Haematologica, Volume 101, issue 6 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)

Editorial Board Omar I. Abdel-Wahab (New York); Jeremy Abramson (Boston); Paolo Arosio (Brescia); Raphael Bejar (San Diego); Erik Berntorp (Malmö); Dominique Bonnet (London); Jean-Pierre Bourquin (Zurich); Suzanne Cannegieter (Leiden); Francisco Cervantes (Barcelona); Nicholas Chiorazzi (Manhasset); Oliver Cornely (Köln); Michel Delforge (Leuven); Ruud Delwel (Rotterdam); Meletios A. Dimopoulos (Athens); Inderjeet Dokal (London); Hervé Dombret (Paris); Peter Dreger (Hamburg); Martin Dreyling (München); Kieron Dunleavy (Bethesda); Dimitar Efremov (Rome); Sabine Eichinger (Vienna); Jean Feuillard (Limoges); Carlo Gambacorti-Passerini (Monza); Guillermo Garcia Manero (Houston); Christian Geisler (Copenhagen); Piero Giordano (Leiden); Christian Gisselbrecht (Paris); Andreas Greinacher (Greifswals); Hildegard Greinix (Vienna); Paolo Gresele (Perugia); Thomas M. Habermann (Rochester); Claudia Haferlach (München); Oliver Hantschel (Lausanne); Christine Harrison (Southampton); Brian Huntly (Cambridge); Ulrich Jaeger (Vienna); Elaine Jaffe (Bethesda); Arnon Kater (Amsterdam); Gregory Kato (Pittsburg); Christoph Klein (Munich); Steven Knapper (Cardiff); Seiji Kojima (Nagoya); John Koreth (Boston); Robert Kralovics (Vienna); Ralf Küppers (Essen); Ola Landgren (New York); Peter Lenting (Le Kremlin-Bicetre); Per Ljungman (Stockholm); Francesco Lo Coco (Rome); Henk M. Lokhorst (Utrecht); John Mascarenhas (New York); Maria-Victoria Mateos (Salamanca); Simon Mendez-Ferrer (Madrid); Giampaolo Merlini (Pavia); Anna Rita Migliaccio (New York); Mohamad Mohty (Nantes); Martina Muckenthaler (Heidelberg); Ann Mullally (Boston); Stephen Mulligan (Sydney); German Ott (Stuttgart); Jakob Passweg (Basel); Melanie Percy (Ireland); Rob Pieters (Rotterdam); Stefano Pileri (Milan); Miguel Piris (Madrid); Andreas Reiter (Mannheim); Jose-Maria Ribera (Barcelona); Stefano Rivella (New York); Francesco Rodeghiero (Vicenza); Richard Rosenquist (Uppsala); Simon Rule (Plymouth); Claudia Scholl (Heidelberg); Martin Schrappe (Kiel); Radek C. Skoda (Basel); Gérard Socié (Paris); Kostas Stamatopoulos (Thessaloniki); David P. Steensma (Rochester); Martin H. Steinberg (Boston); Ali Taher (Beirut); Evangelos Terpos (Athens); Takanori Teshima (Sapporo); Pieter Van Vlierberghe (Gent); Alessandro M. Vannucchi (Firenze); George Vassiliou (Cambridge); Edo Vellenga (Groningen); Umberto Vitolo (Torino); Guenter Weiss (Innsbruck).

Editorial Office Simona Giri (Production & Marketing Manager), Lorella Ripari (Peer Review Manager), Paola Cariati (Senior Graphic Designer), Igor Ebuli Poletti (Senior Graphic Designer), Marta Fossati (Peer Review), Diana Serena Ravera (Peer Review)

Affiliated Scientific Societies SIE (Italian Society of Hematology, www.siematologia.it) SIES (Italian Society of Experimental Hematology, www.siesonline.it)

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

Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies). Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the International edition for the year 2016 are as following: Print edition

Institutional Euro 500

Personal Euro 150

Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature. Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Tipografia PI-ME, via Vigentina 136, Pavia, Italy. Printed in May 2016.

haematologica calendar of events

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

21st Congress of the European Hematology Association European Hematology Association June 9-12, 2016 Copenhagen, Denmark

Anemia Diagnosis and Treatment in the Omics Era Chair: A Iolascon February 2-4, 2017 Barcelona, Spain

Hematology Tutorial on managing complications in patients with hematologic malignancies in the era of new drugs EHA-ROHS-RSH Chairs: E Parovichnikova, I Poddubnaya, R Foà July 1-3, 2016 Moscow, Russian Federation

EHA Hematology Tutorial on Lymphoid malignancies , Multiple myeloma and Bone Marrow Failure February 23-24, 2017 Colombo, Sri Lanka

Summer School of Personalised Medicine for Health Care Professionals European Alliance for Personalised Medicine (EAPM) July 4-7, 2016 Cascais, Portugal

EHA-SAH Hematology Tutorial on Thrombosis, Hemostasis & Myeloid Malignancies Chairs: R Foà, J Korin, G Kusminsky August 27-28, 2016 Buenos Aires, Argentina

EHA Scientific Conference on Bleeding Disorders Scientific Program Committee: C Balduini (Chair), A Falanga (Chair), F Rodeghiero, I Pabinger, M Makris September 14-17, 2016 Barcelona, Spain

2nd International Conference on New Concepts in B-Cell Malignancies European School of Haematology (ESH) Chairs: M Hallek, L Staudt, S Stilgenbauer, A Thomas-Tikhonenko September 9-11, 2016 Estoril, Portugal

12th Edition of the Educational Course of the EBMT Lymphoma Working Party on Treatment of Malignant Lymphoma: State-of-the Art and the Role of Stem Cell Transplantation European Group for Blood and Marrow Transplantation (EBMT) Chairs: S Montoto, P Dreger, A Sureda, E Vandenberghe September 21-23, 2016 Dublin, Ireland

10th Hodgkin Symposium University hospital of Cologne Chairs: A Engert, B von Treskow, B Böll October 22-25, 2016 Cologne, Germany

EHA Hematology Tutorial on EHA Hematology Tutorial on Lymphoid Malignancies March 17-18, 2017 Warsaw, Poland

Advances in Biology and Treatment of B Cell Malignancies, with a Focus on Rare Lymphoma Subtypes Chairs: MJ Kersten and M Dreyling March 10-12, 2017 Barcelona, Spain

Aging and Hematology Chair: D Bron May 4-6, 2017 Barcelona, Spain

22nd Congress of the European Hematology Association European Hematology Association June 22 - 25, 2017 Madrid, Spain

Challenges in the Diagnosis and Management of Myeloproliferative Neoplasms Chairs: JJ Kiladjian and C Harrison October 12-14, 2017 Location: TBC

Shaping the Future of Mesenchymal Stromal Cells Therapy Chair: W Fibbe November 23-24, 2017 Location: TBC

Calendar of Events updated on May 2, 2016

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

Table of Contents Volume 101, Issue 6: June 2016 Cover Figure Chromosome 6 with the location of the HLA genes indicated. (Image created by www.somersault1824.com)

Editorials 653

The optimal dosing of gemtuzumab ozagamicin: where to go from here? Theo de Witte and Sergio Amadori

654

HOXA-activated early T-cell progenitor acute lymphoblastic leukemia: predictor of poor outcome? Jules PP. Meijerink, et al.

657

Innovation in hematology. Perspectives: CML 2016 Rüdiger Hehlmann

Review Articles 660

Bone marrow fibrosis in myelofibrosis: pathogenesis, prognosis and targeted strategies Abdallah Abou Zahr, et al.

672

Wilms tumor 1 mutations in the pathogenesis of acute myeloid leukemia Raajit Rampal and Maria E. Figueroa

680

How to select the best available related or unrelated donor of hematopoietic stem cells? Jean-Marie Tiercy

Articles Red Cell Biology & its Disorders

688

The LSD1 inhibitor RN-1 recapitulates the fetal pattern of hemoglobin synthesis in baboons (P. anubis) Angela Rivers, et al.

Platelet Biology & Its Disorders

698

A distinct plasmablast and naïve B-cell phenotype in primary immune thrombocytopenia Shaun M. Flint, et al.

Myelodysplastic Syndromes

707

Loss of B cells and their precursors is the most constant feature of GATA-2 deficiency in childhood myelodysplastic syndrome Michaela Nováková, et al.

Chronic Myeloid Leukemia

717

Imatinib withdrawal syndrome and longer duration of imatinib have a close association with a lower molecular relapse after treatment discontinuation: the KID study Sung-Eun Lee, et al.

Haematologica 2016; vol. 101 no. 6 - June 2016 http://www.haematologica.org/

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

724

Defining the dose of gemtuzumab ozogamicin in combination with induction chemotherapy in acute myeloid leukemia: a comparison of 3 mg/m2 with 6 mg/m2 in the NCRI AML17 Trial Alan Burnett, et al.

Acute Lymphoblastic Leukemia

732

An early thymic precursor phenotype predicts outcome exclusively in HOXA-overexpressing adult T-cell acute lymphoblastic leukemia: a Group for Research in Adult Acute Lymphoblastic Leukemia study Jonathan Bond, et al.

741

Risk assessment of relapse by lineage-specific monitoring of chimerism in children undergoing allogeneic stem cell transplantation for acute lymphoblastic leukemia Sandra Preuner, et al.

Complications in Hematology

747

Late thyroid complications in survivors of childhood acute leukemia. An L.E.A. study Claire Oudin, et al.

Non-Hodgkin Lymphoma

757

Frequent CTLA4-CD28 gene fusion in diverse types of T-cell lymphoma Hae Yong Yoo, et al.

Stem Cell Transplantation

764

Improved graft-versus-host disease-free, relapse-free survival associated with bone marrow as the stem cell source in adults Rohtesh S. Mehta, et al.

773

Quality of Life

781

Prospective international validation of the Quality of Life in Myelodysplasia Scale (QUALMS) Gregory A. Abel, et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/101/6.toc

e216

Efficient CRISPR-Cas9 mediated gene disruption in primary erythroid progenitor cells Hojun Li, et al. http://www.haematologica.org/content/101/6/e216

e220

Bio-engineered and native red blood cells from cord blood exhibit the same metabolomic profile Dhouha Darghouth, et al. http://www.haematologica.org/content/101/6/e220

e224

The efficacy of current prognostic models in predicting outcome of patients with myelodysplastic syndromes at the time of hypomethylating agent failure Aziz Nazha, et al. http://www.haematologica.org/content/101/6/e224

e228

Two novel germline DDX41 mutations in a family with inherited myelodysplasia/acute myeloid leukemia Ruijuan Li, et al. http://www.haematologica.org/content/101/6/e228

Haematologica 2016; vol. 101 no. 6 - June 2016 http://www.haematologica.org/

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

e232

Retroviral insertional mutagenesis identifies the del(5q) genes, CXXC5, TIFAB and ETF1, as well as the Wnt pathway, as potential targets in del(5q) myeloid neoplasms Angela Stoddart, et al. http://www.haematologica.org/content/101/6/e232

e237

Genetic analysis of five children with essential thrombocytosis identified mutations in cancer-associated genes with roles in transcriptional regulation Nicole Kucine, et al. http://www.haematologica.org/content/101/6/e237

e240

LIN28B is over-expressed in specific subtypes of pediatric leukemia and regulates lncRNA H19 Hetty H. Helsmoortel, et al. http://www.haematologica.org/content/101/6/e240

e245

Mutations of TP53 gene in adult acute lymphoblastic leukemia at diagnosis do not affect the achievement of hematologic response but correlate with early relapse and very poor survival Silvia Salmoiraghi, et al. http://www.haematologica.org/content/101/6/e245

e249

Kr端ppel-like factor 4 (KLF4) inactivation in chronic lymphocytic leukemia correlates with promoter DNA-methylation and can be reversed by inhibition of NOTCH signaling Katharina Filarsky, et al. http://www.haematologica.org/content/101/6/e249

e254

Autoimmune cytopenias in patients with chronic lymphocytic leukemia treated with ibrutinib Candida Vitale, et al. http://www.haematologica.org/content/101/6/e254

e259

Immunohistochemical detection of inhibitor of DNA binding 3 mutational variants in mature aggressive B-cell lymphoma Monika Szczepanowski, et al. http://www.haematologica.org/content/101/6/e259

e262

Allogeneic hematopoietic stem cell transplantation after reduced intensity conditioning regimen for elderly patients (60 years and older) with hematologic malignancies using unrelated donors: a retrospective study from the French society for stem cell transplantation (SFGMTC) Jean El Cheikh, et al. http://www.haematologica.org/content/101/6/e262

Comments Comments are available online only at www.haematologica.org/content/101/6.toc

e266

A milder clinical course for severe hemophilia B: a true or biased effect? Jasper H.A. van Miert, et al. http://www.haematologica.org/content/101/6/e266

e267

Efficacy and safety of second-line ponatinib after failure of a single previous tyrosine kinase inhibitor for chronic myeloid leukemia patients in chronic phase Massimo Breccia, et al. http://www.haematologica.org/content/101/6/e267

e269

Comment on: Frequent CTLA4-CD28 gene fusion in diverse types of T-cell lymphoma, by Yoo et al. Qiang Gong, et al. http://www.haematologica.org/content/101/6/e269

e271

Author reply to Comment on: Frequent CTLA4-CD28 gene fusion in diverse types of T-cell lymphoma, by Yoo et al. Yoo HY, et al. http://www.haematologica.org/content/101/6/e271

Haematologica 2016; vol. 101 no. 6 - June 2016 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. 2014 JCR impact factor = 5.814

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 The optimal dosing of gemtuzumab ozagamicin: where to go from here? Theo de Witte1 and Sergio Amadori2 1

Institute for Molecular Life Sciences, Radboud University Medical Centre, Nijmegen, the Netherlands; and 2Tor Vergata University Hospital, Rome, Italy E-mail: theo.dewitte@radboudumc.nl doi:10.3324/haematol.2016.145763

ncomplete eradication of the malignant clones in high-risk de novo acute myeloid leukemia (AML), myelodysplastic syndrome (MDS) and secondary AML (sAML) is the main cause of treatment failure, demonstrated by a relatively low complete remission (CR) rate and a high early relapse rate of more than 50%, unless treated by allogeneic hematopoietic stem cell transplantation (AHCT). High-risk features include age 60 years or older, co-morbidities, preceding MDS, and adverse risk (cyto)genetic risk factors.1 Therefore, new studies focus on new and better remission-induction and consolidation regimens containing innovative agents. Gemtuzumab ozogamicin (GO) consists of a humanized anti-CD33 monoclonal antibody linked to calicheamicin, a potent antitumor antibiotic.2 GO binds to CD33, an antigen expressed on the surface of more than 90% of AML blast cells. Binding of GO is followed by internalization and toxin release intracellularly leading to DNA damage and cell death.3 In studies of older patients with AML in first relapse, tolerable toxicity and a response rate of 30% was reported following two infusions of GO 9 mg/m2, although full platelet recovery did not occur in roughly half of responders.4 These results led to regulatory approval of the drug in the United States for use in older patients in first relapse for whom standard therapy was unsuitable, setting the stage for its evaluation in patients with newly diagnosed high-risk AML/MDS. However, GO was voluntarily withdrawn from the market in 2010 on the basis of preliminary results from a phase III Southwest Oncology Group (SWOG) randomized study in 673 young adults with untreated AML.5 This study compared the addition of single infusion of 6 mg/m2 on day 4 of the first remission-induction course (daunorubicin (45 mg/m2 on days 1, 2, 3) and cytarabine (100 mg/m2 per day by continuous infusion on days 1-7) versus standard induction therapy with daunorubicin (60 mg/m2 on days 1-3) and cytarabine alone by continuous infusion on days 1 through 7 (DA). The CR rate was 69% for DA+GO and 70% for DA. The overall efficacy, as measured by the relapse-free survival and the overall survival (OS), was similar in both groups. However, the induction mortality was increased in the DA+GO group, at 5% versus 1% in the DA group. Since then, several randomized studies combining GO with intensive chemotherapy in patients with newly diagnosed AML have been reported in the literature providing new evidence on the clinical efficacy and safety of the immunoconjugate. A recent meta-analysis of 5 prospective studies, including the final data from the SWOG5 (total n=3325 patients)6 showed that addition of GO did not increase the proportion of patients entering CR with an odds ratio (OR) of 0.91 and a 95% confidence interval (CI) of 0.771.07 (P=0.3). However, the addition of GO significantly improved survival (OR 0.90; 95%CI: 0.82-0.98, P=0.03), although the 5-year OS difference was only around 3%: 35.5% (GO arm) versus 32.2% (control group). Unfortunately, patients with adverse cytogenetic characteristics did not benefit from the addition of GO in contrast to patients with haematologica | 2016; 101(6)

favorable or intermediate cytogenetic features who did benefit.6 The great majority of patients in this meta-analysis were patients with de novo AML. Only two studies7,8 included patients with sAML and only one study included HR-MDS patients.8 Increased toxicity and an absence of clinical benefit were observed in the EORTC/GIMEMA study, in which single-agent GO administration preceded induction chemotherapy.9 As reported in this issue of the Journal, Burnett and coworkers evaluated two doses of GO (6 vs. 3 mg/m2) in a large prospective, randomized trial: the National Cancer Research Institute (NCRI) AML17 trial.10 GO was usually administered as a single infusion on day 1 of the remission-induction course. They assessed the toxicity profile and antitumor activity of GO in combination with a chemotherapy remission-induction regimen in 673 adults (85%) with untreated de novo AML, 42 patients (5%) with high-risk MDS (HR-MDS) (defined as MDS with BM blasts higher than 10%), and 73 patients (9%) with sAML. There was no difference in overall response rate [defined as complete remission (CR) or CR with incomplete hematopoietic recovery (CRi)] between the two evaluated dose levels. All patients received various schedules of chemotherapy without GO after the first remission-induction course, depending on the risk status after the first remission-induction course and the presence of FLT-3 abnormalities. The overall survival and relapse risk did not differ despite a higher non-relapse (early) mortality and veno-occlusive disease (VOD) in the higher GO group (6 mg/m2). In addition, grade 3-4 serious adverse events (SAEs) were significantly higher in the higher dose GO group. Subgroup analysis did not show any difference in outcome in any subgroup. This large randomized study did not show any significant benefit of using GO at the 6 mg/m2 dose, although there was a possible trend for benefit in the adverse risk patients, who have not been shown to benefit from addition of GO in other trials. The 6 mg/m2 dose did have a detrimental effect with respect to liver toxicity and platelet count recovery; therefore, the outcome of this study suggests that, where a single dose schedule is used, the 3 mg/m2 dose might be preferred. The question remains as to whether fractionated dosing using the lower dose of 3 g/m2 results in better outcome or whether addition of GO to the consolidation course(s) may increase the benefit without additional toxicity, and the authors present a comprehensive discussion of this. It is possible that a more fractionated schedule with a higher total dose of GO might be a more effective strategy, which may take advantage of the CD33-re-expression that occurs after initial exposure to GO.11 The French ALFA group utilized a GO schedule of 3 mg/m2/day on days 1, 4 and 7 during induction chemotherapy, followed by a single dose in each of two post induction courses, in patients aged 50-70 years with untreated de novo AML. Complete response was 81% and event-free survival was 40.8% compared to 17.1% in 653

Editorials

the control group (P=0.003).12 It is plausible that at least some of the benefit was achieved by dosing in consolidation, but also the fractionation of the total GO dose during induction may have contributed to the most prominent effect in this French ALFA group study. This benefit was also apparent in patients with unfavorable cytogenetic characteristics, but the impact of complex karyotype has not been analyzed separately.12 In addition, early mortality seems to be reduced when a dose of 3 g/m2 is used either as a single dose or in a fractionated schedule.6 There was some hematologic toxicity, particularly to platelets. So, while a 3 mg/m2 dose appears adequate, it is still not certain what is the optimal schedule. However, the MRC AML15 trial did not show any additional benefit of adding GO to consolidation irrespective of whether it had been given with the first induction course;7 therefore the urgent issue to be resolved is whether a single dose or a fractionated schedule is to become the standard approach. In an attempt to do this, the NCRI have initiated a direct comparison of a 3 mg/m2 dose on day 1 versus days 1 and 4 in their ongoing trials. In conclusion, data from this study and from studies published after the withdrawal of GO from the market support the need for a re-appraisal of the regulatory approval of GO by the responsible authorities, at least for certain subtypes of AML.

10.

References 1. Dohner H, Weisdorf DJ, Bloomfield CD. Acute Myeloid Leukemia. New Engl J Med. 2015;373(12):1136-1152. 2. Hinman LM, Hamann PR, Wallace R, Menendez AT, Durr FE, Upeslacis J. Preparation and characterization of monoclonal antibody

11.

conjugates of the calicheamicins: a novel and potent family of antitumor antibiotics. Cancer Res. 1993;53(14):3336-3342. Linenberger ML. CD33-directed therapy with gemtuzumab ozogamicin in acute myeloid leukemia: progress in understanding cytotoxicity and potential mechanisms of drug resistance. Leukemia. 2005;19(2): 176-182. Larson RA, Sievers EL, Stadtmauer EA, et al. Final report of the efficacy and safety of gemtuzumab ozogamicin (Mylotarg) in patients with CD33-positive acute myeloid leukemia in first recurrence. Cancer. 2005;104(7):1442-152. Petersdorf S, Kopecky K, Stuart RK, et al. Preliminary results of Southwest Oncology Group study S0106: an international intergroup phase 3 randomized trial comparing the addition of gemtuzumab ozogamicin to standard induction therapy versus standard induction therapy followed by a second randomization to post-consolidation gemtuzumab ozogamicin versus no additional therapy for previously untreated acute myeloid leukemia. Blood. 2009;114:790. Hills RK, Castaigne S, Appelbaum FR, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy in adult patients with acute myeloid leukaemia: a meta-analysis of individual patient data from randomised controlled trials. Lancet Oncol. 2014;15(9):986-996. Burnett AK, Hills RK, Milligan D, et al. Identification of patients with acute myeloblastic leukemia who benefit from the addition of gemtuzumab ozogamicin: results of the MRC AML15 trial. J Clin Oncol. 2011;29(4):369-377. Burnett AK, Russell NH, Hills RK, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy improves survival in older patients with acute myeloid leukemia. J Clin Oncol. 2012;30(32):39243931. Amadori S, Suciu S, Stasi R, et al. Sequential combination of gemtuzumab ozogamicin and standard chemotherapy in older patients with newly diagnosed acute myeloid leukemia: results of a randomized phase III trial by the EORTC and GIMEMA consortium (AML-17). J Clin Oncol. 2013;31(35):4424-4430. van Der Velden VH, te Marvelde JG, Hoogeveen PG, et al. Targeting of the CD33-calicheamicin immunoconjugate Mylotarg (CMA-676) in acute myeloid leukemia: in vivo and in vitro saturation and internalization by leukemic and normal myeloid cells. Blood. 2001;97(10):31973204. Castaigne S, Pautas C, Terre C, et al. Effect of gemtuzumab ozogamicin on survival of adult patients with de-novo acute myeloid leukaemia (ALFA-0701): a randomised, open-label, phase 3 study. Lancet. 2012;379(9825):1508-1516.

