals with a department of hematology in Japan to complete a questionnaire; 99 hospitals returned the questionnaire to the data center. We included patients aged 70 years or younger with aggressive ATL (acute and lymphoma type ATL) who were diagnosed between 2000 and 2013 and received intensive chemotherapy with multiple chemotherapeutic drugs as first-line therapy. We only included patients who received intensive chemotherapy as first-line therapy because ATL patients who are not candidates for intensive chemotherapy are usually not candidates for allo-HSCT. In this study, we defined intensive chemotherapy as chemotherapeutic regimens including at least two intravenous cytotoxic chemotherapeutic drugs. The information about primary induction therapy is shown in the Online Supplementary Table S1. This database included the same cohort of patients who received allo-HSCT as in our previous analysis.10 We also expected that some of the patients in this database were also included in previous national surveys.3,6,7 In this study, the upper age limit was defined as 70 years, as recently the indication of allo-HSCT has been broadened to haematologica | 2017; 102(7)

include patients of age 70 years or above.11,12 As it is still uncommon in Japan that patients aged above 70 years receive allo-HSCT, we set the upper limit of age at 70 years.

Statistical analysis Descriptive statistical analysis was performed to assess the patientsâ&#x20AC;&#x2122; characteristics. Medians and ranges are provided for continuous variables, and percentages are shown for categorical variables. The probability of overall survival (OS) was calculated using the Kaplan-Meier method from date of diagnosis to date of death, or from 180 days after diagnosis to date of death in a landmark analysis. Initially, patients who underwent allo-HSCT were censored on the day of allo-HSCT when developing the prognostic index (PI) in order to reduce the impact of allo-HSCT on OS. We also assessed whether the PI works even when patients who underwent allo-HSCT were not censored on the day of alloHSCT. In the analysis assessing the impact of allo-HSCT, we did not censor allo-HSCT. A Cox proportional hazards regression model was used to analyze OS. The cumulative incidence of relapse or progression was evaluated using the Fine and Gray model in univariate and multivariate analyses.13 In the competingrisks models for relapse and progression, death before these events was defined as a competing risk. Variables included in the analysis were sex, age, clinical subtype (acute or lymphoma type), ECOG PS, Ann Arbor stage, and laboratory data at diagnosis including white blood cell (WBC) count, and levels of serum albumin, blood urea nitrogen (BUN), sIL-2R, adjusted serum calcium (Ca), and C-reactive protein (CRP). Classification of clinical subtype was based on a previously report-

Table 1. Baseline Characteristics of All Patients (n=1,792).

Variable

Age, years median, range 60 (20-70) Sex Female 803 Male 986 Subtype Acute type 1,259 Lymphoma type 533 6 White blood cell count, x 10 /L median, range 9,800 (700-726,000) Serum albumin, g/dL median, range 3.7 (1.6-5.7) Blood urea nitrogen, mg/dL median, range 14 (3-109) Soluble IL2R, U/mL median, range 19,400 (237-3,330,170) Adjusted calcium, mg/dL median, range 9.6 (7.0-26.0) C-reactive protein, mg/dL median, range 0.73 (0.00-34.92) Ann Arbor stage I-II 129 III-IV 1,663 ECOG PS 0-1 1,130 2-4 662 Allogeneic Transplant performed 601

44.9 55.1 70.3 29.7

7.2 92.8 63.1 36.9 33.5

IL2R: interleukine-2 receptor; ECOG PS: Eastern Cooperative Oncology Group performance status.

1259

S. Fuji et al. ed classification system.1,14 We included sIL-2R level in this model, as sIL-2R is a commonly used biomarker in clinical practice for ATL in Japan, and was demonstrated to be useful as a prognostic factor of patients with ATL.3,15,16 The data set was randomly divided into two groups (training and validation sets). The training and validation sets included 907 and 885 patients, respectively. For continuous variables, a cubic spline model with five knots at the 5th, 25th, 50th, 75th, and 95th percentiles was applied prior to data analysis. We evaluated the association between each variable and OS and decided on optimal cutoffs to make the scoring system clinically appropriate, practicable, and easy to use. We then applied a backward elimination variable selection algorithm, and retained only those variables that contained at least one statistically significant category (P<0.05) in the final model. Integer weights for the scoring system were derived from a Cox proportional hazards model applied to the training set. The PI score was the sum of these weights. Finally, PI scores were grouped based on Akaike’s information criterion (AIC). The PI scores were applied to the validation set. To assess the discriminatory capability of the PI score for three-year OS, we computed the c-statistic and Predicted SEParation (PSEP).17,18 All statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

probability of three-year OS was 34.6% (95% CI, 25.5% to 43.9%) in the low-risk group, 18.7% in the intermediate-risk group (95% CI, 14.2% to 23.7%), and 0.0% in the high-risk group (95% CI, 0.0% to 0.0%) (P<0.0001; chisquare, 93.18; log-rank test).

Validation of the Prognostic Index in the Validation Set Patients in the validation set were grouped according to the score from the new PI. As shown in Figure 2A, good separation in the probability of OS was achieved with this PI (P<0.0001; χ2 74.53; AIC, 5,285.12; area under curve [AUC], 0.93 [95% CI, 0.88 to 0.97]; PSEP of 3-year OS, 0.28). Median OS was 626 days (95% CI, 518 to 820 days) in the low risk-group, 322 days (95% CI, 291 to 409 days) in the intermediate-risk group, and 197 days (95% CI, 171 to 278 days) in the high-risk group. The probability of three-year OS was 32.6% (95% CI, 23.8% to 41.7%) in the low-risk group, 18.5% (95% CI, 13.9% to 23.5%) in the intermediate-risk group, and 6.0% in the high-risk

Results

P<0.0001

Patients’ characteristics In total, data from 2,703 patients were obtained from 99 institutions across Japan. Patients who did not meet the inclusion criteria, e.g., >70 years old, no intensive chemotherapy, chronic type ATL, diagnosis in 2014, and lack of data about clinical outcome or variables included in the analyses, were excluded. Consequently, a total of 1,792 patients were included in further analyses. Patients’ characteristics are shown in Table 1. The median age was 60 years (range, 20–70 years), and the median follow up of surviving patients was 1,003 days after diagnosis (range, 7 to 5,302 days). The median OS was 346 days (three-year OS, 19.6%, 95% confidence interval (CI), 17.0 to 22.3).

Development of a Prognostic Index with the Training Set We randomly assigned patients into either the training or validation set. We performed univariate analyses with the variables mentioned above. As shown in Table 2, clinical subtype (lymphoma vs. acute), ECOG PS (0–1 vs. 2–4), sIL-2R (> 5,000 U/mL), adjusted Ca (≥ 12 mg/dL), and CRP (≥ 2.5 mg/dL) were retained as independent prognostic factors in a multivariate analysis using the training set. Based on the regression coefficients for these variables, each variable was assigned a weight of 1 and the PI score was defined as the number of following characteristics present: clinical subtype acute type, ECOG PS 2-4, adjusted Ca ≥ 12 mg/dL, CRP ≥ 2.5 mg/dL and sIL-2R > 5,000 U/mL. OS rates stratified by PI score are shown in Figure 1A. On the basis of the best discrimination according to the AIC in the training set, scores of 0 and 1 were categorized into the low-risk group, 2 and 3 into the intermediate-risk group, and 4 and 5 into the high-risk group. Kaplan-Meier survival curves of the training set grouped by PI are shown in Figure 1B. The median survival was 562 days (95% CI, 436 to 867 days) in the low-risk group, 337 days (95% CI, 307 to 378 days) in the intermediate-risk group, and 206 days (95% CI, 166 to 225 days) in the high-risk group. The 1260

P<0.0001

Figure 1. Overall survival rates in the training set. Overall survival stratified by (A) prognostic index score and (B) prognostic index risk group in the training set. The score was defined as the number of the following characteristics present: acute type, ECOG PS 2–4, adjusted Ca ≥ 12 mg/dL, CRP ≥ 2.5 mg/dL, and soluble interleukin-2 receptor > 5,000 U/mL. ECOG PS: Eastern Cooperative Oncology Group performance status; CRP: C-reactive protein; Ca: calcium.

haematologica | 2017; 102(7)

Modified prognostic index for aggressive ATL

Table 2. Independent Prognostic Factors in a Multivariate Analysis of Overall Survival with a Training Set (n = 907).

Variable

95%CI

Clinical subtype acute type (vs. lymphoma type) ECOG PS 2-4 (vs. 0–1) Adjusted Ca level ≥12 mg/dL (vs.<12 mg/dL) CRP level ≥ 2.5 mg/dL (vs. <2.5 mg/dL) Soluble interleukin-2 receptor > 5,000 U/mL (vs. ≤ 5,000 U/mL)

1.43 1.39 1.50 1.44 1.58

1.15–1.77 1.15–1.68 1.19–1.88 1.18–1.77 1.21–2.06

0.001 0.001 0.001 <0.001 0.001

Variables included in the analysis were as follows: sex, age, clinical subtype (acute or lymphoma type), ECOG PS, Ann Arbor stage and laboratory data at diagnosis including white blood cell counts, serum albumin, blood urea nitrogen, sIL-2R, adjusted serum calcium level and C-reactive protein. HR: hazard ratio; CI: confidence interval; ECOG PS: Eastern Cooperative Oncology Group performance status; Ca: calcium; CRP: C-reactive protein.

group (95% CI, 1.8% to 13.9%). When we did not censor allo-HSCT, median OS was 622 days (95% CI, 485 to 748 days) in the low-risk group, 326 days (95% CI, 296 to 389 days) in the intermediate-risk group, and 208 days (95% CI, 171 to 249 days) in the high-risk group (Figure 2B). The probability of three-year OS was 36.0% (95% CI, 29.2% to 42.9%) in the low-risk group, 22.9% (95% CI, 19.2% to 26.9%) in the intermediate-risk group, and 7.4% in the high-risk group (95% CI, 3.8% to 12.6%). As shown in Figure 2B, good separation was also achieved with this PI (P<0.0001; χ2, 79.90; AIC, 8,143.64; AUC, 0.91 [95% CI, 0.86 to 0.95]; PSEP for three-year OS, 0.28). When the ATL-PI proposed by Katsuya et al. was used,3 median OS was 622 days (95% CI, 460 to 796 days) in the low-risk group, 306 days (95% CI, 291 to 371 days) in the intermediate-risk group, and 182 days (95% CI, 153 to 262 days) in the high-risk group. The probability of three-year OS was 33.5% (95% CI, 25.7% to 41.5%) in the low-risk group, 15.1% (95% CI, 10.9% to 20.0%) in the intermediate-risk group, and not available in the high-risk group. As shown in Figure 2C, good separation was also achieved with ATL-PI (P<0.0001; χ2, 70.31; AIC, 5,288.77; AUC, 0.93 [95% CI, 0.88 to 0.97]; PSEP of three-year OS, 0.22). However, the separation was better with the new PI than with the original ATL-PI, as the new PI was the model with a higher PSEP of three-year OS and a slightly lower AIC than the original ATL-PI.

Relapse or Progression The median time to relapse or progression was 269 days in the low-risk group, 182 days in the intermediate-risk group, and 144 days in the high-risk group in the entire cohort (Figure 2D). The cumulative incidence of threeyear relapse or progression was 78.1% (95% CI, 72.5% to 82.7%) in the low-risk group, 86.3% (95% CI, 83.2% to 88.8%) in the intermediate-risk group, and 92.6% (95% CI, 87.7% to 95.6%) in the high-risk group (P<0.0001).

Impact of Allo-HSCT on the OS in Each Risk Group In order to assess the impact of allo-HSCT in this database, we compared the clinical outcomes between the transplanted and non-transplanted patients. Patient characteristics at diagnosis and the best response to primary induction therapy comparing the transplanted and nontransplanted groups are shown in the Online Supplementary Table S2. Of the 601 patients who received allo-HSCT, information about the intensity of the conditioning regimen was available for 592 of them. Out of 592 patients, 218 patients (36.8%) received a myeloablative conditioning regimen and 374 patients (63.2%) received a reduced haematologica | 2017; 102(7)

intensity conditioning regimen. In terms of stem cell source, 160 patients (27.0%) received stem cells from a human leukocyte antigen (HLA)-matched related donor, 258 (43.6%) from an unrelated volunteer donor, 136 (23.0%) from cord blood and 46 (7.8%) from a HLA-mismatched related donor. Patients in the transplanted group had favorable patient characteristics compared to those in the non-transplant group. In terms of the response to primary induction therapy, patients in the transplanted group had better response rates than those in the non-transplanted group. Stratified according to their response to primary induction therapy, patients in the transplanted group had a significantly better OS compared to those in the nontransplanted group; patients who achieved complete remission (CR) or partial remission (PR) by primary induction therapy (three-year OS 40.7% in transplanted patients and 22.3% in non-transplanted patients, P<0.0001) as shown in the Online Supplementary Figure S1A. Kaplan-Meier estimates of OS based on a landmark analysis at 6 months after diagnosis in patients who achieved CR or PR by primary induction therapy were also assessed, as the median interval from diagnosis to allo-HSCT was approximately 6 months in this cohort. The probability of three-year OS was 43.1% (95% CI, 38.2% to 47.8%) in transplanted patients and 26.1% (95% CI, 22.3% to 30.0%) in non-transplanted patients (P<0.0001, Online Supplementary Figure S1B). Stratified by risk group at diagnosis, the probability of three-year OS in transplanted and non-transplanted patients was 41.4% (95% CI, 33.6% to 49.1%) and 29.2% (95% CI, 23.2% to 35.4%) in the low-risk group, 39.7% (95% CI, 34.4% to 45.0%) and 13.6% (95% CI, 10.9% to 16.6%) in the intermediate-risk group, and 26.7% (95% CI, 17.1% to 37.3%) and 1.4% (95% CI, 0.3% to 4.5%) in the high-risk group, respectively (Figure 3A-C). In the intermediate- and high-risk groups, allo-HSCT was a statistically significant favorable prognostic factor for OS when treated as a time-dependent variable based on the time from date of diagnosis to transplantation and adjusted for the five prognostic factors of the new PI: subtype of ATL, CRP level, ECOG PS, sIL-2R, and adjusted Ca level (low-risk group: hazard ratio (HR), 1.11; 95% CI, 0.86 to 1.43; P=0.443; intermediate-risk group: HR, 0.76; 95% CI, 0.64 to 0.91; P=0.0002; high-risk group: HR, 0.64; 95% CI, 0.45 to 0.91; P=0.0117). Kaplan-Meier estimates of OS based on a landmark analysis at 6 months after diagnosis were also assessed. Stratified by risk group, the probability of three-year OS in transplanted and non-transplanted patients was 43.8% (95% CI, 35.6% to 51.6%) and 33.5% (95% CI, 26.7% to 40.4%) in the low-risk group, 42.9% 1261

S. Fuji et al.

(95% CI, 37.2% to 48.4%) and 19.5% (95% CI, 15.7% to 23.5%) in the intermediate-risk group, and 29.3% (95% CI, 18.8% to 40.6%) and 3.3% (95% CI, 0.7% to 9.7%) in the high-risk group, respectively (Figure 3Dâ&#x20AC;&#x201C;F). Stratified by risk group of original ATL-PI, the probability of threeyear OS in transplanted and non-transplanted patients was 49.6% (95% CI, 42.9% to 56.0%) and 30.6% (95% CI, 24.9% to 36.4%) in the low-risk group, 34.6% (95% CI, 28.8% to 40.4%) and 16.7% (95% CI, 13.0% to 20.7%) in the intermediate-risk group, and 41.0% (95% CI, 24.8% to 56.6%) and 13.5% (95% CI, 5.2% to 25.7%) in the high-risk group, respectively (Online Supplementary Figure S1Câ&#x20AC;&#x201C;E).

Discussion We developed a new PI focusing on patients aged 70 years or younger with aggressive ATL using the largest database of aggressive ATL to date. We were able to stratify patients into three prognostic risk groups (low-risk, intermediate-risk, and high-risk) using this PI. We demonstrated that the new PI has as good a discriminatory capability as the original ATL-PI. To assess the impact of allo-HSCT on clinical outcome, the probability of OS was compared between non-transplanted and transplanted patients in each risk group. In the

P<0.0001

1262

intermediate- and high-risk groups, clinical outcomes in non-transplanted patients were dismal; it was apparent that transplanted patients had better outcomes than nontransplanted patients. Taking these cases of poor clinical outcomes with chemotherapy into consideration, it would be reasonable to consider early allo-HSCT in patients with ATL in the intermediate- and high-risk groups, despite the lack of evidence from prospective randomized clinical trials. Our risk stratification system might be useful in developing a risk-adapted treatment strategy in patients with aggressive ATL, similar to that used for high-risk AML. The difference between our modified ATL-PI and the original ATL-PI should be clarified. In terms of the included variable, albumin was not a significant variable in our modified ATL-PI although it was evaluated in our cohort. In addition, age was not a significant variable although we assessed age with various cutoffs. As age itself is not necessarily related to disease biology, it might be reasonable not to include age in the risk stratification system to evaluate a patient who is potentially eligible for transplant based on the current upper age limit restrictions in alloHSCT. The advantages of this modified PI over the original ATL-PI should be discussed. The original ATL-PI adequately risk stratified in this cohort, as shown in Figure 2C. However, when we stratified patients by the original ATL-PI and assessed the benefit of allo-HSCT, the probabilities of OS in the low-risk group were significantly

P<0.0001

Figure 2. Overall survival rates in the validation set and the relapse rates in the entire cohort. In the validation set, overall survival rates stratified by prognostic index risk group (A) with and (B) without censoring at the time of transplant. (C) Overall survival rates stratified by ATL-PI risk group with censoring at the time of transplant. (D) Cumulative incidence of relapse or progression stratified by prognostic index risk group with censoring at the time of transplant in the entire cohort.

haematologica | 2017; 102(7)

Modified prognostic index for aggressive ATL

higher in transplanted patients in comparison to nontransplanted patients. In addition, the proportion of patients in the high-risk group in whom significant clinical benefits were expected was limited, since age is included as a risk factor in the original ATL-PI. Thus, in order to identify patients in whom upfront allo-HSCT might provide clinical benefits or vice versa, our modified ATL-PI seems to have advantages over the original ATL-PI.

Our group has reported that early allo-HSCT (alloHSCT <100 days after ATL diagnosis) from a related donor might improve the clinical outcomes of patients with aggressive ATL.19 In our current study, median OS and time to relapse or progression was 322 days and 182 days in the intermediate-risk group and 197 days and 144 days in the high-risk group, respectively. Conventional salvage chemotherapy is usually ineffective in relapsed ATL,

P<0.0001

P=0.007

P<0.0001

P=0.058

P<0.0001

Figure 3. Overall survival rates by transplantation status stratified according to the risk group. Overall survival rates by transplantation status in the (A) low-risk, (B) intermediate-risk, and (C) high-risk groups. Overall survival rates based on landmark analysis that included patients who survived at least 6 months after diagnosis by transplantation status in the (D) low-risk, (E) intermediate-risk, and (F) high-risk groups.

haematologica | 2017; 102(7)

1263

S. Fuji et al.

as it is often ineffective in relapsed peripheral T-cell lymphoma in general.20 Therefore, it is important to offer upfront allo-HSCT while ATL is under control with induction chemotherapy, because disease status at the time of allo-HSCT is an important prognostic factor.7,8,10 In terms of AML, median OS in patients with an unfavorable risk profile was reported to be 10.2 months.21 Thus, the expected OS in patients with aggressive ATL in the intermediate-risk and high-risk groups was shorter than that in AML with an unfavorable risk profile. Hence, we should urgently prepare for allo-HSCT in patients with aggressive ATL; initiating HLA typing and donor coordination as soon as possible is strongly recommended. To further confirm the benefit of upfront allo-HSCT in patients with ATL, a prospective clinical trial is desirable. A single-arm confirmatory trial for this strategy that includes allo-HSCT for ATL is ongoing in Japan (Japan Clinical Oncology Group study 0907; UMIN000004147). To optimize the clinical outcomes in patients with ATL who undergo allo-HSCT, more research is needed. For instance, the choice of conditioning regimen is important. In a multivariate analysis of patients who underwent allo-HSCT, incorporating age, stem cell source, disease status at transplant and intensity of conditioning regimen, the latter was not a significant variable (data not shown). Future studies which assess the impact of allo-HSCT in patients with ATL should include the PI at diagnosis. Our current study has several limitations inherent to its retrospective nature and selection bias. We excluded patients who did not receive intensive chemotherapy as first-line treatment, as the inclusion of such critically ill patients would have further worsened the expected clinical outcome in the non-transplant group. Therefore, the clinical outcomes of seriously ill patients who are not candidates for intensive chemotherapy cannot be accurately predicted using our PI. In terms of the factors in the new PI, poor ECOG PS, which is also included in the original ATLPI, seems to be a self-contradictory prognostic factor as patients with a poor ECOG PS would not be transplant candidates. However, in patients with aggressive ATL, those with a poor ECOG PS at diagnosis in general reflect the aggressiveness of the disease, and such patients could experience recovery following primary induction therapy. If patients remain in a poor PS prior to potential transplant, they are not a suitable transplant candidate. As another

References 1. Ishitsuka K, Tamura K. Human T-cell leukaemia virus type I and adult T-cell leukaemia-lymphoma. Lancet Oncol. 2014;15(11):e517-e526. 2. Tsukasaki K, Hermine O, Bazarbachi A, et al. Definition, prognostic factors, treatment, and response criteria of adult T-cell leukemia-lymphoma: a proposal from an international consensus meeting. J Clin Oncol. 2009;27(3):453-459. 3. Katsuya H, Yamanaka T, Ishitsuka K, et al. Prognostic index for acute- and lymphomatype adult T-cell leukemia/lymphoma. J Clin Oncol. 2012;30(14):1635-1640. 4. Tsukasaki K, Utsunomiya A, Fukuda H, et al. VCAP-AMP-VECP compared with biweekly CHOP for adult T-cell leukemialymphoma: Japan Clinical Oncology Group

1264

important limitation of this study, we expected that a proportion of patients in this database have also been included in previous national surveys, as we included patients diagnosed from 2000 to 2013.3,6,7 In addition, a significant percentage of patients in this database did not undergo alloHSCT. We were unable to collect any data as to why these non-transplanted patients did not receive allo-HSCT. Although the reasons why such patients did not receive allo-HSCT were unclear, it might be due to comorbidities that could lead to worse OS irrespective of allo-HSCT, or it may reflect the choices of the physicians and patients. Such factors might attribute to significant selection bias. In addition, we were not able to include lactate dehydrogenase (LDH) levels in this study as the reference range of LDH levels varies significantly among institutes. In the study herein, patients were often transferred from a local hospital to a larger hospital following diagnosis in order to continue intensive chemotherapy. Even at the same hospital, the reference range changed during the study period, thus making it very difficult to include LDH levels in this study. We believe that future studies should assess the importance of LDH levels in aggressive ATL. Finally, our study included only Japanese patients. Therefore, it is important to assess the implications of our modified ATL-PI in different ATL populations, although previous research which reported a prognostic model included similar variables.22 In conclusion, we constructed the largest database of aggressive ATL and developed a new PI focusing on patients who are potential candidates for allo-HSCT. In the intermediate-risk and high-risk groups, transplanted patients had better clinical outcomes than non-transplanted patients. Prospective studies are warranted to confirm the benefit of treatment strategies that include upfront allo-HSCT in patients with aggressive ATL. Funding This research was partially supported by the Practical Research for Innovative Cancer Control from the Japan Agency for Medical Research and Development (15Ack0106136h0002) and the National Cancer Research and Development Fund (26-A-26). Acknowledgments We thank the medical, nursing, data processing, laboratory, and clinical staff at the participating centers for their important contributions to this study and their dedicated care of the patients.