HOXA-activated early T-cell progenitor acute lymphoblastic leukemia: predictor of poor outcome? Jules PP. Meijerink,1* Kirsten Canté-Barrett,1 Eric Vroegindeweij1 and Rob Pieters2 1

Department of Pediatric Oncology/Hematology, Sophia Children's Hospital, Rotterdam and 2The Princess Máxima Center for Pediatric Oncology, Utrecht, the Netherlands. E-mail: j.meijerink@erasmusmc.nl doi:10.3324/haematol.2016.145391

he outcome for T-cell acute lymphoblastic leukemia (T-ALL) has strongly improved over the last decades using high-intensity treatment protocols approaching cure rates of 80% for pediatric patients and 60% for adult patients. Fifteen percent of pediatric ALL patients present with T-ALL, and they represent nearly half of the ALL patients who require the most intensive treatment. Intensive chemotherapy increases the risk for treatment related morbidity and mortality. For relapsed patients, the outcome is poor, as T-ALL cells in those patients are highly resistant to further treatment. Therefore, patient-tailored treatment and the introduction of high precision medicines remain important. Molecular cytogenetic characterization of T-ALL has greatly increased our understanding of the pathogenic events that drive this disease. In contrast to precursor B-ALL, this improved insight into T-ALL has not yet yielded prognostic factors that allow for the iden-

654

tification of patients at high-risk of relapse and who may be eligible to receive alternative treatment, including allogeneic stem cell transplantation. One cytogenetic entity in pediatric and adult T-ALL patients that has been suspected to cause poor outcome include patients bearing a CALM-AF10 (PICALMMLLT10) fusion as a consequence of a t(10;11)(p13.14;q1421) chromosomal translocation.1 A first systematic study comprising unselected pediatric and adult T-ALL patients treated on FRALLE-93, FRALLE 2000 or LALA-94 protocols identified the CALM-AF10 fusion in approximately 9% of patients. This fusion is associated with early and late T-cell developmental arrest in the γδ lineage. In this study, late CALM-AF10+ T-ALL patients responded well to therapy, but 2 out of 12 CALM-AF10+ patients with an immature phenotype did not respond to therapy, and another 8 patients with an immature phenotype haematologica | 2016; 101(6)

Editorials

relapsed.2 Therefore, CALM-AF10 may be associated with poor outcome in T-ALL. The 3 CALM-AF10+ patients treated in the Dutch Childhood Oncology Group (DCOG) ALL-9 protocol demonstrated early relapses during therapy.3 Another study comprising 187 children treated on the AIEOP-BFM ALL 2000 or AIEOP R-2006 protocols identified CALM-AF10 fusions in 14 children, 8 of whom presented with high-risk features, including high white blood cell counts and prednisone poor responses. However, event-free survival and cumulative incidence of relapse were comparable for CALM-AF10+ and CALM-AF10patients.4 In contrast to the earlier study in which immature CALM-AF10+ patients were associated with poor outcome,2 only 1 out of the 14 CALM-AF10+ patients in these protocol studies had an immature T-cell immunophenotype.4 In 2005, two independent studies led to the discovery of a chromosomal inversion on chromosome 7 (inv(7)(p15;q34)) in T-ALL patients.5,6 This fusion leads to ectopic activation of HOXA genes (HOXA5-10 genes in particular) by a cis-acting mechanism due to the close proximity of the TCRB enhancer region.5,6 A high activation of HOXA genes was observed previously in MLLrearranged T- and B-cell neoplasms,7 but gene expression profiling studies revealed that HOXA-deregulation was a more common feature among T-ALL patients bearing inv(7), MLL-rearrangements, CALM-AF10 or SET-NUP214 gene fusions.5,8,9 The activation of HOXA genes therefore seems to play an important role in the cellular transformation of thymocytes. Various other HOXA-activating events have nowadays been identified in T-ALL including a TCRD-HOXA translocation10 and novel MLLT10 gene fusions. MLLT10 was identified fused to XPO1/CRM1, which encodes for a nuclear export protein,11 to NAP1L1 that encodes for a nucleosome assembly protein,12 and to HNRNPH1 and DDX3X genes that are both involved in RNA processing.13 In the latter study by Brandimarte et al.,13 all MLLT10-translocated HOXA+ cases clustered separately from other HOXA+ T-ALL cases including MLLrearranged, SET-NUP214 and inv(7) cases. These MLLT10translocated cases highly expressed the hematopoietic stem cell homeobox HHEX and MEF2C genes, two genes that are commonly expressed in immature T-ALL denoted as early T-cell precursor ALL (ETP-ALL).12,14 Regarding this issue, Bond and colleagues further delineate the ambiguous nature of HOXA-activated T-ALL patients with respect to T-cell developmental arrest and outcome.15 They extended their previous observations regarding CALM-AF10+ cases in a cohort of 190 adult TALL patients that were treated in LALA-94/GRAALL03-05 protocols. In that study, 42 T-ALL patients expressed an ETP-ALL immunophenotypic profile (CD5low, CD1a-, CD8and expression of CD34, CD13, CD33 and/or CD117) whereas 148 patients had arrested at later stages of maturation.16 ETP-ALL patients fared equally well compared to non-ETP-ALL patients. The ETP-ALL group comprised 9 CALM-AF10+ patients who were TCR negative due to absent or incomplete TCRδ or TCRγ recombinations (i.e. IM, IMδ or IMγ). In the non-ETP-ALL group, 2 out of 5 CALM-AF10+ patients had a similar TCR-negative genotype while the remaining 3 had the more mature sCD3+/TCR+ or cortical/pre-αβ TCR-genotype. ETP-ALL haematologica | 2016; 101(6)

patients who were positive for CALM-AF10+ (and were TCR-) had a significantly higher risk for adverse events and a trend toward reduced overall survival compared to CALM-AF10- ETP-ALL patients. CALM-AF10- ETP-ALL patients fared equally well compared to non-ETP-ALL patients regardless of their CALM-AF10 status.16 In the study, Bond and colleagues extended their observations by analyzing the prognostic impact of all HOXApositive cases in relation to ETP-ALL and outcome in a cohort of 209 adult T-ALL treated in the GRAALL2003/2005 protocol.15 Fifty-five T-ALL cases (26%) expressed HOXA9 at levels similar to those in CALMAF10+ patients as defined by RT-QPCR. Apart from 8 patients with CALM-AF10 translocations (1 patient also had an inv(7)), 10 patients had inv(7), 9 were positive for SET-NUP214, and 6 patients had MLL-rearrangements. Further screening for alternative HOXA-activating events revealed XPO1-MLLT10, DDX3X-MLLT10 or NAP1L1MLLT10 fusions in 1 patient each, whereas 1 additional patient had an unresolved MLLT10 rearrangement. Two patients were identified with NUP98-RAP1GDS1 fusions, whereas no HOXA-activating events were found in the remaining 16 cases. HOXA cluster activation in T-ALL patients by these oncogenic fusion products (denoted as trans-acting mechanism) is mostly linked with the presence of an immature TCR-genotype (IM0, IMδ or IMγ) and an ETP-ALL immunophenotype. In contrast, inv(7)positive (cis-HOXA-acting) patients almost exclusively present with a sCD3+/TCR+ or cortical/pre-αβ TCR-genotype. In line with previous observations,4 HOXA+ patients demonstrated increased resistance to corticosteroid treatment and chemotherapy and frequently remained MRD positive (>10-4) after induction therapy.15 Surprisingly, overall outcome (OS, EFS and DFS) for HOXA+ patients was identical to HOXA- T-ALL patients. Further discriminating patients based on ETP-ALL immunophenotype revealed that HOXA+/ETP-ALL patients had a significantly poor outcome as compared to HOXA-/ETP-ALL patients (OS: 31.2% vs. 74.2%; EFS: 25% vs. 60.8%; DFS: 28.6% vs. 64.7% and CIR: 53.7% vs. 29.2%, respectively). The outcome for HOXA-/ETP-ALL patients was as favorable as for non-ETP-ALL patients regardless of the presence of HOXA-activating rearrangements.15 The Children's Oncology Group (COG) has now reported similar findings for pediatric T-ALL patients.17 This study investigated 100 children with T-ALL who were treated in the COG AALL0434 protocol, including 17 patients for whom initial treatment failed. Evidence for MLL-rearrangements were found in 12 patients in addition to 6 CALM-AF10+ patients, 3 DDX3X-AF10+ patients (1 case had a complex CASK-DDX3X-AF10 translocation), 2 patients with NUP98-rearrangements and 3 inv(7)+ patients. MLL- but not AF10-rearrangements were strongly associated with induction failure and inferior EFS in uniand multivariate analyses. Expression of an ETP-ALL expression signature also predicted for inferior EFS, and trended towards enrichment of MLL-rearranged cases. MLL-rearrangements combined with an ETP-ALL expression profile most strongly associated with induction failure, refractory disease and relapse.17 Both studies therefore point to HOXA-activated ETP-ALL cases that are at a higher risk to fail on induction therapy or have inferior survival 655

Editorials

rates. Further studies are needed to investigate whether this can be attributed to specific HOXA-activating events: 13 HOXA-activated adult T-ALL patients that relapsed included 3 patients with SET-NUP214 fusions, 2 patients with MLLT10-rearrangements, 1 MLL-rearranged case and 1 patient with an inv(7),15 while in the pediatric study the MLL-AF6 or Del3â&#x20AC;&#x2122;MLL rearranged patients were at the highest risk to fail on therapy.17 Based on these important findings in both studies, routine screening for HOXA-activation events and ETP-ALL profiles in future T-ALL patients may help to identify patients at risk for induction failure or relapse. For this to happen, several issues need to be resolved: What would be the best detection method to identify HOXA-activated ETP-ALL patients? Gene expression profiling to identify HOXA-activated T-ALL patients failed to classify 3 patients carrying HOXA-activating events in the COG series.17 On the other hand, not all HOXA-activating genetic events have been resolved.15 Also, what is the best method to identify ETP-ALL patients? Will this rely on ETP-ALL immunophenotypic markers18 or on the expression of ETPALL signature genes? Several retrospective studies on historical T-ALL samples did not consistently identify an ETPALL-specific immunophenotype for cases that expressed an ETP-ALL gene signature.19,20 Finally, what alternative treatment should be given? HOXA-activated ETP-ALL patients may receive allogeneic stem cell transplantations if suitable donors are available, or receive precision medicine like the DOT1L inhibitor EPZ-5676 compound that is currently being tested in clinical trials.21 An additional important question that needs to be resolved in the future is which other genetic factors may cause HOXA-activated cases to arrest at the ETP-stage and define poor outcome, while other HOXA-activated T-ALL cases with seemingly identical chromosomal rearrangements arrest at late stages and have a better prognosis? The answer may reveal the true determinant that defines ETP-arrest and the high-risk of treatment failure for HOXA-activated ETP-ALLs. Acknowledgements This study was supported by the Children Cancer Free Foundation (Stichting Kinderen Kankervrij, KiKa) grants KiKa2008-29 and KiKa2013-116 (KC-B) and the Dutch Cancer Society KWF2010-4691 (EV).

References 1. t(10;11)(p13-14;q14-21): a new recurrent translocation in T-cell acute lymphoblastic leukemias. Groupe Francais de Cytogenetique Hematologique (GFCH). Genes Chromosomes Cancer. 1991;3(6):411415.

656

2. Asnafi V, Beldjord K, Boulanger E, et al. Analysis of TCR, pT alpha, and RAG-1 in T-acute lymphoblastic leukemias improves understanding of early human T-lymphoid lineage commitment. Blood. 2003;101(7): 2693-2703. 3. van Grotel M, Meijerink JP, van Wering ER, et al. Prognostic significance of molecular-cytogenetic abnormalities in pediatric T-ALL is not explained by immunophenotypic differences. Leukemia. 2008;22(1): 124-131. 4. Lo Nigro L, Mirabile E, Tumino M, et al. Detection of PICALMMLLT10 (CALM-AF10) and outcome in children with T-lineage acute lymphoblastic leukemia. Leukemia. 2013;27(12):2419-2421. 5. Soulier J, Clappier E, Cayuela JM, et al. HOXA genes are included in genetic and biologic networks defining human acute T-cell leukemia (T-ALL). Blood. 2005;106(1):274-286. 6. Speleman F, Cauwelier B, Dastugue N, et al. A new recurrent inversion, inv(7)(p15q34), leads to transcriptional activation of HOXA10 and HOXA11 in a subset of T-cell acute lymphoblastic leukemias. Leukemia. 2005;19(3):358-366. 7. Ferrando AA, Armstrong SA, Neuberg DS, et al. Gene expression signatures in MLL-rearranged T-lineage and B-precursor acute leukemias: dominance of HOX dysregulation. Blood. 2003;102(1):262-268. 8. Van Vlierberghe P, van Grotel M, Tchinda J, et al. The recurrent SETNUP214 fusion as a new HOXA activation mechanism in pediatric Tcell acute lymphoblastic leukemia. Blood. 2008;111(9):4668-4680. 9. Dik WA, Brahim W, Braun C, et al. CALM-AF10+ T-ALL expression profiles are characterized by overexpression of HOXA and BMI1 oncogenes. Leukemia. 2005;19(11):1948-1957. 10. Bergeron J, Clappier E, Cauwelier B, et al. HOXA cluster deregulation in T-ALL associated with both a TCRD-HOXA and a CALM-AF10 chromosomal translocation. Leukemia. 2006;20(6):1184-1187. 11. Bond J, Bergon A, Durand A, et al. Cryptic XPO1-MLLT10 translocation is associated with HOXA locus deregulation in T-ALL. Blood. 2014;124(19):3023-3025. 12. Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature. 2012;481(7380):157163. 13. Brandimarte L, Pierini V, Di Giacomo D, et al. New MLLT10 gene recombinations in pediatric T-acute lymphoblastic leukemia. Blood. 2013;121(25):5064-5067. 14. Homminga I, Pieters R, Langerak AW, et al. Integrated transcript and genome analyses reveal NKX2-1 and MEF2C as potential oncogenes in T cell acute lymphoblastic leukemia. Cancer Cell. 2011;19(4):484-497. 15. Bond J, Machand T, Touzart A, et al. An early thymic progenitor phentotype predicts outcome exclusively in HOXA-overexpressing adult Tcell acute lymphoblastic leukemia: a group for research in adult acute lymphoblastic leukemia study. Haematologica. 2016;this issue. 16. Ben Abdelali R, Asnafi V, Petit A, et al. The prognosis of CALM-AF10positive adult T-cell acute lymphoblastic leukemias depends on the stage of maturation arrest. Haematologica. 2013;98(11):1711-1717. 17. Matlawska-Wasowska K, Kang H, Devidas M, et al. MLL rearrangements impact outcome in HOXA-deregulated T-lineage acute lymphoblastic leukemia: A children's oncology group study. Leukemia. 2016. 18. Coustan-Smith E, Mullighan CG, Onciu M, et al. Early T-cell precursor leukaemia: a subtype of very high-risk acute lymphoblastic leukaemia. Lancet Oncol. 2009;10(2):147-156. 19. Zuurbier L, Gutierrez A, Mullighan CG, et al. Immature MEF2C-dysregulated T-ALL patients have an ETP-ALL gene signature and typically have non-rearranged T-cell receptors. Haematologica. 2014;99:94-102. 20. Gutierrez A, Dahlberg SE, Neuberg DS, et al. Absence of Biallelic TCR{gamma} Deletion Predicts Early Treatment Failure in Pediatric TCell Acute Lymphoblastic Leukemia. J Clin Oncol. 2010;28(24):38163823. 21. Chen CW, Armstrong SA. Targeting DOT1L and HOX gene expression in MLL-rearranged leukemia and beyond. Exp Hematol. 2015;43(8): 673-684.

haematologica | 2016; 101(6)

Editorials

Innovation in hematology. Perspectives: CML 2016 Rüdiger Hehlmann Heidelberg University, Mannheim, Germany E-mail: r.hehlmann@urz.uni-heidelberg.de doi:10.3324/haematol.2016.142877

ith modern drug treatment [imatinib, tyrosine kinase inhibitors (TKI)] survival with chronic myeloid leukemia (CML) has almost normalized. Results of clinical trials show that survival with CML has improved significantly over the last decades1 (Figure 1) and that life expectancy of patients with CML approaches that of the general population (Figure 2).2 Subgroups of patients, e.g. those with complete cytogenetic remission, may reach normal survival times.3 Nevertheless, survival with CML in population-based registries4,5 is still lower than normal, in a recent analysis of the Dutch CML-registry by 10%-15%.6 Five to 7% of patients progress to advanced phases and blast crisis.7,8 Another 5%-10% may receive suboptimal treatment such as hydroxyurea, particularly in the elderly, or as a consequence of poor adherence. Lack of adherence to prescribed medication is considered to be the main reason for suboptimal treatment and inferior outcome,9 and may result from reduced quality of life in the face of adverse drug effects of life-long treatment. The matter has become a key topic of current research, in particular with regards to attempts to discontinue treatment when durable deep molecular responses have been reached and personalization of treatment according to individual patients’ needs. Various trials try to define conditions which allow treatment discontinuation with the highest chance of success. Duration of TKI treatment, depth of molecular remission and duration of deep remission seem to play a role. Other conditions may be patients’ risk profile at diagnosis and line of therapy. The goal is to improve the proportion of patients who stay free of relapse [i.e. no loss of major molecular remission (MMR)].10 Treatment- and relapsefree survival ranges around 40% in the discontinuation

trial with the longest follow up (STIM-study, median observation 5.5 years).11 Most relapsing patients regain the same depth of response after resumption of pre-discontinuation treatment. Some progress in optimizing treatment discontinuation can be expected from a large ongoing European discontinuation study (Euro-SKI) (Saußele and Mahon, 2016, manuscript in preparation). So far, there is no indication that type or dose of TKI results in differences in discontinuation outcome. The other important approach to improving outcome of CML-treatment is individualization of treatment by considering patients’ variables at diagnosis or response levels at defined milestones to optimize drug dosage and select the right drug for the right patient. Patients’ variables may be the classical risk scores, or individual molecular markers such as transcript type or expression of genes, or gene groups, thought to be of prognostic relevance.12,13 The b2a2 transcript type has been consistently associated with lower response rates and longer times to response. An expression signature at diagnosis of 20 genes has recently been shown to correlate with outcome.14 Whether the early detection of low level resistance mutants with more sensitive detection methods, such as next generation sequencing provides an advantage for treatment choice and outcome, still remains a subject for discussion. Preexisting comorbidities have guided the selection of TKI since the availability of 2nd-generation (2G-) TKI to decrease toxicity and increase efficacy. Epidemiological studies and registries are used to define better patients’ characteristics at diagnosis. Another approach to treatment personalization is optimizing individual drug doses according to blood drug levels or patient tolerability. In contrast to some 2G-TKI, a

Figure 1. Survival with chronic myeloid leukemia in five consecutive randomized studies of the German CML Study Group since 1983; update 2016.

haematologica | 2016; 101(6)

657

Editorials

Figure 2. Relative and overall survival of 2290 chronic myeloid leukemia (CML) patients from the European Treatment and Outcome Study (EUTOS) for CML treated with imatinib in six clinical trials and prospectively enrolled between 2002 and 2006 from Dutch, French, German, Italian, Nordic and Spanish Study Groups2 (courtesy of Dr. M. Pfirrmann).

systematic optimization of imatinib dosage has never been done. A recent study15 has shown that molecular response such as measured by major molecular response (MMR) after 12 months can be improved in up to 80% of patients if the imatinib dose was increased in patients with suboptimal drug levels. Likewise in the German CML-Study, IV imatinib dosage in the 800 mg arm was tailored according to tolerability providing superior responses.8 This agrees with a recent systematic review of 5 randomized trials comparing imatinib 400 mg and 800 mg (or 600 mg), which finds a 30% higher MMR rate at 12 months with imatinib 800 mg similar to that with 2GTKI.16 In no instance do we have convincing evidence that any TKI provides an overall survival advantage over another, at the high survival rate we never might achieve. Patientsâ&#x20AC;&#x2122; factors such as risk profile and TKI resistance seem to overrule choice of TKI, TKI dose or TKI in various combinations with regard to survival. In addition, more patients die in the meantime from comorbidities than from CML.17 Differences in overall survival may thus become too small in relation to the limited power of the trials. Attempts are being made to improve survival further by TKI in combination with interferon (IFN), or IFN maintenance, or by better drugs that overcome resistance and achieve deeper responses faster. Hematopoietic cell transplantation tailored according to patientsâ&#x20AC;&#x2122; and transplantation risks may provide a good chance of cure with minimal transplantation-related mortality.18 A challenges is that a substantial minority of patients still progresses to blast crisis which is only poorly treatable - maybe some new markers, or cytogenetics in the course of disease, provide an earlier clue for progression than rising blasts. Cost, less a problem in Europe than elsewhere, is likely to improve after the imatinib patent expires in most countries in 2016. Most CML survivors are faced with life-long treatment and suffer from reduced quality of life. Attempts at treatment discontinuation result in 60% relapses, with uncertainty as to whether the remainder can be considered cured in the presence of residual disease in most cases. Suboptimal patient education probably 658

contributes to a lack of adherence and to suboptimal outcomes in a considerable portion of patients. Although progress with CML has fundamentally changed the outcome of CML, overall survival is still reduced, quality of life is hampered by potentially life-long treatment, and a minority of patients still die of CML. For most patients with CML, cure, or at least a normal life, has not yet been achieved.

Innovation in education and training to improve knowledge on CML The European Investigators on CML (EI-CML) and the European LeukemiaNet (ELN) have long standing programs on mutual information and training. Examples are the annual EI-CML meetings that have taken place across Europe since 1993, the annual ELN-Symposia in Mannheim (with regular meetings of the CML-working group since 2004), educational meetings on CML and other myeloid neoplasias (ELN-Frontiers Meetings) all over Europe 2006-2014, and educational meetings for young hematologists in Naples (2009-2014). The CML-meetings organized by the European School of Hematology (now the John Goldman conferences) attract young scientists to present their research data to the CML-community. In recent years, the ELN-breakfast meetings at the ASH meeting have been popular, and more weight is put on ELN-workpackage meetings in the context of the EHA-conferences. The sessions of the ELN-EHA scientific working group on CML are the latest addition to the European CML communityâ&#x20AC;&#x2122;s efforts to improve education on CML, primarily for physicians, but also for patients. Since 2006, the CML group of ELN publishes international management recommendations for CML,19 with a new edition planned for 2016.

Innovation in technology Disease monitoring and attempts at treatment discontinuation require specific and sensitive monitoring techniques. To this end, molecular analyses of BCR-ABL 1 have been standardized by an international co-operation within ELN and EUTOS for CML, and an international haematologica | 2016; 101(6)

Editorials

scale has been introduced for standardized BCR-ABL 1 monitoring.20 To achieve a higher sensitivity, allowing more reliable detection of deep responses at the MR4.5 and MR5 levels, digitalized PCR (dPCR) was established. This new modification of PCR-technology reproducibly achieves sensitivities of 10-5 (MR5) in routine analysis. Further topics currently being addressed are the initiating event(s) of CML pathogenesis, the role of CML-stem cells in the maintenance of residual disease, the relevance of genomic changes, gene expression and epigenetics for treatment and outcome, and the optimization of drug treatment with new drugs or combinations, e.g. with IFN. It is hoped that by systematically addressing these questions progress will be made in our understanding of CML, which will enable us to improve management and prognosis. The current goal of the ELN-EHA SWG for CML is the further improvement of survival, achievement of definite cures, and a normal quality of life for all patients with CML. This will be accomplished by continued research programs, clinical trials and educational programs. The ELN-EHA SWG for CML-session in Copenhagen will illustrate and promote this goal.

References 1. Hehlmann R. CML--Where do we stand in 2015? Ann Hematol. 2015;94 Suppl 2:S103-105. 2. Pfirrmann M, Baccarani M, Saussele S, et al. Prognosis of long-term survival considering disease-specific death in patients with chronic myeloid leukemia. Leukemia. 2016;30(1):48-56. 3. Gambacorti-Passerini C, Antolini L, Mahon FX, et al. Multicenter Independent Assessment of Outcomes in Chronic Myeloid Leukemia Patients Treated With Imatinib. J Natl Cancer Inst. 2011;103(7):553561. 4. Hรถglund M, Sandin F, Hellstrom K, et al. Tyrosine kinase inhibitor usage, treatment outcome, and prognostic scores in CML: report from the population-based Swedish CML registry. Blood. 2013;122(7):12841292. 5. 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. 6. Thielen N, Visser O, Ossenkoppele G, Janssen J. Chronic myeloid leukemia in the Netherlands: a population-based study on incidence, treatment, and survival in 3585 patients from 1989 to 2012. Eur J Haematol. 2015 Oct 31. [Epub ahead of print] 7. Deininger M, O'Brien SG, Guilhot F, et al. International Randomized

haematologica | 2016; 101(6)

10.

11. 12. 13.

14.

15.

16.

17. 18.

19. 20.

Study of Interferon Vs STI571 (IRIS) 8-Year Follow up: Sustained Survival and Low Risk for Progression or Events in Patients with Newly Diagnosed Chronic Myeloid Leukemia in Chronic Phase (CML-CP) Treated with Imatinib. Blood. 2009;114(22):1126. 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. Noens L, Hensen M, Kucmin-Bemelmans I, Lofgren C, Gilloteau I, Vrijens B. Measurement of adherence to BCR-ABL inhibitor therapy in chronic myeloid leukemia: current situation and future challenges. Haematologica. 2014;99(3):437-447. Rousselot P, Charbonnier A, Cony-Makhoul P, et al. Loss of Major Molecular Response As a Trigger for Restarting Tyrosine Kinase Inhibitor Therapy in Patients With Chronic-Phase Chronic Myelogenous Leukemia Who Have Stopped Imatinib After Durable Undetectable Disease. J Clin Oncol. 2014;32(5):424-430. Etienne G, Rea D, Guilhot J, et al. Long-Term Follow-up of the French 1 Stop Imatinib Study (STIM1) in Chronic Myeloid Leukemia Patients. ASH Annual Meeting Abstracts. 2015. (Abstract 345). Hanfstein B, Lauseker M, Hehlmann R, et al. Distinct characteristics of e13a2 versus e14a2 BCR-ABL1 driven chronic myeloid leukemia under first-line therapy with imatinib. Haematologica. 2014;99(9):1441-1447. Jain P, Kantarjian HM, Luthra R, et al. Impact of the Type of BCR-ABL Fusion Transcripts on Response and Survival in Patients with Chronic Phase CML Treated with Four Frontline TKI Modalities. Blood 2016;126(23):598. Kok CH, Leclercq TM, Watkins D, et al. A 20 Gene Expression Signature That Predicts Early Molecular Response Failure in Chronic Phase CML Patients Treated with Frontline Imatinib. Blood. 2015;126:(23):596. Rousselot P, Johnson-Ansah H, Huguet F, et al. Personalized Daily Doses of Imatinib By Therapeutic Drug Monitoring Increase the Rates of Molecular Responses in Patients with Chronic Myeloid Leukemia. Final Results of the Randomized OPTIM Imatinib Study. Blood. 2015;126(23):133. Hoffmann VS, Hasford J, Hehlmann R. Systematic Review and MetaAnalysis of Randomized Trials Comparing Imatinib 400 Mg/d Vs. Imatinib 800 Mg/d, and Imatinib 400 Mg/d Vs. Second Generation TKIs in Chronic Phase CML-Patients. Blood. 2015;126(23):2787. Saussele S, Krauss MP, Hehlmann R, et al. Impact of comorbidities on overall survival in patients with chronic myeloid leukemia: results of the randomized CML Study IV. Blood. 2015;126(1):42-49. Gratwohl A, Pfirrmann M, Zander A, et al. Long-term outcome of patients with newly diagnosed chronic myeloid leukemia: a randomized comparison of stem cell transplantation with drug treatment. Leukemia. 2016; 30:562-569. Baccarani M, Deininger MW, Rosti G, et al. European LeukemiaNet recommendations for the management of chronic myeloid leukemia: 2013. Blood. 2013;122(6):872-884. Cross NCP, White HE, Colomer D, et al. Laboratory recommendations for scoring deep molecular responses following treatment for chronic myeloid leukemia. Leukemia. 2015;29(5):999-1003.

659

REVIEW ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Bone marrow fibrosis in myelofibrosis: pathogenesis, prognosis and targeted strategies Abdallah Abou Zahr,1 Mohamed E. Salama,2 Nicole Carreau,3 Douglas Tremblay,3 Srdan Verstovsek,4 Ruben Mesa,5 Ronald Hoffman,6 and John Mascarenhas6

Division of Hematology Oncology, Mount Sinai Beth Israel, New York, NY; 2Associated Regional University Pathologists Laboratories, Department of Pathology, University of Utah, Salt Lake City, UT ; 3Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, NY; 4Department of Leukemia, MD Anderson Cancer Center, Houston, TX; 5Division of Hematology & Medical Oncology, Mayo Clinic Cancer Center, Scottsdale, AZ; and 6Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA 1

Haematologica 2016 Volume 101(6):660-671

ABSTRACT

Correspondence: john.mascarenhas@mssm.edu

Received: January 6, 2016. Accepted: February 8, 2016. Pre-published: no prepublication. doi:10.3324/haematol.2015.141283

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/660

ÂŠ2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.

660

one marrow fibrosis is a central pathological feature and World Health Organization major diagnostic criterion of myelofibrosis. Although bone marrow fibrosis is seen in a variety of malignant and non-malignant disease states, the deposition of reticulin and collagen fibrosis in the bone marrow of patients with myelofibrosis is believed to be mediated by the myelofibrosis hematopoietic stem/progenitor cell, contributing to an impaired microenvironment favoring malignant over normal hematopoiesis. Increased expression of inflammatory cytokines, lysyl oxidase, transforming growth factor-Î˛, impaired megakaryocyte function, and aberrant JAK-STAT signaling have all been implicated in the pathogenesis of bone marrow fibrosis. A number of studies indicate that bone marrow fibrosis is an adverse prognostic variable in myeloproliferative neoplasms. However, modern myelofibrosis prognostication systems utilized in riskadapted treatment approaches do not include bone marrow fibrosis as a prognostic variable. The specific effect on bone marrow fibrosis of JAK2 inhibition, and other rationally based therapies currently being evaluated in myelofibrosis, has yet to be fully elucidated. Hematopoietic stem cell transplantation remains the only curative therapeutic approach that reliably results in resolution of bone marrow fibrosis in patients with myelofibrosis. Here we review the pathogenesis, biological consequences, and prognostic impact of bone marrow fibrosis. We discuss the rationale of various anti-fibrogenic treatment strategies targeting the clonal hematopoietic stem/progenitor cell, aberrant signaling pathways, fibrogenic cytokines, and the tumor microenvironment.

Introduction Bone marrow fibrosis (BMF) is characterized by the increased deposition of reticulin fibers and in some cases collagen fibers. The currently accepted methods of evaluating and scoring BMF are primarily dependent on manual grading by the hematopathologist based on the density and type of fibrosis (Table 1).1,2 There are a number of hematologic and non-hematologic disorders that are associated with increased BMF (Table 2).3 Myelofibrosis (MF) refers to the Philadelphia chromosome (BCR-ABL1)-negative myeloproliferative neoplasm (MPN) originating at the level of the multipotent hematopoietic stem cell. MF can present as primary myelofibrosis (PMF), or arise from a pre-existing diagnosis of polycythemia vera or essential thrombocythemia. MF is characterized by variable degrees of cytopenias, a leukoerythroblastic blood picture, and extramedullary hematopoiesis resulting in progressive splenomegaly and debilitating disease-related constitutional symptoms, compromising quality of life.4,5 In addition to increased disease-related morhaematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

bidity, MF results in early death with the median survival of affected patients being approximately 6 years.6 Causes of early death include leukemic transformation, complications arising from progressive bone marrow failure, portal/pulmonary hypertension, infections, thrombosis and bleeding.7 Pathologically, MF is characterized by thickening and distortion of bony trabeculae, deposition of reticulin and collagen fibers, and megakaryocytic hyperplasia with atypical features.8 The exact pathogenesis of MF is not fully understood. However, better understanding of the role of increased JAK-STAT signaling [either through activating mutations (JAK2V617F, MPL515L/K) within the signaling pathway, or mutations involving CALR], the role of deregulated proinflammatory cytokine expression, and the impaired bone marrow microenvironment is transforming the treatment approach for MF. Here we review the pathogenesis of MF and the prognostic impact of BMF, and highlight the potential prospects for anti-fibrogenic strategies that can be utilized in the treatment of MF and other hematologic malignancies.

Pathogenesis Cytokines A major biological hallmark of MF is a significant elevation in circulating pro-inflammatory cytokines. The MF inflammatory cytokine signature is believed to be both a consequence of the malignant clone as well an integral modifier of the bone marrow microenvironment, thereby, promoting malignant hematopoiesis.9 Transforming growth factor beta (TGF-β) is a pleiotropic cytokine that potently stimulates fibroblasts to produce extracellular matrix.10-13 It also increases the expression of proteases that inhibit enzymes involved in the degradation of extracellular matrix. Experimental studies have demonstrated that TGF-β1 is important in the development of BMF in animal models.14 Chagraoui et al. compared the pathological changes in irradiated wild-type recipient mice repopulated with thrombopoietin-overexpressing hematopoietic stem cells (HSC) from homozygous TGF-β1-/- or wild-type littermates.15 Mice engrafted with wild-type and TGFβ1-/cells developed thrombocytosis, leukocytosis, and increased numbers of progenitor cells in the blood and spleen. However, BMF and reticulin deposition in the spleen occurred only in mice reconstituted with wild-type cells. Additionally, osteosclerosis was seen only in mice engrafted with wild-type cells.15 The Gata1low murine model of MF is characterized by normal or mild elevation in TGF-β1 levels. Elevated levels Table 1. European consensus on the grading of bone marrow fibrosis.1 0 1 2

Scattered linear reticulin with no intersection (cross-overs) corresponding to normal bone marrow Loose network of reticulin with many intersections, especially in perivascular areas Diffuse and dense increase in reticulin with extensive intersections, occasionally with only focal bundles of collagen and/or focal osteosclerosis Diffuse and dense increase in reticulin with extensive intersections with coarse bundles of collagen, often associated with significant osteosclerosis haematologica | 2016; 101(6)

of TGF-β1 have also been documented in patients with MF.16 In order to evaluate the pathological contribution of TGF-β in MF, Zingariello et al. utilized a Gata1low mouse model of MF.17 First, the investigators demonstrated that despite normal or mildly elevated plasma concentrations of total and bioactive TGF-β1 in PMF patients and Gata1low mice, the TGF-β1 content in megakaryocytes was increased 5- to 10-fold in both. In Gata-1low mice, this increase was accompanied by an increase in TGF-β1, hedgehog, and p53 signaling pathways in the spleen and bone marrow. The mammalian target of rapamycin (mTOR) signaling pathway was found to be increased in the spleen only. The biological consequences of altered gene expression were predicted using the David bioinformatics database and are in concordance with the GATA1low phenotype. Increased TGF-β signaling results in increased levels of osteoblast differentiation in bone marrow but not in the spleen, increased apoptosis, G1 arrest in bone marrow and spleen, reduced ubiquitinmediated proteolysis in the bone marrow only. Ubiquitin-mediated proteolysis in the bone marrow is important for erythroid maturation. The mTOR signaling pathway is important for erythropoiesis and its selective increase in the spleen may lead to erythroblast maturation and promote extramedullary hematopoiesis. Importantly, the investigators demonstrated that inhibition of TGF-β1 signaling in Gata-1low mice resulted in reduced BMF, neovascularization/osteogenesis and

Table 2. Conditions associated with bone marrow fibrosis.3

Reticulin fibrosis only

Reticulin and collagen

Hairy cell leukemia Human immunodeficiency virus Pulmonary arterial hypertension Visceral leishmaniasis Treatment with hematopoietic growth factors

Primary/secondary myelofibrosis Chronic myeloid leukemia Acute megakaryoblastic leukemia Acute myeloid leukemia Acute lymphocytic leukemia Systemic mastocytosis Myelodysplastic syndromes Paroxysmal nocturnal hemoglobinuria Hodgkin lymphoma Non-Hodgkin lymphoma Multiple myeloma Metastatic tumors Osteopetrosis Primary and secondary hyperparathyroidism Nutritional and renal rickets Osteomalacia Primary hypertrophic osteoarthropathy Tuberculosis Granulomatous diseases Grey platelet syndrome Systemic lupus erythematosus Systemic sclerosis Sjogren syndrome Anti-phospholipid syndrome 661

A.A. Zahr et al.

increased hematopoiesis in the bone marrow while a reduction of hematopoiesis in the spleen was seen.17 TGF-β1 also stimulates fibroblasts to produce bone morphogenetic protein-6 in vitro.18 Additionally, the expression of bone morphogenetic protein-4, -2, -5 and -6 genes is increased in GATA-1low mice. Bone morphogenetic proteins are a group of proteins involved in extracellular matrix synthesis, formation, and homeostasis. Although the exact role of bone morphogenetic proteins in MF is not clear, these studies suggest that they have an integral role in the pathogenesis of BMF.18 Advanced MF is not only typified by BMF, but also by osteosclerosis with an increase and thickening of the bone trabeculae. These bone trabeculae are characterized by the absence or sparse presence of osteoclasts. Bock et al. demonstrated that this pathological finding may be related to increased expression of osteoprotegerin, a known inhibitor of osteoclast formation. Levels of osteoprotegerin expression were significantly increased up to 71-fold in bone marrow samples from patients with severe BMF and documented osteosclerosis.19 Studies of cardiac fibrosis in animal models have demonstrated that TGF-β upregulates expression of Lysl oxidase (LOX).20 The LOX family contains genes encoding copper-dependent enzymes that are important for covalent cross-linking of collagen and elastin, ultimately leading to stabilization of extracellular matrix.21 Hypoxia induces LOX expression in different types of cancer cells.22 This increase in expression of LOX is associated with proliferation, invasion, and a worse prognosis.21 In vitro studies and animal models suggest an intricate interplay between LOX, megakaryocytes and the bone marrow matrix.23 LOX, by enhancing the binding of platelet-derived growth factor to its receptor, promotes expansion of the megakaryocyte lineage. Although LOX is not critical for the induction of megakaryocyte ploidy, its expression is increased in low ploidy megakaryocytes and decreased in mature and high ploidy megakaryocytes.23 Eliades et al. showed that LOX expression is increased in megakaryocytes derived from the GATA-1low murine model. Importantly, inhibiting LOX led to a decrease in BMF in these GATA-1low mice.23 Several other studies of MF found increases in various inflammatory cytokines. Tefferi et al. analyzed plasma levels of 30 cytokines/chemokines in 127 patients with PMF, identifying a common cytokine signature. Among the 20 cytokines whose levels were statistically different from those in control samples, interleukin (IL)-8 and IL-2R had the strongest correlation with phenotype and prognosis. Increased plasma levels of these two cytokines were associated with the presence of constitutional symptoms, transfusion need, leukocytosis, and inferior overall and leukemiafree survival.24 IL-8 is a pleiotropic pro-inflammatory cytokine released by numerous cells and has angiogenic, mitogenic and growth factor activity. IL-8 expression has been associated with worse prognosis in numerous solid malignancies. The pathogenic role of IL-8 in MF is not yet fully clear, but it may be involved in leukemic transformation through its growth factor and mitogenic activities.25 An increase in IL-2R levels may be a reflection of either immune activation or an increase in tumor cell burden.24 Lipocalin-2 is a potent inflammatory cytokine that is markedly elevated in the plasma of patients with MPN (particularly MF) and has been shown to preferentially promote proliferation of MF CD34+ cells, induce breaks in 662

double-stranded DNA and cause apoptosis of normal bone marrow cells, as well as enhance stromal cell proliferation through the production of reactive oxygen species.26 A lipocalin-2-induced increase in the expression of the extracellular matrix protein collagen type 1 (COL1A1) by mesenchymal stem cells as well as other disturbances in the bone marrow microenvironment likely contribute to the progression of the malignant HSC. Some studies have demonstrated dysregulation in the levels of various other circulating cytokines, although the findings have not been reproducible in all studies.9 This may reflect differences in stage of disease, prior treatments, and technical aspects. However, despite these inconsistencies, cytokines are postulated to play an important role in the initiation, progression, and phenotypic presentation of MF.

Cellular interactions transforming the bone marrow niche Osteoblastic lineage cells constitute the endosteal niche which is defined anatomically by its close proximity to trabecular or cortical bone. Osteoblastic lineage cells are derived from multipotent stromal cells and are important for the support and maintenance of normal HSC.27,28 Scheper et al. eloquently demonstrated the potential of BCR-ABL+ clonal cells to transform the endosteal bone marrow niche into a “self-reinforcing leukemic niche”. BCR-ABL1 transgenic mice have increased expansion of osteoblastic lineage cells. This increase is associated with an increase in BMF and trabecular thickening. The expansion of osteoblastic lineage cells is driven by an interaction between the MPN HSC and multipotent stromal cells. MPN HSC stimulate production of multipotent stromal cells of osteoblastic lineage through cytokines such as thrombopoietin and CCL3, as well as direct cell contact. Notably, the expansion of osteoblastic lineage cells and the associated BMF are reversible. It was demonstrated that by blocking BCR-ABL expression, the numbers of osteoblastic lineage cells decreased and BMF resolved.28 The expanded osteoblastic lineage cells have increased expression of genes involved in the regulation of extracellular matrix, cell adhesion, and inflammatory responses. These genes include targets of TGF-β1, suggesting increased TGF-β1 signaling in the osteoblastic lineage cells. Importantly the ability of the expanded populations of osteoblastic lineage cells to support normal HSC is compromised, as evidenced by reduced expression of HSC retention factors. Additionally, these cells have increased expression of pathways such as the TGF-β1 pathway, which may promote myeloid neoplastic differentiation. It was demonstrated that a malignant HSC clone can transform the bone marrow niche into a pathological environment preferentially supporting the neoplasm rather than normal HSC.27,28

Genetics Since the discovery of the Janus kinase 2 (JAK2)V617 mutation in 2005, several other clonal markers have been identified including mutations in MPL and CALR.8,29,30 Understanding genotype-phenotype associations in MPN will likely allow for better prognostication and potentially provide novel targets for therapeutic intervention. JAK2 is a member of the JAK family kinases (JAK1, JAK3, and TYK2). These cytoplasmic tyrosine kinases are associated with transmembrane class receptors of a number of cytokines (e.g. erythropoietin, thrombopoietin, haematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

granulocyte-stimulating factor) involved in cellular growth, differentiation, and survival of hematopoietic and immune cells. JAK2 is important in mediating the signaling pathways of these cytokines.2 JAK2-signal transducers and activators of transcription (STAT) dysregulation plays a central role in the pathogenesis of MPN. JAK2V617F has been identified in approximately 96% of patients with polycythemia vera, 50% of patients with essential thrombocythemia and 60% of patients with PMF. JAK2 signaling is initiated by ligand binding to the cognate cytokine receptor leading to dimerization, which in turn leads to JAK autophosphorylation. The phosphorylated kinases then phosphorylate intracellular receptor tyrosine residues creating a binding side for SH2 domain-containing proteins such as STAT. STAT phosphorylation leads to dimerization and translocation to the nucleus, affecting the transcription of genes important for cell-cycle regulation, apoptosis, and proteasomal degradation.2 Several other signaling pathways are also activated by JAK2 including the PI3K/Akt and mitogenactivated protein kinase (MAPK) signaling pathways. JAKSTAT signaling is negatively regulated by suppressor of cytokine signaling proteins (SOCS), Casitas B-cell lymphoma (CBL) and by protein tyrosine phosphatases (PTP).2 JAK2V617F is constitutively active resulting in chronic activation of the JAK-STAT pathway. This is achieved by a gain of function in the JH1 domain and loss of function in the auto-inhibitory JH2 domain. Furthermore, JAK2V617F escapes negative feedback by SOCS3.2,31 The exact mechanism by which JAK2V617F contributes to the pathogenesis of MF is not fully known. However, as mentioned above, JAK2V617F may promote BMF by transforming the bone marrow microenvironment through various cellular and cytokine-mediated mechanisms.27,28 Alhough Schepers et al. did not observe a defect in megakaryocytopoiesis, earlier studies suggested that the effects of JAK2V617F are in part mediated by its impact on platelet function and structure. Using a JAK2V617F knock-in mouse model of essential thrombocythemia, Hobbs et al. showed that the proportion of JAK2V617F-positive megakaryocytes forming proplatelets is greater than the proportion of control megakaryocytes. Moreover, the JAK2V617F knock-in megakaryocyte has increased fos gene expression. Fos is a regulator of TGF-Î˛ signaling.32 Previously Muth et al. studied patients with MPN and reported that proplatelet numbers were higher in patients with essential thrombocythemia and pre-fibrotic MF than in controls.33 Proplatelets were even more abundant in patients with fibrotic MF compared to those with pre-fibrotic MF or essential thrombocythemia.34 Exaggerated production of proplatelets may be associated with increased thrombospondin production. Thrombospondin 1 in PMF is thought to be fibrogenic by activating latent TGF-Î˛1 and inhibiting metalloproteinases.33,35 Calreticulin (CALR) is a multifunctional calcium-binding protein that is located primarily in the endoplasmic reticulum. CALR is important for intracellular calcium homeostasis and intracellular protein chaperoning.36 Recently two groups identified CALR mutations in approximately 20 to 25% of patients with essential thrombocythemia and PMF.29,37 All mutations identified so far are either insertions or deletions in exon 9 of the gene. CALR mutations are mutually exclusive with mutations in both JAK2 and MPL.29,37 Clinically, CALR mutations in PMF are associated with younger age, a higher incidence of thrombocytosis, haematologica | 2016; 101(6)

and lower leukocyte count.36,37 The prognostic impact of CALR is discussed later. CALR mutations lead to impaired calcium binding and cellular dislocation. This in turn may lead to activation of several pathways including IL-3 and JAK-STAT signaling.30 The functional consequences of CALR mutations have been recently elucidated.38,39 In murine transplant studies, mutant CALR alone was sufficient to induce an MPN phenotype. In vitro, mutant CALR leads to transformation and activation of the JAK-STAT pathway only in MPL-expressing cell lines. Moreover, mutant, but not wild-type, CALR was shown to bind MPL. These findings suggest that mutant CALR is sufficient to induce MPN through an MPL-dependent pathway. Somatic mutations of JAK2, MPL, and CALR behave as founding driver mutations responsible for the MPN phenotype. As the disease progresses, the founding malignant clone acquires additional subclonal mutations. There are a number of subclonal mutations which have been reported such as those in ASXL1, EZH2, CBL, IDH1/IDH2, TP53 and SRSF2. These mutations also have prognostic impact as discussed later.30

Prognosticators and predictors beyond the IPSS and DIPSS scores Risk stratification is at the core of current MF management. The International Prognostic Scoring System (IPSS) and the Dynamic International Prognostic Scoring System (DIPSS) are the two main prognostication schemes used to guide risk-adapted treatment strategies for patients with MF.7 These scores combine patient and MF-specific clinical variables to provide a risk category with an associated median survival. However, these scoring systems do not always reflect the genetic heterogeneity of MF. Several groups have now identified mutational profiles associated with poor prognosis in MF. Vannucchi et al. showed that patients harboring mutations in any one of the following genes, ASXL1, EZH2, SRSF2 or IDH1/ IDH2, have a shorter overall survival and a higher risk of their disease transforming into acute myeloid leukemia compared to MF patients who do not have mutations in any of these genes.40,41 The prognostic impact of mutations in CALR versus JAK2 versus MPL mutations versus triple negativity was assessed in a cohort of 254 patients with PMF. Triple negative status and CALR-/ASXL+ patients had the shortest median overall survival of 2.5 years and 2.3 years, respectively.36 Low JAK2V617F allele burden at diagnosis, homozygosity for JAK2V617F, as well as elevated levels of IL-8 and/or IL-2R are also associated with worse overall survival.42,43,24 Other molecular markers have also been associated with poor prognosis; however, they have not yet been validated in large prospective studies.