Study JCOG9801. J Clin Oncol. 2007;25 (34):5458-5464. 5. Yamada Y, Tomonaga M, Fukuda H, et al. A new G-CSF-supported combination chemotherapy, LSG15, for adult T-cell leukaemia-lymphoma: Japan Clinical Oncology Group Study 9303. Br J Haematol. 2001;113(2):375-382. 6. Fukushima T, Nomura S, Shimoyama M, et al. Japan Clinical Oncology Group (JCOG) prognostic index and characterization of long-term survivors of aggressive adult T-cell leukaemia-lymphoma (JCOG0902A). Br J Haematol. 2014;166 (5):739-748. 7. Hishizawa M, Kanda J, Utsunomiya A, et al. Transplantation of allogeneic hematopoietic stem cells for adult T-cell leukemia: a nationwide retrospective study. Blood. 2010;116(8):1369-1376.

8. Ishida T, Hishizawa M, Kato K, et al. Allogeneic hematopoietic stem cell transplantation for adult T-cell leukemia-lymphoma with special emphasis on preconditioning regimen: a nationwide retrospective study. Blood. 2012;120(8):1734-1741. 9. Kanda J, Hishizawa M, Utsunomiya A, et al. Impact of graft-versus-host disease on outcomes after allogeneic hematopoietic cell transplantation for adult T-cell leukemia: a retrospective cohort study. Blood. 2012;119(9):2141-2148. 10. Fuji S, Inoue Y, Utsunomiya A, et al. Pretransplantation anti-CCR4 antibody mogamulizumab against adult T-cell leukemia/lymphoma is associated with significantly increased risks of severe and corticosteroid-refractory graft-versus-host disease, nonrelapse mortality, and overall mortality. J Clin Oncol. 2016;34(28):3426-3433.

haematologica | 2017; 102(7)

Modified prognostic index for aggressive ATL

11. Fenske TS, Hamadani M, Cohen JB, et al. Allogeneic hematopoietic cell transplantation as curative therapy for patients with non-Hodgkin lymphoma: increasingly successful application to older patients. '. Biol Blood Marrow Transplant. 2016; 22(9):1543-1551. 12. McClune BL, Weisdorf DJ, Pedersen TL, et al. Effect of age on outcome of reducedintensity hematopoietic cell transplantation for older patients with acute myeloid leukemia in first complete remission or with myelodysplastic syndrome. J Clin Oncol. 2010;28(11):1878-1887. 13. Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J Am Stat Assoc. 1999;94(446):496509. 14. Shimoyama M. Diagnostic criteria and classification of clinical subtypes of adult T-cell leukaemia-lymphoma. A report from the

haematologica | 2017; 102(7)

15.

16.

17.

18.

Lymphoma Study Group (1984-87). Br J Haematol. 1991;79(3):428-437. Tokunaga M, Uto H, Takeuchi S, et al. Newly identified poor prognostic factors for adult T-cell leukemia-lymphoma treated with allogeneic hematopoietic stem cell transplantation. Leuk Lymphoma. 2017; 58(1):37-44. Shigematsu A, Kobayashi N, Yasui H, et al. High level of serum soluble interleukin-2 receptor at transplantation predicts poor outcome of allogeneic stem cell transplantation for adult T cell leukemia. Biol Blood Marrow Transplant. 2014;20(6):801-805. Harrell FE, Jr., Lee KL, Califf RM, Pryor DB, Rosati RA. Regression modelling strategies for improved prognostic prediction. Stat Med. 1984;3(2):143-152. Altman DG, Royston P. What do we mean by validating a prognostic model? Stat Med. 2000;19(4):453-473.

19. Fuji S, Fujiwara H, Nakano N, et al. Early application of related SCT might improve clinical outcome in adult T-cell leukemia/lymphoma. Bone Marrow Transplant. 2016;51(2):205-211. 20. Mak V, Hamm J, Chhanabhai M, et al. Survival of patients with peripheral T-cell lymphoma after first relapse or progression: spectrum of disease and rare long-term survivors. J Clin Oncol. 2013; 31(16):19701976. 21. Patel JP, Gonen M, Figueroa ME, et al. Prognostic relevance of integrated genetic profiling in acute myeloid leukemia. N Engl J Med. 2012;366(12):1079-1089. 22. Phillips EH, Hodson A, Hermine O, Bazarbachi A, Cwynarski K. Striving to cure adult T-cell leukaemia/lymphoma: a role for allogeneic stem cell transplant? Bone Marrow Transplant. 2016; 51(12): 1549-1555.

1265

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Plasma Cell Disorders

Ferrata Storti Foundation

Monitoring multiple myeloma by quantification of recurrent mutations in serum Even Holth Rustad,1 Eivind Coward,1,2 Emilie R Skytøen,1 Kristine Misund,1 Toril Holien,1 Therese Standal,1,3 Magne Børset,1Vidar Beisvag,1 Ola Myklebost,2,4,5 Leonardo A Meza-Zepeda,2,5 Hong Yan Dai,6 Anders Sundan1,3 and Anders Waage1,2,7

Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, NTNU, Trondheim; 2Norwegian Cancer Genomics Consortium; 3 CEMIR – Center for Molecular Inflammation Research, Norwegian University of Science and Technology, NTNU, Trondheim; 4Institute for Clinical Science, University of Bergen; 5 Institute for Cancer Research, Oslo University Hospital; 6Department of Pathology and Medical Genetics, St. Olav’s University Hospital, Trondheim and 7Department of Hematology, St. Olav’s University Hospital, Trondheim, Norway 1

Haematologica 2017 Volume 102(7):1266-1272

ABSTRACT

C Correspondence: anders.waage@ntnu.no

Received: November 28, 2016. Accepted: March 31, 2017. Pre-published: April 6, 2017. doi:10.3324/haematol.2016.160564 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1266 ©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.

1266

irculating tumor DNA is a promising biomarker to monitor tumor load and genome alterations. We explored the presence of circulating tumor DNA in multiple myeloma patients and its relation to disease activity during long-term follow-up. We used digital droplet polymerase chain reaction analysis to monitor recurrent mutations, mainly in mitogen activated protein kinase pathway genes NRAS, KRAS and BRAF. Mutations were identified by next-generation sequencing or polymerase chain reaction analysis of bone marrow plasma cells, and their presence analyzed in 251 archived serum samples obtained from 20 patients during a period of up to 7 years. In 17 of 18 patients, mutations identified in bone marrow during active disease were also found in a time-matched serum sample. The concentration of mutated alleles in serum correlated with the fraction in bone marrow plasma cells (r=0.507, n=34, P<0.002). There was a striking covariation between circulating mutation levels and M protein in ten out of 11 patients with sequential samples. When relapse evaluation by circulating tumor DNA and M protein could be directly compared, the circulating tumor DNA showed relapse earlier in two patients (3 and 9 months), later in one patient (4 months) and in three patients there was no difference. In three patients with transformation to aggressive disease, the concentrations of mutations in serum increased up to 400 times, an increase that was not seen for the M protein. In conclusion, circulating tumor DNA in myeloma is a multi-faceted biomarker reflecting mutated cells, total tumor mass and transformation to a more aggressive disease. Its properties are both similar and complementary to M protein.

Introduction Multiple myeloma is caused by proliferation of monoclonal plasma cells in the bone marrow and is the second most common hematologic malignancy.1 The treatment options have improved markedly in recent years and led to prolonged survival.2 However, the disease is still considered to be incurable. The typical course of multiple myeloma is repeated treatment responses followed by increasingly aggressive relapses. Ultimately, the disease becomes refractory to all treatment and the patient dies. To assess disease progression and treatment response, clinicians rely on monitoring of the monoclonal immunoglobulin (M protein) secreted by the tumor cells as a biomarker for tumor mass.3,4 However, some patients escape the traditional monitoring. Between 1-3% of patients have non-secretory multiple myeloma and no haematologica | 2017; 102(7)

Monitoring myeloma by mutations in serum

detectable M protein.5,6 Furthermore, 10% of newly diagnosed myeloma patients have oligo-secretory disease, defined as a baseline level of M protein that is too low to evaluate treatment response reliably by traditional methods.3,7 These patients are challenging to monitor and are, therefore, often denied access to clinical trials.7 A promising new cancer biomarker is circulating tumor DNA (ctDNA), which may be extracted from serum or plasma.8 DNA fragments are released from cancer cells as well as normal cells in the body during apoptosis and necrosis.9,10 The cancer-derived fragments may be identified if they contain tumor-specific mutations or other genetic aberrations.8 In studies of solid tumors, ctDNA has provided information about tumor mass and residual disease, as well as information about the tumor genome that could otherwise only have been obtained by a tumor biopsy.11-15 Information about ctDNA in multiple myeloma lags behind as only a single study has so far been published.16 The somatic mutational landscape of multiple myeloma has been described in several studies.17-21 Out of more than 6,000 genes in which coding mutations have been identified, 13 are mutated more frequently than predicted from the background mutation rate, suggesting that they are implicated in the development of the disease.17,18 Among these recurrently mutated genes, NRAS, KRAS and BRAF in the mitogen activated protein (MAP) kinase pathway are most frequently mutated, occurring in bone marrow plasma cells from approximately 50% of patients at diagnosis. Moreover, activating mutations in the MAP kinase pathway are of interest because they are potential therapeutic targets.22-25 In this study, we explored ctDNA as a biomarker of multiple myeloma and focused on mutations in recurrently mutated genes including NRAS, KRAS and BRAF. We measured the concentrations of specific mutations in serum through several responses and relapses for up to 7 years in 20 patients and found a remarkable covariation with the concentration of M protein. However, in terminal aggressive disease, ctDNA appears to reflect the development of the disease better.

Details about the following experimental procedures are provided in the Online Supplementary Methods.

Detection of mutations in serum by digital droplet polymerase chain reaction Serum (n=249) and citrate-plasma (n=2) samples were obtained from the Norwegian Multiple Myeloma Biobank. DNA was extracted from a median sample volume of 1.8 mL (range, 0.4-3 mL) using a QiaAmp Circulating Nucleic Acid kit (Qiagen, Hilden, Germany). To detect mutations, we used the ddPCR system QX100/200 from Bio-Rad Laboratories (Hercules, CA, USA).27 Detailed assay information is presented in Online Supplementary Table S1 and raw data examples in Online Supplementary Figure S2. Patientsâ&#x20AC;&#x2122; samples were considered to be mutation-positive if the mutant concentration in the sample was higher than the 95% confidence interval of the assay-specific false positive rate (Online Supplementary Table S2, Online Supplementary Figure S3). The estimated number of mutant copies required in a sample to be considered mutation-positive ranged from 0.84 to 2.96 copies of mutated DNA (median 1.4). The quantity of mutated DNA in positive samples was reported in copies per mL of serum.

Whole exome sequencing WES of purified plasma cells and matched germline controls was performed as previously described.26 The target coverage of >100x was achieved for 85% of exonic regions. The limit of detection of WES was a mutated allele fraction of 2-4 % in the bone marrow sample.

Statistical analysis Bivariate correlations were performed by the Spearman correlation rank test. The level of statistical significance with two-tailed P-values was P<0.05. Statistical analyses were carried out in SPSS v. 21 (IBM Corporation, Armonk, NY, USA).

Results A summary of clinical and mutational data for each patient is given in Table 1.

Methods Study design and patients We conducted a retrospective study measuring ctDNA in archived serum samples from patients with multiple myeloma. Mutations of interest were identified in a bone marrow biopsy or purified bone marrow plasma cells and subsequently measured in serum by mutation-specific digital droplet polymerase chain reaction (ddPCR). Patients were included based on the following criteria: (i) presence of one or more mutations in genes recurrently mutated in myeloma17,18 and (ii) availability of relevant serum or plasma samples. Twenty patients from two sources were included in this study: one previously published study of the BRAFV600E mutation in myeloma and an on-going whole exome sequencing (WES) study.26 A flowchart describing the patientsâ&#x20AC;&#x2122; inclusion in detail is presented in the Online Supplementary Material (Online Supplementary Figure S1). Clinical data were obtained from the patientsâ&#x20AC;&#x2122; records and archived blood smears were evaluated for the presence of plasma cells. All patients had given written consent. The study was approved by the Regional Committee for Medical and Health Research Ethics (2016/821). haematologica | 2017; 102(7)

Figure 1. Correlation between mutation levels in bone marrow plasma cells and serum samples. Time-matched bone marrow and serum samples were obtained within 10 days of each other. Purified bone marrow plasma cells were analyzed by WES. In three cases in which WES was negative, positive results from the more sensitive ddPCR of bone marrow plasma cells were reported instead. Serum samples were analyzed by ddPCR.

1267

E.H. Rustad et al.

Relation between tumor mutations in serum and bone marrow plasma cells We started by determining whether mutations found in bone marrow plasma cells could be detected in timematched serum samples by ddPCR, and found that this was the case for 17 of 18 patients (34 of 35 mutations). We examined the quantitative relationship between the concentrations of circulating mutated DNA and the allele fractions of the same mutations in bone marrow plasma cells. There was a moderate positive correlation between the two (r=0.507, n=34, P<0.002) (Figure 1). Thus, the concentration of a mutation in serum reflects the fraction of tumor cells harboring the same mutation.

Relation between levels of recurrent mutations and M protein in serum Eleven patients had sequential serum samples available, spanning a median of 50 months (range, 8-90). In these patients, we monitored the concentration of mutated DNA over time in relation to tumor mass and treatment response as evaluated by M protein concentration. All 11 patients had a MAP kinase pathway mutation and two had at least one additional mutation (Table 1, patients 111). Most of these mutations were highly present in the bone marrow at diagnosis, with 75-100% mutation-positive plasma cells by immunohistochemistry or >50% mutated allele fraction by WES. Slightly lower MAP

kinase mutated allele fractions of 34% and 26% were found at diagnosis in patients 2 and 5, respectively, and patient 11 had 25-50% BRAFV600E-mutated cells by immunohistochemistry. No diagnostic bone marrow samples were available from patients 9 and 10. The concentrations of MAP kinase mutations in serum showed marked covariation with M protein levels. For example, patient 1 (Figure 2A) was monitored by M protein as well as circulating BRAFV600E mutation during 51 months, from diagnosis through five relapses until death. Every change in disease activity, as reflected by the M protein level, was accompanied by similar changes in serum BRAFV600E mutation levels. Similar observations were made in ten of the 11 patients with available sequential samples (Figures 2 and 3 and Online Supplementary Figure S4). The observed covariation in ten patients was confirmed by a formal correlation analysis of 210 time-matched measurements of M protein and circulating MAP kinase pathway mutation with correlation coefficients ranging from 0.63 to 0.96 (Online Supplementary Table S3). Only in patient 10 (Online Supplementary Figure S4B) was there no correlation. In this patient the BRAFV600E mutation became undetectable after being present at a very low concentration (<10 copies/mL) at an early time point. An important aspect of ctDNA analysis in myeloma is its sensitivity, compared to conventional methods, to detect low levels of disease. When looking at the ability to

Table 1. Summary of clinical data and mutations.

Patient Mutation(s)

Sex

Age (years)

Survival (months)

N. of treatments

M-Protein

ISS-stage

1 2 3 4

M F F F

69 66 77 54

51 40 35* 52

6 4 2 6

IgA kappa IgA kappa IgG kappa IgG kappa

3 2 2

8.7 11.9 11.9

IgG kappa

F M F F M M F M M

81 57 68 61 54 75 48 67 73

64 77 104* 107 79 22 58* 55* 24

2 10 3 9 6 3 4 2 3

IgG kappa IgG lambda IgG lambda IgG lambda Lambda Lambda IgA kappa IgA kappa IgA kappa

M M M F F

68 61 50 83 69

23* 33* 32* 28* 46*

2 2 2 3 1

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

BRAF V600E NRAS Q61K KRAS Q61H FAM46C S27Y, IRF4 K123R, KRAS A146P KRAS Q61R, TP53 Y236N BRAF V600E BRAF V600E BRAF V600E KRAS Q22K BRAF V600E BRAF V600E NRAS G12D NRAS Q61K DIS3 H788R, NRAS Q61R NRAS G12A NRAS Q61R NRAS Q61K KRAS Q22K BRAF V600E, NRAS Q61K BRAF V600E

Hb Ca-corr (g/dL) (mmol/L )

Creatinine (Îźmol/L)

Bone disease

2.53 2.71 2.31

105 58 71

Yes No -

15.3

2.42

3 1 2 1 3 3 1

12.8 11.9 14.2 10.9 14.1 7.8 10.9 9.9 14.5

2.51 2.26 2.36 2.53 3.53 3.2 2.31

94 107 60 61 70 248 82 81 93

Yes No Yes Yes Yes No Yes Yes

IgG kappa IgA lambda IgA kappa IgA kappa IgA lambda

1 2 2 3 2

14.1 9.4 12.3 12.6 10

2.23 2.76 2.27 2.44

30 97 63 79

Yes No Yes Yes

Lambda

11.3

2.85

125

Yes

Clinical parameters are reported from the time of diagnosis of multiple myeloma. Survival is calculated from the date of diagnosis to the date of death or last observation. M: male; F: female; *: patient still alive; -: missing data; ISS: International staging System; Ca-corr: albumin-corrected serum calcium; Hb: hemoglobin.

1268

haematologica | 2017; 102(7)

Monitoring myeloma by mutations in serum

predict relapse, we found that serum mutation levels tended to increase before or at the same time as M protein in most cases in which the two methods could be compared (Figures 2 and 3 and Online Supplementary Figure S4). Notably, a relapse from complete remission in patient 2 (Figure 2B) was detected by ctDNA 9 months before M protein became detectable. Relapses were also detected earlier by ctDNA in patient 1 (Figure 2A), although with somewhat shorter lead-times. On the other hand, in patient 11 (Online Supplementary Figure S4F) the second relapse was heralded by an increase in urine M protein 4 months before ctDNA became detectable. Furthermore, ctDNA often became undetectable during periods of remission even though low levels of M protein were still detectable, or the concentration of ctDNA would fluctuate around the limit of detection. In summary, ctDNA showed relapse earlier in two patients (3 and 9 months), later in one patient (4 months) and in three patients there was no difference. In patient 3 (Figure 2C), we initially detected a KRASQ61H mutation in plasma cells by WES as well as by ddPCR of serum. Light chain escape occurred at the second relapse, 38 months after the start of treatment, when the tumor cell secretion converted from IgA-Îş to Îş chains only. Despite this change, we could monitor the disease by the serum concentration of KRASQ61H mutation. In two patients, we monitored one or two recurrent mutations in addition to the MAP kinase pathway muta-

tions (Figure 3A,B). Patient 4 (Figure 3A) had an IRF4 mutation highly present in bone marrow plasma cells at the last relapse. At diagnosis, this mutation was not detected in plasma cells by WES, but a few copies were found by ddPCR of plasma cells and serum. The concentration of the mutation in serum increased abruptly after initiation of therapy and covaried with M protein level for the rest of the disease course. Conversely, a FAM46C mutation present at a 60% allele fraction in plasma cells at diagnosis became undetectable in both serum and plasma cells during the disease course. In patient 5 (Figure 3B), the concentrations of M protein and KRAS and TP53 mutations followed similar patterns in serum, despite a plasma cell allele fraction of only 4% for the TP53 mutation at diagnosis. Altogether, we monitored 14 mutated clones in 11 patients. Twelve of the mutations were detectable in serum at each relapse and covaried with M protein, whereas two mutations became undetectable during the disease course (FAM46CS279N in patient 4 and BRAFV600E in patient 10). These observations suggest that the serum concentration of recurrent mutations over long periods of time reflect the changes in total tumor mass in most myeloma patients.