Impact of bone marrow fibrosis on the prognosis of myeloproliferative neoplasms and other hematologic malignancies Myelofibrosis As discussed above, the IPSS and DIPSS are currently the two commonly used scales to assess prognosis in MF and the prognostication may now be further refined through the incorporation of gene mutation profiling. 663

A.A. Zahr et al.

BMF is not at present included in risk stratification schemes for MF or other hematologic malignancies. However, studies suggest a correlation between the grade of BMF and prognosis in MF.44 Several studies have evaluated the prognostic impact of BMF in MF. In most of these studies, high grade BMF was associated with worse outcome.45,46 In a retrospective analysis that included 131 patients with PMF, Lekovic et al. found that BMF grade >1 was associated with shorter overall survival (median: 51 months) compared to grade â&#x2030;¤1 (median: 147 months). Importantly the grade of BMF was an independent risk factor for overall survival when analyzed with IPSS and competing comorbidities.45 In a study that included 196 patients with PMF, higher grade BMF was an independent predictor of poor survival.46 Moreover, higher grade BMF can further refine the prognosis of PMF patients with low and intermediate risk IPSS.46 However, Nazha et al. reported similar rates of overall survival, event-free survival, and disease transformation in MF patients with different grades of BMF. Their study included 512 patients with MF: 1% had MF grade 0, 9% MF grade 1, 33% MF grade 2, and 57% MF grade 3. Higher grade of BMF correlated with clinical manifestations of MF such as lower hemoglobin level, higher percentage of peripheral blood blasts and larger spleen. Although grade of BMF was demonstrated to be correlated with IPSS and DIPSS risk, an impact on overall survival was not evident in this analysis. This may in part be explained by higher numbers of patients with MF grade 2 and MF grade 3 in this cohort.47 In a recent retrospective Italian study of 540 bone marrow specimens from time of PMF diagnosis, higher grade BMF was found to be associated with more constitutional symptoms (P<0.0001), larger splenomegaly (P<0.0001), greater risk of developing anemia (P<0.0001) and thrombocytopenia (P=0.003), as well as with higher risk IPSS score.48 Although there was no association between driver mutation status and grade of BMF, high molecular risk mutations, such as ASXL1 and EZH2, were seen more frequently in PMF patients with higher grade BMF. Importantly, grades 2 and 3 BMF maintained their adverse prognostic values for overall survival [HR 3.9 (1.4-10.8) and HR 4.2 (1.5-12.0), respectively], independently of IPSS score, driver mutation and high-risk molecular status. Pre-transplant MF grade 3 was associated with worse overall survival in a univariate analysis of patients with MF who underwent hematopoietic stem cell transplantation (HSCT) with reduced intensity conditioning. However, this was not confirmed in multivariate analysis.49 Interestingly, the regression of BMF after HSCT was associated with a better overall survival, as discussed in more detail in the treatment section.44 With the advent of mutation-based prognostication in MF, it remains to be seen whether BMF grade will have an additional prognostic impact.

Essential thrombocythemia/polycythemia vera Increased reticulin BMF has been reported in approximately 20% of patients with essential thrombocythemia and can also be seen to various degrees in the bone marrow of patients with polycythemia vera.50,51 BMF was associated with an increased risk of transformation to MF or acute myeloid leukemia (HR 1.89, P=0.0359 for essential thrombocythemia; HR 1.71, P=0.0164 for polycythemia vera). However, BMF grade was not associated with 664

worse overall survival.52 These results were replicated in another study by Campbell et al. in which 311 patients with essential thrombocythemia were included. Again, increased reticulin BMF grade at diagnosis was not significantly associated with poorer overall survival. Interestingly, increased grade of reticulin BMF was associated with an increased rate of arterial thrombosis, major hemorrhage and progression to MF.50 The presence of BMF in polycythemia vera has been associated with an increased risk of transformation to MF, but this has not been shown to affect overall survival.51

Chronic myelogenous leukemia BMF is not specific to MPN and can be seen in many other hematologic malignancies, solid malignancies, and non-malignant conditions, as summarized in Table 2. Patients with chronic myelogenous leukemia have been shown to have increased BMF compared to their normal counterparts53 and, in the pre-imatinib era, an increased degree of BMF prior to starting treatment was associated with a worse outcome.54 Treatment with imatinib was associated with complete reversal of pathological BMF in 73% of treated patients (14/19 patients) in one series.53 The degree of BMF in chronic myelogenous leukemia at baseline in imatinib-treated patients was not shown to be an adverse prognostic factor.55 However, emergence of small foci with abnormal fiber increase or BMF during imatinib treatment is associated with a lower probability of achieving complete cytogenetic response or major molecular response. Furthermore, emergence of abnormal fiber increase can occur in patients who have achieved complete cytogenetic or major molecular response preceding the loss of molecular response. This is believed to be a consequence of the ability of the chronic myelogenous leukemia clone to produce fibrogenic cytokines even when the quantity of the clone is below the threshold of cytogenetic or molecular detection.53,56 Importantly emerging or relapsing abnormal fiber increase was an independent predictor of failure of imatinib treatment. Advanced BMF in chronic myelogenous leukemia has also been associated with an increased risk of developing accelerated phase and blast phase disease.53

Myelodysplastic syndromes Significant BMF has been reported in approximately 1020% of patients with myelodysplastic syndromes (MDS). BMF in MDS is associated with profound cytopenias and increased red cell/platelet transfusion dependence.57,58 In a study of 301 patients with MDS, patients with grades 2 and 3 BMF had shorter overall and leukemia-free survival compared to those with grade 0 or 1 BMF. The impact of BMF on overall and leukemia-free survival was seen in WHO subgroups without excess of blasts (overall survival, HR=2.89, P=0.001; leukemia-free survival, HR=2.21, P=0.006) and in refractory anemia with excess blasts type 1 and type 2 (overall survival, HR=2.25, P=0.004; leukemia-free survival, HR=2.03, P=0.01). In MDS patients stratified by the IPSS and WPSS (WHO classification-based Prognostic Scoring System) risk score, grades 2 and 3 BMF maintained their prognostic significance. Furthermore, developing BMF during the course of MDS was shown to be associated with a worse outcome.59 These results held true even when applying the revisedIPSS. Advanced BMF is an adverse risk factor that is not captured by the revised-IPSS.58 In this study, BMF grade 2haematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

3 was an independent variable that negatively influenced overall and leukemia-free survival in MDS patients treated with a non-transplant modality. Moderate/severe BMF was identified in 17% of cases of therapy-related MDS, but was not associated with an increased risk of leukemic transformation or inferior outcome.57 This may in part be explained by the intrinsic high-risk features of therapyrelated MDS resulting from genetic and epigenetic alterations rather than histopathomorphological aspects. The presence of severe BMF was associated with reduced overall survival even in MDS patients who underwent allogeneic HSCT.58,60 It has also been shown that in MDS patients treated with azacitdine, grade 3 BMF is associated with a poor response to treatment.61 The potential prognostic impact of BMF extends beyond myeloid malignancies. Among patients with chronic lymphocytic leukemia, those with high grade BMF (grades 2 and 3) had a worse 5-year overall survival rate than those with grade 0-1 BMF (51.9% versus 86.92 %, respectively). Fluorescence in situ hybridization analysis was performed in only a few patients in this study and none was tested for P53 status. It, therefore, remains unknown whether BMF has true prognostic significance independent of modern prognostic features.62

Grading of bone marrow fibrosis Increased reticulin and collagen deposition in the bone marrow is the hallmark of MF. Reticulin and collagen are connective tissue fibers that provide the structural framework of bone marrow stroma. BMF is routinely assessed and graded in core biopsies from patients with a known or suspected MPN, and is a major diagnostic criterion in the 2008 WHO classification system of MPN. Reticulin is detectable by silver staining methods (Figure 1) while collagen deposition is detected by trichrome staining (Figure 2). Several reporting systems have been proposed to quantify reticulin and collagen deposition in the bone marrow.

Figure 1. Photomicrograph showing reticulin (silver) stain in a specimen from a patient with myelofibrosis. Note the diffuse and dense increase in reticulin with extensive intersections which would be graded as 2+ myelofibrosis. The left upper corner is a mark up image of post-image analysis processing with the reticulin fibers being de-convoluted from background tissue. The area occupied by the fibers, as well as branching points, could be objectively quantified using computer assisted image analysis (original magnification, 400X).

haematologica | 2016; 101(6)

The most frequently used grading systems are based on reticulin grade scored using the Bauermeister system from 1971 (0-4 scale) and the recently revised European consensus system (0-3 scale) (Table 1 and Figure 1). All accepted traditional grading systems are dependent on manual grading of reticulin by a histopathologist. These traditional grading systems are semi-quantitative, and suffer major limitations related to subjectivity.63 This problem is further confounded by the heterogeneity of fibrosis within a given sample, variability of pre-analytical processing, staining inconsistency and subjective assessment from lack of a positive staining internal standard, and inadequate guidance to disregard lymphoid nodules, vessels as well as fibers framing adipocytes. Reports that address these limitations are scarce to non-existent. Whole slide imaging with computer-assisted image analysis has emerged recently as a more objective approach to provide numerical assessment of both fibrosis and osteosclerosis, offering a more precise way of following patients over time.63 The most significant advantages of this approach are its contribution to standardizing assessment of fibrosis among histopathologists and hematologists and that it makes it possible to assess disease progression or to serially assess the impact of a therapy on marrow histology as a surrogate endpoint of efficacy of therapy. The microscopic assessment of BMF by skilled hematopathologists remains the standard of care in the diagnosis MPN. However, computer-assisted image analysis is being increasingly utilized in clinical trials for objective assessment of BMF as an adjunct to semi-quantitative grading. Computer-assisted image analysis correlates well with morphology but was more sensitive than histopathology for detecting BMF level changes in MF patients undergoing treatment.64 Recently, a stereology-based computer-assisted image analysis method using systematic uniform random sampling and line counting to calculate length density of the reticulin network as well as to measure heterogeneity across the bone marrow sample was developed by Salama et al.65 Computer-based stereology proved to be more reproducible than manual scoring at predicting a therapeutic actionable cut-point. This novel stereology-based method is fast and can be easily implemented in the clini-

Figure 2. Trichrome stain showing deposits of collagen fibers. The fibers are arranged in bundles that are marked with red arrows (original magnification 400X).

665

A.A. Zahr et al.

cal laboratory with a high level of reproducibility. A major advantage of this approach is that it can provide a measure of heterogeneity of the BMF. Another promising stain-free modality that employs two photon excitation and second harmonic generation phenomena has great potential as an application for quantification of BMF where pre-analytical variability related to stain limitation can be eliminated. Accurate scoring of BMF remains a major area of academic interest and an unmet clinical need.

Treatment strategies for bone marrow fibrosis Most of the data regarding treatments targeting BMF has been garnered from MF. However, because of a potential overlap in the pathogenesis of BMF among different disease entities, some treatment strategies can be used in diseases other than MF. Historically, BMF has been considered a secondary or reactive process driven by the clonal malignancy and not produced by the malignant cell population itself. In an early, small study, the incidence of cytogenetic abnormalities in fibroblasts from patients with MF was comparable to that in patients with non-malignant causes of anemias. Importantly, the karyotypic abnormalities seen in fibroblasts were often different from those seen in the hematopoietic cells. This finding supports the hypothesis that the fibroblast population in MF is the result of a reactive phenomenon rather than a malignant outgrowth.66 However, there are evolving data to suggest clonal involvement of fibroblasts in PMF as evidenced by the detection of JAK2V617F and other chromosomal abnormalities in fibroblasts that are concordant with the MPN cell population.67 In general, the goals of MF-directed treatment can be divided into two broad categories: clonal eradication and treatments targeting various signaling pathways and mediators implicated in BMF. As discussed in the section on pathogenesis, eradicating the BCR-ABL1+ clonal cells in animal models of chronic myelogenous leukemia was associated with reversal of pathological BMF.28 In PMF, allogeneic HSCT is the only therapeutic intervention with curative potential, effectively eradicating the malignant MPN HSC clone, leading to reversal of BMF and MPNassociated histomorphological features of the bone marrow. In a study by Rondelli et al., 21 patients with intermediate/high risk MF underwent reduced intensity conditioning-HSCT. Nineteen had grade 3 or 4 BMF at baseline. After transplantation, all evaluable patients had BMF grade 0 - 2 for at least 12 months of follow up.68 Importantly, regression of BMF at day +100 after HSCT in patients with MF is associated with improved survival, independently of IPSS score at the time of transplantation, as demonstrated by Kroger et al.44 This conclusion was reached in a study that included 57 patients with PMF or MF evolving from polycythemia vera or essential thrombocythemia who underwent reduced intensity conditioning-HSCT. Bone marrow histology was available for 35 patients at days +30 and +100, while 13 patients only had a day +30 bone marrow biopsy, and nine patients only had a day +100 bone marrow biopsy. Patients with grade 2/3 BMF at day +100 had a higher rate of dependence on red blood cell and platelet transfusions compared to patients who had grade 0/1 BMF at day +100. Importantly, patients with grade 0/1 BMF at day +100 after HSCT had 666

a 5-year overall survival rate of 96% compared to 57% in those with persistent grade 2/3 BMF (P=0.04). This survival advantage was attributed to a lower risk of treatment-related mortality and a decrease in relapse rate. Additionally, patients with a reduction of BMF by 2 or 3 grades at day +100 after HSCT had a trend towards a better overall survival rate than that of patients with unchanged BMF (95% versus 71%, respectively; P=0.19). The regression of BMF achieved in patients after reduced intensity conditioning-HSCT did not correlate with JAK2V617F allele burden or with IPSS risk group at the time of transplantation.44 This study highlights the prognostic significance of BMF in MF not accounted for by the current prognostic scoring systems. It also raises the potential of improving outcomes of HSCT by combining anti-fibrogenic strategies prior to or even after the transplant. Apart from HSCT, none of the conventional therapies utilized in the treatment of MF is considered curative. However, there are several published reports describing reversibility of BMF with interferon-α (IFN-α) based therapy. Early data from patients with MPN did not show significant changes in grade of reticulin BMF with the use of recombinant IFN-α.69 However, in a subsequent, small, prospective study, 17 patients with early stage MF, without grade 3/4 BMF, were treated with IFN-α (n=14) or pegylated IFN-α (n=3) and four of the patients had documented improvements in bone marrow reticulin and collagen fibrosis. Two of these patients had complete resolution of BMF and megakaryocytic atypia after 1-4 years of treatment. The median duration of bone marrow responses was 1.9 years.70 A large retrospective analysis by Ianotto et al. suggests that IFN-α2a can be clinically effective in patients with MF who do not have massive splenomegaly (spleen size <6 cm), marked leukopenia or thrombocytopenia, and BMF <grade 3.71 However, there were no data regarding the effect of IFN-α treatment on BMF in this study. Evidence of reduction in grade of BMF with IFN-α therapy in patients with MF is mostly restricted to case reports.72 Collectively, these reports suggest IFN-α therapy may have the most beneficial impact on BMF early on in the disease course. However, the clinical significance of this finding and the impact on overall survival of reducing/eliminating BMF with IFN-α-based therapy remain uncertain.

Impact of ruxolitinib on bone marrow fibrosis With the introduction of selective JAK2 inhibitors into clinical investigation over the last 10 years, initial expecta-

Table 3. Ruxolitinib therapy and effect on bone marrow fibrosis.

24 months 77

Kvasnicka et al. Number of patients Improvement, n. Worsening, n. Thiele et al.78 Number of patients Stabilization, n. Improvement, n. Worsening, n.

RUX 67 10 19 RUX 68 39 10 19

HU 31 3 12 BAT 97 52 6 39

48 months

60 months

RUX HU 17 20 4 0 4 13 RUX BAT 38 63 19 29 9 1 10 33

RUX BAT 26 32 11 9 9 1 6 22

Rux: ruxolitinib; HU: hydroxyurea; BAT: best available therapy.

haematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

tion of clonal suppression as measured by elimination of molecular and karyotypic abnormalities was dampened by the observations of persistent clonal hematopoiesis and unaltered MPN bone marrow pathological features. None of the approximately dozen oral JAK2 tyrosine kinase inhibitors evaluated in the preclinical and clinical MF setting has recapitulated the success of the BCR-ABL1 inhibitors in chronic myelogenous leukemia. An analysis of results from bone marrow biopsy specimens obtained from MF patients enrolled in the pivotal phase III COMFORT studies at 6 and 12 months of therapy with ruxolitinib failed to show improvements in histopathological abnormalities and did not confirm reductions in grade of BMF.73,74 However, several case reports and retrospective analyses suggest that longer treatment duration with ruxolitinib may have a modest impact on BMF in a subset of patients. In several published case reports, resolution of BMF was noted after 17 and 48 weeks of ruxolitinib treatment.75,76 Recently, reports of long-term follow-up of bone marrow responses in MF patients treated in the phase I/II trial of ruxolitinib have suggested modest responses in BMF in selected cases. The changes in grade of BMF at 24 and 48 months of ruxolitinib treatment were analyzed and compared to those in a similar, matched cohort of MF patients treated with hydroxyurea (Table 3). At 24 months, BMF had stabilized in 57% and 52% of the ruxolitinib and hydroxurea treatment groups, respectively and 15% and 10%, respectively, had improvements in grade of BMF. At 48 months, stabilization was achieved in 53% of ruxolitinib-treated patients versus 35% of the hydroxyurea-treated patients. Improvements in BMF grade were achieved in 24% of the ruxolitinib group, but in none of those treated with hydroxyurea.77 These results were further validated in a larger cohort with a longer follow-up. In this study, changes in the grade of BMF were determined at 24, 48 and 60 months of ruxolitinib treatment and compared with those in a group of MF patients treated with “best available therapy”. The majority of patients treated with “best available therapy” received hydroxyurea or various sequential therapies, or were

being observed; only a few patients were treated with IFN-α. As summarized in Table 3, compared to this control cohort, a higher percentage of patients in the ruxolitinib group had stabilization or improvement of BMF. Ruxolitinib-treated patients who achieved stabilization or improvement in grade of BMF at 24 months had a reduced relative risk of death in this analysis.78

Targeting fibrogenic cytokines Pirfenidone is an anti-fibrotic agent that inhibits fibrogenic cytokines including platelet-derived growth factor, tumor necrosis factor-α and TGF-β and was studied in 28 patients with MF. None of the patients had improvement in BMF or osteosclerosis. Additionally, there were no significant clinical benefits seen in terms of improvement in anemia or reduction in splenomegaly.79 Targeting fibrogenic cytokine expression either alone or in combination with a JAK2 inhibitor is an active area of research in MF. As previously mentioned, the TGF-β signaling pathway plays a central role in the pathogenesis of BMF in MF. Monoclonal antibodies antagonizing TGF-β are being evaluated in various fibrotic conditions including pulmonary fibrosis, glomerulosclerosis, and MF. GC1008 (fresolimumab, Genzyme) is a human IgG4 monoclonal antibody capable of neutralizing TGF-β isoforms 1, 2, and 3. The relative tolerability of GC1008 was demonstrated in a phase I trial in advanced melanoma and renal cell cancer.80 In an aborted phase I trial, three MF patients were treated with GC1008 and changes in bone marrow pathology were assessed after 6 and 12 months.14 There were no observed changes in bone marrow reticulin or collagen fibrosis after six or 12 cycles of treatment and no appreciable decrease in spleen size after six cycles of treatment. Interestingly, two treated patients obtained responses in anemia that were durable, including transfusion independence for over 16 months in one patient (unpublished observation, JM). Of interest, plasma levels of TGF-β were significantly higher in the three patients than in normal controls and after treatment with GC1008 were undetectable in two evaluable patients. Further clinical evaluation of

Figure 3. Four images from unstained bone marrow biopsies obtained from patients with myeloproliferative neoplasms with a spectrum of marrow fibrosis ranging from 0-3 according to the revised European consensus system for grading bone marrow fibrosis. The red is a pseudo color that highlights all tissue elements in the bone marrow according to two-photon excitation. The technology utilizes the second harmonic generation phenomenon to highlight fibrillar collagen with high specificity as highlighted by the green fluorescent colored structures.

haematologica | 2016; 101(6)

667

A.A. Zahr et al.

TGF-β inhibition alone or in combination with other agents will be necessary for a complete assessment of this targeted therapeutic approach.14 The Myeloproliferative Disorder Research Consortium (MPD-RC) will be evaluating the oral selective TGF-β receptor 1 kinase inhibitor, galunisertib (LY2157299, Lilly), in patients with MF and at least MF grade 2 BMF. Pentraxin (PTX) family proteins include C-reactive protein (CRP; PTX1), serum amyloid P (SAP; PTX2), and pentraxin-3 (PTX3).81 PTX2 is a circulating plasma acute phase response protein made by the liver which localizes to the site of injury and affects monocyte differentiation and function in the removal of damaged tissue. Physiologically, PTX2 may act as a powerful antifibrotic agent because of its ability to inhibit human derived fibrocyte differentiation.82 Injections of PTX2 reduce fibrosis significantly in several animal models (pulmonary fibrosis, ischemic cardiac fibrosis, renal fibrosis).83-88 On the other hand, PTX3 promotes human and murine fibrocyte differentiation.81 PRM-151 is a recombinant PTX2 being actively investigated in the treatment of various fibrotic diseases including MF (NCT01981850).89 In a phase II study that included 26 patients with MF, PRM-151 (alone or in combination with ruxolitinib) was associated with symptom improvement and/or BMF reduction in 43% of the treated patients. Specifically, a ≥1 grade reduction of BMF was seen in 35% of patients at 24 weeks of treatment. The therapy was well tolerated without significant infusion reactions or grade 3/4 treatment-associated adverse events.90,91 Longer term follow-up of 13 patients treated for at least 72 weeks with PRM-151 continued to demonstrate reductions of ≥1 grade of BMF in nearly 70% of patients, and improvements in spleen size, anemia and thrombocytopenia, as well as symptom burden.92 The second stage of this trial is now evaluating PRM-151 monotherapy in patients who are intolerant of, refractory to or ineligible for ruxolitinib therapy. The relative pre-treatment levels of PTX2 and PTX3 in the bone marrow may serve as a predictor of response to PRM-151. As mentioned above, LOX levels are increased in GATA-1low mice and LOX inhibition led to a decrease in BMF in these animal models. Bone marrow samples from patients with PMF overexpress all LOX genes. Furthermore, serum LOX levels were significantly higher in PMF patients than in a control group.93 Simtuzumab is a humanized monoclonal antibody that binds to and inhibits LOXL2. A phase II trial of simtuzumab monotherapy (stage 1) and combination therapy with ruxolitinib (stage 2) in patients with intermediate-1 or higher risk MF has completed recruitment and the results were recently presented (NCT01369498).94 A total of 24 patients were treated in the stage 1 part of the study and 30 in the stage 2 part, for a minimum median duration of 22 weeks, without a clear signal of response in terms of reduction of BMF. Larger studies are required to determine whether serum levels of LOX can be used to refine diagnosis and prognosis, and to determine the predictive potential of response to LOX-directed therapy in patients with MF. In the Gata1low mouse model of MF, the animals have increased expression of the hedgehog pathway in both bone marrow and spleen.17 An increase in the expression of hedgehog target genes has also been observed in granulocytes isolated from MPN patients.95 The exact role of the hedgehog pathway in MF and its contribution to BMF are not fully understood. However, preclinical and clinical data 668

suggest that hedgehog pathway inhibitors have therapeutic activity in MF. Combining sonidegib (LDE225, Novartis), a hedgehog inhibitor, with ruxolitinib in a murine transplant model of essential thrombocythemia/MF resulted in a reduction of mutant allele burden in the bone marrow and a significant reduction in BMF compared to that achieved by ruxolitinib alone. Sonidegib, saridegib, and PF-04449913 are currently under clinical investigation, alone or in combination with other agents, in MF. Seven patients with MF were included in a phase I study investigating PF-04449913 in myeloid malignancies. Two of these patients achieved durable clinical responses (>50% reduction in spleen size) and one patient had a significant reduction in BMF. The therapy was relatively well tolerated and is currently being investigated as a second-line agent in MF (NCT02226172).96,97 Sonidegib in combination with ruxolitinib was tested in a multicenter, phase Ib/II study: the maximum tolerated dose was not reached during the dose escalation phase, but myelosuppression and elevation of creatinine kinase levels were reported.98 A total of 27 MF patients were treated at the recommended phase II dose of 400 mg once daily with ruxolitinib 20 mg twice daily. Although the combination was relatively well tolerated and resulted in approximately 50% of patients achieving a ≥35% decrease in spleen volume, only two treated patients had at least a one grade reduction in BMF.99

Disrupting the myelofibrosis bone marrow microenvironment As discussed above, modulation of the bone marrow microenvironment by the neoplastic hematopoietic clone plays an integral role in the pathogenesis of myeloid malignancies. Disrupting the tumor-microenvironment interaction is a potential therapeutic strategy in MPN.28 The effects of disrupting the tumor microenvironment by targeting Eph receptor tyrosine kinases are currently being studied in both solid and hematologic malignancies.100,101 EphA3 is important in cell positioning during fetal development and is not expressed in normal adult tissues. However, EphA3 expression has been demonstrated in various hematologic and solid tumors.88,89 In solid malignancies, EphA3 is preferentially expressed in tumor stroma, vasculature, and bone marrow-derived mesenchymal stem cells.101 Data from Vail et al. suggest that Eph is not ligated in stromal cells derived from the solid tumor so it is kinase-dormant. In solid tumors, kinase-dormant Eph lead to cell-cell adhesion, invasion and tumor maintenance. Targeting Epha3 by ChIIIA4 led to EphA3 kinase activation, cell contraction and apoptosis of solid tumorresident multipotent stromal cells. In turn, this led to disruption of the integrity of tumor stromal architecture, microvasculature, and ultimately to inhibition of tumor growth. Epha3 is expressed in blood and mone marrow leukemic HSC as well as the stromal compartment.102 Humaneered® IIIA4 (KB004, Kalobios) is currently being evaluated in the treatment of hematologic malignancies including MF (NCT01211691). Early data from a phase I/II trial of 58 patients (the majority with acute myeloid leukemia) suggest a well-tolerable side effect profile with infections (41.4%), febrile neutropenia (20.7%), and infusion-related reactions (13.8%) being the most frequent side effects.103 Efficacy data are also encouraging. The potential impact of KB004 on BMF in responding patients is of interest. In a single patient with acute myeloid leukemia, a sushaematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

tained complete remission was observed for over 1 year and this clinical response was also accompanied by a decrease in reticulin and collagen BMF. In another patient with MF, clinical improvement (decrease in spleen size, transfusion independence) was achieved and accompanied by a significant improvement in BMF.100 The anti-fibrotic mechanism of action of KB004 is unknown.

Conclusion BMF is seen in many hematologic and non-hematologic conditions and is a prominent pathologic feature of MF. The extent to which BMF contributes to a disorganized bone marrow microenvironment and promotes disease progression rather than serving as a biomarker reflecting disease activity remains incompletely understood. BMF is

References 12. 1. Thiele J, Kvasnicka HM, Facchetti F, Franco V, van der Walt J, Orazi A. European consensus on grading bone marrow fibrosis and assessment of cellularity. Haematologica. 2005;90(8):1128-1132. 2. Meyer SC, Levine RL. Molecular pathways: molecular basis for sensitivity and resistance to JAK kinase inhibitors. Clin Cancer Res. 2014;20(8):2051-2059. 3. Kuter DJ, Bain B, Mufti G, Bagg A, Hasserjian RP. Bone marrow fibrosis: pathophysiology and clinical significance of increased bone marrow stromal fibres. Br J Haematol. 2007;139(3):351-362. 4. Reilly JT, McMullin MF, Beer PA, et al. Guideline for the diagnosis and management of myelofibrosis. Br J Haematol. 2012;158(4):453-471. 5. Mesa RA, Niblack J, Wadleigh M, et al. The burden of fatigue and quality of life in myeloproliferative disorders (MPDs): an international internet-based survey of 1179 MPD patients. Cancer. 2007;109(1):68-76. 6. Cervantes F, Dupriez B, Pereira A, et al. New prognostic scoring system for primary myelofibrosis based on a study of the International Working Group for Myelofibrosis Research and Treatment. Blood. 2009;113(13):2895-2901. 7. Gangat N, Caramazza D, Vaidya R, et al. DIPSS plus: a refined Dynamic International Prognostic Scoring System for primary myelofibrosis that incorporates prognostic information from karyotype, platelet count, and transfusion status. J Clin Oncol. 2011;29(4):392-397. 8. Tabarroki A, Tiu RV. Molecular genetics of myelofibrosis and its associated disease phenotypes. Transl Med UniSa. 2014;8:53-64. 9. Hasselbalch HC. The role of cytokines in the initiation and progression of myelofibrosis. Cytokine Growth Factor Rev. 2013;24(2):133-145. 10. Roberts AB, Sporn MB, Assoian RK, et al. Transforming growth factor type beta: rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Natl Acad Sci USA. 1986;83 (12):4167-4171. 11. Kimura A, Katoh O, Hyodo H, Kuramoto A. Transforming growth factor-beta regulates growth as well as collagen and fibronectin

haematologica | 2016; 101(6)

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

a culminating effect of a complex interplay between MPN cells and supporting stromal cells through the interaction of various inflammatory cytokines such as TGF-β. Increased reticulin and collagen fibrosis may have a prognostic significance that is not accounted for by current prognostic scoring systems. Better understanding of the molecular and cellular mechanisms governing MF will potentially lead to more effective therapies targeting the MPN HSC. Anti-fibrotic strategies are currently being evaluated in a number of therapeutic trials aimed at disrupting the malignant bone marrow niche and promoting normal over malignant hematopoiesis. In the future, validating the prognostic impact of BMF, identifying soluble surrogates to allow for easier monitoring of changes in BMF, and combining anti-fibrogenic strategies with currently available treatments, including HSCT, will likely result in improved clinical outcomes.

synthesis of human marrow fibroblasts. Br J Haematol. 1989;72(4):486-491. Massagué J. The transforming growth factor-beta family. Annu Rev Cell Biol. 1990;6:597-641. Ignotz RA, Endo T, Massagué J. Regulation of fibronectin and type I collagen mRNA levels by transforming growth factor-beta. J Biol Chem. 1987;262(14):6443-6446. Mascarenhas J, Li T, Sandy L, et al. Antitransforming growth factor beta (TGF-β) therapy in patients with myelofibrosis. Leuk Lymphoma. 2014;55(2):450-452. Chagraoui H, Komura E, Tulliez M, Giraudier S, Vainchenker W, Wendling F. Prominent role of TGF-beta 1 in thrombopoietin-induced myelofibrosis in mice. Blood. 2002;100(10):3495-3503. Chou JM, Li CY, Tefferi A. Bone marrow immunohistochemical studies of angiogenic cytokines and their receptors in myelofibrosis with myeloid metaplasia. Leuk Res. 2003;27(6):499-504. Zingariello M, Martelli F, Ciaffoni F, et al. Characterization of the TGF-β1 signaling abnormalities in the Gata1low mouse model of myelofibrosis. Blood. 2013;121 (17):3345-3363. Bock O, Höftmann J, Theophile K, et al. Bone morphogenetic proteins are overexpressed in the bone marrow of primary myelofibrosis and are apparently induced by fibrogenic cytokines. Am J Pathol. 2008;172(4):951-960. Bock O, Loch G, Schade U, et al. Osteosclerosis in advanced chronic idiopathic myelofibrosis is associated with endothelial overexpression of osteoprotegerin. Br J Haematol. 2005;130(1):76-82. Voloshenyuk TG, Landesman ES, Khoutorova E, Hart AD, Gardner JD. Induction of cardiac fibroblast lysyl oxidase by TGF-β1 requires PI3K/Akt, Smad3, and MAPK signaling. Cytokine. 2011;55 (1):90-97. Nishioka T, Eustace A, West C. Lysyl oxidase: from basic science to future cancer treatment. Cell Struct Funct. 2012;37(1): 75-80. Schietke R, Warnecke C, Wacker I, et al. The lysyl oxidases LOX and LOXL2 are necessary and sufficient to repress E-cadherin in hypoxia: insights into cellular transformation processes mediated by HIF-1. J Biol Chem. 2010;285(9):6658-6669.

23. 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):27630-27638. 24. Tefferi A, Vaidya R, Caramazza D, Finke C, Lasho T, Pardanani A. Circulating interleukin (IL)-8, IL-2R, IL-12, and IL-15 levels are independently prognostic in primary myelofibrosis: a comprehensive cytokine profiling study. J Clin Oncol. 2011;29(10): 1356-1363. 25. Zarogoulidis P, Katsikogianni F, Tsiouda T, Sakkas A, Katsikogiannis N, Zarogoulidis K. Interleukin-8 and interleukin-17 for cancer. Cancer Invest. 2014;32(5):197-205. 26. Lu M, Xia L, Liu YC, et al. Lipocalin produced by myelofibrosis cells affects the fate of both hematopoietic and marrow microenvironmental cells. Blood. 2015;126(8):972-982. 27. Mullally A, Ebert BL. Sinister symbiosis: pathological hematopoietic-stromal interactions in CML. Cell Stem Cell. 2013;13(3): 257-258. 28. Schepers K, Pietras EM, Reynaud D, et al. Myeloproliferative neoplasia remodels the endosteal bone marrow niche into a selfreinforcing leukemic niche. Cell Stem Cell. 2013;13(3):285-299. 29. Nangalia J, Massie CE, Baxter EJ, et al. Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N Engl J Med. 2013;369(25):2391-2405. 30. Cazzola M, Kralovics R. From Janus kinase 2 to calreticulin: the clinically relevant genomic landscape of myeloproliferative neoplasms. Blood. 2014;123(24):3714-3719. 31. Bandaranayake RM, Ungureanu D, Shan Y, Shaw DE, Silvennoinen O, Hubbard SR. Crystal structures of the JAK2 pseudokinase domain and the pathogenic mutant V617F. Nat Struct Mol Biol. 2012;19(8):754-759. 32. Hobbs CM, Manning H, Bennett C, et al. JAK2V617F leads to intrinsic changes in platelet formation and reactivity in a knockin mouse model of essential thrombocythemia. Blood. 2013;122(23):3787-3797. 33. Muth M, Büsche G, Bock O, Hussein K, H. K. Aberrant proplatelet formation in chronic myeloproliferative neoplasms. Leuk Res. 2010;34(11):1424-1429. 34. Kreipe H, Büsche G, Bock O, Hussein K. Myelofibrosis: molecular and cell biological aspects. Fibrogenesis Tissue Repair. 2012;5(Suppl 1):S21.

669

A.A. Zahr et al. 35. Muth M, Engelhardt BM, Kröger N, et al. Thrombospondin-1 (TSP-1) in primary myelofibrosis (PMF) - a megakaryocytederived biomarker which largely discriminates PMF from essential thrombocythemia. Ann Hematol. 2010;90(1):33-40.36. 36. Tefferi A, Lasho TL, Finke CM, et al. CALR vs JAK2 vs MPL-mutated or triple-negative myelofibrosis: clinical, cytogenetic and molecular comparisons. Leukemia. 2014;28 (7):1472-1477. 37. Klampfl T, Gisslinger H, Harutyunyan AS, et al. Somatic mutations of calreticulin in myeloproliferative neoplasms. N Engl J Med. 2013;369(25):2379-2390. 38. Elf S, Abdelfattah N, Chen E, et al. Physical interaction between mutant calreticulin and the thrombopoietin receptor is required for hematopoietic transformation. ASH. 2015;57th Annual Meeting and Exposition (LBA-4). 39. Chachoua I, Pecquet C, El-Khoury M, et al. Thrombopoietin receptor activation by myeloproliferative neoplasm associated calreticulin mutants. Blood. 2015 Dec 14. [Epub ahead of print] 40. Guglielmelli P, Lasho TL, Rotunno G, et al. The number of prognostically detrimental mutations and prognosis in primary myelofibrosis: an international study of 797 patients. Leukemia. 2014;28(9):1804-1810. 41. Vannucchi AM, Lasho TL, Guglielmelli P, et al. Mutations and prognosis in primary myelofibrosis. Leukemia. 2013;27(9):18611869. 42. Guglielmelli P, Barosi G, Specchia G, et al. Identification of patients with poorer survival in primary myelofibrosis based on the burden of JAK2V617F mutated allele. Blood. 2009;114(8):1477-1483. 43. Barosi G, Poletto V, Massa M, et al. JAK2 V617F genotype is a strong determinant of blast transformation in primary myelofibrosis. PLoS One. 2013;8(3):e59791. 44. 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. 45. Lekovic D, Gotic M, Perunicic-Jovanovic M, et al. Contribution of comorbidities and grade of bone marrow fibrosis to the prognosis of survival in patients with primary myelofibrosis. Med Oncol. 2014;31(3):869. 46. Gianelli U, Vener C, Bossi A, et al. The European Consensus on grading of bone marrow fibrosis allows a better prognostication of patients with primary myelofibrosis. Mod Pathol. 2012;25(9):1193-1202. 47. Nazha A, Estrov Z, Cortes J, Bueso-Ramos CE, Kantarjian H, Verstovsek S. Prognostic implications and clinical characteristics associated with bone marrow fibrosis in patients with myelofibrosis. Leuk Lymphoma. 2013;54(11):2537-2539. 48. Guglielmelli P, Rotunno G, Pacilli A, et al. Prognostic impact of bone marrow fibrosis in primary myelofibrosis: a study of Agimm Group on 540 patients. ASH. 2015;57th Annual Meeting and Exposition:634. 49. Alchalby H, Yunus DR, Zabelina T, et al. Risk models predicting survival after reduced-intensity transplantation for myelofibrosis. Br J Haematol. 2012;157(1): 75-85. 50. Campbell PJ, Bareford D, Erber WN, et al. Reticulin accumulation in essential thrombocythemia: prognostic significance and relationship to therapy. J Clin Oncol. 2009;27(18):2991-2999. 51. Barbui T, Thiele J, Passamonti F, et al. Initial

670

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

bone marrow reticulin fibrosis in polycythemia vera exerts an impact on clinical outcome. Blood. 2012;119(10):2239-2241. Abdulkarim K, Ridell B, Johansson P, Kutti J, Safai-Kutti S, Andréasson B. The impact of peripheral blood values and bone marrow findings on prognosis for patients with essential thrombocythemia and polycythemia vera. Eur J Haematol. 2011;86(2): 148-155. Buesche G, Ganser A, Schlegelberger B, et al. Marrow fibrosis and its relevance during imatinib treatment of chronic myeloid leukemia. Leukemia. 2007;21(12):24202427. Kvasnicka HM, Thiele J, Schmitt-Graeff A, et al. Bone marrow features improve prognostic efficiency in multivariate risk classification of chronic-phase Ph(1+) chronic myelogenous leukemia: a multicenter trial. J Clin Oncol. 2001;19(12):2994-3009. Kantarjian HM, Bueso-Ramos CE, Talpaz M, et al. The degree of bone marrow fibrosis in chronic myelogenous leukemia is not a prognostic factor with imatinib mesylate therapy. Leuk Lymphoma. 2005;46(7):993997. Kimura A, Katoh O, Hyodo H, Kuramoto A, Satow Y. Platelet derived growth factor expression, myelofibrosis and chronic myelogenous leukemia. Leuk Lymphoma. 1995;18(3-4):237-242. Fu B, Ok CY, Goswami M, et al. The clinical importance of moderate/severe bone marrow fibrosis in patients with therapy-related myelodysplastic syndromes. Ann Hematol. 2013;92(10):1335-1343. Fu B, Jaso JM, Sargent RL, et al. Bone marrow fibrosis in patients with primary myelodysplastic syndromes has prognostic value using current therapies and new risk stratification systems. Mod Pathol. 2014;27(5):681-689. Della Porta MG, Malcovati L, Boveri E, et al. Clinical relevance of bone marrow fibrosis and CD34-positive cell clusters in primary myelodysplastic syndromes. J Clin Oncol. 2009;27(5):754-762. Kröger N, Zabelina T, van Biezen A, et al. Allogeneic stem cell transplantation for myelodysplastic syndromes with bone marrow fibrosis. Haematologica. 2011;96(2): 291-297. Sanna A, Gozzini A, Donnini I, et al. Influence of mild bone marrow fibrosis on response of int-2/high risk MDS patients to 5-azacitidine. Leuk Res. 2011;35:S126-S127. Tadmor T, Shvidel L, Aviv A, et al. Significance of bone marrow reticulin fibrosis in chronic lymphocytic leukemia at diagnosis: a study of 176 patients with prognostic implications. Cancer. 2013;119(10):18531859. Teman CJ, Wilson AR, Perkins SL, Hickman K, Prchal JT, Salama ME. Quantification of fibrosis and osteosclerosis in myeloproliferative neoplasms: a computer-assisted image study. Leuk Res. 2010;34(7):871-876. Pozdnyakova O, Hasserjian R, Salama M, et al. Bone marrow fibrosis by WHO Grade and quantitative image analysis is reduced by PRM-151 in patients with myelofibrosis and associated with improved bone marrow morphology and increased platelet counts. Haematologica. 2015;100(s1):677. Salama ME, Hagendorn E, Perkins SL, et al. Stereology and computer-based image analysis quantifies heterogeneity and improves reproducibility for grading reticulin In: Bone Marrow. Histopathology in Pharmaceutical Product Development. 1st edition. Eberhard DA, Potts SJ, Young GD,

eds. Springer Publishing, NY, USA. 2014. 66. Wang JC, Lang HD, Lichter S, Weinstein M, Benn P. Cytogenetic studies of bone marrow fibroblasts cultured from patients with myelofibrosis and myeloid metaplasia. Br J Haematol. 1992;80(2):184-188. 67. Mascarenhas J. Rationale for combination therapy in myelofibrosis. Best Pract Res Clin Haematol. 2014;27(2):197-208. 68. Rondelli D, Barosi G, Bacigalupo A, et al. Allogeneic hematopoietic stem-cell transplantation with reduced-intensity conditioning in intermediate- or high-risk patients with myelofibrosis with myeloid metaplasia. Blood. 2005;105(10):4115-4119. 69. Silver RT. Long-term effects of the treatment of polycythemia vera with recombinant interferon-alpha. Cancer. 2006;107(3):451458. 70. Silver RT, Vandris K, Goldman JJ. Recombinant interferon-α may retard progression of early primary myelofibrosis: a preliminary report. Blood. 2011;117(24): 6669-6672. 71. Ianotto JC, Boyer-Perrard F, Gyan E, et al. Efficacy and safety of pegylated-interferon α-2a in myelofibrosis: a study by the FIM and GEM French cooperative groups. Br J Haematol. 2013;162(6):783-791. 72. Hebballi S, Akiki S, Bareford D. Resolution of post-polycythaemic myelofibrosis with a combination of thalidomide and interferon. J Clin Pathol. 2012;65(8):762-763. 73. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med. 2012;366(9):799-807. 74. Harrison C, Kiladjian JJ, Al-Ali HK, et al. JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N Engl J Med. Mar 1 2012;366(9):787-798. 75. Wilkins BS, Radia D, Woodley C, Farhi SE, Keohane C, Harrison CN. Resolution of bone marrow fibrosis in a patient receiving JAK1/JAK2 inhibitor treatment with ruxolitinib. Haematologica. 2013;98(12):18721876. 76. Molica M, Serrao A, Saracino R, et al. Disappearance of fibrosis in secondary myelofibrosis after ruxolitinib treatment: new endpoint to achieve? Ann Hematol. 2014;93(11):1951-1952. 77. Kvasnicka HM, Thiele J, Bueso-Ramos C, et al. Exploratory analysis of the effect of ruxoli, tinib on bone marrow morphology in patients with myelofibrosis. J Clin Oncol. 2013;31(Suppl): [abstract 703]. 78. Thiele J, Bueso-Ramos CE, Sun W, et al. Effects of five-years of ruxolitinib therapy on bone marrow morphology in patients with myelofibrosis and comparison with best available therapy. Blood. 2013;122 (21):4055 [abstract]. 79. Mesa RA, Tefferi A, Elliott MA, et al. A phase II trial of pirfenidone (5-methyl-1phenyl-2-[1H]-pyridone), a novel anti-fibrosing agent, in myelofibrosis with myeloid metaplasia. Br J Haematol. 2001;114(1):111113. 80. Morris JC, Shapiro GI, Tan AR, et al. Phase I/II study of GC1008: a human anti-transforming growth factor-beta (TGFβ) monoclonal antibody (MAb) in patients with advanced malignant melanoma (MM) or renal cell carcinoma (RCC). J Clin Oncol. 2008;26(15S):9028. 81. Pilling D, Cox N, Vakil V, Verbeek JS, Gomer RH. The long pentraxin PTX3 promotes fibrocyte differentiation. PLoS One. 2015;10(3):e0119709. 82. Crawford JR, Pilling D, Gomer RH. FcγRI mediates serum amyloid P inhibition of

haematologica | 2016; 101(6)

Bone marrow fibrosis in myelofibrosis

83.

84.

85.

86. 87.

88.

89.

90.

fibrocyte differentiation. J Leukoc Biol. 2012;92(4):699-711. Haudek SB, Xia Y, Huebener P, et al. Bone marrow-derived fibroblast precursors mediate ischemic cardiomyopathy in mice. Proc Natl Acad Sci USA. 2006;103(48):1828418289. Pilling D, Roife D, Wang M, et al. Reduction of bleomycin-induced pulmonary fibrosis by serum amyloid P. J Immunol. 2007;179 (6):4035-4044. Casta単o AP, Lin SL, Surowy T, et al. Serum amyloid P inhibits fibrosis through Fc gamma R-dependent monocytemacrophage regulation in vivo. Sci Transl Med. 2009;1(5):5ra13. Maharjan AS, Roife D, Brazill D, Gomer RH. Serum amyloid P inhibits granulocyte adhesion. Fibrogenesis Tissue Repair. 2013;6(1):2. Murray LA, Rosada R, Moreira AP, et al. Serum amyloid P therapeutically attenuates murine bleomycin-induced pulmonary fibrosis via its effects on macrophages. PLoS One. 2010;5(3):e9683. Dillingh MR, van den Blink B, Moerland M, et al. Recombinant human serum amyloid P in healthy volunteers and patients with pulmonary fibrosis. Pulm Pharmacol Ther. 2013;26(6):672-676. Duffield JS, Lupher MLJ. PRM-151 (recombinant human serum amyloid P/pentraxin 2) for the treatment of fibrosis. Drug News Perspect. 2010;23(5):305-315. Verstovsek S, Mesa RA, Foltz LM, et al.

haematologica | 2016; 101(6)

91.

92.

93.

94.

95.

96.

Phase 2 trial of PRM-151, an antifibrotic agent, in patients with myelofibrosis: Stage 1 results. J Clin Oncol. 2014;32:5s (abstr 7114). Verstovsek S, Pozdnyakova O, Hasserjian R, et al. Outcomes in myelofibrosis patients completing 24 and 36 weeks of treatment with PRM-151. Clin Lymphoma Myeloma Leuk. 2015;15:S58-S59. Verstovsek S, Mesa RA, Foltz LM, et al. PRM-151 in myelofibrosis: durable efficacy and safety at 72 weeks. ASH. 2015;57th Annual Meeting and Exposition. Tadmor T, Bejar J, Attias D, et al. The expression of lysyl-oxidase gene family members in myeloproliferative neoplasms. Am J Hematol. 2013;88(5):355-358. Verstovsek S, Savona MR, Mesa RA, et al. A phase 2 study to evaluate the efficacy and safety of simtuzumab in adult subjects with primary, post polycythemia vera (PV) or post essential thrombocythemia (ET) myelofibrosis. ClinicalTrials.gov. 2015; NCT01369498. Bhagwat N, Keller MD, Rampal R, et al. Improved efficacy of combination of JAK2 and hedgehog inhibitors in myelofibrosis. Blood. 2013;122:s666. Martinelli G, Oehler VG, Papayannidis C, et al. Treatment with PF-04449913, an oral smoothened antagonist, in patients with myeloid malignancies: a phase 1 safety and pharmacokinetics study. Lancet Haematol. 2015;2(8):e339-e346.