Serum mutation levels in aggressive disease In patients 1, 4, and 5 (Figure 2A, Figure 3 A,B) we noticed a marked increase in serum mutation levels in the

Figure 2. Sequential levels of M protein and circulating MAP kinase pathway mutations. Sequential serum concentrations of M protein and MAP kinase pathway mutations through the disease course are shown for three patients (panels A-C depict patients 1-3) along with the type and duration of treatment. For patient 3, there were no serum samples collected between 10 and 38 months. Axis legends for all panels are the same as for panel A. X, time of death. Treatments: M: melphalan; P: prednisone; T: thalidomide; V: bortezomib; D: dexamethasone; L: lenalidomide; C: cyclophosphamide; HDT-ASCT: high-dose melphalan therapy with autologous stem cell transplant; Pom: pomalidomide.

haematologica | 2017; 102(7)

1269

E.H. Rustad et al.

terminal phase of the disease. At that time the patients had treatment refractory disease and remained alive only for a few weeks or months. To further analyze the dynamics of ctDNA over time, we compared the peak levels of mutations and M protein at each relapse (Figure 4). To facilitate the comparison between patients, we normalized the concentrations of M protein and the ctDNA as indicated in the legend to Figure 4. Only one mutation per patient is shown in Figure 4, however, in patients 4 and 5, mutations in IRF4, TP53 and KRAS behaved in the same manner indicating that they were all characteristics of the same aggressive clone. The discrepancy between ctDNA and M protein in patients 1, 4 and 5 was particularly evident in the terminal phase when the ratio of ctDNA to M protein was up to 400-fold higher than at the start of treatment. We analyzed several aspects of these patients which can contribute to the marked increase in serum concentrations of mutations. Patient 1 and 5 had plasma cells with immature morphology, whereas patients 1 and 4 had secondary plasma cell leukemia, with >20% plasma cells in blood. No plasma cells were found in blood from patient 5. In two of the patients there was >10% increase in mutated allele fraction in bone marrow plasma cells from the start of treatment to the time of terminal disease (34-49% in

patient 2 and 26-52% in patient 5). Thus, several factors may have contributed to the increased concentrations of ctDNA that were evident after transformation to a more aggressive disease.

Discussion We studied the serum concentrations of recurrent mutations identified in bone marrow plasma cells from 20 patients with multiple myeloma. Our comprehensive series of samples covering the entire disease course from diagnosis to death of several patients provides a unique insight into the dynamics of ctDNA in relation to disease activity. The most striking findings were a marked covariation with the concentration of M protein, the gold standard biomarker to monitor tumor mass in multiple myeloma, and increasing concentrations of ctDNA relative to M protein as the disease became more aggressive. To explain the increase of ctDNA, it is useful to discern between tumor mass and activity of the cells. The latter includes a number of functional aspects, such as proliferative rate and degree of adherence to the bone marrow environment. M protein is a typical tumor mass marker as long as the mechanisms of production and secretion of

Figure 3. Monitoring of three or two mutations in serum. Here, we monitored three (panel A) or two (panel B) recurrent mutations by ddPCR of serum from the start of treatment until the terminal phase. Treatments: M: melphalan; P: prednisone; T: thalidomide; V: bortezomib; D: dexamethasone; L: lenalidomide; C: cyclophosphamide; HDT-ASCT: high-dose melphalan therapy with autologous stem cell transplant

1270

haematologica | 2017; 102(7)

Monitoring myeloma by mutations in serum

immunoglobulin are intact. As shown, ctDNA and M protein seem to reflect tumor mass equivalently during long periods of the disease when the secretory mechanisms are operative and cellular functions relatively stable. This picture changes, however, when there is transformation to a more proliferative disease with high turnover of cells and perhaps a larger fraction of non-secretory cells. Furthermore, myeloma cells may be present in the circulation as shown in two of our patients. Although its clinical significance is unclear, ctDNA seems to reflect disease activity and progression differently from M protein. Our serum samples were stored for up to 11 years before analysis. Despite reports of DNA degradation during protracted sample storage,28 we found no statistically significant correlation between DNA yield and storage time, as shown in the methods section. Furthermore, it is recommended that ctDNA is analyzed in plasma rather than serum because of DNA released from leukocytes during sample preparation.29 However, to our knowledge, serum and plasma have not been directly compared in a clinical setting, and previous studies have successfully used stored serum samples.30,31 The close covariation between ctDNA and M protein found in our study adds to the evidence that stored serum can provide meaningful results and is a valuable material for the study of ctDNA. A weakness of this study was the low number of patients, limiting the generalizability of our results. Another weakness was the low and variable volume of serum available for analysis at each time-point, as reported in the methods section and elaborated in the Online Supplementary Methods. Because the ability of ddPCR to detect low levels of mutations is primarily limited by the sample volume and concentration of DNA, the sensitivity of our ctDNA measurements varied and was suboptimal in many samples. The potential to detect early relapse and minimal residual disease by ctDNA was, therefore, most likely under-estimated in our study. Mithraprabhu et al. recently reported the detection and monitoring of ctDNA in myeloma patients.16 Their design

differed from ours as they sequenced DNA from plasma as well as bone marrow plasma cells, targeting recurrently mutated regions in the NRAS, KRAS, BRAF and TP53 genes. Interestingly, they found 24% of mutations exclusively in plasma, consistent with the spatial heterogeneity of multiple myeloma previously demonstrated by multiregion DNA sequencing of bone marrow plasma cells.32,33 They also monitored specific mutations by ddPCR in three to six sequential samples from seven patients16 and our results are essentially in agreement with their observations. There are also apparent discrepancies between the studies. We detected 97% of mutations in serum when they had been identified in a time-matched bone marrow sample, whereas the corresponding number was only 39% (38/97) in the study by Mithraprabhu et al.16 This may be explained by the high sensitivity of their procedure as the majority of mutations they detected in bone marrow plasma cells had a mutated allele fraction between 0.01 and 1%. In comparison, the limit of detection by WES of bone marrow plasma cells in our study was 2-4 % mutated allele fraction, which is in line with previous studies using WES.17,18 There are several potential applications of ctDNA in multiple myeloma. The mechanisms by which M protein and ctDNA are released into the bloodstream appear to be independent of each other. Thus, monitoring the disease using ctDNA may be possible in situations in which M protein is not a reliable biomarker, such as in light chain escape and non-secretory or oligo-secretory disease.6,7,16 Furthermore, non-invasive detection of specific mutations may be useful to guide the use of targeted drugs such as BRAF or MEK inhibitors in patients with BRAF, NRAS or KRAS mutations.22-24 In principle, any tumor-specific DNA sequence such as a somatic mutation or a translocation breakpoint could be monitored by ddPCR.12,34,35 Alternatively, targeted sequencing may be applied directly to plasma or serum DNA to detect several targets simultaneously.13,30,36 This approach has the potential to describe tumor clonal evolution over time and its relation to clinical phenomena such as drug resistance37,38 and may be preferred in many situations. The choice of method will depend on the purpose. Altogether, this study provides detailed insight into the development of ctDNA levels over long periods of time in a limited number of patients. Circulating tumor DNA appears to be a multi-faceted biomarker of mutated cells, total tumor mass and transformation to a more aggressive disease in patients with multiple myeloma. However, several important questions remain unanswered, including the potential of ctDNA in minimal residual disease assessment and early detection of relapse.

Acknowledgments Figure 4. Ratio between serum levels of recurrent mutations and M protein during long-term follow-up. Diagnostic serum concentrations of M protein and MAP kinase mutations were normalized to one for each patient. Subsequent data are from the time points of peak M protein level at relapses, before a new treatment was started. The ratio of normalized mutation level to M protein was calculated for each data point. Patients were included in the figure if the diagnostic and at least one peak value of ctDNA and M protein were available. Patient 3 was excluded because of light chain escape, and patient 9 (Online Supplementary Figure S4A) was excluded because most peak values of ctDNA were too low to be confidently quantified.

haematologica | 2017; 102(7)

The authors would like to thank Lill Anny GrĂ¸seth and Solveig Kvam for managing the myeloma biobank in Biobank1. The project was supported by grants n. 218241 and 221580 from the Norwegian Research Council, as well as grants from the Norwegian Cancer Society, the Central Norway Regional Health Authority, the K. G. Jebsen Foundation and the St. Olavâ&#x20AC;&#x2122;s Hospital Cancer Foundation. The exome sequencing was performed in collaboration with the Genomics Core Facility (GCF), Norwegian University of Science and Technology (NTNU). GCF is funded by the Faculty of Medicine at NTNU and the Central Norway Regional Health Authority. 1271

E.H. Rustad et al.

References 1. Palumbo A, Anderson K. Multiple myeloma. N Engl J Med. 2011;364(11):1046-1060. 2. Naymagon L, Abdul-Hay M. Novel agents in the treatment of multiple myeloma: a review about the future. J Hematol Oncol. 2016;9(1):52. 3. Durie BG, Harousseau JL, Miguel JS, et al. International uniform response criteria for multiple myeloma. Leukemia. 2006;20(9): 1467-1473. 4. Rajkumar SV, Dimopoulos MA, Palumbo A, et al. International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. Lancet Oncol. 2014;15(12):e538-548. 5. Kyle RA, Gertz MA, Witzig TE, et al. Review of 1027 patients with newly diagnosed multiple myeloma. Mayo Clin Proc. 2003;78(1):21-33. 6. Chawla SS, Kumar SK, Dispenzieri A, et al. Clinical course and prognosis of non-secretory multiple myeloma. Eur J Haematol. 2015;95(1):57-64. 7. Larson D, Kyle RA, Rajkumar SV. Prevalence and monitoring of oligosecretory myeloma. N Engl J Med. 2012;367(6):580-581. 8. Heitzer E, Ulz P, Geigl JB. Circulating tumor DNA as a liquid biopsy for cancer. Clin Chem. 2015;61(1):112-123. 9. Jahr S, Hentze H, Englisch S, et al. DNA fragments in the blood plasma of cancer patients: quantitations and evidence for their origin from apoptotic and necrotic cells. Cancer Res. 2001;61(4):1659-1665. 10. Snyder MW, Kircher M, Hill AJ, Daza RM, Shendure J. Cell-free DNA comprises an in vivo nucleosome footprint that informs its tissues-of-origin. Cell. 2016;164(1-2):57-68. 11. Dawson SJ, Tsui DW, Murtaza M, et al. Analysis of circulating tumor DNA to monitor metastatic breast cancer. N Engl J Med. 2013;368(13):1199-1209. 12. Garcia-Murillas I, Schiavon G, Weigelt B, et al. Mutation tracking in circulating tumor DNA predicts relapse in early breast cancer. Sci Transl Med. 2015;7(302):302ra133. 13. Newman AM, Bratman SV, To J, et al. An ultrasensitive method for quantitating circulating tumor DNA with broad patient coverage. Nat Med. 2014;20(5):548-554. 14. Tie J, Wang Y, Tomasetti C, et al. Circulating tumor DNA analysis detects minimal residual disease and predicts recurrence in patients with stage II colon cancer. Sci Transl Med. 2016;8(346):346ra392. 15. Bettegowda C, Sausen M, Leary RJ, et al.

1272

16.

17.

18.

19.

20.

21.

22. 23.

24.

25.

26.

27.

Detection of circulating tumor DNA in early- and late-stage human malignancies. Sci Transl Med. 2014;6(224):224ra224. Mithraprabhu S, Khong T, Ramachandran M, et al. Circulating tumour DNA analysis demonstrates spatial mutational heterogeneity that coincides with disease relapse in myeloma. Leukemia. 2017 Jan 3. [Epub ahead of print] Lohr JG, Stojanov P, Carter SL, et al. Widespread genetic heterogeneity in multiple myeloma: implications for targeted therapy. Cancer Cell. 2014;25(1):91-101. Walker BA, Boyle EM, Wardell CP, et al. Mutational spectrum, copy number changes, and outcome: results of a sequencing study of patients with newly diagnosed myeloma. J Clin Oncol. 2015;33(33):39113920. Bolli N, Avet-Loiseau H, Wedge DC, et al. Heterogeneity of genomic evolution and mutational profiles in multiple myeloma. Nat Commun. 2014;5:2997. Kortum KM, Mai EK, Hanafiah NH, et al. Targeted sequencing of refractory myeloma reveals a high incidence of mutations in CRBN and Ras pathway genes. Blood. 2016;128(9):1226-1233. Lionetti M, Barbieri M, Todoerti K, et al. Molecular spectrum of BRAF, NRAS and KRAS gene mutations in plasma cell dyscrasias: implication for MEK-ERK pathway activation. Oncotarget. 2015;6(27): 24205-24217. Heuck CJ, Jethava Y, Khan R, et al. Inhibiting MEK in MAPK pathway-activated myeloma. Leukemia. 2016;30(4):976-980. Hyman DM, Puzanov I, Subbiah V, et al. Vemurafenib in multiple nonmelanoma cancers with BRAF V600 mutations. N Engl J Med. 2015;373(8):726-736. Andrulis M, Lehners N, Capper D, et al. Targeting the BRAF V600E mutation in multiple myeloma. Cancer Discov. 2013;3(8): 862-869. Sharman JP, Chmielecki J, Morosini D, et al. Vemurafenib response in 2 patients with posttransplant refractory BRAF V600Emutated multiple myeloma. Clin Lymphoma Myeloma Leuk. 2014;14(5): e161-163. Rustad EH, Dai HY, Hov H, et al. BRAF V600E mutation in early-stage multiple myeloma: good response to broad acting drugs and no relation to prognosis. Blood Cancer J. 2015;5:e299. Huggett JF, Cowen S, Foy CA. Considerations for digital PCR as an accu-

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

rate molecular diagnostic tool. Clin Chem. 2015;61(1):79-88. Sozzi G, Roz L, Conte D, et al. Effects of prolonged storage of whole plasma or isolated plasma DNA on the results of circulating DNA quantification assays. J Natl Cancer Inst. 2005;97(24):1848-1850. El Messaoudi S, Rolet F, Mouliere F, Thierry AR. Circulating cell free DNA: preanalytical considerations. Clin Chim Acta. 2013;424: 222-230. Roschewski M, Dunleavy K, Pittaluga S, et al. Circulating tumour DNA and CT monitoring in patients with untreated diffuse large B-cell lymphoma: a correlative biomarker study. Lancet Oncol. 2015;16(5):541549. Diaz LA, Jr., Williams RT, Wu J, et al. The molecular evolution of acquired resistance to targeted EGFR blockade in colorectal cancers. Nature. 2012;486(7404):537-540. Weinhold N, Chavan SS, Heuck C, et al. High risk multiple myeloma demonstrates marked spatial genomic heterogeneity between focal lesions and random bone marrow; implications for targeted therapy and treatment resistance. Blood. 2015;126 (23):20. Melchor L, Jones JR, Lenive O, et al. Spatiotemporal analysis of intraclonal heterogeneity in multiple myeloma: unravelling the impact of treatment and the propagating capacity of subclones using whole exome sequencing. Blood. 2015;126 (23):371. Krumbholz M, Hellberg J, Steif B, et al. Genomic EWSR1 fusion sequence as highly sensitive and dynamic plasma tumor marker in Ewing sarcoma. Clin Cancer Res. 2016;22(17):4356-4365. Reinert T, Scholer LV, Thomsen R, et al. Analysis of circulating tumour DNA to monitor disease burden following colorectal cancer surgery. Gut. 2016;65(4):625-634. Schreuer M, Meersseman G, Van Den Herrewegen S, et al. Quantitative assessment of BRAF V600 mutant circulating cellfree tumor DNA as a tool for therapeutic monit oring in metastatic melanoma patients treated with BRAF/MEK inhibitors. J Transl Med. 2016;14:95. Russo M, Siravegna G, Blaszkowsky LS, et al. Tumor heterogeneity and lesion-specific response to targeted therapy in colorectal cancer. Cancer Discov. 2016;6(2):147-153. Murtaza M, Dawson SJ, Tsui DW, et al. Non-invasive analysis of acquired resistance to cancer therapy by sequencing of plasma DNA. Nature. 2013;497(7447):108-112.

haematologica | 2017; 102(7)

ARTICLE

Plasma Cell Disorders

Resistin induces multidrug resistance in myeloma by inhibiting cell death and upregulating ABC transporter expression

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Jianan Pang,1,2* Qiaofa Shi,2,3* Zhiqiang Liu,2 Jin He,2 Huan Liu,2 Pei Lin,4 Jiuwei Cui1 and Jing Yang2,5

Cancer Center, The First Hospital of Jilin University, Changchun, Jilin Province, China; Department of Lymphoma/Myeloma, Division of Cancer Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX. USA; 3Immunology Department of Medical College, Nanchang University, Jiangxi, China; 4Department of Hematopathology, Division of Pathology and Laboratory Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 5Cancer Research Institute and Cancer Hospital, Guangzhou Medical University, China 1 2

*JP and QS contributed equally to the study.

Haematologica 2017 Volume 102(7):1273-1280

ABSTRACT

espite advances in therapy, multiple myeloma remains incurable, with a high frequency of relapse. This suggests the need to identify additional factors that contribute to drug resistance. Our previous studies revealed that bone marrow adipocytes promote resistance to chemotherapy in myeloma through adipocyte-secreted adipokines, but the mechanism underlying this effect and the specific adipokines involved are not well understood. We proposed to determine the role of resistin, an adipokine that is secreted by adipocytes, in chemotherapy resistance in myeloma. We found that resistin abrogated chemotherapyinduced apoptosis in established myeloma cell lines and primary myeloma samples. Resistin inhibited chemotherapy-induced caspase cleavage through the NF-κB and PI3K/Akt pathways. Resistin also increased the expression and drug efflux function of ATP-binding cassette (ABC) transporters in myeloma cells through decreasing the expression of both DNA methyltransferases DNMT1 and DNMT3a and the methylation levels of ABC gene promoters. In vivo studies further demonstrated the protective effect of resistin in chemotherapy-induced apoptosis. Our study thus reveals a new biological function of resistin in the pathogenesis of myeloma, with the implication that targeting resistin could be a potential strategy to prevent or overcome multidrug resistance in myeloma.

Introduction Multiple myeloma, a cancer of long-lived plasma cells,1 is the second most frequent hematologic malignancy in the United States, after non-Hodgkin lymphoma.2 Despite progress in the development of treatment, myeloma remains incurable, with the median survival of affected patients being 5–6 years.3 In most patients, myeloma develops resistance to a wide spectrum of anticancer agents, leading to failure of chemotherapy. In order to achieve a cure for multiple myeloma, we must determine the mechanism underlying the development of multidrug resistance in this disease. One well-known mechanism of drug resistance is the overexpression of ATPbinding cassette (ABC) transporters.4 The 49 human ABC transporters are classified into seven subfamilies, from ABCA to ABCG, based on their sequence similarities.5 ABCG2, also known as breast cancer resistance protein, is a 655-amino-acid polypeptide transporter with a wide range of substrates.5,6 Its expression is upregulated in a variety of malignancies, in which it may produce resistance to chemotherapeutic agents.6-8 ABCC5, also known as multidrug resistance protein 5, belongs to haematologica | 2017; 102(7)

Correspondence: jiyang@mdanderson.org or cuijw@jlu.edu.cn

Received: August 8, 2016. Accepted: March 23, 2017. Pre-published: March 30, 2017. doi:10.3324/haematol.2016.154062 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1273 ©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.

1273

J. Pang et al.

the largest sub-family, the ABCC family. ABCC5 has been shown to transport nucleosides and antifolates.9 Increased ABCC5 expression has been associated with breast cancer, hepatocellular carcinoma, and pancreatic ductal adenocarcinoma.10-12 In addition, myeloma cells grow and expand almost exclusively within the bone marrow, which plays a pivotal role in the pathogenesis of multiple myeloma. A number of studies have demonstrated that the interactions of myeloma cells with bone marrow stromal cells and with the extracellular matrix enhance the growth and survival of myeloma cells and induce drug resistance.3,13-17 Bone marrow stromal cells produce a large amount of soluble cytokines and chemokines, which can bind to their receptors on myeloma cells, activate the nuclear factor kappa B (NF-κB), phosphoinositide 3 kinase (PI3K)/Akt, mitogen activated protein kinase (MAPK) signaling pathways, and thereby inhibit chemotherapy-induced caspase cleavage and apoptosis in myeloma cells. In previous studies we found that the adipocytes derived from bone marrow confer chemotherapy resistance in myeloma through their secreted soluble adipokines.18 One such adipokine, resistin, is a 12.5-kDa hormone that is secreted not only by adipocytes but also by monocytes, macrophages, and spleen and bone marrow cells.19 It was first discovered as providing a link between obesity and insulin resistance,20 but its physiological role is much more complex than originally thought. Resistin has been shown to participate in inflammatory processes and cancer development through induction of inflammatory cytokines, such as interleukin (IL)-1β, IL-6, IL-8, IL-12 and tumor necrosis factor alpha, some of which can activate the Janus kinase/signal transducers and activators of transcription pathway.21,22 It also has protective effects against acute myocardial infarction and 6-hydroxydopamine–induced neuronal cell death.23,24 However, its role in the pathogenesis of myeloma is unknown. In this study, we hypothesized that the adipokine resistin has the capacity to induce multidrug resistance in myeloma. We added recombinant resistin to cultures of human myeloma cell lines and primary myeloma cells isolated from patients’ bone marrow aspirates, and we observed that resistin protects these tumor cells against chemotherapy by reducing tumor apoptosis through the NF-κB and PI3K/Akt signaling pathways and by enhancing the expression of ABC transporters in myeloma cells through demethylation of ABC gene promoters. We also observed a protective effect of resistin on myeloma cells against melphalan treatment in vivo. These findings indicate that resistin is a new factor contributing to myeloma cell growth and survival.

Table 1. Primer sequences for the real-time polymerase chain reaction.