97. Tibes R, Mesa RA. Targeting hedgehog signaling in myelofibrosis and other hematologic malignancies. J Hematol Oncol. 2014;7:18. 98. Gupta V, Koschmieder S, Harrison CN, et al. Phase 1b dose-escalation study of sonidegib (LDE225) in combination with ruxolitinib (INC424) in patients with myelofibrosis. Blood. 2014;124(21):712. 99. Gupta V, Harrison CN, Hasselbalch H, et al. Myeloproliferative ssyndromes: clinical: combination therapy in MPN. ASH. 2015;57th Annual Meeting and Exposition:634. 100. Lancet J, Wei AH, Durrant ST, et al. A phase I study of KB004, a novel non-fucosylated humaneered速 antibody, targeted against the receptor tyrosine kinase EphA3, in advanced hematologic malignancies. ASH. 2013:3838 [abstract]. 101. Vail ME, Murone C, Tan A, et al. Targeting EphA3 inhibits cancer growth by disrupting the tumor stromal microenvironment. Cancer Res. 2014;74(16):4470-4481. 102. Slape CI. EphA3 is expressed on leukemia stem cells, and eph/ephrin signalling features in the remodelling of the leukemia stem cell niche. Blood. 2014;124(21):3756. 103. Swords RT, Wei AH, Durrant S, et al. KB004, a novel non-fucosylated humaneered速 antibody, targeting EphA3, is active and well tolerated in a phase I/II study of advanced hematologic malignancies. Blood. 2014;124 (21):3756.

671

REVIEW ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2016 Volume 101(6):672-679

Wilms tumor 1 mutations in the pathogenesis of acute myeloid leukemia Raajit Rampal1 and Maria E. Figueroa2 1 2

Leukemia Service, Memorial Sloan Kettering Cancer Center, New York, NY; and Department of Pathology, University of Michigan Medical School, Ann Arbor, MI, USA

ABSTRACT

ilms tumor 1 (WT1) has long been implicated in acute myeloid leukemia. It has been described to be both overexpressed and mutated in different forms of acute myeloid leukemia, and overexpression has been reported to play a prognostic role in this disease. However, the precise mechanism through which WT1 may play a role in leukemogenesis has remained elusive. In recent years, new evidence has emerged that points towards a novel role of WT1 mutations in the deregulation of epigenetic programs in leukemic cells through its interaction with TET proteins. Herein we review the current status of the field and its therapeutic and prognostic implications in acute myeloid leukemia. Introduction

Correspondence: marfigue@med.umich.edu or rampalr@mskcc.org

Received: December 31, 2015. Accepted: April 5, 2016. Pre-published: no prepublication. doi:10.3324/haematol.2015.141796

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/672

ÂŠ2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is required for any other use.

672

The observations that the Wilms tumor 1 (WT1) protein is expressed in the majority of cases of acute myeloid leukemia (AML), and in addition, is mutated in a proportion of AML cases has led to extensive work over the last several decades to determine the mechanistic role of WT1 in AML. The fact that WT1 may be either overexpressed or mutated has given rise to the concept that it may act as both a tumor suppressor and oncogene, depending on the context. To date, much of the mechanistic work on WT1 has centered on its function as a transcription factor and the role of its many isoforms. However, more recent work stemming from largescale genomic studies in AML has revealed a new role for WT1 in epigenetic regulation. Indeed, work by several groups has demonstrated that WT1 appears to play a role along with TET proteins in mediating 5-hydroxymethylation of cytosines. These novel observations have given rise to an enhanced understanding of the complexity of this protein and its pathogenic role in leukemogenesis. Herein, we seek to review the major concepts and findings with regard to WT1 in oncogenesis and normal hematopoiesis, as well as any therapeutic strategies that may arise from these observations.

WT structure and function The WT1 gene was initially identified as an inherited predisposition allele in probands with familial Wilms tumor.1,2 Numerous WT1 missense and nonsense mutations have been described in inherited and sporadic Wilms tumors, a pediatric malignancy affecting the kidneys.3,4 WT1 mutations are identified in approximately 20% of Wilms tumors, and are seen in patients with a familial disposition to these tumors. Wilms tumor is also associated with the Denys-Drash syndrome (DDS), encompassing Wilms tumor, congenital nephrotic syndrome, and XY pseudohermaphroditism, which results from point mutations in WT1.5,6 Mutations in the intron 9 splice site have been identified in patients with Frasier syndrome, a developmental condition that affects kidneys and genitalia.7 In addition, WT1 has been implicated in a number of other malignancies, including desmoplastic small cell tumor, breast cancer, retinoblastoma, and lung carcinoma.8,9 In lung cancer, Wt1 has been shown to play a role in murine lung cancer models driven by Kras mutations where deletion or suppression of Wt1 resulted in senescence of primary murine cells expressing Kras, but had no effect on wild type cells.10 Furthermore, loss of Wt1 led to decreased proliferation and tumor burden in Kras-driven lung carcinoma, haematologica | 2016; 101(6)

WT1 mutations in AML

thus implicating Wt1 as a necessary component of Krasdriven oncogenesis. Taken together, these observations indicate that a spectrum of mutations in WT1 can lead to both inherited developmental syndromes and cancer predisposition. The WT1 protein contains an N-terminal transactivation domain and a C-terminus with four zinc-fingers that share DNA binding motifs with EGR11 (Figure 1). Four major WT1 isoforms result from differences in alternative splicing. These isoforms arise through two major splice events. Isoforms resulting through alternative splicing of exon 5, which causes a 17 amino acid insertion between the transregulatory domain and the zinc finger domain, have been found more commonly in relapsed AML samples11 (we will refer to the isoform containing this 17 amino acid region as 17+). The second splice event inserts three amino acids –lysine (K), threonine (T), and serine (S)– between exons 9 and 10, resulting in a significant reduction of the DNA binding ability, while enriching RNA binding12,13 (we will refer to the isoform containing the KTS sequence as KTS+). A fifth WT1 isoform produced by an alternative initiation codon within exon 1 lacks the extreme N-terminal repressive domain and augments the protein’s transactivation function.14 WT1 has been demonstrated to interact with a variety of other proteins. Among these are p53, which WT1 stabilizes and plays a role in preventing apoptosis.15 WT1 also binds to chaperone heat shock protein 90, resulting in WT1 stabilization16 and STAT3, resulting in enhanced cell proliferation of Wilms tumor cells.17 Most recently, WT1 has been demonstrated to interact with the epigenetic modifiers TET2 and TET3 in acute myeloid leukemias. This phenomenon will be discussed in detail later on in this review.18,19

WT1 in normal hematopoiesis The initial assessment of WT1 expression in normal hematopoiesis was based on observations in normal donors demonstrating that WT1 was expressed in CD34+ bone marrow-derived cells, but not in those lacking CD34 expression. Further analysis demonstrated that WT1 expression occurred in the CD34+ CD38- population.20 Notably, single-cell RT-PCR studies revealed that only a small proportion (1.2%) of bone marrow CD34+ cells actually expressed WT1.21 Further studies demonstrated expression of WT1 in some quiescent primitive progenitors as well as more differentiated populations, while it was undetectable in lineage-committed progenitors.22 In murine hematopoiesis, Wt1 expression has been noted in embryonic murine liver at a time point when the liver is the principle site of hematopoiesis (day 12.5 post conception).23 Attempts to further define the role of WT1 in normal hematopoiesis have been performed by observing the effects of alterations of WT1 levels in hematopoietic cells. Overexpression of WT1 in human umbilical cord derived CD34+ cells –in particular the isoform lacking a KTS sequence– resulted in enhanced differentiation of CD34+ cells in culture conditions, whereas overexpression in CD34+ CD38- cells resulted in a relative increase in this population attributed to quiescence.22 In methylcellulose assays, forced overexpression of WT1 in umbilical cord derived CD34+ cells resulted in reduced myeloid and erythroid colony formation. This observation did not correlate with any differences in cell cycle or viability.24 Finally, haematologica | 2016; 101(6)

overexpression of Wt1 in 32D cl3 cells, an IL-3-dependant myeloid progenitor murine cell line, blocked G-CSF induced differentiation of these cells in culture.25 Several murine studies have been carried out to better characterize the role of Wt1 in normal hematopoiesis in vivo. The complete absence of Wt1 results in embryonic lethality due to failure of renal and gonad development.26 Thus, several alternative approaches, including generating chimeric mice through the injection of normal C57BL6 blastocysts with embryonic stem (ES) cells lacking Wt1, as well as reconstitution of normal mice with hematopoietic cells from embryos lacking Wt1, have been implemented to overcome this limitation. Using these approaches, the investigators demonstrated that while embryonic stem cells lacking Wt1 show a competitive disadvantage in the chimeric context, failing to contribute to mature hematopoiesis, fetal liver cells lacking Wt1 were able to repopulate the hematopoietic system of lethally irradiated recipients, at least for 2 months. In addition, despite Wt1progenitors displaying multipotent potential on methylcellulose, their colony-forming ability was reduced compared to Wt1+ progenitors.27,28 In a Wt1-GFP transgenic murine model, Wt1 was not found to be expressed in long term-hematopoietic stem cells, and infrequently in multipotent progenitors.29 Deletion of Wt1 in young and adult mice using an inducible system results in the death of the animals within approximately 10 days and results in glomerulosclerosis, atrophy of the pancreas, and diminished extramedullary hematopoiesis.30 Notably, the phenotype of conditional deletion of Wt1 in the adult hematopoietic compartment has not been reported to date.

WT1 in acute myeloid leukemia (AML) WT1 is overexpressed in both myeloid and lymphoid leukemias. Overexpression has been demonstrated in acute lymphoblastic leukemia (both B cell and T cell) primary patient samples and cell lines.31-33 Likewise, overexpression has been described in both myeloid and lymphoid blast crisis of chronic myeloid leukemia (CML) (but not in chronic phase CML),31 and myelodysplastic syndrome (MDS).34 However, despite this broad range of hematological malignancies harboring overexpression of WT1, this phenomenon has been most extensively studied in acute myeloid leukemia (AML). Indeed, WT1 expression levels have been demonstrated to be higher in both primary human AML samples as well as human leukemia cell lines.32,33,35 Furthermore, recurrent somatic mutations in WT1 have also been described in AML.36,37 The initial observations of WT1 involvement in leukemia have given rise to work examining the potential pathogenic role of WT1 in leukemogenesis. Below, we focus specifically on the clinical implications, pathogenesis and therapeutic strategies with regards to WT1 in AML. WT1 is overexpressed in the majority of AML patients.35,38 In MDS the expression of WT1 is associated with higher blast counts and an increased risk of progression to AML.34 Several studies have demonstrated that increased levels of WT1 in AML are associated with resistance to therapy, a higher incidence of relapse, and poor overall survival.39,40,41 Furthermore, a failure to reduce WT1 transcript levels to below detectable limits has been associated with a higher incidence of relapse in AML42 (Table 1). Significant strides in understanding the role of WT1 673

R. Rampal and M.E. Figueroa

overexpression in AML have been made by studies in human leukemia cell lines and murine models. Notably, downregulation of WT1 expression in the K562 cell line, as well as a proportion of tested primary AML and CML cases (chronic phase and blast crisis) resulted in inhibition of cell growth.43 Further studies have demonstrated that downregulation of the 17+ isoform specifically, but not the 17- isoform, was able to induce apoptosis in leukemia cell lines K562, HL-60, and Kasumi-1. Conversely, the expression of the 17+ isoform led to a decrease in the expression of proapoptotic Bak.44 In order to further define the role of WT1 overexpression in leukemogenesis, a transgenic murine model of WT1 overexpression in the hematopoietic system was established by Nishida and colleagues.45 Bone marrow from transgenic mice overexpressing WT1 was transduced with a construct expressing the AML1-ETO fusion product, which by itself is insufficient to produce AML. By contrast, in cooperation with WT1 overexpression AML1ETO led to rapid onset leukemia, with a median time to leukemia development of 50 days. To further characterize WT1’s contribution to leukemogenesis, Hosen and colleagues generated a knock-in reporter WT1-Green Fluorescent Protein (WT1GFP/+) murine model that allowed them to track WT1 expression in normal and leukemic bone marrow cells.29 Using this model they characterized the expression of WT1 in an AML1-ETO+TEL-PDGRFB leukemia model. Recipient mice transplanted with bone marrow transduced with AML1-ETO+TEL-PDGRFB developed a rapid leukemia and displayed higher expression of WT1 in the lineage negative, Sca+,cKit+ (LSK) fraction compared to normal WT1GFP/+ mice. The same observation was made in WT1GFP/+ mice transduced with BCRABL. WT1 was found to be expressed in a subset of LSK cells capable of transplanting the disease phenotype into recipient animals; however, a fraction of GFP negative LSK cells was also able to transplant the leukemia phenotype, indicating that WT1 was not required for the propagation of leukemia. Also of note, enforced expression of WT1 in hematopoietic stem cells did not result in proliferation or expansion of these cells or their downstream progeny, nor did this overexpression result in a differentiation block in these cells. Thus, the precise role of WT1 overexpression in leukemogenesis remains elusive. The disparate results in these models may be due to cooperating oncogenes or other factors. Further evaluation in models using other cooperating genomic alterations, based on the current understanding of the spectrum and co-occurrence of mutational events in human AML may be necessary, as it is possible that the pathogenic role of WT1 is context-specific.

WT1 mutations in AML Recurrent somatic mutations in WT1 appear to occur in approximately 6-15% of de novo AML.46,47 WT1 mutations associate with younger age,47-49 and the presence of FLT3ITD mutations48,49 and CEBPA mutations.49 The clinical impact of WT1 in AML has been assessed in several studies, with conflicting results. Multivariate analysis of 470 de novo AML patients (excluding Acute Promyelocytic Leukemia) demonstrated that WT1 mutations were associated with worsened overall survival (OS) and relapsefree survival (RFS) both in the entire cohort as well as when the analysis was restricted to normal karyotype AML patients.47 These results were supported by other 674

Table 1. Clinical associations with WT1 overexpression and mutation in AML and MDS.

WT1 Overexpression Clinical finding Higher blast count Progression to AML from MDS Reduced overall survival Increased incidence of relapse Higher relapse incidence if failure to reduce WT1 transcript levels below detection Associated with FLT3 mutations

Ref. 39 39 40,41 40 65 40

WT1 mutation Chemotherapy resistance Associated with FLT3-ITD and CEBPA mutations Associated with younger age Impaired overall survival and relapse-free survival*

51 49,50 48-50 48, 49, 51,52

*No association with relapse-free or overall survival found in one study.54

studies of adult AML patients.48,50,51 By contrast, a study including patients from three German-Austrian AML study protocols demonstrated no association with RFS or OS.49 A possible explanation for this difference in findings may reside in the fact that patients in the GermanAustrian trials received higher doses of cytarabine during consolidation (18-54 g/m2 vs. 6-25 g/m2) than those in the other studies. In addition to the impact of WT1 mutation on OS and RFS, the presence of WT1 mutations has also been associated with resistance to induction chemotherapy50 (Table 1). The functional effects of WT1 mutations have been assessed in various studies. Several different WT1 mutations have been described in AML, which occur primarily in exons 1, 7 and 9 (Figure 1). These include base substitutions, deletions, and insertions. Significantly, however, the vast majority of mutations result in the creation of stop codons and reading frame shifts, resulting in loss-of-function and expression of a truncated protein lacking the zincfinger domain.52,37 Additionally, more recent data indicate that WT1 mutant transcripts with frameshift mutations are subject to nonsense-mediated RNA decay without expression of the truncated protein.53,35 Thus, these WT1 mutations may result in the loss of DNA binding ability due to loss of the zinc-finger domain, or result in loss of expression of the WT1 protein altogether. The fact that WT1 is overexpressed in many patients with AML, and that overexpression of WT1 can contribute to driving leukemogenesis in a murine model, yet it is also mutated –with presumed loss-of-function– in a significant proportion of AML cases presented a paradox as to the role of this protein in AML. Microarray profiling has identified a gene expression signature characterizing WT1-mutant disease, including overexpression of GTSF1, CD96, MLL, PML, and MACC1, and down-regulation of SNRPN, SNURF, FCHO2, MTX3, INSR, and IRS2. In total, the WT1-mutant gene expression signature included 74 upregulated and 40 downregulated genes. When analyzed from a functional point of view, this signature included genes involved in gene regulation, cell proliferation and metabolic homeostasis.54 Beyond this, however; little was known regarding the mechanism by which WT1 mutations might contribute to leukemogenesis. Recent largehaematologica | 2016; 101(6)

WT1 mutations in AML

scale genetic studies in AML have begun to shed light on this problem. Mutational analysis of a large cohort of AML cases who were treated on the ECOG 1900 clinical trial55,56 revealed that TET2 and IDH1/2 mutations are mutually exclusive in AML, and that TET2/IDH mutant AML is characterized by a common DNA hypermethylation phenotype.57 TET2 is a member of the TET protein family, a group of three iron(II)/αKG-dependent dioxygenases that catalyze the conversion of 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5-hmC).58,59 AMLs carrying mutations in IDH1 or IDH2 mutant produce the oncometabolite 2-hydroxyglutarate, which inhibits TET2 activity.60 Thus, both TET2 loss-of-function mutations and IDH1/2 mutations result in inhibition of the DNA demethylation pathways with accumulation of 5-mC and decrease of 5-hmC, thus providing a convergent mechanism for TET2 and IDH1/2 mutations in AML57 (Figure 2).

Interestingly, analysis of the ECOG 1900 cohort also demonstrated that WT1 mutations are strongly anticorrelated with TET2 and IDH1/2 mutations – a finding confirmed in the TCGA AML dataset–, leading to the hypothesis that WT1 mutations may also correspond to a subgroup of AMLs characterized by low levels of 5-hmC, and further, that WT1 may have a previously unrecognized role in this epigenetic pathway. Examination of primary AML samples from the ECOG 1900 cohort using promoter DNA methylation microarrays demonstrated a hypermethylation signature when compared to AMLs that were wild type for all known epigenetic modifier mutations. This hypermethylation signature very strongly overlapped with the TET2 hypermethylation signature. In turn, the TET2 and WT1 hypermethylated loci showed a near complete overlap with the IDH1/2 hypermethylation signature, demonstrating a site-specific overlap of the epigenet-

Table 2. Overview of Select Trials of WT1 vaccines in AML and MDS patients.

Vaccine WT1 peptide WT1 peptide WT1 peptide WT1 and PR1 peptides WT1 peptide

MDS/AML patient population

Notable outcomes

Ref.

AML in CR High-risk MDS with at least 1 line of therapy AML in hematologic CR MDS

CR> 1 year (AML), transfusion dependence (MDS) Decrease in blast count in MDS and AML proportion of patients. Undetectable WT1 levels in 3 patients treated for > 8 years WT1 –specific T-cell responses noted; median DFS not reached Decrease in WT1 mRNA expression

Brayer, 201574

AML in CR with evidence of WT1 transcripts AML in CR MDS (RA, RARS) AML, high-risk MDS, with relapse, incomplete response, or not a candidate for induction therapy

Stable disease, blast count reduction, 1 CR. Decrease in WT1 mRNA levels

Tsuboi, 201275 Maslak, 201076 Rezvani, 200877 Keilholz, 200978

CR: complete remission; DFS: disease-free survival; RA: refractory anemia; RARS: refractory anemia with ringed sideroblasts.

Figure 1. Schematic representation of the WT1 protein. Top: arrows indicate mutations in both AML (red) and those associated with Wilms tumor (green). Bottom: schematic representation of WT1 interactions with other proteins. haematologica | 2016; 101(6)

675

R. Rampal and M.E. Figueroa

Figure 2. The effects of mutations in mediators of the 5-hmC pathway. (A) pathway under normal conditions; (B) IDH mutations lead to production of 2-hydroxyglutarate which inhibits TET2 function; (C) TET2 mutations disrupt interaction with WT1; (D) WT1 mutations fail to properly direct TET2 to its target sites, either by disruption of the interaction itself or by failing to bind to DNA.

ic abnormalities in all three types of AMLs. Furthermore, analysis of both global 5-hmC levels in primary AML patient samples as well as genome-wide distribution of 5hmC by next-generation sequencing demonstrated that WT1-mutant AMLs presented a global reduction in 5hmC levels comparable to that seen in TET2 and IDH1/2 mutant AMLs when compared with AMLs wild type for WT1/TET2/IDH1/2.18 The observations that WT1-mutant AMLs are characterized by similar epigenetic alterations as found in TET2 and IDH1/2 mutant AMLs raised the question as to whether WT1 plays a direct role in DNA hydroxymethylation. Knockdown of Wt1 in murine mesonephron cells (which overexpress Wt1) resulted in a significant reduction in 5hmC levels, while overexpression of WT1 in 32D cells (which do not express endogenous Wt1) led to increases in 5-hmC, thus demonstrating a dynamic response of 5-hmC 676

to WT1 levels. In vitro characterization of the effects of Wt1 downregulation in murine bone marrow cells demonstrated that such downregulation led to increased ckit expression and an increase in the LSK stem/progenitor population, similar to prior observations with loss of Tet2 in vivo and in vitro.61-63 These observations raised the question as to the mechanism by which WT1 contributes to alteration in 5-hmC levels. Analysis of primary patient samples and hematopoietic cell lines demonstrated that the presence or absence of WT1 mutations had no bearing on TET1/TET2/TET3 expression levels, or on 2-hydroxyglutarate levels. WT1 was subsequently found to bind to TET2 in co-immunoprecipitation assays in transfected cell lines, as well as in AML cell lines. Furthermore, WT1 overexpression was able to attenuate the biochemical and immunophenotypic alterations observed in Tet2 knockout murine hematopoietic cells. However, the ability of WT1 haematologica | 2016; 101(6)

WT1 mutations in AML

to alter 5-hmC levels was noted to be attenuated by exposure to a cell-permeable form of 2-HG, which argued that WT1 likely mediates its effect via TET enzymes, and that in the absence of Tet2, WT1 might exert its effects through other TET enzymes. Subsequent experiments demonstrated that WT1 also binds to TET3, but not TET1.18 Work by Wang et al.19 confirmed and extended these findings. These authors likewise demonstrated that WT1 mutations are exclusive of TET2 and IDH1/2 mutations in AML using a meta-analysis of six studies, and also demonstrated that WT1 indeed binds to TET2. Moreover, overexpression of a catalytically active form of TET2 in cell lines resulted in the activation of known WT1-target genes; this activation required the presence of WT1. Further analysis using chromatin immunoprecipitation and PCR demonstrated that TET2 binds to transcription start sites and CpG islands of WT1 target genes, and that overexpression of both WT1 and TET2 led to increases in 5-hmC levels near the transcription start site of WT1-target genes. Finally these authors demonstrated that a proportion of TET2 mutations commonly found in AML result in the loss of the ability of WT1 to bind to TET2 (Figure 2). Collectively, these studies have identified a previously unrecognized role for WT1 as an epigenetic modifier and place WT1 mutations in the mutational class defined by alterations in TET2 function in AML. Much work remains to fully describe the mechanism of leukemogenesis in WT1 mutant AMLs. For example, the precise subsets of genes that are regulated by the TET2/WT1 interaction and which become deregulated by a mutation in either remain an area of investigation. Toward this end, work by Sinha et al.64 also identified an association between WT1 mutations and DNA CpG hypermethylation using an algorithm based on Boolean implications on the TCGA AML patient data set. In particular, hypermethylation of polycomb repressive complex 2 (PRC2) target genes marked by H3K27 trimethylation, in both embryonic stem cells and the leukemia cell lines was observed. Gene expression analysis further revealed overexpression of EZH2, the catalytic subunit of the PRC2 complex, in WT1-mutant AMLs compared to normal karyotype AMLs. Taken together these two findings imply that silencing of polycomb target genes may be a necessary event in WT1mediated leukemogenesis. Further delineation of the pathways involved in WT1 mutant mediated leukemogenesis is under way. In addition, no murine models of leukemogenesis that are driven by Wt1 mutations have yet been published. Such studies will likely give rise to a more precise understanding of the mechanism through which WT1 mutations mediate leukemogenesis.

Therapeutic implications The observation that failure to reduce high levels of WT1 transcript after therapy was associated with a higher incidence of relapse in AML42 spurred efforts to utilize WT1 as a biomarker of minimal residual disease (MRD) in AML. Earlier efforts to use WT1 were met with mixed results as some investigators utilized primers targeting the 3â&#x20AC;&#x2122; region, which is where many WT1 mutations cluster.65 To address these discrepancies, Cilloni et al. performed an evaluation of nine published and other real-time quantitative polymerase chain reaction assays for WT1 quantification.66 The authors noted marked variation between haematologica | 2016; 101(6)

assays. Using the best-performing assay to assess WT1 levels in diagnostic and follow-up samples they demonstrated that the kinetics of WT1 transcript reduction following induction was prognostically significant in terms of prediction of relapse.67 Patients in whom WT1 expression is not normalized following consolidation were at particular risk of relapse. By contrast, pretreatment WT1 levels were not found to be predictive of outcome. Of additional note, WT1 levels were not significantly modulated upon marrow regeneration following chemotherapy. Moreover, correcting for background WT1 expression in control peripheral blood and bone marrow, the degree of WT1 expression was not sufficiently robust to be used as a marker for sequential MRD analysis. Thus, the optimal utilization of WT1 in disease response assessment remains to be determined. The involvement of both WT1 overexpression and WT1 mutations in AML has given rise to both clinical and preclinical therapeutic strategies. Perhaps the most prominent strategy in this regard has been the attempt to use peptide vaccines against WT1, given its overexpression in AML. Numerous studies have been carried out to date using a variety of different vaccination strategies (HLA-restricted versus non HLA-restricted peptides, for example). Clinically meaningful responses have been reported in several trials in both AML and MDS patients, with associated increases in WT1-specific T-cell frequencies (Table 2).68 Given the fact that WT1 is normally expressed in several tissues and plays a role in hematopoiesis, concern had been raised about the potential for WT1-vaccine therapies to elicit autoimmune reactions. However, this has not been reported thus far.69 This approach has thus demonstrated clinical activity, but still requires further large-scale evaluation. Another approach has been the development of a monoclonal antibody that recognizes a peptide fragment of WT1 complexed with HLA-A0201. This antibody demonstrated efficacy in a NOD/SCID mouse xenotransplanted with human leukemias.70 Given the epigenetic alterations catalogued in WT1 mutant (as well as TET2 and IDH mutant AMLs), epigenetic-targeted therapy has been explored as a potential mechanism to deal with this subgroup of leukemias. Indeed retrospective data has demonstrated an increased response rate with the hypomethylating agent azacitidine in MDS and AML patients with TET2 mutations compared with those without TET2 mutations. However, this increased response rate did not translate to improved overall survival of TET2-mutant cases.71,72 Finally, AML patients with IDH1/2 mutations were noted to have a higher response rate compared to those without mutations (71% vs. 23%, P=0.01) in a retrospective analysis of 42 patients receiving either decitabine or azacitidine. However, conclusions from this study are limited by the fact that it only included 42 patients, 7 of whom had mutations in IDH1/2, and larger studies have failed to detect this association.72,73 Thus, while there is evidence of a possible increased response to hypomethylating agents in patients bearing mutations in the DNA hydroxymethylation pathway, this still requires further exploration. Whether this holds true for WT1 mutations has not been established yet, and prospective evaluations of these findings are required. Finally, as previously discussed, WT1 mutations are associated with higher expression levels of EZH2 and appear to impact PRC2 target genes. In vitro genetic downregulation of the PRC2 member EZH2 in a 677

R. Rampal and M.E. Figueroa

WT1 mutant AML cell line, as well as pharmacologic inhibition of this enzyme in WT1 mutant cell lines or primary AML specimens resulted in increased differentiation of these cells.64 Thus, there are several treatment strategies, both clinical and preclinical, that warrant further testing both for WT1 overexpressing and WT1 mutant AML cases.

Conclusions WT1 appears to be an important factor in both normal development and oncogenesis. However, the precise role of this protein has been difficult to pinpoint, owing to its complexity, both in terms of protein isoforms as well as the fact that it may participate in oncogenesis by way of either overexpression or loss-of-function mutations. This is a particular challenge in AML, where both scenarios are

References 1. Call KM, Glaser T, Ito CY, et al. Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms' tumor locus. Cell. 1990;60(3):509-520. 2. Gessler M, Poustka A, Cavenee W, et al. Homozygous deletion in Wilms tumours of a zinc-finger gene identified by chromosome jumping. Nature. 1990;343(6260): 774-778. 3. Huff V. Wilms tumor genetics. Am J Med Genet. 1998;79(4):260-267. 4. Ruteshouser EC, Robinson SM, Huff V. Wilms tumor genetics: mutations in WT1, WTX, and CTNNB1 account for only about one-third of tumors. Genes Chromosomes Cancer. 2008;47(6):461-470. 5. Little M, Holmes G, Bickmore W, et al. DNA binding capacity of the WT1 protein is abolished by Denys-Drash syndrome WT1 point mutations. Hum Mol Genet. 1995;4(3):351-358. 6. Little MH, Williamson KA, Mannens M, et al. Evidence that WT1 mutations in DenysDrash syndrome patients may act in a dominant-negative fashion. Hum Mol genet. 1993;2(3):259-264. 7. Klamt B, Koziell A, Poulat F, et al. Frasier syndrome is caused by defective alternative splicing of WT1 leading to an altered ratio of WT1 +/-KTS splice isoforms. Hum Mol genet. 1998;7(4):709-714. 8. Yang L, Han Y, Saiz FS, Minden MD. A tumor suppressor and oncogene: the WT1 story (vol 21, pg 868, 2007). Leukemia. 2007;21(5):1603-1603. 9. Little M, Wells C. A clinical overview of WT1 gene mutations. Hum Mutat. 1997;9(3):209-225. 10. Vicent S, Chen R, Sayles LC, et al. Wilms tumor 1 (WT1) regulates KRAS-driven oncogenesis and senescence in mouse and human models. J Clin Invest. 2010;120(11): 3940-3952. 11. Gu W, Hu S, Chen Z, et al. High expression of WT1 gene in acute myeloid leukemias with more predominant WT1+17AA isoforms at relapse. Leuk Res. 2010;34(1):46-49. 12. Bor YC, Swartz J, Morrison A, et al. The Wilms' tumor 1 (WT1) gene (+KTS isoform) functions with a CTE to enhance

678

13.

14.

15.

16.

17.

18.

19.

20. 21.

22.

23.

found. Recent work on WT1 mutations in AML have shed new light on the function of this protein, as there is a clear role for WT1 in epigenetic modifications. Thus far, this has only been studied in the context of AML and in the context of WT1 mutations. It remains to be determined what epigenetic role WT1 plays in normal development and other WT1-associated cancers. Likewise, whether the overexpression of WT1 in AML does or doesn't lead to the perturbation of the epigenome on a genome-wide manner still remains to be explored. In addition, it is likely that the effects of mutations or overexpression are in part context dependent, and that cooperating genomic alterations modulate and are modulated by WT1â&#x20AC;&#x2122;s effects on gene expression. As such, new answers revealed by recent work have given rise to new questions. Importantly, this recent work has given vital clues and brought us one step closer to unraveling this enigma wrapped in a mystery.

translation from an unspliced RNA with a retained intron. Genes Dev. 2006;20(12): 1597-1608. Morrison AA, Venables JP, Dellaire G, Ladomery MR. The Wilms tumour suppressor protein WT1 (+KTS isoform) binds alpha-actinin 1 mRNA via its zinc-finger domain. Biochem Cell Biol. 2006;84(5):789798. Hossain A, Nixon M, Kuo MT, Saunders GF. N-terminally truncated WT1 protein with oncogenic properties overexpressed in leukemia. J Biol Chem. 2006;281(38): 28122-28130. Maheswaran S, Park S, Bernard A, et al. Physical and functional interaction between WT1 and p53 proteins. Proc Natl Acad Sci USA. 1993;90(11):5100-5104. Bansal H, Bansal S, Rao M, et al. Heat shock protein 90 regulates the expression of Wilms tumor 1 protein in myeloid leukemias. Blood. 2010;116(22):4591-4599. Rong Y, Cheng L, Ning H, et al. Wilms' tumor 1 and signal transducers and activators of transcription 3 synergistically promote cell proliferation: a possible mechanism in sporadic Wilms' tumor. Cancer Res. 2006;66(16):8049-8057. 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. Wang Y, Xiao M, Chen X, et al. WT1 recruits TET2 to regulate its target gene expression and suppress leukemia cell proliferation. Mol Cell 2015;57(4):662-673. Baird PN, Simmons PJ. Expression of the Wilms' tumor gene (WT1) in normal hemopoiesis. Exp Hematol. 1997;25(4):312-320. Hosen N, Sonoda Y, Oji Y, et al. Very low frequencies of human normal CD34+ haematopoietic progenitor cells express the Wilms' tumour gene WT1 at levels similar to those in leukaemia cells. Br J Naematol. 2002;116(2):409-420. Ellisen LW, Carlesso N, Cheng T, Scadden DT, Haber DA. The Wilms tumor suppressor WT1 directs stage-specific quiescence and differentiation of human hematopoietic progenitor cells. EMBO J. 2001;20(8): 1897-1909. Fraizer GC, Patmasiriwat P, Zhang X,

24.

25.

26. 27.

28. 29.

30.

31.

32. 33.

34.

35.

Saunders GF. Expression of the tumor suppressor gene WT1 in both human and mouse bone marrow. Blood. 1995;86(12): 4704-4706. Svedberg H, Richter J, Gullberg U. Forced expression of the Wilms tumor 1 (WT1) gene inhibits proliferation of human hematopoietic CD34(+) progenitor cells. Leukemia. 2001;15(12):1914-1922. Inoue K, Tamaki H, Ogawa H, et al. Wilms' tumor gene (WT1) competes with differentiation-inducing signal in hematopoietic progenitor cells. Blood. 1998;91(8):29692976. Kreidberg JA, Sariola H, Loring JM, et al. Wt-1 Is Required for Early Kidney Development. Cell. 1993;74(4):679-691. Alberta JA, Springett GM, Rayburn H, et al. Role of the WT1 tumor suppressor in murine hematopoiesis. Blood. 2003;101(7): 2570-2574. King-Underwood L, Little S, Baker M, et al. Wt1 is not essential for hematopoiesis in the mouse. Leuk Res. 2005;29(7):803-812. Hosen N, Shirakata T, Nishida S, et al. The Wilms' tumor gene WT1-GFP knock-in mouse reveals the dynamic regulation of WT1 expression in normal and leukemic hematopoiesis. Leukemia. 2007;21(8): 1783-1791. Chau YY, Brownstein D, Mjoseng H, et al. Acute Multiple Organ Failure in Adult Mice Deleted for the Developmental Regulator Wt1. Plos Genet. 2011;7(12):e1002404. Miyagi T, Ahuja H, Kubota T, et al. Expression of the candidate Wilm's tumor gene, WT1, in human leukemia cells. Leukemia. 1993;7(7):970-977. Miwa H, Beran M, Saunders GF. Expression of the Wilms' tumor gene (WT1) in human leukemias. Leukemia. 1992;6(5):405-409. Menssen HD, Renkl HJ, Rodeck U, et al. Presence of Wilms' tumor gene (wt1) transcripts and the WT1 nuclear protein in the majority of human acute leukemias. Leukemia. 1995;9(6):1060-1067. Tamaki H, Ogawa H, Ohyashiki K, et al. The Wilms' tumor gene WT1 is a good marker for diagnosis of disease progression of myelodysplastic syndromes. Leukemia. 1999;13(3):393-399. Miyagi T, Ahuja H, Kubota T, et al. Expression of the Candidate Wilms-Tumor

haematologica | 2016; 101(6)

WT1 mutations in AML

36.

37.

38. 39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

Gene, Wt1, in Human Leukemia-Cells. Leukemia. 1993;7(7):970-977. King-Underwood L, Pritchard-Jones K. Wilms' tumor (WT1) gene mutations occur mainly in acute myeloid leukemia and may confer drug resistance. Blood. 1998;91(8): 2961-2968. KingUnderwood L, Renshaw J, PritchardJones K. Mutations in the Wilms' tumor gene WT1 in leukemias. Blood. 1996;87(6):2171-2179. Miwa H, Beran M, Saunders GF. Expression of the Wilms-Tumor Gene (Wt1) in Human Leukemias. Leukemia. 1992;6(5):405-409. Barragan E, Cervera J, Bolufer P, et al. Prognostic implications of Wilms' tumor gene (WT1) expression in patients with de novo acute myeloid leukemia. Haematologica. 2004;89(8):926-933. Bergmann L, Miething C, Maurer U, et al. High levels of Wilms' tumor gene (wt1) mRNA in acute myeloid leukemias are associated with a worse long-term outcome. Blood. 1997;90(3):1217-1225. Karakas T, Miething CC, Maurer U, et al. The coexpression of the apoptosis-related genes bcl-2 and wt1 in predicting survival in adult acute myeloid leukemia. Leukemia. 2002;16(5):846-854. Cilloni D, Renneville A, Hermitte F, et al. Real-Time Quantitative Polymerase Chain Reaction Detection of Minimal Residual Disease by Standardized WT1 Assay to Enhance Risk Stratification in Acute Myeloid Leukemia: A European LeukemiaNet Study. J Clin Oncol. 2009;27(31):5195-5201. Yamagami T, Sugiyama H, Inoue K, et al. Growth inhibition of human leukemic cells by WT1 (Wilms tumor gene) antisense oligodeoxynucleotides: Implications for the involvement of WT1 in leukemogenesis. Blood. 1996;87(7):2878-2884. Ito K, Oji Y, Tatsumi N, et al. Antiapoptotic function of 17AA(+)WT1 (Wilms' tumor gene) isoforms on the intrinsic apoptosis pathway. Oncogene. 2006;25(30):42174229. Nishida S, Hosen N, Shirakata T, et al. AML1-ETO rapidly induces acute myeloblastic leukemia in cooperation with the Wilms tumor gene, WT1. Blood. 2006;107(8):3303-3312. King-Underwood L, Renshaw J, PritchardJones K. Mutations in the Wilms' tumor gene WT1 in leukemias. Blood. 1996;87(6):2171-2179. Hou HA, Huang TC, Lin LI, et al. WT1 mutation in 470 adult patients with acute myeloid leukemia: stability during disease evolution and implication of its incorporation into a survival scoring system. Blood. 2010;115(25):5222-5231. Renneville A, Boissel N, Zurawski V, et al. Wilms tumor 1 gene mutations are associated with a higher risk of recurrence in young adults with acute myeloid leukemia: a study from the Acute Leukemia French Association. Cancer. 2009;115(16):37193727. Gaidzik VI, Schlenk RF, Moschny S, et al. Prognostic impact of WT1 mutations in cytogenetically normal acute myeloid leukemia: a study of the German-Austrian AML Study Group. Blood. 2009; 113(19):4505-4511. Virappane P, Gale R, Hills R, et al. Mutation of the Wilms' tumor 1 gene is a poor prog-

haematologica | 2016; 101(6)

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

nostic factor associated with chemotherapy resistance in normal karyotype acute myeloid leukemia: the United Kingdom Medical Research Council Adult Leukaemia Working Party. J Clin Oncol. 2008;26(33):5429-5435. Paschka P, Marcucci G, Ruppert AS, et al. Wilms' tumor 1 gene mutations independently predict poor outcome in adults with cytogenetically normal acute myeloid leukemia: a cancer and leukemia group B study. J Clin Oncol. 2008;26(28):4595-4602. Pritchardjones K, Renshaw J, Kingunderwood L. The Wilms-Tumor (Wt1) Gene Is Mutated in a Secondary Leukemia in a Wagr Patient. Hum Mol Genet. 1994;3(9):1633-1637. Abbas S, Erpelinck-Verschueren CAJ, Goudswaard CS, Lowenberg B, Valk PJM. Mutant Wilms' tumor 1 (WT1) mRNA with premature termination codons in acute myeloid leukemia (AML) is sensitive to nonsense-mediated RNA decay (NMD). Leukemia. 2010;24(3):660-663. Becker H, Marcucci G, Maharry K, et al. Mutations of the Wilms tumor 1 gene (WT1) in older patients with primary cytogenetically normal acute myeloid leukemia: a Cancer and Leukemia Group B study. Blood. 2010;116(5):788-792. Fernandez HF, Sun Z, Yao X, et al. Anthracycline dose intensification in acute myeloid leukemia. N Engl J Med. 2009;361(13):1249-1259. 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. Figueroa ME, Abdel-Wahab O, Lu C, et al. Leukemic IDH1 and IDH2 Mutations Result in a Hypermethylation Phenotype, Disrupt TET2 Function, and Impair Hematopoietic Differentiation. Cancer Cell. 2010;18(6):553-567. Ito S, D'Alessio AC, Taranova OV, et al. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature. 2010;466(7310): 1129-1233. Tahiliani M, Koh KP, Shen YH, et al. Conversion of 5-Methylcytosine to 5Hydroxymethylcytosine in Mammalian DNA by MLL Partner TET1. Science. 2009;324(5929):930-935. Ward PS, Patel J, Wise DR, et al. The Common Feature of Leukemia-Associated IDH1 and IDH2 Mutations Is a Neomorphic Enzyme Activity Converting alpha-Ketoglutarate to 2-Hydroxyglutarate. Cancer Cell. 2010;17(3):225-234. Moran-Crusio K, Reavie L, Shih A, et al. Tet2 Loss Leads to Increased Hematopoietic Stem Cell Self-Renewal and Myeloid Transformation. Cancer Cell. 2011;20(1):11-24. Quivoron C, Couronne L, Della Valle V, et al. TET2 Inactivation Results in Pleiotropic Hematopoietic Abnormalities in Mouse and Is a Recurrent Event during Human Lymphomagenesis. Cancer Cell. 2011;20 (1):25-38. Li Z, Cai XQ, Cai CL, et al. Deletion of Tet2 in mice leads to dysregulated hematopoietic stem cells and subsequent development of myeloid malignancies. Blood. 2011;118 (17):4509-4518. Sinha S, Thomas D, Yu L, et al. Mutant WT1 is associated with DNA hypermethy-

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

lation of PRC2 targets in AML and responds to EZH2 inhibition. Blood. 2015;125(2):316-326. Grimwade D, Vyas P, Freeman S. Assessment of minimal residual disease in acute myeloid leukemia. Curr Opin Oncol. 2010;22(6):656-663. Cilloni D, Renneville A, Hermitte F, et al. Real-time quantitative polymerase chain reaction detection of minimal residual disease by standardized WT1 assay to enhance risk stratification in acute myeloid leukemia: a European LeukemiaNet study. J Clin Oncol. 2009;27(31):5195-5201. Van Dijk JP, Knops GH, Van De Locht LT, et al. Abnormal WT1 expression in the CD34negative compartment in myelodysplastic bone marrow. Br J Haematol. 2002;118(4): 1027-1033. Di Stasi A, Jimenez AM, Minagawa K, AlObaidi M, Rezvani K. Review of the Results of WT1 Peptide Vaccination Strategies for Myelodysplastic Syndromes and Acute Myeloid Leukemia from Nine Different Studies. Front Immunol. 2015;6:36. Van Driessche A, Berneman ZN, Van Tendeloo VF. Active specific immunotherapy targeting the Wilms' tumor protein 1 (WT1) for patients with hematological malignancies and solid tumors: lessons from early clinical trials. Oncologist. 2012;17(2):250-259. Dao T, Yan S, Veomett N, et al. Targeting the intracellular WT1 oncogene product with a therapeutic human antibody. Sci Transl Med. 2013;5(176):176ra133. Itzykson R, Kosmider O, Cluzeau T, et al. Impact of TET2 mutations on response rate to azacitidine in myelodysplastic syndromes and low blast count acute myeloid leukemias. Leukemia. 2011;25(7):11471152. Bejar R, Lord A, Stevenson K, et al. TET2 mutations predict response to hypomethylating agents in myelodysplastic syndrome patients. Blood. 2014;124(17):2705-2712. Emadi A, Faramand R, Carter-Cooper B, et al. Presence of isocitrate dehydrogenase mutations may predict clinical response to hypomethylating agents in patients with acute myeloid leukemia. Am J Hematol. 2015;90(5):E77-79. Brayer J, Lancet JE, Powers J, et al. WT1 vaccination in AML and MDS: A pilot trial with synthetic analog peptides. Am J Hematol. 2015;90(7):602-607. Tsuboi A, Oka Y, Kyo T, et al. Long-term WT1 peptide vaccination for patients with acute myeloid leukemia with minimal residual disease. Leukemia. 2012;26(6): 1410-1413. Maslak PG, Dao T, Krug LM, et al. Vaccination with synthetic analog peptides derived from WT1 oncoprotein induces Tcell responses in patients with complete remission from acute myeloid leukemia. Blood. 2010;116(2):171-179. Rezvani K, Yong AS, Mielke S, et al. Leukemia-associated antigen-specific T-cell responses following combined PR1 and WT1 peptide vaccination in patients with myeloid malignancies. Blood. 2008;111(1): 236-242. Keilholz U, Letsch A, Busse A, et al. A clinical and immunologic phase 2 trial of Wilms tumor gene product 1 (WT1) peptide vaccination in patients with AML and MDS. Blood. 2009;113(26):6541-6548.

679

REVIEW ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2016 Volume 101(6):680-687

How to select the best available related or unrelated donor of hematopoietic stem cells? Jean-Marie Tiercy

National Reference Laboratory for Histocompatibility, Department of Genetic and Laboratory Medicine, University Hospitals Geneva, Switzerland

ABSTRACT

This review article is based on an educational manuscript from the EHA20 meeting (Vienna, June 2015)

Correspondence: jean-marie.tiercy@unige.ch

Received: February 29, 2016. Accepted: March 9, 2016. Pre-published: no prepublication. doi:10.3324/haematol.2015.141119

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/680

©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.

680

ecognition of HLA incompatibilities by the immune system represents a major barrier to allogeneic hematopoietic stem cell transplantation. HLA genotypically identical sibling donors are, therefore, the gold standard for transplantation purposes, but only 30% patients have such a donor. For the remaining 70% patients alternative sources of stem cells are a matched unrelated adult volunteer donor, a haploidentical donor or a cord blood unit. The definition of ‘HLA matching’ depends on the level of resolution and on which loci are tested. The development of HLA molecular typing technologies and the availability of more than 27 million donors in the international database has greatly facilitated unrelated donor searches. The gold standard is high resolution typing at the HLA-A, -B, -C, -DRB1, and -DQB1 loci (10/10 match). Single disparities for HLA-A, -B, - C, or -DRB1 are associated with increased risk of post-transplant complications, but less so in patients with advanced disease, and in those undergoing Tcell-depleted allografting. HLA-DQB1 mismatches seem to be better tolerated and some HLA-C, -DRB1 and -DPB1 disparities are potentially less immunogenic. HLA typing by next-generation sequencing methods is likely to change matching algorithms by providing full sequence information on all HLA loci in a single step. In most European populations a 10/10 matched donor can be found for at least 50% of patients and an additional 20-30% patients may have a 9/10 matched donor. Genetic factors that help in identifying donors with less immunogenic mismatches are discussed. Haploidentical donors are increasingly used as an alternative source of stem cells for those patients lacking a matched unrelated donor.