Name ABCB1 ABCB3 ABCC1 ABCC3 ABCC4 ABCC5 ABCG2 DNMT1 DNMT3a DNMT3b GAPDH

Forward

Reverse

TCGGACCACCATTGTGATAG GTAAGGAGGGTGCTGCACTT TGGACTAACGGCTAACCTGGA CTGCTTCAGGGAGAAACCTC CATCCGAAGAATCCAGACCT ACTGTGGCAAGAAGAGCTGA CAGCAGGTCAGAGTGTGGTT CCCCTGAGCCCTACCGAAT TTCTACCGCCTCCTGCATGAT GAATTACTCACGCCCCAAGGA CTGGGCTACACTGAGCACC

CATTTCCTGCTGTCTGCATT CACGCTCTCCTGGTAGATCA TAAGCAACCAACACTGCTTTG CCATCTTTGTGAACCACCAG GGTCTCTGATGCCTTATCCC AAATTTGGTCCACTGAAGCC TGCAAAGCCGTAAATCCATA CTCGCTGGAGTGGACTTGTG GCGAGATGTCCCTCTTGTCACTA ACCGTGAGATGTCCCTCTTGTC AAGTGGTCGTTGAGGGCAATG

Aldrich (St. Louis, MO, USA), all antibodies for flow cytometry analysis were purchased from BD Biosciences (San Jose, CA, USA), and all antibodies for western blot analysis were purchased from Cell Signaling Technology (Danvers, MA, USA). The short interfering RNA (siRNA) against human ABCC5 and ABCG2 genes as well as the non-target control siRNA were purchased from Santa Cruz Technologies (Dallas,TX, USA).

Flow cytometry analysis of cell apoptosis Apoptosis of treated cells was detected by annexin V–fluorescein isothiocyanate/propidium iodide staining. For details see the Online Supplementary Methods.

Western blot analysis Cells were lysed and, in some instances, the nuclear and cellular fractions of cells were isolated using NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher, Waltham, MA, USA). Cell lysates were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, and immunoblotted with specific antibodies.

Quantitative real-time polymerase chain reaction Assessment of mRNA of treated cells was performed as described in the Online Supplementary Methods. Primer sequences are shown in Table 1.

eFluxx-ID gold uptake assay for ABC transporter activity The fluorescent probe eFluxx-ID gold was used to monitor the activity of ABC transporters (ENZO Life Sciences, Inc., Farmingdale, NY, USA). For details see the Online Supplementary Methods.

DNA methylation analysis Methods Cell lines, primary myeloma cells, and reagents This study was approved by the institutional review board of The University of Texas MD Anderson Cancer Center (Houston, TX, USA). ARP-1 and ARK cells were kindly provided by Arkansas Cancer Research Center (AR, Usa). Others were purchased from the American Type Culture Collection (ATCC). Primary myeloma cells were isolated from bone marrow aspirates of myeloma patients using anti-CD138 antibody-coated magnetic beads (Miltenyi Biotec, Inc. San Diego, CA, USA). Recombinant human resistin was purchased from PeproTech (Rocky Hill, NJ, USA). Except where specified, all chemicals were purchased from Sigma1274

Genomic DNA was obtained from treated cells using the QIAamp Tissue kit (Qiagen, Germantown, MD, USA) according to the manufacturer’s protocol. DNA was processed by bisulfite modification using the Zymo EZ DNA methylation kit (Zymo Research, Irvine, CA, USA). Specific primers for methylation-specific polymerase chain reaction (MS-PCR, Table 2) were designed using MethylPrimer software (http://www.urogene.org/methprimer/index1.html). PCR products were subjected to electrophoresis on a 2% agarose gel.

In vivo mouse model Six- to 8-week-old CB.17 SCID mice were obtained from Charles River Laboratories and maintained in facilities accredited haematologica | 2017; 102(7)

Resistin-induced multidrug resistance in myeloma

Figure 1. Resistin protects myeloma cells from chemotherapy-induced apoptosis. (A) Human myeloma cell lines ARP-1, MM.1S, and U266 were cultured in medium containing melphalan (25 μM) plus resistin (0, 10, 25, 50, 100, or 200 ng/mL) for 24 h; cells without melphalan treatment served as a control. Apoptosis in the cultured cells was determined by using an annexin V binding assay. The percentages of apoptotic cells in each of the three cell lines are shown. (B, C, D) ARP-1, MM.1S, and RPMI8226 cells were cultured in medium containing melphalan (Mel; 25 μM), bortezomib (BTZ; 5 nM), or carfilzomib (CFZ; 20 nM) with or without resistin (50 ng/mL) for 24 h. Cells cultured without the chemotherapy agents or resistin served as controls. Percentages of apoptotic cells are shown. (E) CD138+ plasma cells were isolated from bone marrow aspirates of five patients with multiple myeloma and cultured with melphalan (25 μM) without or with resistin (50 ng/mL) for 24 h. Percentages of apoptotic myeloma cells are shown. Results shown represent three to five independent experiments. *P<0.05; **P<0.01.

by the American Association of Laboratory Animal Care. The studies were approved by the Institutional Animal Care and Use Committee of MD Anderson Cancer Center. ARP-1 myeloma cells (5×105 cells/mouse) were injected into mouse femora. After 3 weeks, treatment with melphalan (50 μg/mouse) or resistin (50 µg/mouse), singly or in combination, was begun; each mouse received intraperitoneal treatment every 3 days for 3 weeks. Control mice received equal amounts of phosphate-buffered saline solution (PBS). After treatment, mouse sera were collected and serum M-protein levels were measured using the Human Kappa ELISA kit (Bethyl Laboratories, Montgomery, TX/US). Bone marrow cells were flushed from mouse femora and the human CD138+ subset was isolated by using anti-human CD138coated magnetic beads. Cells were then labeled with antibodies and analyzed with an LSRFortessa flow cytometer (BD Biosciences, San Jose, CA, USA).

Results Resistin inhibits chemotherapy-induced apoptosis of myeloma cells

In situ tumor cell apoptosis was determined by a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay kit (Roche Life Sciences, Indianapolis, IN, USA). For details see the Online Supplementary Methods.

To assess the effects of resistin on myeloma’s response to chemotherapy, the human ARP-1, MM.1S, and U266 myeloma cell lines were cultured in medium containing melphalan (25 μM), with or without resistin (10–200 ng/mL), for 24 h. After culture, an annexin V binding assay was performed to quantify cell apoptosis. As shown in Figure 1A, melphalan treatment increased apoptosis in all three cell lines, a result that is consistent with previous reports.18 The addition of resistin reduced melphalaninduced apoptosis of myeloma cells in a dose-dependent manner (Figure 1B). Similarly, resistin significantly reduced apoptosis induced in ARP-1, MM.1S, and RPMI8226 cells by bortezomib (Figure 1C) or by carfilzomib (Figure 1D). The experiment was repeated in CD138+ malignant plasma cells isolated from the bone marrow aspirates of five patients with newly diagnosed multiple myeloma. Resistin also reduced melphalan-induced apoptosis in the patients’ myeloma cells (Figure 1E).

Statistical analysis

Resistin activates anti-apoptotic signaling pathways

All data are shown as means ± standard deviation for at least three independent experiments performed in triplicate. The Student t-test was used to compare experimental groups. A P value <0.05 was considered statistically significant.

Caspase cascades are the functional regulators and executioners of apoptosis.25 To explore the effect of resistin on caspase cascades in myeloma cells, we cultured ARP-1 or MM.1S cells with melphalan in the presence or absence of

In situ TUNEL assay

haematologica | 2017; 102(7)

1275

J. Pang et al. A

Figure 2. Resistin activates anti-apoptotic signaling pathways in myeloma cells. (A, B) ARP-1 and MM.1S myeloma cells were treated with melphalan (Mel; 25 μM) and/or resistin (50 ng/mL) for 24 h. Western blot analysis shows (A) the levels of cleaved (c) caspase (Cas)-9, Cas-3, and PARP, and (B) the expression of the mitochondria-related anti-apoptotic proteins Bcl-2 and Bcl-xL and the pro-apoptotic protein Bax in the cells. Cells cultured without treatment served as controls. (C) ARP1 and MM.1S cells were cultured in medium with or without resistin (0, 50 ng/mL, or 100 ng/mL) for 12 h. Western blot analysis shows the levels of non-phosphorylated and phosphorylated (p) IκBα, Akt, and ERK1/2 in the cells treated with resistin. GAPDH served as a protein loading control. (D-F) ARP-1 or MM.1S cells were pretreated with (D) 1 μM NF-κB inhibitor Ro106, (E) 0.5 μM PI3K inhibitor LY294002, or (F) 1 μM MEK1/2 inhibitor U0126 for 1 h, followed by treatment with melphalan (25 μM) and/or resistin (50 ng/mL) for 24 h. The annexin V binding assay shows the percentages of apoptotic cells for each treatment condition. Cells cultured with none of the inhibitors served as controls. Results shown represent three independent experiments. *P<0.05.

resistin. Western blot analysis showed that ARP-1 and MM.1S cells treated with melphalan alone had significantly higher levels of cleaved caspase-9, caspase-3, and poly (ADP-ribose) polymerase (PARP) than untreated cells, while addition of resistin significantly decreased the levels of these proteins in the presence of melphalan (Figure 2A). We also tested the effects of melphalan and resistin on Bcl2 family members known to suppress apoptosis, such as Bcl-2 and Bcl-xL, or to promote apoptosis, such as Bax.26,27 Treatment with melphalan resulted in significant decreases of Bcl-2 and Bcl-xL expression in myeloma cells, while addition of resistin increased their expression. In contrast, melphalan treatment enhanced Bax expression in ARP-1 and MM.1S cells, while addition of resistin reduced its expression (Figure 2B). To further elucidate the means by which resistin regulates myeloma cell apoptosis, we focused on the NF-κB, PI3K/Akt, and MAPK/ERK1/2 signaling pathways, which are essential to myeloma growth and survival. As shown in Figure 2C, adding resistin increased the levels of phosphorylated IκBα, Akt, and ERK1/2 in ARP-1 and MM.1S cells in a dose-dependent manner, but not the non-phosphorylated levels. We further analyzed the nuclear and cytosolic components of the NF-κB pathways by western blotting. Online Supplementary Figure S1 shows that addition of resistin to the myeloma cells ARP-1 and MM.1S resulted in an increase in the levels of nuclear p65 and p50, while the levels of p52 and relB in the nucleus and those 1276

Table 2. Primer sequences for the methylation-specific polymerase chain reaction.

Name ABCG2 M F ABCG2 M R ABCG2 U F ABCG2 U R ABCC5 M F ABCC5 M R ABCC5 U F ABCC5 U R

Sequence

Size (bp)

TGAGGTAGGAGAATGGTATGAATTC TTTAAAATAAAATCTCGCTTTATCG GAGGTAGGAGAATGGTATGAATTTG TTAAAATAAAATCTCACTTTATCACC GGAAGATATTACGTTAAAGGGATACG AACAAATAAACACAAAAACGACGA AAGATATTATGTTAAAGGGATATGT AACAAATAAACACAAAAACAACAAA

103 101 170 168

M: methylation-specific primer; F: forward; R, reverse; U: unmethylation-specific primer.

in the cytoplasm remained unchanged. The levels of phosphorylated p65 in nuclei were also increased in resistintreated myeloma cells (Online Supplementary Figure S1). To define the importance of these signaling pathways in resistin-induced protection from apoptosis, specific inhibitors against NF-κB (Ro106), PI3K (LY294002), or MEK1/2 (U0126), the upstream kinase of ERK1/2, were added to cultures of ARP-1 and MM.1S cells. Treatment with either the NF-κB inhibitor or the PI3K inhibitor could abrogate resistin-induced protection of ARP-1 and MM.1S cells from melphalan-induced apoptosis (Figure 2D,E), whereas the MEK1/2 inhibitor had no effect (Figure 2F). haematologica | 2017; 102(7)

Resistin-induced multidrug resistance in myeloma

Figure 3. Resistin increases the expression of ABC transporters in myeloma cells. ARP-1 and MM.1S myeloma cells were cultured with resistin (50 ng/mL) for 12 h. Some of the cultured cells were labeled with eFluxx-ID gold fluorescent dye and further analyzed by flow cytometry. Others were subjected to RNA or protein extraction for real-time PCR or western blot analysis. (A) Intracellular eFluxx-ID gold fluorescence intensity was quantified. PBS, phosphate-buffered saline solution (controls). (B) Real-time PCR shows relative mRNA expression of ABC transporter genes. (C) Western blot analysis shows expression of ABCG2 and ABCC5 proteins. Cells cultured without resistin served as controls. GAPDH served as a protein loading control. (D) Real-time PCR analysis shows relative expression levels of ABCC5 and ABCG2 mRNA in ARP-1 or MM.1S cells bearing non-targeted siRNA (siCtrl) or the pooled siRNA of ABCC5 (siC5) and ABCG2 (siG2). (E) The percentages of apoptotic cells in siCtrl- or both siC5- and siG2-expressing ARP-1 or MM.1S cells treated with or without resistin or melphalan (Mel) are shown. Results are representative of three independent experiments. *P<0.05; **P<0.01.

These results indicate that resistin inhibits myeloma cell apoptosis through the NF-ÎşB and PI3K/Akt signaling pathways.

Resistin increases expression of ABC transporters in myeloma cells Overexpression of ABC transporters has been shown to increase ATP-driven efflux of chemotherapy drugs from tumor cells and thus decrease intracellular drug accumulation, leading to drug resistance.7 To investigate the impact of resistin on ABC-driven efflux of chemotherapy drugs in myeloma cells, ARP-1 or MM.1S cells were cultured with or without the addition of resistin. The cells were then labeled with the eFluxx-ID gold fluorescent dye and analyzed by flow cytometry. Our results showed that the intracellular accumulation of eFluxx-ID gold fluorescence was reduced by 46% in ARP-1 cells and by 33% in MM.1S cells treated with resistin as compared to that in cells not treated with resistin (Figure 3A). We then examined the expression of seven cell membraneâ&#x20AC;&#x201C;bound ABC transporters in resistin-treated myeloma cells using real-time PCR analysis. As shown in Figure 3B, the mRNA levels of ABCB3, ABCC1, ABCC3, ABCC5, and ABCG2, but not ABCB1 and ABCC4, were markedly higher in resistintreated ARP-1 and MM.1S cells than in non-treated cells. In addition, we checked the mRNA levels of ABC transporters in four more myeloma cell lines in the absence or presence of resistin. Although the patterns of increased levels of ABC transporter genes varied somewhat in differhaematologica | 2017; 102(7)

ent cell lines in response to resistin, the up-regulation of ABCC5, and ABCG2 persisted throughout (Online Supplementary Figure S2). Western blot analysis also showed that resistin increased the protein levels of two ABC, ABCG2 and ABCC5, in ARP-1 and MM.1S cells (Figure 3C). We further determined the importance of these two transporters in resistin-mediated protection. Using pool siRNA specific for human ABCC5 and ABCG2, we knocked down the expression of these genes in the myeloma cells ARP-1 and MM.1S (Figure 3D). We observed an increased percentage of apoptotic cells in the presence of resistin and the chemotherapy drugs melphalan (Figure 3E), bortezomib (Online Supplementary Figure S3A) or carfilzomib (Online Supplementary Figure S3B), indicating that resistin inhibits myeloma cell apoptosis through ABCC5, and ABCG2. We then assessed whether resistin increases ABC expression through regulation of CpG methylation in the gene promoters. Focusing on the promoters of the ABCG2 and ABCC5 genes, we used methylation-specific PCR analysis to determine the effect of resistin in cultures of ARP-1 and MM.1S cells. Resistin significantly decreased the levels of methylated ABCG2 and ABCC5 gene promoters while concomitantly increasing the levels of unmethylated ABCG2 and ABCC5 promoters (Figure 4A,B). Previous studies had shown that DNA methylation is catalyzed by DNA methyltransferases (DNMTs), including DNMT1, DNMT3a, and DNMT3b.28,29 Real-time PCR analysis showed that addition of resistin greatly reduced 1277

J. Pang et al. A

Figure 4. Resistin reduces the methylation of ABCG2 and ABCC5 gene promoters. ARP-1 and MM.1S myeloma cells were cultured with resistin (50 ng/mL) for 12 h. (A, B) Genomic DNA was extracted from cultured cells for methylation-specific PCR analysis. (A) Representative images of the methylated (M) and un-methylated (U) CpG sites and (B) quantitative data of M/U ratios in the promoters of ABCG2 or ABCC5 genes. PBS, phosphate-buffered saline solution (controls). (C, D) Total RNA and total proteins were extracted from cultured cells for real-time reverse transcriptase-PCR or western blot analysis. (C) Relative DNMT1, DNMT3a, and DNMT3b mRNA expression and (D) DNMT1 and DNMT3a protein expression. Cells cultured without resistin served as controls. GAPDH served as a protein loading control in western blot analysis. Results shown are representative of three independent experiments. *P<0.05; **P<0.01.

the mRNA levels of DNMT1 and DNMT3a but did not alter the levels of DNMT3b in ARP-1 and MM.1S cells (Figure 4C). Consistent with the real-time PCR results, western blot analysis showed decreased levels of DNMT1 and DNMT3a proteins in ARP-1 and MM.1S cells cultured with resistin compared to cells cultured without resistin (Figure 4D). These results clearly suggest that resistin downregulates DNMT1 and DNMT3a expression, reduces methylation in the promoters of ABC genes such as ABCG2 and ABCC5, and enhances their expression, leading to resistance to drug treatments for myeloma.

Annexin V binding assay of these CD138+ cells demonstrated that the percentage of apoptotic CD138+ cells from the bone marrow of mice receiving both melphalan and resistin was lower than that from the bone marrow of mice receiving melphalan alone (Figure 5C). Similar results were obtained from the staining (Figure 5D) and quantification (Figure 5E) of TUNEL in mouse femora. These results indicate that resistin protects myeloma cells against chemotherapy in vivo.

Discussion In vivo validation of the protective effects of resistin in myeloma To validate the protective effects of resistin in myeloma chemotherapy in vivo, we injected ARP-1 cells into the femora of SCID mice. When myeloma was established, the mice were treated with melphalan or resistin, alone or in combination. Myeloma-bearing mice that received neither treatment served as controls. Our results showed that melphalan treatment reduced the serum levels of M-proteins (Figure 5A), which are secreted from myeloma cells and the levels of which reflect tumor burden. The mice that received both resistin and melphalan had much higher levels of serum M-proteins than had the mice that received melphalan alone (Figure 5A). Flow cytometry analysis of bone marrow cells from these mice showed that the percentages of CD138+ cells in the mice treated with both melphalan and resistin were higher than those in the mice treated with melphalan alone (Figure 5B). 1278

Multidrug resistance is a major cause of chemotherapy failure in myeloma. We found that the adipokine resistin protected both established myeloma cell lines and primary myeloma cells from chemotherapy-induced apoptosis. This effect was verified in the mouse model of myeloma, in which resistin abrogated apoptosis and enhanced tumor growth. The mechanism underlying this inhibition involved activation of anti-apoptotic signaling pathways, including NF-ÎşB and PI3K/Akt. Resistin also upregulated the expression of ABC transporters through demethylation of their gene promoters, indicating that resistin is responsible for multidrug resistance in myeloma. DNA methylation of CpG dinucleotides is the most common epigenetic modification in mammalian cells.30 Demethylation in ABC transporter promoters has been shown to induce re-expression and overexpression of ABC transporters, leading to chemotherapy resistance.8, 30 haematologica | 2017; 102(7)

Resistin-induced multidrug resistance in myeloma

Figure 5. Resistin protects myeloma from chemotherapy in vivo. SCID mice were injected with ARP-1 myeloma cells (5×105 cells per mouse) directly into the femur (n=5 mice per group). Three weeks after ARP-1 cell injection, mice began intraperitoneal treatment with melphalan (Mel; 50 μg/mouse), resistin (20 μg/mouse), or both every 3 days for 3 weeks. After treatment, the mouse sera were subjected to enzyme-linked immusorbent assay to measure M-protein levels. After the mice had been euthanized, the cells flushed from each mouse’s femoral bone marrow cavity were labeled with an antibody against human CD138, and the CD138+ cells were sorted by flow cytometry. CD138+ cells were subjected to an annexin V binding assay to determine cell apoptosis. The mouse femora were analyzed with an in situ TUNEL assay. Mice that received neither melphalan nor resistin served as controls. (A) Relative levels of M-proteins. (B) Percentages of CD138+ cells. (C) Percentages of apoptotic CD138+ cells. (D) Representative images of TUNEL+ cells in bone marrow. (E) Quantitative analysis of TUNEL staining. Bar: 20 μm. Original magnification × 200. The results shown represent averages ± SD (n = 5 mice/group, 3 replicate studies). *P<0.05.

DNMTs are responsible for maintenance and de novo methylation of genes.31 We found that resistin decreased methylation in the promoters of the ABCG2 and ABCC5 genes, resulting in high protein expression, and greatly decreased the expression of DNMT1 and DNMT3a, but not DNMT3b in myeloma cells, indicating that demethylation of ABC transporters is one of the mechanisms in resistin-induced myeloma drug resistance. Further studies will be needed to identify additional mechanisms underlying resistin-induced ABC gene overexpression in myeloma cells. Previous studies showed that resistin can activate PI3K/Akt and MAPK signaling pathways and promote cell proliferation and survival in myocardial infarction and in some tumors, such as prostate cancer.23,32 Resistin indirectly upregulates the expression of pro-inflammatory genes via the NF-κB signaling pathways.21 Our results showed that resistin activated the PI3K/Akt, ERK1/2, and NF-κB signaling pathways and that inhibitors of PI3K or NF-κB, but not of ERK1/2, abrogated resistin-mediated inhibition of apoptosis in myeloma cells. Akt activation has been reported to suppress the pro-apoptotic function of Bax33 and NF-κB activation has been reported to promote expression of anti-apoptotic Bcl-2 family members (Bcl-2 and Bcl-xL).26,34 Our observations that resistin decreased the cleavage of caspase-9, caspase-3, and PARP and the expression of apoptotic proteins such as Bax, while it increased the expression of anti-apoptotic proteins such as Bcl-2 and Bcl-xL, in myeloma cells are conhaematologica | 2017; 102(7)

sistent with these previous reports. Together, our findings indicate the importance of the PI3K/Akt and NF-κB signaling pathways in resistin-induced protection of myeloma cells against chemotherapy-induced apoptosis. Further investigations are necessary to identify the receptor of resistin and to determine whether resistin activates other signaling pathways and whether the activated signaling mediates resistin-induced inhibition of myeloma cell apoptosis. The expression of resistin was initially defined in adipocytes. Our studies have demonstrated that marrow adipocytes can protect myeloma cells against chemotherapy drug-induced apoptosis through adipocyte-secreted adipokines,18 and resistin can inhibit myeloma cell apoptosis and promote ABC expression. However, other stromal cells can also produce resistin, such as macrophages which are heavily infiltrated into bone marrow of myeloma patients.35 Further studies are, therefore, needed to determine the effects of resistin, secreted from different cell types of bone marrow stromal cells, on myeloma growth and survival. In addition, we observed that resistin inhibits apoptosis in human myeloma cell lines which reflect the advanced stages of myeloma. Using primary CD138+ malignant cells isolated from bone marrow aspirates of two patients with plasma cell leukemia disease, we obtained a similar result (data not shown). We still need to investigate the effects of resistin on more samples from patients with relapsed and refractory disease who have already developed drug resistance. 1279

J. Pang et al.

In summary, our studies focusing on the adipokine resistin give new insight into the pathogenesis of multiple myeloma and suggest a potential approach to prevent chemotherapy resistance in this disease. Funding This work was supported by the U.S. National Cancer Institute through R01 awards CA190863 and CA193362 and

References 1. Kuehl WM, Bergsagel PL. Molecular pathogenesis of multiple myeloma and its premalignant precursor. J Clin Invest. 2012;122(10):3456-3463. 2. Zhou W, Yang Y, Xia JL, et al. NEK2 induces drug resistance mainly through activation of efflux drug pumps and is associated with poor prognosis in myeloma and other cancers. Cancer Cell. 2013;23(1):48-62. 3. Rollig C, Knop S, Bornhauser M. Multiple myeloma. Lancet. 2015;385(9983):21972208. 4. Gottesman MM, Fojo T, Bates SE. Multidrug resistance in cancer: role of ATPdependent transporters. Nat Rev Cancer. 2002;2(1):48-58. 5. Chen Z, Shi T, Zhang L, et al. Mammalian drug efflux transporters of the ATP binding cassette (ABC) family in multidrug resistance: a review of the past decade. Cancer Lett. 2016;370(1):153-164. 6. Robey RW, Medina-Perez WY, Nishiyama K, et al. Overexpression of the ATP-binding cassette half-transporter, ABCG2 (Mxr/BCrp/ABCP1), in flavopiridol-resistant human breast cancer cells. Clin Cancer Res. 2001;7(1):145-152. 7. Liu L, Zuo LF, Guo JW. ABCG2 gene amplification and expression in esophageal cancer cells with acquired adriamycin resistance. Mol Med Rep. 2014;9(4):1299-1304. 8. Turner JG, Gump JL, Zhang C, et al. ABCG2 expression, function, and promoter methylation in human multiple myeloma. Blood. 2006;108(12):3881-3889. 9. Wielinga P, Hooijberg JH, Gunnarsdottir S, et al. The human multidrug resistance protein MRP5 transports folates and can mediate cellular resistance against antifolates. Cancer Res. 2005;65(10):4425-4430. 10. Park S, Shimizu C, Shimoyama T, et al. Gene expression profiling of ATP-binding cassette (ABC) transporters as a predictor of the pathologic response to neoadjuvant chemotherapy in breast cancer patients. Breast Cancer Res Treat. 2006;99(1):9-17. 11. Borel F, Han R, Visser A, et al. Adenosine triphosphate-binding cassette transporter genes up-regulation in untreated hepatocellular carcinoma is mediated by cellular microRNAs. Hepatology. 2012;55(3):821832. 12. Mohelnikova-Duchonova B, Brynychova V,

1280

13.

14.

15. 16.

17.

18.

19.

20. 21.

22.

23.

the MD Anderson Cancer Center SPORE in Multiple Myeloma Career Development Award (CDP-060315) and Developmental Research Program (DRP-00013585); the American Cancer Society Research Scholar Grant (127337RSG-15-069-01-TBG); the Leukemia Research Foundation; the National Natural Science Foundation of China (Grant N. 81470356); and an MD Anderson Institutional Research Grant for Basic Research.

Oliverius M, et al. Differences in transcript levels of ABC transporters between pancreatic adenocarcinoma and nonneoplastic tissues. Pancreas. 2013;42(4):707-716. Hideshima T, Mitsiades C, Tonon G, Richardson PG, Anderson KC. Understanding multiple myeloma pathogenesis in the bone marrow to identify new therapeutic targets. Nat Rev Cancer. 2007; 7(8):585-598. Gorgun GT, Whitehill G, Anderson JL, et al. Tumor-promoting immune-suppressive myeloid-derived suppressor cells in the multiple myeloma microenvironment in humans. Blood. 2013;121(15):2975-2987. Roodman GD. Targeting the bone microenvironment in multiple myeloma. J Bone Miner Metab. 2010;28(3):244-250. Chauhan D, Singh AV, Brahmandam M, et al. Functional interaction of plasmacytoid dendritic cells with multiple myeloma cells: a therapeutic target. Cancer Cell. 2009;16 (4):309-323. Yang J, He J, Wang J, et al. Constitutive activation of p38 MAPK in tumor cells contributes to osteolytic bone lesions in multiple myeloma. Leukemia. 2012;26(9):21142123. Liu ZQ, Xu JD, He J, et al. Mature adipocytes in bone marrow protect myeloma cells against chemotherapy through autophagy activation. Oncotarget. 2015; 6(33):34329-34341. Fain JN, Cheema PS, Bahouth SW, Hiler ML. Resistin release by human adipose tissue explants in primary culture: Biochem Biophys Res Commun. 2003;300(3):674678. Erratum in: Biochem Biophys Res Commun. 2003;302(4):917-918. Steppan CM, Bailey ST, Bhat S, et al. The hormone resistin links obesity to diabetes. Nature. 2001;409(6818):307-312. Kim HJ, Lee YS, Won EH, et al. Expression of resistin in the prostate and its stimulatory effect on prostate cancer cell proliferation. BJU Int. 2011;108(2 Pt 2):E77-E83. Kuo CH, Chen KF, Chou SH, et al. Lung tumor-associated dendritic cell-derived resistin promoted cancer progression by increasing Wolf-Hirschhorn syndrome candidate 1/Twist pathway. Carcinogenesis. 2013;34(11):2600-2609. Gao J, Chang Chua C, Chen Z, et al. Resistin, an adipocytokine, offers protection against acute myocardial infarction. J

Mol Cell Cardiol. 2007;43(5):601-609. 24. Lu DY, Chen JH, Tan TW, Huang CY, Yeh WL, Hsu HC. Resistin protects against 6hydroxydopamine-induced cell death in dopaminergic-like MES23.5 cells. J Cell Physiol. 2013;228(3):563-571. 25. Ding AX, Sun G, Argaw YG, Wong JO, Easwaran S, Montell DJ. CasExpress reveals widespread and diverse patterns of cell survival of caspase-3 activation during development in vivo. Elife. 2016;5. 26. Reed JC. Bcl-2-family proteins and hematologic malignancies: history and future prospects. Blood. 2008;111(7):3322-3330. 27. Lindqvist LM, Vaux DL. BCL2 and related prosurvival proteins require BAK1 and BAX to affect autophagy. Autophagy. 2014; 10(8):1474-1475. 28. Okano M, Xie S, Li E. Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat Genet. 1998;19(3):219-220. 29. Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell. 1999; 99(3):247-257. 30. Nakano H, Nakamura Y, Soda H, et al. Methylation status of breast cancer resistance protein detected by methylation-specific polymerase chain reaction analysis is correlated inversely with its expression in drug-resistant lung cancer cells. Cancer. 2008;112(5):1122-1130. 31. Ellis L, Atadja PW, Johnstone RW. Epigenetics in cancer: targeting chromatin modifications. Mol Cancer Ther. 2009 ;8(6):1409-1420. 32. Danese E, Montagnana M, Minicozzi AM, et al. The role of resistin in colorectal cancer. Clin Chim Acta. 2012;413(7-8):760-764. 33. Gardai SJ, Hildeman DA, Frankel SK, et al. Phosphorylation of Bax Ser184 by Akt regulates its activity and apoptosis in neutrophils. J Biol Chem. 2004;279(20):2108521095. 34. Huang Z. Bcl-2 family proteins as targets for anticancer drug design. Oncogene. 2000;19(56):6627-6631. 35. Zheng Y, Cai Z, Wang S, et al. Macrophages are an abundant component of myeloma microenvironment and protect myeloma cells from chemotherapy druginduced apoptosis. Blood. 2009; 114(17): 3625-3628.

haematologica | 2017; 102(7)

ARTICLE

Plasma Cell Disorders

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Martin Granzow,1 Ute Hegenbart,2 Katrin Hinderhofer,1 Dirk Hose,2 Anja Seckinger,2 Tilmann Bochtler,2,3 Kari Hemminki,4 Hartmut Goldschmidt,2,5 Stefan O. SchĂśnland2* and Anna Jauch1*

1 Institute of Human Genetics, University of Heidelberg; 2Department of Internal Medicine V, Hematology/Oncology, Amyloidosis Center, University of Heidelberg; 3Clinical Cooperation Unit Molecular Hematology/Oncology, German Cancer Research Center (DKFZ) and Department of Internal Medicine V, University of Heidelberg; 4Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Heidelberg and 5National Center for Tumor Diseases, Heidelberg, Germany

*SOS and AJ contributed equally to this work

Haematologica 2017 Volume 102(7):1281-1290

ABSTRACT

mmunoglobulin light chain (AL) amyloidosis is a rare plasma cell dyscrasia characterized by the deposition of abnormal amyloid fibrils in multiple organs, thus impairing their function. In the largest cohort studied up to now of 118 CD138-purified plasma cell samples from previously untreated immunoglobulin light chain amyloidosis patients, we assessed in parallel copy number alterations using high-density copy number arrays and interphase fluorescence in situ hybridization (iFISH). We used fluorescence in situ hybridization probes for the IgH translocations t(11;14), t(4;14), and t(14;16) or any other IgH rearrangement as well as numerical aberrations of the chromosome loci 1q21, 8p21, 5p15/5q35, 11q22.3 or 11q23, 13q14, 15q22, 17p13, and 19q13. Recurrent gains included chromosomes 1q (36%), 9 (24%), 11q (24%), as well as 19 (15%). Recurrent losses affected chromosome 13 (29% monosomy) and partial losses of 14q (19%), 16q (14%) and 13q (12%), respectively. In 88% of patients with translocation t(11;14), the hallmark chromosomal aberration in AL amyloidosis, a concomitant gain of 11q22.3/11q23 detected by iFISH was part of the unbalanced translocation der(14)t(11;14)(q13;q32) with the breakpoint in the CCND1/MYEOV gene region. Partial loss of chromosome regions 14q and 16q were significantly associated to gain 1q. Gain 1q21 detected by iFISH almost always resulted from a gain of the long arm of chromosome 1 and not from trisomy 1, whereas deletions on chromosome 1p were rarely found. Overall and event-free survival analysis found a potential adverse prognostic effect of concomitant gain 1q and deletion 14q as well as of deletion 1p. In conclusion, in the first whole genome report of clonal plasma cells in AL amyloidosis, novel aberrations and hitherto unknown potential adverse prognostic effects were uncovered. Introduction AL amyloidosis is characterized by the deposition of abnormal amyloid fibrils in multiple organs, thereby impairing their function. Plasma cells that undergo clonal alterations produce amyloid fibrils emanating from misfolding of the native protein.1 The panel of iFISH probes in AL amyloidosis originated from the diagnostic management of multiple myeloma (MM). At our center, we use a comprehensive haematologica | 2017; 102(7)

Correspondence: granzow@uni-heidelberg.de

Received: November 28, 2016. Accepted: March 15, 2017. Pre-published: March 24, 2017. doi:10.3324/haematol.2016.160721 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1281 ÂŠ2017 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.

1281

M. Granzow et al.

probe set for chromosome regions 1q21, 5p15/5q35, 8p21, 9q34, 11q22.3 or 11q23, 13q14, 15q22, 17p13, and 19p13 as well as the IgH translocations t(11;14), t(4;14), and t(14;16) or any other IgH rearrangement. Indeed, the iFISH probe set has shown a very similar aberration pattern in both AL and MM.2-8 In analogy to MM patients, the oncogenetic tree model9 distinguished AL into different subgroups: (i) hyperdiploid (HD), (ii) translocation t(11;14), (iii) non-hyperdiploid (NHD) with deletion of 13q14 / t(4;14), and (iv) IgH translocation with an unknown partner.6 The only difference was demonstrated for gain of 1q21 showing an association with the hyperdiploid subgroup in AL patients, whereas it was linked to the NHD group with deletion 13q14 / t(4;14) in MM patients. Within the group of patients with gain of 11q23, a dichotomy was observed which split into t(11;14) positive and hyperdiploid karyotypes. Using different microarray platforms, genome-wide screening for copy number (CN) aberrations has been done in MM10-16 as well as in monoclonal gammopathy of unknown significance (MGUS)17 and smoldering MM,17 the precursors of symptomatic MM. These studies confirmed the aberrations detected by routine iFISH, with the exception of balanced translocations that cannot be identified by CN array. Furthermore, several other aberrant regions were identified, some of which are associated with prognosis or the stage of plasma cell dyscrasia (MGUS, smoldering MM, or MM). Given the overall genomic similarity of AL amyloidosis and MM, it seemed obvious to perform a similar study in this plasma cell disease. In the study herein, we analyzed 118 CD138-purified plasma cell samples from AL amyloidosis patients by high-density CN array in order to detect novel CN alterations and relate these findings to known molecular entities, in particular to translocation t(11;14), the hallmark molecular alteration of AL amyloidosis.

Methods Patients One hundred and eighteen AL amyloidosis patients presenting at the Amyloidosis Center Heidelberg between 2005 and 2014 were included in the study, which was approved by the Ethics Committee (#123/2006) following written informed consent in concordance with the Declaration of Helsinki. iFISH results and clinical correlation of 44 of these patients have been published previously.5,6 Clinical characteristics of the patients including distribution of age, sex, number of involved organs, underlying clonal plasma cell dyscrasia (AL with less than 10% and AL-MM with 10% or more plasma cells in bone marrow cytology), light chain type, clinical scores, AL-specific serum parameters, and therapy regimen are summarized in Table 1. The distributions of values are typical and representative for AL amyloidosis patients in general.

Interphase FISH diagnostics For all 118 patients, iFISH was performed on CD138-positive bone marrow plasma cells purified by auto-magnetic-activated cell sorting with anti-CD138 immunobeads as described previously.18,19 Purity of sorted plasma cells ranged from 75-99% with a median of 90%. Results were available for numerical chromosome aberrations at the loci 1q21, 5p15/5q35, 8p21, 9q34, 11q22.3 or 11q23, 13q14, 15q22, 17p13, and 19p13 as well as the IgH translo1282

cations t(11;14)(q13;q32), t(4;14)(p16;q32), t(14;16)(q32;q23), and an IgH break apart probe. Hyperdiploidy was defined according to Wuilleme et al.,20 determining HD-iFISH by gains of at least two of the three iFISH probes for chromosomes 5, 9, and 15. Only patients with a minimum of one aberration detectable by iFISH in at least 60% of cells were included in the study.

Copy number array analysis For each patient, 50 ng of DNA from CD138-positive plasma cells (see above) was used. Hybridization to an AffymetrixÂŽ CytoScan HD Oligo/SNP-array was performed according to the manufacturer's instructions. Arrays were scanned with the

Table 1. Clinical characteristics of AL amyloidosis patients and iFISH results.

Number of patients

118

Age median (range) in years Sex: female / male Involved organs: median number (range) Heart Liver Kidney GI tract Soft tissue Peripheral neuropathy Underlying disease49 AL AL-MM Percent PCs median (range) Ig intact present Light chain type: Îş /l Clinical scores Karnofsky index in percent: median (range) Mayo score median (range) Mayo I, II, III NYHA classification median (range) Renal stage51 (median, range) Serum marker median (range) NT-BNP ng/l Creatinine mg/dl dFLC mg/l iFISH (No. of pts. (percentage)) NHD HD del 13q14 gain 1q21 del 17p13 t(11;14) t(4;14) t(14;16) Therapy None High-dose Len-Mel-Dex Mel-Dex Velcade

65 (41-87) 47 / 71 pts 2 (1-6) 89 pts 25 pts 66 pts 38 pts 57 pts 23 pts 48 pts 70 pts 10 (1-58) 49 pts 23 pts / 95 pts 80 (50-100) 2 (1-3) 24, 43, 48 pts 2 (0-4) 2 (1-3) 3075 (22-165677) 1.03 (0.06-10.77) 232 (1-12078) 95 (81) 23 (19) 52 (43) 43 (36) 6 (5) 73 (62) 3 (3) 2 (2) 5 24 18 33 38

The table summarizes age, sex, number of involved organs, underlying disease, percentage of plasma cells, heavy and light chain type, Karnofsky index, Mayo score,50 NYHA classification and several serum markers. AL, stage of clonal disease is amyloidosis; AL- MM, stage of clonal disease is AL+MM. NYHA: New York Heart Association; NT-BNP: N-terminal pro brain natriuretic peptide; AL: immunoglobulin light chain; MM: multiple myeloma; GI: gastrointestinal; pts, patients; FLC: free light chain; Ig: immunoglobulin; PC: plasma cell; HD: hyperdiploid; NHD: Non-hyperdiploid; del: deletion; t: translocation; Len: lenalidomid; Mel: melphalan; Dex: dexamethasone.

haematologica | 2017; 102(7)

High-density copy number array in AL amyloidosis

Figure 1. Overview of results from genome-wide copy number array analysis in AL amyloidosis patients. (A) Karyogram depicting each chromosome and each patient sample as gain and loss (right and left of each chromosome, respectively). The karyogram style allows for the clear visualization of gains and losses belonging to each single patient in addition to providing an exact insight of the overall distribution of aberrations and furthermore, trisomies are easily recognizable, which is more difficult to achieve in frequency plots. Different colors indicate the following iFISH defined subgroups: all patients with a translocation t(11;14) (n=73; green) were selected into one group, of the remaining patients, those showing a hyperdiploid karyotype (HD) according to Wuilleme et al.,20 (n=18; gray) were grouped together, again of the remaining patients, those with gain of chromosome region 1q21 (gain 1q; n=14; blue) were put into the respective group. Finally, the still remaining patients were split into those with high-risk aberrations (n=5; black) and a non-hyperdiploid group (NHD; n=8; red). This categorization was chosen in order to visualize the underlying structure according to cytogenetic risk groups. Additionally, the chromosomal location of the iFISH probes is highlighted by yellow rectangles. Genomic Recurrent Event ViEwer was used to visualize aberrations.52 (B) Frequency plot providing an overview of the distributions of gains (green) and losses (red) as percentages of all 118 patients for chromosomes 1-22.

haematologica | 2017; 102(7)

1283

M. Granzow et al.

Affymetrix GeneChip® Scanner 3000 7G and CN analysis was done with Affymetrix Chromosome Analysis Suite software version 2.1.0.16(r6634) and Annotation NetAffx Build 33. Interpretation was based on human reference sequence GRCh37/hg19, February 2009. Data were deposited at the Gene Expression Omnibus (accession GSE89616). The complete data set was visually analyzed. Gains and losses had to meet three criteria to be reported: a minimum of 25 markers per segment, a minimum genome length of 100 kb, and less than 50% overlap with known CN variants from the Database of Genomic Variants21 and/or in-house data obtained from healthy parents of patients with non-syndromic mental retardation. Copy number aberrations located at 2p11.2 (IgK@), 14q32.33 (IgH@), and 22q11.22 (IgL@), which might stem from rearrangements of the B-cell receptor gene, were excluded from further analysis as well as aberrations on the sex chromosomes. Based on the results from CN analysis, HD-CN was defined by trisomy of at least two chromosomes with respect to all analyzed chromosomes, to compare the difference to the score of Wuilleme et al.20

Statistical analysis

The χ2 test was used to analyze correlations of aberrations detected by CN array and the different cytogenetic groups. Differences were considered statistically significant at P<0.05.

To correlate CN aberrations with categorical clinical and hematological variables we used the χ2 test, and for the correlation with continuous clinical and hematological variables, we used an analysis of variance model (ANOVA). For all statistical computation, R version 3.2.2 was used with library 'stats' version 3.2.2. The library 'copynumber’22 (version 1.8.0) was used to prepare the frequency plot in Figure 1B.

Results Overview of CN aberrations detected in 118 AL amyloidosis patients In the following analysis, CN aberrations affecting whole chromosomes are not included in counts of smaller alterations, e.g., chromosome arms or regional bands. In total, the range of aberrant chromosomes per patient varied between zero (11% of patients) and 16 (one patient) with an average and median of five and three aberrant chromosomes, respectively. An overview of the distribution of all CN aberrations is presented in Figure 1. Table 2 comprises - for the sake of clarity - all aberrations detected in at least 5% of patients. The most prevalent gain detected in the study cohort was gain of chromosome

Table 2. Frequency in % of CN aberrations and concordance with iFISH results.

Chr 1q 3p 3q 4p 4q 5p 5q 6p 6q 7p 7q 8p 9p 9q 11p 11q 13q 14q 15q 16p 16q 17p 17q 18p 18q 19p 19q 20p

Trisomy 9 5 14 8 24 11 14 5 13 15 -

Gain 36 5 5 10 5 7 31 5 5 6 -

Monosomy

Loss

7 10 8 12 19 5 14 5

29 5 -

Concordance with iFISH (median: 98%) 95.7 100 99.2 98.3 94.1 95.8 92.4 100 97.5 99.2 -

Note that the numbers include not only trisomies and aberrations affecting whole chromosome arms but also smaller aberrant regions. Chromosomes with no aberration above 4% are not shown in the table. Chr: chromosome.

1284

haematologica | 2017; 102(7)

High-density copy number array in AL amyloidosis

1q (36%, including one trisomy 1), followed by trisomy 9 (24%), gain of chromosome arm 11q (24%), and trisomy 19 (15%). Trisomy of chromosomes 15 (14%), 5 (14%), 18 (13%), and 11 (11%) was observed less frequently. The most common deletion affected chromosome 13 in 40% of patients, with 70% of them showing a monosomy and 30% presenting only partial loss of chromosome 13. The

most commonly affected chromosomal region was 13q21.32-q21.33 (chr13:67,533,438-70,847,141) in 88% of all patients with deletions affecting chromosome 13. This region includes only two genes, PCDH9 (Protocadherin 9) and LINC00550, a non-coding ribonucleic acid. Deletions affected chromosome 14 in a total of 25% of all patients. In contrast to the aberrations on chromosome 13, mono-

A other HR+del1p P=0.22 (3 vs. 15 months)

other HR+del1p P=0.01 (3 vs. 28 months)

C other gain 1q+del14q P=0.06 (13 vs. 15 months)

other gain 1q+del14q P=0.35 (24 vs. 28 months)

E other gain 1q+del16q P=0.5 (12 vs. 20 months)

other gain 1q+del16q P=0.23 (27 vs. 40 months)

Figure 2. EFS and OS of AL amyloidosis patients. Event-free survival (A, C, E) and overall survival curves (B, D, F) are provided for patients with high-risk aberrations or deletion 1p vs. other patients (A, B), patients with gain 1q and concomitant partial deletion 14q vs. other patients (C, D), and patients with deletion of chromosome arm 16q vs. patients without this aberration (E, F). HR: high-risk aberrations; del1p: deletion 1p; gain1q: gain chromosome 1q; del14q: partial deletion chromosome 14q; del16q: deletion chromosome 16q; pts.: patients.

haematologica | 2017; 102(7)

1285

M. Granzow et al. Table 3. P-values (medians) of associations of chromosome aberrations with clinical parameters.

Percentage of plasma cells t(11;14) vs. no t(11;14) HD-CN vs. NHD

<0.05 (10 vs. 13) <0.005 (17 vs. 10) HD-iFISH vs. NHD <0.005 (13.5 vs. 10) Gain 1q21 vs. no gain <0.005 1q21 (13.5 vs. 10) Partial deletion 14q <0.05 vs. other (12 vs. 10) Gain 1q21 & partial <0.05 loss chr. 14q vs. other (21 vs. 10) Gain 1q21 & partial loss chr. 16q vs. other HR & del 1p vs. other Gain 1q21 & HD-CN vs. gain 1q21 & NHD

Age

MDRD

dFLC

Intact Ig

<0.001

<0.005

<0.05

<0.005

<0.01 (73 vs. 65) <0.01 (73 vs. 65) -

<0.05 (55 vs. 69) -

<0.05 (75.5 vs. 65)

<0.005 (32.6 vs. 69)

<0.05 (488 vs. 198) -

Plasma cell dyscrasia = AL+MM

t(11;14): Patients with translocation t(11;14); HD-CN: hyperdiploid group determined by CN array; HD-iFISH: hyperdiploid group determined by iFISH; NHD: Non-hyperdiploid group; Gain 1q21: Patients with gain 1q21; Gain 1q21 & partial loss chr. 14: Patients with gain 1q21 and concomitant partial loss of chromosome 14; Gain 1q21 & partial loss chr. 16q: Patients with gain 1q21 and concomitant partial loss of chromosome 16q; Gain 1q21 & HD-CN: Patients with gain 1q21 and concomitant hyperdiploid karyotype determined by CN analysis; MDRD: Modification of Diet in Renal Disease; dFLC: difference between involved FLC and uninvolved FLC; FLC: serum free light chains; HR: high-risk aberrations; del 1p: deletion chromosome 1p; -: not significant.

somy 14 was detected in only 23% of the patients with deletions concerning chromosome 14, with the remaining 77% bearing partial loss. The minimal common region of the partial deletions of chromosome 14q was 14q24.1q31.1 (chr14: 69,416,523-82,198,988) spanning 14 Mb and containing 122 genes. Seventy-one genes in this region are annotated in the Online Mendelian Inheritance in Man (OMIM) database, including the tumor suppressor genes JDP2 (Jun dimerization protein 2) and MLH3 (mutL homolog 3, E. coli), which play a role in the apoptosis signaling pathway and DNA mismatch repair mechanism, respectively. Additionally, two oncogenes are located in this region, namely, FOS (FBJ murine osteosarcoma viral oncogene homolog) and ESRRB (estrogen-related receptor-Î˛). Chromosomal regions less frequently affected by deletions were located on chromosome 16q (14%), 6q (10%), and 8p (8%).

Chromosome aberrations in molecular entities as defined by iFISH Translocation t(11;14) group: In total, our cohort contained 73 patients (62%) with a translocation involving the genes CCND1/MYEOV on chromosomes 11q13 and IgH in 14q32 detected by iFISH. Copy number array detected gain of chromosomal material of 11q in 34 of these 73 patients (47%), in 88% the breakpoint localized within the genes CCDN1/MYEOV, the defined breakpoints of the translocation t(11;14) (data not shown). Thus, in the majority of patients with a t(11;14), gain of 11q22.3 or 11q23 detected by iFISH was the result of the unbalanced translocation der(14)t(11;14)(q13;q32). This group also included four hyperdiploid patients defined by CN array and 13 patients with gain of chromosome 1q (18%). 1286

Monosomy 13 was the most common deletion (16 patients, 22%). There was an association of t(11;14) with a lower total number of aberrations detected in the patients [median of 2 vs. 5; P<0.001; all aberrations used for the HD score of Wuilleme et al.20 were counted as one (see Bochtler et al.6 for details)]. For every cytogenetic group significant associations of chromosome aberrations with clinical parameters are summarized in Table 3. Hyperdiploid group: CN array results identified 25% of the patients as HD, in addition, four patients had only one trisomy (all of chromosome 9). Most frequent gains were trisomy of chromosomes 9 (83%), 19 (59%), 15 (55%), 5 (52%), 18 (48%), 11 (45%), 7 (34%), 3 (31%), 17 (21%), 4 (21%), 21 (17%), and 14 (14%). Furthermore, gain of chromosome arm 1q was detected in 52% and partial gain of chromosome arm 6p in 14%. Concerning chromosomal losses, monosomy 13 was the most frequent aberration, found in 31% of patients, followed by deletion of material on chromosome arm 8p (14%). No association of monosomy 13 with the HD group was detected (P>0.05). Comparing HD-iFISH to HD-CN, six patients were not classified as HD by iFISH. Five of these patients showed trisomy of chromosomes 11 and 18 together with additional trisomies of either chromosome 4, 9, 14, 17, or 19, and the sixth patient was without a third trisomy. Only four of these HD patients carried a t(11;14). Interestingly, the HD-iFISH group showed a significant association with gain 1q21 (P<0.01), which was not the case for the HDCN group. Clinical associations were similar to the HDCN group (cf. Table 3). Non-hyperdiploid group: This group included 95 patients (81%) when iFISH was used to distinguish HD and NHD patients. According to CN array results, 75% of our haematologica | 2017; 102(7)

High-density copy number array in AL amyloidosis

patient cohort presented a NHD karyotype. The most frequent gains concerned chromosome arm 11q (40%), which included the localization of the breakpoint between genes CCND1 and MYEOV in 36%, followed by gains of chromosome 1q (27%). Deletions affecting chromosome 13 and 14 were detected in 40% and 28% of NHD patients, respectively. This included monosomy 13 (69%) and 14 (28%) as well as partial deletions on 13q (31%) and 14q (72%). Additionally, deletions occurred in this group pertaining to chromosome arms 16q (13%) and 6q (10%). An association of NHD karyotype to translocation t(11;14) was significant as compared to HD patients (P<0.001). In the NHD group, neither an association to IgH translocations with an unknown partner nor monosomy 13 was found (P>0.05). Patients with gain 1q21: Gain of chromosomal region 1q21 was detected in a total of 45 patients (36%; in one patient the iFISH result for 1q21 was missing where CN array detected gain of chromosome arm 1q), including four patients with subclonal presence of the respective aberration, i.e., here <35% by iFISH analysis that were not detected by CN array. In one patient, gain of chromosome region 1q21 was due to a trisomy 1 in the context of a HD karyotype. According to the number of fluorescence signals, five patients showed four copies of 1q21, indicating a possible additional isochromosome 1q, and in one patient five copies were detected, presumably resulting from unbalanced translocations involving chromosome 1q. Loss of chromosome arm 1p was detected in 3% overall, and only once together with gain 1q. Monosomy of chromosome 13 was the most frequent loss in 41% of these patients. Fifteen of the patients (32%) in this group showed a concomitant partial loss of chromosome 14q. Only three of 45 AL amyloidosis patients with gain 1q21 showed a concomitant gain of 11q. Focusing on the most commonly affected region of all patients with partial loss of chromosome 14q there was a significant association with gain 1q21 (33% vs. 12%, P<0.005). Furthermore, we also detected a significant association between patients with a deletion in this region in combination with gain 1q21 and clinical parameters (Table 3). Two patients in our cohort carried a t(4;14), together with a partial deletion 14q. The other chromosomal region frequently deleted in the gain 1q21 group concerned 16q, which was affected in 11 patients (24%) compared to 1% in the group of patients without gain 1q (P<0.001). Again, there were two patients with translocation t(14;16) in our cohort, none showed a partial deletion of 14q and only one had a deletion 16q. No significant association of this patient group to clinical parameters was observed. Gain 1q21 in association with HD vs. NHD karyotype: The group of 45 patients with gain 1q21 comprised 16 patients belonging to the HD-CN group (36%) and 29 to the NHD group (64%). These HD-CN patients with gain 1q21 frequently showed trisomy of chromosomes 9 (94%), 19 (69%), 5 (63%), 15 (63%), and 3 (50%). In the NHD group with gain of 1q21, a partial loss of 14q and monosomy 13 was detected in eleven patients each (38%) and loss of 16q in nine patients (31%). High-risk patients: Defined high-risk aberrations deletion 17p13, t(4;14), and t(14;16) were detected in six, three, and two patients, respectively. Three patients with deletion 17p13 also carried a translocation t(4;14) and two concomitantly showed a t(14;16). This group showed associhaematologica | 2017; 102(7)

ations to gain 1q and monosomy 13 in four and seven of the eight patients, respectively, whereas monosomy or partial loss of chromosome 14 was detected in three patients.

Potential prognostic role of the new CN array findings We analyzed the following genetic aberrations in terms of hematologic remission, event-free survival (EFS) and overall survival (OS): loss of 1p (together with high-risk aberrations), gain of 1q with partial deletion 14q, and deletion 16q (Figure 2). We could not find any associations with hematologic remission defined as very good partial response or better after end of treatment (data not shown). However, we found a potential negative prognostic role of loss of 1p (EFS: median of 2 months vs. 14.5, P<0.05 and OS: median of 3 vs. 28 months, P<0.05) which was maintained for OS when we grouped these patients with our high-risk aberrations (median of 3 vs. 28 months, P<0.01; Figure 2B). In addition there was a borderline significant adverse effect for deletion 14q in patients with gain 1q regarding EFS (median of 13 vs. 15 months, P=0.06, Figure 2C).

Concordance of aberrations detected by iFISH and CN array Of all 118 AL amyloidosis patients, iFISH analysis results were available for genomic regions 1q21, 5p15/5q35, 8p21, 9q34, 11q22.3 or 11q23, 13q14, 15q22, 17p13, and 19q13 in 99.8% of cases (1178 of 1180 single iFISH results). Balanced translocations detected by iFISH are not detectable by CN analysis. Overall concordance between iFISH and CN array results was 98% (range: 92100%, Table 2). In 28 patients, a discordance was observed in 32 single iFISH results attributable almost exclusively to subclone aberrations (<35%) detectable by iFISH, but hardly perceptible by CN array.

Discussion To our knowledge, this is the first report on a genomewide CN survey of a large cohort of AL amyloidosis patients. The frequencies of aberrations detected by iFISH in the study herein were comparable to those in previous patient cohorts of studies from our group and others.26,23 The most frequent chromosomal aberration detected in AL amyloidosis was translocation t(11;14), followed by gain of 11q22.3 or q23, deletion 13q14, gain of 1q21, and gain 19q13. The seemingly higher frequency of 36% of gain 1q in our patient cohort was comparable to iFISH results of probe 1q21 in MM,9-15 yet slightly higher than previously published results regarding AL amyloidosis patients ranging between 23% and 28%.6,23,24 This as well as the slightly higher frequencies of translocation t(11;14) and deletion 13q14 are partly attributable to our inclusion criteria, which only considered patients with at least one cytogenetic aberration detectable by iFISH and patients with a high enough number of purified clonal (i.e., aberrant) plasma cells, namely of 60% and above, for performing CN array analysis. Hyperdiploidy and translocation t(11;14) were shown to be mutually exclusive in AL5,6 and MM.9,25-30 CN analysis grouped more patients into the HD group as compared to iFISH. So obviously, the rate of non-hyperdiploidy karyotypes of 75% was lower than that detected in previous 1287

M. Granzow et al. iFISH trials.6 Nevertheless, the significant association between NHD status and translocation t(11;14) and the absence of association to monosomy 13 and IgH translocation with an unknown partner fully confirmed previous iFISH results. Furthermore, our results support the concept of hyperdiploidy or non-hyperdiploidy as the two major pathogenetic pathways in these plasma cell dyscrasias.28,30 The previously described association of gain 1q21 with hyperdiploidy was found in our cohort only when the HD group was defined by iFISH (comparable to the classification in Bochtler et al.6). When CN results defined HD, this association was no longer significant. As far as associations of clinical parameters are concerned, the higher frequency of intact immunoglobulin as well as a higher degree of plasmacytosis and greater age in HD as compared to NHD patients, as reported earlier,5,6 could also be confirmed by our study.

Translocation t(11;14) group Our CN analysis showed that gain of 11q detected by iFISH in patients not belonging to the HD group with a translocation t(11;14) are almost always due to an unbalanced translocation der(11)t(11;14)(q13;q32), and are not caused by trisomy of chromosome 11. In the majority of patients, the translocation breakpoint was localized within the region of CCDN1 and MYEOV. Patients in this group showed less genomic complexity indicated by a lower number of aberrations, i.e., genetic instability, which is comparable to findings in MM6,9 and reinforces the concept that the occurrence of t(11;14) as an initiating event in plasma cell ontogenesis leads to a more stable phenotype.

Hyperdiploid karyotype

found no prognostic significance of partial deletion of chromosome 16q.11,13,36-38 Two tumor suppressor genes are located in the minimal common region of the deletions on chromosome 14q: JDP2 has been described as a MYC collaborating gene that has been implicated in suppressing p53 function.39 Furthermore, it is downregulated in many human cancers40 and plays a pivotal role during myeloid and lymphoid commitment from hematopoietic progenitors.41 MLH3 is involved in DNA mismatch repair, however, data on the role of the mismatch repair pathway in MM is scarce and discordant.42,43 Interestingly, Walker et al.12 also described a deletion of region 14q24.1-q31.1 in 15% of the MM patients in their study in which gene expression differences were also analyzed. Within the deleted region 14q24.1-q31.1 a total of 33 genes were underexpressed, although it was not documented whether these included JDP2 and MLH3. As MLH3 is also involved in the pathogenesis of hereditary tumor syndromes we wanted to rule out germline deletions of MLH3 in our patients. Therefore, we checked for homozygous deletions in this region as most of the AL patients in our study were included in a recent genome-wide association study.44 Using genotype data on blood DNA from 15 patients with partial loss of chromosome 14q, we checked the MLH3 locus on 14q24 for stretches of homozygous single nucleotide polymorphisms between rs1548807 (at nucleotide 75,479,582, genome build 37) and rs7303 (75,520,065) spanning the gene. None of the 15 patients were homozygous over the MLH3 locus ruling monoallelic deletion.

Comparison of the iFISH probe set with genome-wide CN array analysis results

Detecting hyperdiploidy by CN array identified six additional patients, five of whom had not less than trisomies of chromosome 11 and 18, whereas the sixth of these patients showed a trisomy of chromosomes 9 and 17. Thus, the percentage of HD patients in our AL amyloidosis cohort is slightly underestimated by the score of Wuilleme et al.,20 as already considered by Avet-Loiseau et al. for MM.31 When CN aberrations were used to define hyperdiploidy, no significant difference could be detected between the appearance of gain 1q21 in HD and NHD patients. Future studies that include trisomies of chromosomes 11 and 18 to define HD patients could investigate if the association of gain 1q to HD is more a matter of bias due to the probes for chromosomes used for classification or actually exists.

We were able to show the reliability of the iFISH panel that is used in the diagnostics of AL amyloidosis patients by the high concordance with the CN results (median: 98%). This high concordance is in line with similar studies investigating MM and chronic lymphocytic leukemia patients by array-based Comparative Genomic Hybridization that report a concordance of 91.8% and 95.5% for overall results, respectively.15,45 As demonstrated by the CN array analysis data of this study, our iFISH panel covers most relevant aberrations. CN analysis, in contrast to iFISH, is able to clearly distinguish whole chromosome gains from partial gains and thus accurately identify trisomies, which is particularly relevant in the context of hyperdiploid karyotypes. Additionally, the use of magnetic-activated cell sorting of the CD138-positive plasma cells to conduct iFISH and CN array analysis contributed to the high concordance of CN and iFISH results in our study.

Gain of 1q21

Potential prognostic role of the new CN array findings

Gain of the region investigated by the iFISH probe for chromosome 1, i.e., 1q21, almost exclusively resulted from a gain of the whole long arm of chromosome 1 but not due to trisomy 1. Gain 1q21 has been demonstrated to be of adverse prognostic impact in melphalan/dexamethasone treated AL amyloidosis patients,25 and is considered by some authors to be a progression marker in AL23,24 as well as in MM.32-35 The association of partial deletions of chromosomes 14q and 16q - not attributable to t(4;14) or t(14;16), respectively - with gain 1q21 has not been noted before in clonal plasma cells. Deletions of material of chromosome 14q and 16q have been described in MM with significant adverse prognostic effect, but others

Our survival analyses imply a possible prognostically adverse effect of loss of 1p and deletion 14q. However, these results should be interpreted with caution as our patients were not treated homogenously (see Table 1), and the cohort size as well as the number of patients with the respective CN aberration are rather small for prognostic assessments.24,46,47 Deletions on chromosome 1p were rarely detectable in our AL cohort (3%), which is contrary to the findings in MM, where it was found in up to 25% and 30% in a cohort of newly diagnosed and relapsed MM patients,13,15 respectively, and has been shown to be associated with adverse prognosis.13,48 Deductive reasoning suggests that

1288

haematologica | 2017; 102(7)

High-density copy number array in AL amyloidosis

the next step will be to evaluate our findings in a larger cohort of homogenously treated AL amyloidosis patients, e.g., by suitable iFISH probes, and evaluate a potential additional prognostic effect of the deletions on chromosome 1p and 14q. In conclusion, we report for the first time on a genomewide CN array analysis of a large cohort of 118 AL amyloidosis patients. We were able to detect hitherto unrecognized associations between prognostically relevant chromosomal aberrations in addition to the confirmation of known associations between CN aberrations. With respect to a potentially adverse prognostic effect, the concomitant partial deletion of 14q and gain 1q as well as the observed effect of a partial deletion of chromosome 1p should be further investigated to validate the results of this study.

References 11. 1. Comenzo RL, Zhang Y, Martinez C, et al. The tropism of organ involvement in primary systemic amyloidosis: contributions of Ig VL germ line gene use and clonal plasma cell burden. Blood. 2001;98(3):714-720. 2. Hayman SR, Bailey RJ, Jalal SM, et al. Translocations involving the immunoglobulin heavy-chain locus are possible early genetic events in patients with primary systemic amyloidosis. Blood. 2001;98(7):22662268. 3. Harrison CJ, Mazzullo H, Ross FM, et al. Translocations of 14q32 and deletions of 13q14 are common chromosomal abnormalities in systemic amyloidosis. Br J Haematol. 2002;117(2):427-435. 4. Bryce AH, Ketterling RP, Gertz MA, et al. Translocation t(11;14) and survival of patients with light chain (AL) amyloidosis. Haematologica. 2009;94(3):380-386. 5. Bochtler T, Hegenbart U, Cremer FW, et al. Evaluation of the cytogenetic aberration pattern in amyloid light chain amyloidosis as compared with monoclonal gammopathy of undetermined significance reveals common pathways of karyotypic instability. Blood. 2008;111(9):4700-4705. 6. Bochtler T, Hegenbart U, Heiss C, et al. Hyperdiploidy is less frequent in AL amyloidosis compared with monoclonal gammopathy of undetermined significance and inversely associated with translocation t(11;14). Blood. 2011;117(14):3809-3815. 7. Zhou P, Hoffman J, Landau H, Hassoun H, Iyer L, Comenzo RL. Clonal plasma cell pathophysiology and clinical features of disease are linked to clonal plasma cell expression of cyclin D1 in systemic lightchain amyloidosis. Clin Lymphoma Myeloma Leuk. 2012;12(1):49-58. 8. Warsame R, Kumar SK, Gertz MA, et al. Abnormal FISH in patients with immunoglobulin light chain amyloidosis is a risk factor for cardiac involvement and for death. Blood Cancer J. 2015;5:e310. 9. Cremer FW, Bila J, Buck I, et al. Delineation of distinct subgroups of multiple myeloma and a model for clonal evolution based on interphase cytogenetics. Genes Chromosomes Cancer. 2005;44(2):194-203. 10. Carrasco DR, Tonon G, Huang Y et al. High-resolution genomic profiles define distinct clinico- pathogenetic subgroups of

haematologica | 2017; 102(7)

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

Funding This work was supported by the Federal Ministry of Education (BMBF) in the context of the German AL Amyloidosis Consortium GERAMY to UH, SOS, KH, and AJ (grant 01GM1107), the German Federal Ministry of Education (BMBF) within the framework of the e:Med research and funding concept “CLIOMMICS” (01ZX1309) and “CAMPSIMM” (01ES1103) to DH, AS, HG, and the Deutsche Forschungsgemeinschaft (SFB/TRR79) to DH, AS, HG. The authors thank Michelle Ebentheuer, Evelyn Fey, Alexandra Koeppel, Kristin Schmitt, Annkathrin Borowski, Michaela Brough, and Stephanie Pschowski-Zuck for their excellent technical assistance in iFISH and/or CN analysis as well as MarieLouise Brygider, Maria Dörner, Ewelina Nickel, and Hendrike Seidt for plasma cell enrichment.

multiple myeloma patients. Cancer Cell. 2006; 9(4):313-325. Avet-Loiseau H, Li C, Magrangeas F et al. Prognostic significance of copy-number alterations in multiple myeloma. J Clin Oncol. 2009;27(27):4585-4590. Martínez-Climent JA, Fontan L, Fresquet V, Robles E, Ortiz M, Rubio A. Integrative oncogenomic analysis of microarray data in hematologic malignancies. Methods Mol Biol. 2010;576:231-77. Walker BA, Leone PE, Chiecchio L, et al. A compendium of myeloma-associated chromosomal copy number abnormalities and their prognostic value. Blood. 2010; 116(15):e56-65. Kamada Y, Sakata-Yanagimoto M, Sanada M, et al. Identification of unbalanced genome copy number abnormalities in patients with multiple myeloma by singlenucleotide polymorphism genotyping microarray analysis. Int J Hematol. 2012;96(4):492-500. Smetana J, Frohlich J, Zaoralova R, et al. Genome-wide screening of cytogenetic abnormalities in multiple myeloma patients using array-CGH technique: a Czech multicenter experience. Biomed Res Int. 2014;2014:209670. Chretien ML, Corre J, Lauwers-Cances V, et al. Understanding the role of hyperdiploidy in myeloma prognosis: which trisomies really matter? Blood. 2015;126(25):2713-2719. López-Corral L, Sarasquete ME, Beà S, et al. SNP-based mapping arrays reveal high genomic complexity in monoclonal gammopathies, from MGUS to myeloma status. Leukemia. 2012;26(12):2521-2529. Hose D, Moreaux J, Meissner T, et al. Induction of angiogenesis by normal and malignant plasma cells. Blood. 2009; 114(1):128-143. Seckinger A, Meissner T, Moreaux J et al. Clinical and prognostic role of annexin A2 in multiple myeloma. Blood. 2012; 120(5):1087-1094. Wuilleme S, Robillard N, Lode L, et al. Ploidy, as detected by fluorescence in situ hybridization, defines different subgroups in multiple myeloma. Leukemia. 2005;19(2):275-278. Zhang J, Feuk L, Duggan GE, Khaja R, Scherer SW. Development of bioinformatics resources for display and analysis of copy number and other structural variants

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

in the human genome. Cytogenet Genome Res. 2006;115(3-4):205-214. Nilsen G, Liestol K, Van Loo P et al. Copynumber: Efficient algorithms for single- and multi-track copy number segmentation. BMC Genomics. 2012;13(1):591. Bochtler T, Hegenbart U, Kunz C, et al. Gain of chromosome 1q21 is an independent adverse prognostic factor in light chain amyloidosis patients treated with melphalan/dexamethasone. Amyloid. 2014; 21(1):9Bochtler T, Hegenbart U, Kunz C, et al. Translocation t(11;14) is associated with adverse outcome in patients with newly diagnosed AL amyloidosis when treated with bortezomib-based regimens. J Clin Oncol. 2015; 33(12):1371-1378. Fonseca R, Barlogie B, Bataille R, et al. Genetics and cytogenetics of multiple myeloma: a workshop report. Cancer Res. 2004;64(4):1546-1558. Smadja NV, Leroux D, Soulier J, et al. Further cytogenetic characterization of multiple myeloma confirms that 14q32 translocations are a very rare event in hyperdiploid cases. Genes Chromosomes Cancer. 2003;38(3):234-239. Gutierrez NC, Garcia JL, Hernandez JM, et al. Prognostic and biologic significance of chromosomal imbalances assessed by comparative genomic hybridization in multiple myeloma. Blood. 2004; 104(9):2661-2666. Chiecchio L, Protheroe RK, Ibrahim AH,et al. Deletion of chromosome 13 detected by conventional cytogenetics is a critical prognostic factor in myeloma. Leukemia. 2006; 20(9):1610-1617. Fonseca R, Debes-Marun CS, Picken EB, et al. The recurrent IgH translocations are highly associated with nonhyperdiploid variant multiple myeloma. Blood. 2003; 102(7):2562-2567. Chng WJ, Santana-Davila R, Van Wier SA, et al. Prognostic factors for hyperdiploidmyeloma: effects of chromosome 13 deletions and IgH translocations. Leukemia. 2006;20(5):807-813. Avet-Loiseau H, Attal M, Moreau P, et al. Genetic abnormalities and survival in multiple myeloma: the experience of the Intergroupe Francophone du Myelome. Blood. 2007;109(8):3489-3495. Hanamura I, Stewart JP, Huang Y, et al. Frequent gain of chromosome band 1q21 in

1289

M. Granzow et al.

33.

34.

35.

36.

37.

38.

1290

plasma-cell dyscrasias detected by fluorescence in situ hybridization: incidence increases from MGUS to relapsed myeloma and is related to prognosis and disease progression following tandem stem-cell transplantation. Blood. 2006; 108(5):1724– 1732. Fonseca R, Van Wier SA, Chng WJ, et al. Prognostic value of chromosome 1q21 gain by fluorescent in situ hybridization and increase CKS1B expression in myeloma. Leukemia. 2006; 20(11):2034–2040. Chang H, Yeung J, Xu W, Ning Y, Patterson B. Significant increase of CKS1B amplification from monoclonal gammopathy of undetermined significance to multiple myeloma aet al. Further cytogenetic characterization of multiple myeloma confirms that 14q32 translocations are a very rare event in hyperdiploid cases. Genes Chromosomes Cancer. 2003;38(3):234-239. López-Corral L, Gutierrez NC, Vidriales MB, et al. The progression from MGUS to smoldering myeloma and eventually to multiple myeloma involves a clonal expansion of genetically abnormal plasma cells. Clin Cancer Res. 2011;17(7):1692–1700. Jenner MW, Leone PE, Walker BA, et al. Gene mapping and expression analysis of 16q loss of heterozygosity identifies WWOX and CYLD as being important in clinical outcome in multiple myeloma. Blood. 2007;110(9):3291–3300. Kjeldsen E. Identification of prognostically relevant chromosomal abnormalities in routine diagnostics of multiple myeloma using genomic profiling. Cancer Genomics Proteomics. 2016;13(2):91-127. Reindl L, Bacher U, Dicker F, et al. Biological and clinical characterization of

39.

40.

41.

42.

43.

44.

45.

46.

recurrent 14q deletions in CLL and other mature B-cell neoplasms. Br J Haematol. 2010;151(1):25-36. Heideman MR, Wilting RH, Yanover E, et al. Dosage-dependent tumor suppression by histone deacetylases 1 and 2 through regulation of c-Myc collaborating genes and p53 function. Blood. 2013; 121(11):2038-2050. Heinrich R, Livne E, Ben-Izhak O, Aronheim A. The c-Jun dimerization protein 2 inhibits cell transformation and acts as a tumor suppressor gene. J Biol Chem. 2004;279(7):5708-5715. Ji H, Ehrlich LIR, Seita J, et al. Comprehensive methylome map of lineage commitment from haematopoietic progenitors. Nature. 2010; 467(7313): 338342. Kotoula V, Hytiroglou P, Kaloutsi V, Barbanis S, Kouidou S, Papadimitriou CS. Mismatch repair gene expression in malignant lymphoproliferative disorders of B-cell origin. Leuk Lymphoma. 2002;43(2):393399. Martin P, Santón A, García-Cosio M, Bellas C. hMLH1 and MGMT inactivation as a mechanism of tumorigenesis in monoclonal gammopathies. Mod Pathol. 2006; 19(7):914-921. Ouyang 1, Gou X, Ma Y, Huang Q, Jiang T. Prognostic value of 1p deletion for multiple myeloma: a meta-analysis. Int J Lab Hematol. 2014;36(5):555-565. Gunn SR, Mohammed MS, Gorre ME et al. Whole-genome scanning by array comparative genomic hybridization as a clinical tool for risk assessment in chronic lymphocytic leukemia. J Mol Diag 2008; 10(5):442-451. Muchtar E, Dispenzieri A, Kumar SK et al.

47.

48.

49.

50.

51.

52.

Interphase fluorescence in situ hybridization in untreated AL amyloidosis has an independent prognostic impact by abnormality type and treatment category. Leukemia. 2016 Dec 16. doi: 10.1038/leu.2016.369. [Epub ahead of print]. Bochtler T, Hegenbart U, Kunz C et al. Prognostic impact of cytogenetic aberrations in AL amyloidosis patients after highdose melphalan: a long-term follow-up study. Blood. 2016; 128(4):594-602. da Silva Filho MI, Försti A, Weinhold N et al. Genome-wide association study of immunoglobulin light chain amyloidosis in three patient cohorts: comparison to myeloma. Leukemia 2017; advance online publication 13 January 2017; doi: 10.1038/leu.2016.387 Kourelis TV, Kumar SK, Gertz MA, et al. Coexistent multiple myeloma or increased bone marrow plasma cells define equally high-risk populations in patients with immunoglobulin light chain amyloidosis. J Clin Oncol. 2013;31(34):4319-4324. Dispenzieri A, Gertz MA, Kyle RA, et al. Serum cardiac troponins and N-terminal pro-brain natriuretic peptide: a staging system for primary systemic amyloidosis. J Clin Oncol. 2004;22(18):3751-3757. Palladini G, Hegenbart U, Milani P, et al. A staging system for renal outcome and early markers of renal response to chemotherapy in AL amyloidosis. Blood. 2014; 124(15):2325-2332. Cazier JB, Holmes CC, Broxholme J. GREVE. Genomic Recurrent Event ViEwer to assist the identification of patterns across individual cancer samples. Bioinformatics. 2012;28(22):2981-2982.

haematologica | 2017; 102(7)

ARTICLE

Cell Therapy & Immunotherapy

Effect of antithymocyte globulin source on outcomes of bone marrow transplantation for severe aplastic anemia

Natasha Kekre,1 Ying Zhang,2 Mei-Jie Zhang,2,3 Jeanette Carreras,3 Parvez Ahmed,4 Paolo Anderlini,5 Elias Hallack Atta,6 Mouhab Ayas,7 Jaap Jan Boelens,8 Carmem Bonfim,9 H. Joachim Deeg,10 Neena Kapoor,11 Jong-Wook Lee,12 Ryotaro Nakamura,13 Michael A. Pulsipher,11 Mary Eapen3 and Joseph H. Antin14

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(7):1291-1298

Division of Hematology, The Ottawa Hospital, ON, Canada; 2Division of Biostatistics, Institute for Health and Society, Medical College of Wisconsin, Milwaukee, WI, USA; 3 Center for International Blood and Marrow Transplant Research, Department of Medicine, Medical College of Wisconsin, Milwaukee, WI, USA; 4Armed Forces Bone Marrow Transplant Center, Rawalpindi, Pakistan; 5Department of Stem Cell Transplantation and Cellular Therapy, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; 6Instituto Nacional de Câncer, Rio de Janeiro, Brazil; 7 Department of Pediatric Hematology/Oncology, King Faisal Specialist Hospital & Research Center, Riyadh, Saudi Arabia; 8Department of Pediatrics, University Medical Center Utrecht, The Netherlands; 9Hospital de Clinicas, Federal University of Parana, Curitiba, Brazil; 10Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; 11Division of Hematology, Oncology and Blood & Marrow Transplantation, Children’s Hospital Los Angeles, CA, USA; 12BMT Center, Seoul St. Mary’s Hospital, The Catholic University of Korea, Seoul, South Korea; 13Department of Hematology & Hematopoietic Cell Transplantation, City of Hope, Duarte, CA, USA and 14 Division of Hematologic Malignancies, Dana-Farber Cancer Institute, Boston, MA, USA 1

ABSTRACT

or treatment of severe aplastic anemia, immunosuppressive therapy with horse antithymocyte globulin results in superior response and survival compared with rabbit antithymocyte globulin. This relative benefit may be different in the setting of transplantation as rabbit antithymocyte globulin results in more profound immunosuppression. We analyzed 833 severe aplastic anemia transplants between 2008 and 2013 using human leukocyte antigen (HLA)-matched siblings (n=546) or unrelated donors (n=287) who received antithymocyte globulin as part of their conditioning regimen and bone marrow graft. There were no differences in hematopoietic recovery by type of antithymocyte globulin. Among recipients of HLA-matched sibling transplants, day 100 incidence of acute (17% versus 6%, P<0.001) and chronic (20% versus 9%, P<0.001) graft-versus-host disease were higher with horse compared to rabbit antithymocyte globulin. There were no differences in 3-year overall survival, 87% and 92%, P=0.76, respectively. Among recipients of unrelated donor transplants, acute graft-versus-host disease was also higher with horse compared to rabbit antithymocyte globulin (42% versus 23%, P<0.001) but not chronic graft-versus-host disease (38% versus 32%, P=0.35). Survival was lower with horse antithymocyte globulin after unrelated donor transplantation, 75% versus 83%, P=0.02. These data support the use of rabbit antithymocyte globulin for bone marrow transplant conditioning for severe aplastic anemia.

Introduction Aplastic anemia is a bone marrow failure syndrome that is almost always associated with an aberrant immune response that leads to activated type 1 cytotoxic T cells which destroy hematopoietic stem cell progenitors.1 The current standard approach to severe aplastic anemia (SAA) includes immunosuppressive therapy haematologica | 2017; 102(7)

Correspondence: meapen@mcw.edu

Received: January 12, 2017. Accepted: March 16, 2017. Pre-published: March 24, 2017. doi:10.3324/haematol.2017.164459 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/7/1291 ©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.

1291

N. Kekre et al. (IST) and/or allogeneic transplantation.2 Studies of IST in SAA have shown that horse antithymocyte globulin (hATG) is superior to rabbit ATG (r-ATG; thymoglobulin), with both better response rates and survival.3-6 This observation is surprising as r-ATG has been successfully used in patients who fail IST with h-ATG.7,8 However, r-ATG is associated with a more effective depletion of lymphocytes,9 which may be the reason for a delayed time to remission after IST with r-ATG.10 In addition to lymphocyte depletion, r-ATG and not h-ATG enhances the number and function of regulatory T cells11,12 which are important in suppressing immune response and maintaining tolerance. The preservation or permissive expansion of regulatory T cells could be beneficial in limiting graft-versushost disease (GvHD) after allogeneic transplantation as these cells are needed for tolerance, controlling alloreactive donor lymphocytes involved in GvHD as well as innate and adaptive immune responses. These mechanistic differences between r-ATG and h-ATG may not lead to the same results following allogeneic transplantation as reported after IST, as regulatory T-cell induction and T-cell depletion may be more relevant to GvHD and graft rejection in the setting of allogeneic transplantation. We therefore sought to determine the difference, if any, in outcomes between h- and r-ATG in HLA-matched sibling and HLA-matched or mismatched unrelated donor bone marrow transplantation in SAA.

Methods Patients Data were reported to the Center for International Blood and Marrow Transplant Research (CIBMTR), a voluntary group of over 350 transplant centers that contribute data on consecutive transplants performed at each center with longitudinal follow up until death or loss to follow up. Eligible patients were aged 1 to 71 years old with acquired SAA, and received h- or r-ATG (thymoglobulin) transplanted with bone marrow grafts from a HLAmatched sibling or unrelated donor between 2008 and 2013 at 145 centers. Recipients of peripheral blood grafts were excluded (HLAmatched sibling n=136; unrelated donor n=140). The analysis was restricted to bone marrow transplants (BMT) as we previously showed that bone marrow is the preferred graft for SAA transplants.13,14 The Institutional Review Board of the National Marrow Donor Program approved this study.

Endpoints The primary endpoint was overall survival. Death from any cause was considered an event and surviving patients were censored at last follow up. Neutrophil recovery was defined as achieving an absolute neutrophil count of ≥ 0.5 × 109/L for 3 consecutive days, and platelet recovery as achieving platelet count ≥ 20 × 109/L, unsupported by transfusion for 7 days. Incidences of grades 2 to 4 acute GvHD and chronic GvHD were based on reports from each transplant center using standard criteria.15,16

Statistical Methods Analyses were undertaken separately by donor type. Within each donor group, characteristics and outcomes were compared by the formulation of ATG: h-ATG or r-ATG. Patient, disease and transplant-related characteristics were compared using the χ2 statistic. The probability of overall survival was calculated with the Kaplan-Meier estimator,17 and the incidence of hematopoietic recovery, infections and acute and chronic GvHD were deter1292

mined using the cumulative incidence estimator18 to accommodate competing risks. The 95% confidence interval was generated by log transformation. Cox regression models were built for acute and chronic GvHD and mortality to identify factors associated with these outcomes.19 Variables tested included a term for ATG type (h-ATG vs. r-ATG), patient age (<20 vs. ≥20 years), sex (male vs. female), performance score (90 – 100 vs. <90), comorbidity index (0-2 vs. ≥3), interval between diagnosis and BMT (< 3 vs. ≥3 months for HLA-matched sibling and <12 months vs. ≥12 months for unrelated donor BMT), conditioning regimen (total body irradiation (TBI) containing regimen vs. non-TBI containing regimen for unrelated donor BMT and cyclophosphamide vs. cyclophosphamide + fludarabine for HLA-matched sibling BMT), donorrecipient HLA match (8/8 vs. 8/8 HLA match for unrelated donor BMT only), GvHD prophylaxis (calcineurin inhibitor with mycophenolate vs. with methotrexate) and BMT period (20082010 vs. 2011-2013). The final model included type of ATG regardless of level of significance and other factors that attained a P-value ≤0.05. All P-values are two-sided. There were no firstorder interactions between ATG type and other factors held in the final model. Analyses were carried out using SAS software, Version 9.3 (Cary, NC, USA).

Table 1A. Characteristics of patients undergoing HLA-matched sibling transplant. Number of patients Patient age, years Median <20 20-39 ≥40 Sex Male Female Performance score <90% 90-100% Missing Comorbidity index 0-2 3+ Not reported Time from diagnosis to transplant <6 months ≥6 months Conditioning Regimen Cyclophosphamide + fludarabine Cyclophosphamide alone GvHD prophylaxis Tacrolimus or cyclosporine + methotrexate Tacrolimus or cyclosporine + mycophenolate Year of transplant 2008-2010 2011-2013 Median follow up, median (range), months

h-ATG

r-ATG

278

268

16 (1-69) 176 (63) 65 (23) 37 (13)

20 (1-67) 133 (50) 75 (28) 60 (22)

149 (54) 129 (46)

135 (50) 133 (50)

68 (24) 201 (72) 9 ( 3)

54 (20) 210 (78) 4 ( 1)

218 (83) 44 (17) 16

197 (93) 15 ( 7) 56

223 (80) 55 (20)

188 (70) 80 (30)

P 0.003

0.45

0.18

0.001

0.006

<0.001 29 (10) 249 (90)

89 (33) 179 (67)

249 (90)

248 (93)

29 (10)

20 ( 7)

130 (47) 148 (53) 35 (3-82)

158 (59) 110 (41) 26 (1-77)

0.22

0.004

ATG: antithymocyte globulin; h-ATG: horse derived ATG; r-ATG: rabbit derived ATG; GvHD: graft-versus-host disease

haematologica | 2017; 102(7)

Rabbit vs. horse ATG in BMT for aplastic anemia

Results Patient, disease and transplant characteristics Patient, disease and transplant characteristics of the study population are shown in Tables 1A (HLA-matched sibling donors) and 1B (unrelated donors) by type of ATG. No patient received ATG-Fresenius or Lymphoglobuline h-ATG. HLA-matched sibling BMT recipients of h-ATG were younger, more likely to have a higher comorbidity index, receive a transplant within 6 months from diagnosis, receive cyclophosphamide as the sole chemotherapeutic agent and to be transplanted after 2010. There were no differences in regards to sex, performance score, conditioning regimen and GvHD prophylaxis. Table 1B shows recipients of HLA-matched or mismatched unrelated donor BMT. The only difference in characteristics was the BMT conditioning regimen. Although most recipients of unrelated donor BMT received low-dose TBI, recipients of h-ATG were more likely to receive cyclophosphamide with TBI, and recipients of r-ATG were more likely to

receive cyclophosphamide and fludarabine with TBI. The dose of ATG was available for approximately half of HLAmatched sibling and unrelated donor transplants; the median dose of h-ATG was 90 mg/kg (range 60 mg/kg – 150 mg/kg) and that for r-ATG was 9 mg/kg (range 6 mg/kg – 15 mg/kg) for both donor types.

Hematopoietic Recovery The probability of neutrophil recovery at day 28 was not different for h-ATG and r-ATG for HLA-matched sibling BMT, 86% (95% confidence interval (CI) 82-90) and 89% (95% CI 85-92), P=0.43 or unrelated donor BMT, 90% (95% CI 84-94) and 88% (82-92), P=0.59, respectively. The corresponding median time to neutrophil recovery for h-ATG and r-ATG for HLA-matched sibling BMT was 18 and 17 days and for unrelated donor BMT, it was 19 days, regardless of type of ATG. Similarly, the probability of platelet recovery at 100 days was not different for hATG and r-ATG for HLA-matched sibling BMT, 95% (95% CI 92-97) and 92% (95% CI 87-95), P=0.14 and

Table 1B. Characteristics of patients undergoing unrelated donor transplant Number of patients Patient age, years Median <20 ≥20 Sex Male Female Performance score <90% 90-100% Missing Comorbidity index 0-2 3+ Missing Time from diagnosis to transplant Median (range), months <12 months ≥12 months Conditioning Regimen Cyclophosphamide + fludarabine + TBI 200 cGy Cyclophosphamide + fludarabine Cyclophosphamide +TBI 200 cGy Cyclophosphamide alone Donor-Recipient HLA match 8/8 HLA match 7/8 HLA match GvHD prophylaxis Tacrolimus or cyclosporine + methotrexate Tacrolimus or cyclosporine + mycophenolate Year of transplant 2008-2010 2011-2013 Median follow up, median (range), months

h-ATG

r-ATG

126

161

21 (2-67) 57 (45) 69 (55)

20 (<1-66) 82 (51) 79 (49)

68 (54) 58 (46)

81 (50) 80 (50)

34 (27) 91 (73) 1

35 (22) 122 (78) 4

92 (74) 32 (26) 2

122 (77) 36 (23) 3

9 (2-210) 78 (62) 48 (38)

10 (1-298) 93 (58) 68 (42)

39 (31) 15 (12) 57 (45) 15 (12)

99 (61) 29 (18) 21 (13) 12 ( 7)

101 (80) 25 (20)

121 (75) 40 (25)

108 (86) 18 (14)

143 (89) 18 (11)

52 (41) 74 (59) 35 (11-73)

60 (37) 101 (63) 26 (4-77)

P 0.33

0.53

0.34

0.55

0.47

<0.001

0.31

0.43

0.49

ATG: antithymocyte globulin; h-ATG: horse derived ATG; r-ATG: rabbit derived ATG; GvHD: graft-versus-host disease; HLA: human leukocyte antigen; TBI: total body irradiation.

haematologica | 2017; 102(7)

1293

N. Kekre et al.

unrelated donor BMT, 81% (95% CI 74-88) and 88% (95% CI 83-93), P=0.10. The corresponding median time to platelet recovery for h-ATG and r-ATG for HLAmatched sibling BMT was 24 and 25 days and for unrelated donor BMT it was 27 and 26 days, respectively.

patients after r-ATG. Among recipients of unrelated donor BMT, 4 patients developed EBV-associated lymphoproliferative disease after h-ATG compared to 13 patients after r-ATG. This limited number of events prevented us from calculating the incidence of EBV-associated lymphoproliferative disease.

Infections Data on infections post-transplant was available for approximately 25% of HLA-matched siblings and 50% of unrelated donor transplant recipients. Those infections considered included bacterial, viral, fungal and parasitic within the first 100 days after transplantation. The incidence of any infection did not differ by type of ATG. Among recipients of HLA-matched sibling donor transplants, the day 100 cumulative incidence of infection was 67% (95% CI 53 – 77) and 74% (58 – 85) after h-ATG and r-ATG, respectively (P=0.41). The corresponding rates following unrelated donor transplantation were 72% (95% CI 56 -83) and 84% (95% CI 73 – 91), P=0.13. Regardless of donor type, bacterial and viral infections were predominant, and there were no differences in the proportion of bacterial, viral and fungal infection by type of ATG (data not shown). There were no parasitic infections reported.

Epstein-Barr virus (EBV) associated lymphoproliferative disease EBV-associated lymphoproliferative disease was uncommon. Among recipients of HLA-matched sibling donor BMT, only 1 patient developed EBV-associated lymphoproliferative disease after h-ATG compared to 6

Acute and Chronic GvHD Grade II-IV acute GvHD was higher with h-ATG compared to r-ATG after HLA-matched sibling BMT (Table 2A). Acute GvHD risk was lower for patients aged less than 20 years and for males. Recipients of h-ATG reported grade II (n=24) and grade III (n=21) acute GvHD whereas r-ATG only reported grade II (n=16) acute GvHD. The day 100 incidence of acute grade II-IV GvHD adjusted for patient age and sex was 17% (95% CI 13 – 21) with hATG and 6% (95% CI 3 – 9) with r-ATG, P<0.001 (Figure 1A). Chronic GvHD was also higher with h-ATG compared to r-ATG after HLA-matched sibling BMT and it was lower for those aged less than 20 years and for males, independent of the type of ATG (Table 2A). The severity of chronic GvHD did not differ by type of ATG (P=0.15). Among recipients of h-ATG, 47 reported chronic GvHD, and severity was reported as limited (n=18) and extensive (n=29). Among recipients of r-ATG, 21 reported chronic GvHD, and severity was reported as limited (n=12) and extensive (n=9). The 3-year incidence of chronic GvHD adjusted for patient age and sex was 20% (95% CI 15 – 25) with h-ATG and 9% (95% CI 6 – 14) with r-ATG, P<0.001 (Figure 1B).

Figure 1. Acute and chronic graft-versus-host disease (GvHD). The adjusted cumulative incidence of grade II – IV acute GvHD (A) and chronic GvHD (B) after HLAmatched sibling transplant, and grade II – IV acute GvHD (C) and chronic GvHD (D) after unrelated donor transplant. h-ATG: horse derived ATG; r-ATG: rabbit derived ATG.

1294

haematologica | 2017; 102(7)

Rabbit vs. horse ATG in BMT for aplastic anemia

Among recipients of unrelated donor BMT, grade II-IV acute GvHD was higher with h-ATG compared to r-ATG (Table 2B). The risk of acute GvHD was higher for recipients of a single HLA locus mismatched unrelated donor BMT. Recipients of h-ATG reported grade II (n=23) and grade III-IV (n=32) and those who received r-ATG reported grade II (n=25) and grade III-IV (n=14) acute GvHD. The day 100 incidence of acute grade II-IV GvHD adjusted for donor-recipient HLA match was 42% (95% CI 34 – 50) with h-ATG and 23% (95% CI 17 – 29) with r-ATG, P<0.001 (Figure 1C). Restricting the population to 8/8 HLA-matched transplants, the day 100 incidence of grade II-IV acute GvHD was 20% (95% CI 12-30%) with hATG and 10% (95% CI 5-16%) with r-ATG (P=0.05). The risk of chronic GvHD after unrelated donor BMT did not differ by type of ATG (Table 2B). However, the severity of chronic GvHD differed by ATG type in that recipients of h-ATG were more likely to report extensive chronic GvHD (P=0.01). Of the 43 patients with chronic GvHD who received h-ATG, the severity was reported as limited (n=14) and extensive (29); 49 patients with chronic GvHD who received r-ATG reported its severity as limited (n=29)

and extensive (n=20). Patients aged 20 years or older, and the addition of mycophenolate to tacrolimus or cyclosporine rather than methotrexate were factors associated with a higher risk of chronic GvHD (Table 2B). Donor-recipient HLA match was not associated with a risk of chronic GvHD. The 3-year incidence of chronic GvHD adjusted for patient age and GvHD prophylaxis was 38% (95% CI 29 – 46) with h-ATG and 32% (95% CI 25 – 40) with r-ATG, P=0.35 (Figure 1D).

Overall Survival Among recipients of HLA-matched sibling BMT, there was no difference in survival by type of ATG (Table 2A). Mortality risk was higher for patients aged 40 years or older and when the interval from diagnosis to BMT was longer than 3 months. The 3-year probability of overall survival adjusted for age and the interval between diagnosis and BMT was 90% (95% CI 85 – 93) and 89% (95% CI 85 – 92), P=0.67, with h-ATG and r-ATG, respectively (Figure 2A). There were no differences in the causes of death by ATG type (P=0.11). Graft failure and infection were the predominant cause of death after h-ATG and r-

Table 2A. Risk factors outcomes after HLA-matched sibling transplant.

Acute graft-versus-host disease Type of ATG r-ATG h-ATG Age at transplant, years <20 20-39 ≥40 Sex Female Male Chronic graft-versus-host disease Type of ATG r-ATG h-ATG Age at transplant, years <20 20-39 ≥40 Sex Female Male Overall survival Type of ATG r-ATG h-ATG Age at HCT, years <20 20-39 ≥40 Time from diagnosis to transplant <3 months ≥3 months

Number

Hazard Ratio (95% Confidence Interval)

267 277

1.00 3.34 (1.87-5.96)

<0.001

308 140 96

1.00 2.38 (1.33- 4.28) 2.83 (1.47-5.42)

0.004 0.002

261 283

1.00 1.82 (1.07-3.07)

0.026

259 270

1.00 2.55 (1.51-4.31)

<0.001

303 135 91

1.00 4.38 (2.53-7.58) 3.49 (1.77-6.85)

<0.001 <0.001

252 277

1.00 2.05 (1.23-3.39)

0.006

268 278

1.00 0.92 (0.53-1.59)

0.76

309 140 97

1.00 1.56 (0.72-3.40) 4.90 (2.49-9.66)

0.26 <0.001

350 196

1.00 2.43 (1.34-4.40)

0.004

ATG: antithymocyte globulin; h-ATG: horse derived ATG; r-ATG: rabbit derived ATG; HLA: human leukocyte antigen; HCT: hematopoietic cell transplantation.

haematologica | 2017; 102(7)

1295

N. Kekre et al.

ATG. There were 23 deaths among recipients of h-ATG; graft failure (n=8), infection (n=5), GvHD (n=4), pneumonitis/organ failure (n=4), and the cause of death was not reported for 2 patients. There were 30 deaths among recipients of r-ATG; graft failure (n=16), infection (n=8), GvHD (n=3), myelodysplastic syndrome (n=1), bleeding (n=1), and the cause of death was not reported for 2 patients. Among recipients of unrelated donor BMT, the risk of mortality was higher with h-ATG compared to r-ATG (Table 2B). Mortality risk was higher for patients with a comorbidity score of 3 or greater and 1 HLA locus mismatched unrelated donor BMT. The 3-year probability of overall survival adjusted for the comorbidity score and HLA match was 75% (95% CI 67 – 81) and 83% (95% CI 76 – 88) with h-ATG and r-ATG, respectively (Figure 2B). There were no differences in the causes of death by ATG type (P=0.26). There were 32 deaths among recipients of h-ATG; graft failure (n=6), infection (n=7), GvHD (n=13) and organ failure (n=6). There were 25 deaths among recipients of r-ATG; graft failure (n=7), infection (n=8), GvHD (n=4), pneumonitis/organ failure (n=5), and EBVassociated lymphoproliferative disease (n=1).

Discussion This analysis of a large, prospectively reported cohort of patients directly compared the utility of h-ATG with

r-ATG in transplant conditioning regimens for SAA. Data reported to the CIBMTR in recent years show that 75% of HLA-matched sibling donor and 77% of unrelated donor transplants include ATG in the transplant conditioning regimen. As we have previously demonstrated the superiority of bone marrow to peripheral blood as a graft source in SAA,13,14 this study was restricted to bone marrow grafts. Our findings support r-ATG (thymoglobulin) as the preferred type of ATG compared to h-ATG for HLA-matched sibling and unrelated donor BMT for SAA. With r-ATG there was less acute GvHD with both HLAmatched sibling and unrelated donor BMT and less chronic GvHD after HLA-matched sibling BMT. This is not surprising as r-ATG has a more potent immune suppressive effect and spares T-regulatory cells more effectively compared to h-ATG. Although survival was similar with hATG and r-ATG in HLA-matched sibling BMT, there was a survival advantage with r-ATG in the setting of unrelated donor BMT. The observed survival rate of 83% in the current analysis with r-ATG after unrelated donor BMT is consistent with that reported from the Blood and Marrow Transplant Clinical Trials Network (BMT CTN 0301) in which about 75% of patients received r-ATG.20 In that trial, h-ATG was only used for patients who could not tolerate r-ATG. Our findings are in keeping with smaller studies that have assessed differences in ATG formulations in allogeneic transplantation. Atta et al. also showed that r-ATG was associated with less GvHD, but more fungal infec-

Table 2B. Risk factors associated with Risk factors outcomes after unrelated donor transplant.

Acute graft-versus-host disease Type of ATG r-ATG h-ATG Donor recipient HLA match Matched 1 locus Mismatch Chronic graft-versus-host disease Type of ATG r-ATG h-ATG Age at transplant, years <20 ≥20 GvHD prophylaxis Tacrolimus or cyclosporine + methotrexate Tacrolimus or cyclosporine + mycophenolate Overall survival Type of ATG r-ATG h-ATG Comorbidity index 0-2 3+ Donor-recipient HLA match Matched 1 locus Mismatch

Number

Hazard Ratio (95% Confidence Interval)

160 124

1.00 2.20 (1.47-3.35)

<0.001

220 64

1.00 1.91 (1.23-2.97)

0.004

158 120

1.00 1.36 (0.90-2.04)

0.147

138 140

1.00 1.58 (1.05-2.39)

0.030

245 33

1.00 2.31 (1.39-3.84)

0.001

158 124

1.00 1.90 (1.12-3.24)

0.0183

214 68

1.00 4.85 (2.85-8.27)

<0.001

219 63

1.00 1.96 (1.09-3.54)

0.0250

ATG: antithymocyte globulin; h-ATG: horse derived ATG; r-ATG: rabbit derived ATG; HLA: human leukocyte antigen; GvHD: graft-versus-host disease.

1296

haematologica | 2017; 102(7)

Rabbit vs. horse ATG in BMT for aplastic anemia

Figure 2. Overall survival. The probability of overall survival after HLA-matched sibling transplant adjusted for age and time from diagnosis to transplant (A) and unrelated donor transplant adjusted for comorbidity index and donor-recipient HLA match (B). h-ATG: horse derived ATG; r-ATG: rabbit derived ATG.

tions and cytomegalovirus (CMV) reactivation.21 The overall mortality rate, however, was not different between rATG and h-ATG, with more infectious related deaths with r-ATG and more GvHD related deaths with h-ATG. As the current analyses used data reported to an observational registry, data on infectious complications were not collected consistently. However, infection was reported as the primary cause of death in 22% of patients who received h-ATG and 25% of patients who received r-ATG following HLA-matched sibling and unrelated donor BMT. One death was attributed to EBV-associated posttransplant lymphoproliferative disorder in a patient who received r-ATG and unrelated donor BMT. In addition to the effect of the type of ATG on acute and chronic GvHD and survival, the current analyses identified several modifiable factors that may improve outcomes both after HLA-matched sibling and unrelated donor BMT. In the setting of HLA-matched sibling BMT, delaying transplant beyond 3 months from diagnosis, regardless of patient age, results in higher mortality.22 Therefore, initiating a donor search at diagnosis is highly desirable and unlikely to delay the initiation of IST by more than 2-3 weeks in the event that a matched sibling is not available. Older patients are less likely to have a sustained response to IST, thus offering BMT early may mitigate some of the mortality risks associated with a longer waiting period to BMT. In the setting of unrelated donor BMT, selecting a HLA-matched donor lowered the risk for acute GvHD and mortality.23 Consistent with reports after reduced intensity conditioning transplants for hematologic malignancies, the addition of mycophenolate rather than methotrexate to a calcineurin inhibitor for GvHD prophylaxis resulted in a two-fold increase in the risk of contracting chronic GvHD.24 A limitation of the current study is that the choice of treatment strategy, including whether to use h-ATG or rATG, and the dose and timing of ATG, was at the discrehaematologica | 2017; 102(7)

tion of the treating physician and/or transplantation center and therefore subject to bias. This may have also been influenced by the availability of different ATG formulations at various centers. Others have noted that the timing of ATG and dose have been noted to be important in outcomes of engraftment, infection rates and survival postBMT.25,26 These studies have shown that higher doses of ATG are associated with lower rates of GvHD but higher rates of infection, including post-transplant lymphoproliferative disorder. A limitation of the current analyses is the lack of information on timing and dose. In addition, there may likely be other unmeasured or unknown factors that may have affected GvHD rates and survival. Due to the prohibitive costs of conducting multi-site trials, we often rely on observational registry data, such as that used in the study herein, to address some of the issues that may or may not be associated with outcomes. Nevertheless, we performed carefully controlled comparisons of the effects of r-ATG and h-ATG considering known prognostic factors. In patients undergoing HLA-matched sibling and unrelated donor BMT, survival was excellent regardless of type of ATG, and these findings may stimulate trials that test ATG type and dose for SAA transplants. The higher rates of GvHD associated with h-ATG support using rATG as opposed to h-ATG in order to lower the burden of morbidity in bone marrow transplantation for SAA. Funding The Center for International Blood and Marrow Transplant Research is supported primarily by U24-CA76518 from the National Cancer Institute, the National Heart, Lung, and Blood Institute, and the National Institute of Allergy and Infectious Diseases and HHSH234200637015C from the Health Resources and Services Administration (HRSA/DHHS). The content is solely the responsibility of the authors and does not represent the official policy of the National Institutes of Health or the Health Resources and Services Administration. 1297

N. Kekre et al.

References 1. Young NS, Calado RT, Scheinberg P. Current concepts in the pathophysiology and treatment of aplastic anemia. Blood. 2006;108(8):2509-2519. 2. Marsh JC, Ball SE, Cavenagh J, et al. Guidelines for the diagnosis and management of aplastic anaemia. Br J Haematol. 2009;147(1):43-70. 3. Afable MG, 2nd, Shaik M, Sugimoto Y, et al. Efficacy of rabbit anti-thymocyte globulin in severe aplastic anemia. Haematologica. 2011;96(9):1269-1275. 4. Scheinberg P, Nunez O, Weinstein B, et al. Horse versus rabbit antithymocyte globulin in acquired aplastic anemia. N Engl J Med. 2011;365(5):430-438. 5. Marsh JC, Bacigalupo A, Schrezenmeier H, et al. Prospective study of rabbit antithymocyte globulin and cyclosporine for aplastic anemia from the EBMT Severe Aplastic Anaemia Working Party. Blood. 2012;119(23):5391-5396. 6. Shin SH, Yoon JH, Yahng SA, et al. The efficacy of rabbit antithymocyte globulin with cyclosporine in comparison to horse antithymocyte globulin as a first-line treatment in adult patients with severe aplastic anemia: a single-center retrospective study. Ann Hematol. 2013;92(6):817-824. 7. Scheinberg P, Nunez O, Young NS. Retreatment with rabbit anti-thymocyte globulin and ciclosporin for patients with relapsed or refractory severe aplastic anaemia. Br J Haematol. 2006;133(6):622627. 8. Di Bona E, Rodeghiero F, Bruno B, et al. Rabbit antithymocyte globulin (r-ATG) plus cyclosporine and granulocyte colony stimulating factor is an effective treatment for aplastic anaemia patients unresponsive to a first course of intensive immunosuppressive therapy. Gruppo Italiano Trapianto di Midollo Osseo (GITMO). Br J Haematol. 1999;107(2):330-334.

1298

9. Scheinberg P, Fischer SH, Li L, et al. Distinct EBV and CMV reactivation patterns following antibody-based immunosuppressive regimens in patients with severe aplastic anemia. Blood. 2007;109(8):3219-3224. 10. Vallejo C, Montesinos P, Polo M, et al. Rabbit antithymocyte globulin versus horse antithymocyte globulin for treatment of acquired aplastic anemia: a retrospective analysis. Ann Hematol. 2015;94(6):947-954. 11. Feng X, Kajigaya S, Solomou EE, et al. Rabbit ATG but not horse ATG promotes expansion of functional CD4+CD25highFOXP3+ regulatory T cells in vitro. Blood. 2008;111(7):3675-3683. 12. Lopez M, Clarkson MR, Albin M, Sayegh MH, Najafian N. A novel mechanism of action for anti-thymocyte globulin: induction of CD4+CD25+Foxp3+ regulatory T cells. J Am Soc Nephrol. 2006;17(10):28442853. 13. Eapen M, Le Rademacher J, Antin JH, et al. Effect of stem cell source on outcomes after unrelated donor transplantation in severe aplastic anemia. Blood. 2011;118(9):26182621. 14. 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. 15. Flowers ME, Kansu E, Sullivan KM. Pathophysiology and treatment of graftversus-host disease. Hematol Oncol Clin North Am. 1999;13(5):1091-1112, viii-ix. 16. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus conference on acute GVHD grading. Bone Marrow Transplant. 1995;15(6):825-828. 17. Klein J, Moeschberger M. Survival analysis: statistical methods for censored and truncated data. Springer-Verlag, New York, NY. 2003. 18. Gooley TA, Leisenring W, Crowley J, Storer BE. Estimation of failure probabilities in the

19. 20.

21.

22.

23.

24.

25.

26.

presence of competing risks: new representations of old estimators. Stat Med. 1999; 18(6):695-706. Cox DR. Regression models and life-tables. J R Stat Soc Series B Stat Methodol. 1972;34(2):187-220. Anderlini P, Wu J, Gersten I, et al. Cyclophosphamide conditioning in patients with severe aplastic anaemia given unrelated marrow transplantation: a phase 1-2 dose de-escalation study. Lancet Haematol. 2015;2(9):e367-375. Atta EH, de Sousa AM, Schirmer MR, Bouzas LF, Nucci M, Abdelhay E. Different outcomes between cyclophosphamide plus horse or rabbit antithymocyte globulin for HLA-identical sibling bone marrow transplant in severe aplastic anemia. Biol Blood Marrow Transplant. 2012;18(12):18761882. Gupta V, Eapen M, Brazauskas R, et al. Impact of age on outcomes after bone marrow transplantation for acquired aplastic anemia using HLA-matched sibling donors. Haematologica. 2010;95(12):2119-2125. Pidala J, Lee SJ, Ahn KW, et al. Nonpermissive HLA-DPB1 mismatch increases mortality after myeloablative unrelated allogeneic hematopoietic cell transplantation. Blood. 2014;124(16):25962606. Eapen M, Logan BR, Horowitz MM, et al. Bone marrow or peripheral blood for reduced-intensity conditioning unrelated donor transplantation. J Clin Oncol. 2015;33(4):364-369. Remberger M, Svahn BM, Mattsson J, Ringden O. Dose study of thymoglobulin during conditioning for unrelated donor allogeneic stem-cell transplantation. Transplantation. 2004;78(1):122-127. Bacigalupo A, Lamparelli T, Gualandi F, et al. Prophylactic antithymocyte globulin reduces the risk of chronic graft-versus-host disease in alternative-donor bone marrow transplants. Biol Blood Marrow Transplant. 2002;8(12):656-661.

haematologica | 2017; 102(7)