Introduction Among the many factors that influence the outcome of hematopoietic stem cell transplantation (HSCT) polymorphism of the classical human leukocyte antigen (HLA) genes represents the most important barrier.1,2 The number of known HLA alleles is still growing and this trend will become even more pronounced with the wider use of high throughput sequencing methods3 in clinical laboratories that perform histocompatibility testing. Allorecognition of HLA allelic differences by T lymphocytes confers a higher risk of acute graft-versus-host disease (GVHD) and mortality. HLA genotypically matched sibling donors are, therefore, the gold standard source of stem cells for allogeneic HSCT. However, since about 70% of patients do not have an available HLA-identical sibling, at least in Western countries, alternative sources have to be considered, such as HLA-‘matched’ unrelated adult donors, cord blood units, or haploidentical donors. Since 2007, the number of transplants with stem cells from an unrelated donor has been higher than the number from matched sibling donors, reaching 53% in 2013.4 The developments of molecular typing technologies and the continuous increase in the number of volunteer donors in the Bone Marrow Donor Worldwide (BMDW) database have undoubtedly improved the identification of well-matched, unrelated donors and contributed to the impressive expansion of HSCT programs worldwide.1,4-7 Over 27 million donors are now registered in the international database (www.bmdw.org) and an increasing proportion of these donors are typed by haematologica | 2016; 101(6)

How to select the best available donor

molecular techniques at all HLA loci and at the allele level. Nevertheless, despite these achievements, many patients will still not have a fully matched donor because of the extremely great diversity of HLA alleles and haplotypes.1,47 As of January 2016 more than 14,000 HLA alleles have been assigned, accounting for more than 10,000 different HLA proteins (www.ebi.ac.uk/ipd/imgt/hla) (Figure 1). This increasing level of complexity has negative consequences for patient/unrelated donor matching. Thus for many patients a challenge for the histocompatibility laboratory is to identify mismatched donors or cord blood units with the lowest potential for recognition by the immune system, in particular by the direct T-cell allorecognition mode. A better characterization of ‘permissive’ mismatches would undoubtedly allow increased access to HSCT for many patients. These strategies should be weighted for by donor-associated non-HLA criteria that also affect clinical outcome. This review focuses on essential immunogenetic parameters that have been reported to be relevant in unrelated and haploidentical donor search algorithms. The selection of cord blood units is not discussed here. Issues that are less familiar to the clinicians are reviewed, such as the impact of HLA typing resolution on matching criteria, the clinical relevance of so-called ‘permissive’ mismatches, and the impact of high throughput sequencing techniques on donor selection. A few simple recommendations for unrelated donor search algorithms are summarized in the conclusion.

Searching for ‘HLA-compatible’ unrelated donors What do we mean by ‘HLA-compatible’? The compatibility status of each patient/donor pair

depends on the level of resolution of HLA typing and on which loci are or are not tested. ‘High resolution’ typing is often described as ‘four-digit typing’ or ‘allele-level typing’ and may therefore not have exactly the same meaning in all centers. In order to establish a common language for histocompatibility terms a group of experts defined the following levels of resolution:8 (i) low resolution, or ‘first-field level’ typing, by reference to the two digits preceding the first separator, or antigen level typing, e.g. A*02; (ii) high resolution typing, which is defined by one or a set of alleles that share the same antigen binding site formed by the α1/α2 domains of class I alleles (encoded by exons 2+3), and by the α1 domain of class II alleles (encoded by exon 2), and that exclude null alleles (i.e. alleles not expressed at the cell surface). For example A*02:01:01G includes all the alleles (n=52, based on the IMGT/HLA 3.23.0 release, January 2016) sharing the same exons 2+3 nucleotide sequence as A*02:01:01). The alleles or groups of alleles are designated by ‘second- and thirdfield level’ names, referring to the ≥2 digit numbers preceding, respectively, the second and third separators. Alleles with nucleotide sequences encoding the same protein sequence for the antigen binding site are designated by the suffix ‘P’, e.g. A*02:01P. (iii) allele level typing, which corresponds to a unique nucleotide sequence for an HLA gene, as defined by using all digits in the first, second, third and fourth fields, e.g. A*02:01:01:01. Functionally the third and fourth fields which characterize alleles that differ, respectively, by silent substitutions in the coding sequence and by substitutions in the non-coding sequence, are irrelevant, except when substitutions prevent the expression of HLA alleles (e.g. the null allele B*15:01:01:02N). Missing a null allele

Figure 1. Schematic map of the 4 Mb human major histocompatibility complex. The map is not drawn to scale, double separators (//) indicate larger distances and correspond to the regions where recombinations occur most frequently. C: centromere. The first row below the map indicates the number of well-defined serotypes for each locus. The DR serotypes include the antigens (heterodimers) encoded by the DRA/DRB1 (DR1-DR18), DRA/DRB3 (DR52), DRA/DRB4 (DR53) and DRA/DRB5 (DR51) genes. In the second row the total number of alleles is given (IMGT/HLA database version 3.23.0). The third row indicates the total number of different proteins. A total of 387 class I alleles with no surface expression (null alleles) and 90 class II null alleles have been described. Alleles that share identical sequences in the peptide-binding region represent 9% of the A, 6% of the B, 7.7% of the C, 1.7% of the DRB1, 17.2% of the DQB1 and 6.2% of the DPB1 alleles. Matching for the HLA-A, -B,- C,- DRB1 and -DQB1 loci is referred to as a 10/10 match, when HLA-DPB1 is included it becomes a 12/12 match. Matching for HLA-A, -B, -C, and -DRB1 loci is an 8/8 match. There is still no international standard for reporting DRB3/4/5 as well as DQA1 and DPA1 mismatches. Donor search algorithms do not include DQA1 and DPA1 testing because of strong linkage disequilibrium with the corresponding DQB1 and DPB1 loci.

haematologica | 2016; 101(6)

681

J.-M. Tiercy

will lead to a mismatch that is very likely to be recognized by alloreactive T cells and have a deleterious clinical impact.9,10 It should be mentioned here that substitutions in non-coding sequences may influence the level of expression (e.g. the A24low allele A*24:02:01:02L). Such variability may also have an impact on anti-HLA allorecognition; (iv) other levels of resolution, usually referred to as intermediate-level resolution, include any typing results that fall between low and high resolution. This term is used when the technique resolves a group of alleles, usually defined at the second-field level and irrespectively of the site of the polymorphisms. For example, DRB1*11:01/11:09/11:28, a string of three alleles that are depicted by the National Marrow Donor Program (NMDP) code DRB1*11:BYCC. Depending on the number and the nature of the unresolved ambiguities, the ‘intermediate’ level of resolution can be quite heterogeneous. It has practical relevance for donor selection when it allows discrimination between frequent alleles, as shown in the example given in Table 1 (DRB1*04:04 absent in the string of alleles assigned in the donor under the code DRB1*04:VN). Examples with these different levels of resolution and their impact on matching status are presented in Table 1. In most European centers the gold standard is to look for an HLA-A, -B, -C, -DRB1, and -DQB1-matched donor, a so-called 10/10 match. An alternative matching algorithm, which is recommended by the NMDP, is to look for an HLA-A, -B, -C, and -DRB1-compatible donor (8/8 match). When HLA-DPB1 typing is included, donors with a 12/12 match are sought. Although less polymorphic the DRB3/B4/B5 loci may lead to additional HLA class II mismatches. However, there is neither a common practice nor any international recommendation on how to count these mismatches in the 10/10 matching algorithm. In clinical studies, there is still some confusion in using the histocompatibility terms for reporting HLA compatibility. For example, the German5 and the Center for International Blood and Marrow Transplant Research (CIBMTR) studies11,12 refer to high-resolution typing, whereas the International Histocompatibility Working Group (IHWG)13 and a recent CIBMTR14 study refer to allele-level matching and Japanese15 and Swiss16 studies refer to allele-level typing at the second-field level.

ty of finding a matched donor,18 with the lowest probability assigned to patients of African ancestry. In European countries 45-65% patients will eventually have a 10/10 matched donor, and a 9/10 matched donor may be identified for an additional 20-30% of patients (Table 2).

Impact of single mismatches There is now a general consensus that single HLA mismatches at the HLA-A, -B, -C and -DRB1 loci are clinically relevant.1,5-7,11,19,20 A comprehensive review of the impact of HLA-C incompatibilities on clinical outcome has been published recently.21 In a large CIBMTR study on patients with chronic myeloid leukemia no significant difference in overall survival was noted between patients transplanted with HLA class I or class II mismatched grafts.22 With regards to HLA class II disparities, several studies indicated that HLADQB1 disparities are not associated with mortality.5,11,20 Because of the high priority given to HLA-DRB1 matching and the strong DRB1-DQB1 linkage disequilibrium, studies are often underpowered to reveal the clinical relevance of DQB1 disparities. Evidence for a role of HLA-DPB1 mismatches is now well documented.6,13,23-25 In 10/10matched HSCT, DRB3 and DRB4 mismatches were associated with poorer outcome, although most donor-recipient pairs were also incompatible for the DPB1 locus.16,26 In

Table 1. Examples of patient/donor matching status as a function of HLA typing resolution levels.

Patient

Probability of identifying an HLA-identical sibling donor or a highly matched unrelated donor The probability of identifying an HLA-identical sibling donor depends only on the number of siblings and is 25% for patients with one sibling, 44% for those with two, 58% for those with three, 68% for those with four, and up to 90% for patients with eight siblings. On the other hand the probability of identifying a highly matched unrelated donor depends on the frequency of the patient’s HLA haplotypes. As shown in Table 2, the average probability of identifying a matched unrelated donor differs greatly depending on the ethnic origin of the patients and on the matching grade required by the transplant center (8/8 or 10/10). Indeed, depending on ethnic origin, 1-5% of patients do not have a single potentially matched donor upon direct interrogation of the BMDW database,7,17 because the large majority of donors registered in the database are of Western European ancestry. The ethnic origin of the patient strongly influences the probabili682

Donor

Compatibility Compatibility high resolution allele level/ 2nd field exons 2+3 (cl.I) exon 2 (cl. II)

A*02 A*02:01:01Ga A*02:01Pb A*02:01 A*02:06 A*02:01:01G A*02:01 DRB1*14:01:01 A*02:01:01:01 A*02:01:01:01 A*02:01:01:01 DRB1*11:BYCC

A*02 A*02:01:01G A*02:01P A*02:06 A*02:126c A*02:09 A*02:09 DRB1*14:54:01 A*02:01:01:01 A*02:26 A*02:01:01:02N DRB1*11:RDPB

(11:01/11:09/11:28)

(11:01/11:95/11:97/11:100/11:117)d

DRB1*04:04

potential match potential match match mismatch match potential match match match match mismatch mismatch potential match

potential match potential match potential match mismatch mismatch potential match mismatch mismatch match mismatch mismatch potential match

mismatch

C*07:02 C*07:FEAU

match

potential match

(07:02/07:50/07:66/07:74)f

match

potential match

DRB1*04:VN (04:01/04:13/04:16/04:21)e

C*07:02:01G C*07:02:01G

Nomenclature: G marks all the alleles with the same nucleotide sequence in the peptide binding site (including null alleles). Since A*02:01:01G includes five null alleles (A*02:43N/02:83N/02:305N/02:356N/02:608N) the patient and donor should be tested for these alleles -unless cell surface expression has been established by serological typing- before being categorized as matched at a high resolution level. bP denotes a string of alleles that encode the same protein sequence in the peptide binding site (α1/α2 domains for class I and α1 for class II alleles) as the first numbered allele in the group: cA*02:126 differs from A*02:06 by a residue outside the peptide binding site. dthe DRB1*11:01, *11:95, *11:97 and *11:100 share the same α1 domain but not DRB1 *11:117. ethis string of four alleles does not include DRB1*04:04, this donor is therefore incompatible at locus DRB1. fthe C*07:02, *07:50, *07:66 and *07:74 alleles share the same α1/α2 domains protein sequence, as does C*07:02:01G, these four alleles are included in the C*07:02:01G group of alleles. a

haematologica | 2016; 101(6)

How to select the best available donor

7/8-matched HSCT more than two mismatches at the DP, DQ and DRB3/4/5 loci were associated with an increased risk of mortality.27 The Japan Marrow Donor Program (JMDP) studies showed that the impact of single HLA incompatibilities has changed over time because of multiple factors, such as varying clinical protocols (GVHD prophylaxis, treatments for infections), HLA mismatches readily available in the latest period, and more intensive GVHD prophylaxis in patients with HLA-DRB1 mismatches in the earlier period.28 Furthermore, an initial observation that HLA-B and -C incompatibilities were better tolerated than -A or -DRB1 mismatches11 has not been confirmed in more recent studies.5,29 In their latest report Morishima et al.15 provided evidence that single HLA-A, -B or -C allele mismatches and double HLA-DRB1/DQB1 mismatches are associated with increased mortality in non-T-cell-depleted bone marrow transplantation. Interestingly single HLADRB1, -DQB1 or -DPB1 mismatches did not significantly affect overall survival rate. Indeed HLA-DPB1 incompatibilities were associated with both increased acute GVHD and lower relapse rates.15 The impact of HLA mismatches on overall mortality has been reported to be most apparent in patients with early disease.12 In HSCT for non-malignant disorders single HLA-A, -B, -C or -DRB1 mismatches were not associated with acute or chronic GVHD but were associated with graft failure.19 A large study of unrelated donor reduced-intensity conditioning HSCT based on 8/8 matching recently found that single HLA-A, -B, -C or -DRB1 mismatches were associated with a higher incidence of acute GVHD and a lower disease-free survival rate without differences in relapse rate or chronic GVHD.30 In this study HLAC*03:03/03:0428 and HLA-DPB131 permissive mismatches were not associated with better outcome. The key findings reported in recent publications (2013-2016) are summarized in Table 3. Direct allorecognition of mismatched HLA antigens is mediated at least partly by cross-reactive viral peptide-

specific memory T cells.32 Mismatches that are characterized by changes in amino acid residues not seen by the Tcell receptor, i.e. outside the α1/α2 domains for class I antigens, are not expected to be recognized and could, therefore, be considered as acceptable mismatches. However indirect allorecognition of HLA allopeptides could also play a role. Indeed the number of peptides derived from incompatible HLA molecules presented by HLA antigens shared between the recipient and the donor can be predicted using the PIRCHE (predicted indirectly recognizable HLA epitopes) model.33,34 In unrelated HSCT with HLA-C or -DPB1 mismatches, the number of PIRCHE has been reported to correlate with clinical outcome.33,34

First-field versus second-field (antigen versus allele, or low versus high) mismatches Comparisons of the impact of single allele and single antigen mismatches on clinical outcome did not reveal significant differences,5,11,12 except, possibly, for the HLA-C locus for which allele mismatches have been reported to be less detrimental than antigen mismatches.18,20 However this could possibly be explained by the very high frequency (68.7%) of C*03:03/03:04 mismatches in the NMDP study.29 This incompatibility had previously been reported to be more permissive, on the basis of in vitro assays measuring direct cytotoxic T-lymphocyte alloreactivity.35,36

Searching for ‘permissive’ mismatches The identification of so-called permissive mismatches has been a long-lasting challenge. As a first approach, determination of the frequency of cytotoxic T-lymphocyte precursors revealed a number of HLA class I incompatibilities that had not previously been recognized and that could be considered as more permissive.36,37 However it was not possible to reliably predict this lack of recognition by looking at the structural differences between the mismatched alleles.38 It seems reasonable to predict that HLA disparities characterized by substitutions in the peptide binding site which significantly alter the set of pep-

Table 2. Overall probabilities of identifying a 7/8, 8/8, 9/10 and 10/10 matched unrelated donor.

Ethnic origin (country)a) European (NL) European (UK) European (A) European (D) European (CH) European (NL) European (IT) European (HR) European (USA) African (USA) ME/ NA (USA)b Asian (USA)c Hispanic (USA)d

Match 8/8

Match ≥7/8

Match 9/10

20% 24% 31% 32% 30% 75% 18% 46% 27-42% 34%

97% 71% 90% 76-88% 80%

Match 10/10

61% 58% 48% 43% 65%

Match 9-10/10

Reference

69%e 72% 80%f

62 63 64 17 7 46 65 66 18 18 18 18 18

NL: the Netherlands; UK: United Kingdom; A: Austria; D: Germany; CH: Switzerland; HR: Croatia; USA: United States of America; bME: Middle Eastern; NA: North African; cAsian: Chinese, Korean, South Asian, Japanese, Southeast Asian,Vietnamese; dHispanic: South/Central American; e<9/10 in 13% patients; fexceptionally 8/10 matched donors.

haematologica | 2016; 101(6)

683

J.-M. Tiercy Table 3. Impact of specific HLA locus or allele mismatches as reported in recent (2013-2016) multicenter studies of unrelated HSCT.

Ref.

N. of patients

2,646

13 30 29

8,539 3,853 7,349

8,003

7,898

2,588

803

50 44 16

2,029 6,967 11,039

Main conclusions Single HLA-A,B,C,DRB1 MM (either antigen or allele) associated with increased mortality, additional risk with <9/10 matched (including DQB1) donors Non-permissive DPB1 MM associated with increased mortality in 9-10/10 matched HSCT In 7/8 matched HSCT : >2 MM at DRB3/4/5, DQB1 or DPB1 loci associated with lower survival C*03:03/03:04 MM better tolerated, lower impact of C-locus MM explained by the high frequency of C*03:03/03:04 MM in the 7/8 matched group Single HLA-A,B,C,DRB1 MM associated with increased mortality, DQB1 MM associated with increased acute GVHD, non-permissive DPB1 MM associated with increased mortality in 10/10 or 8/8 matched cases Single HLA-A,B,C and double HLA-DRB1-DQB1 MM associated with increased mortality, HLA-A,B,C,DPB1 MM associated with higher risk of acute GVHD, reduced relapse only with C,DPB1 MM Reduced intensity conditioning HSCT: increased mortality in 7/8 matched HSCT, no impact of C*03:03/03:04 or permissive DPB1 MM Single HLA-A,B,C MM (9/10) associated with higher mortality, HLA-DRB1/DQB1 MM more permissive (high ratio of DRB1*11:01/11:04 and DQB1*03:01/03:02 MM) In 11/12 matched HSCT: single nucleotide polymorphism in the regulatory region of DPB1 locus associated with acute GVHD Patient and/or donor B*51:01 and patient C*14:02 associated with increased acute GVHD and mortality Donor age (>32 years) and 7/8, 6/8 mismatched donors associated with lower overall survival

MM: mismatch.

tides presented by the HLA molecules will be more efficiently recognized by alloreactive T cells, whereas mismatches involving residues outside the peptide binding site are not expected to be recognized. Indeed a semiquantitative, in vitro measurement of CD8+CD137+ alloreactive T cells in mixed lymphocyte reactions demonstrated that such a mismatch in the B44 serotype (i.e. B*44:02/44:27) was not recognized by cytotoxic T-lymphocytes and could possibly be considered as permissive.39 Based on in vitro assays set up to detect anti-HLA-DP alloreactive T cells, a new algorithm has been proposed for the identification of non-permissive HLA-DPB1 disparities, as defined by the presence of T-cell epitope mismatching.31,40 Two groups of alleles with high (DPB1*09:01, 10:01, 17:01) and intermediate (DPB1*03:01, 14:01, 45:01, 86:01) immunogenicity have been assigned, whereas all the remaining most frequent HLA-DPB1 alleles are classified in a third group. Each patient/donor pair in which an HLA-DPB1 allele of the high or intermediate immunogenicity groups is present in the patient or the donor only is classified as a non-permissive mismatch (graft-versushost or host-versus-graft direction). Of the total donor pool 70% consisted of either HLA-DPB1-matched donors or donors with a permissive HLA-DPB1 mismatch. Non-permissive HLA-DPB1 mismatches were associated with increased hazards for acute GVHD and transplant-related mortality, but not for relapse.31 In the IHWG study,13 HLADPB1 non-permissive mismatches were associated with increased risks of overall mortality in both 10/10- and 9/10-matched transplants. In contrast to these findings, a recent study12 found that any HLA-DPB1 mismatch was associated with acute GVHD. However the adverse impact of non-permissive HLA-DPB1 mismatches on transplant-related mortality and overall mortality was confirmed in 8/8- and in 10/10-, but not in 7/8- or in 9/10matched cases. Based on a retrospective clinical study, HLADRB1/DQB1 mismatches have recently been reported to be more permissive than HLA-A, -B or -C disparities in 9/10-matched HSCT.16 This was correlated with a preferential selection of donors with DRB1*11:01/11:04, 684

DRB1*14:01/14:54 and DQB1*03:01/03:02 mismatches, which might be associated with weaker immunogenicity, as suggested by the results of previous in vitro assays.16 A few reports have described the role of individual amino acids on clinical outcome. The impact of individual HLA amino acid mismatches, such as those reported in the JMDP study41 may not be applicable in other populations which show greater heterogeneity in HLA disparities and, therefore, fewer mismatches of similar nature. A large scale analysis evaluated the clinical impact of specific amino acid substitutions in HSCT patients with single class I mismatches.42 It was found that patients with mismatched donors lacking an amino acid (aa) substitution at aa116 and aa99 of HLA-C and aa9 of HLA-B alleles had outcomes similar to those of patients grafted with hematopoietic stem cells from 8/8-matched donors. In particular substitutions at aa116 and aa99 were both associated with increased transplant-related mortality in the multivariate analysis.42 The importance of aa116 of HLAC had been observed previously.43 Supporting the hypothesis that levels of expression of target antigens in HLAincompatible combinations might also affect allorecognition, two studies found that patients with HLA-C*14:02 (high expression allele) had increased risks of acute GVHD and mortality.44,45

Immunogenetics of HLA haplotypes There is growing evidence that, in addition to individual HLA allele disparities, non-HLA polymorphisms in the major histocompatibility complex have an impact on clinical outcome. Indeed, a haplotype effect has been reported by several groups.23,46-50 In one study5 the survival of patients transplanted with 10/10-matched donors from the national registry was better than that of patients transplanted with 10/10-matched donors from the international registry. This could reflect the impact of non-HLA polymorphisms which may vary in populations from different origins and may be linked to different HLA-A, -B, and -DRB1 haplotypes. The translation of major histocompatibility-resident DNA variations into clinical practice still requires careful assessment of the relative importance of haematologica | 2016; 101(6)

How to select the best available donor

such polymorphisms on outcome. A promising candidate non-HLA polymorphism is rs9277534, a single nucleotide polymorphism associated with HLA-DPB1 expression and reported to be correlated with increased risk of acute GVHD.50 Appropriate selection of donors based on rs9277534 typing could potentially lead to a decrease in the incidence of acute GVHD.

Haploidentical donors For patients lacking a highly matched unrelated donor the choice between a mismatched unrelated donor, a haploidentical donor, or a cord blood unit largely depends on the centers’ expertise.51 These different graft sources have not yet been compared directly in a randomized trial. High-dose cyclophosphamide treatment after non-myeloablative conditioning and T-cell-replete haploidentical HSCT has been shown to result in acceptable rates of graft rejection and acute GVHD.52 This new protocol has provided a valuable alternative for adult patients with hematologic malignancies who lack a matched related or unrelated donor and has drastically increased patients’ chance of access to allogeneic transplantation. Indeed, a trend towards increased use of haploidentical donors and a concomitant decrease of cord blood unit transplantation have been confirmed recently.4 In selecting the best haploidentical donor, the number of mismatched HLA antigens on the non-shared haplotype does not seem to play a role.53 On the other hand, evaluation of the anti-HLA immunization of the patient should be performed systematically. Indeed the presence of donor-specific antibodies has been shown to be associated with an increased risk of primary graft failure54,55 The anti-tumor effect of alloreactive natural killer cells has been well documented for more than a decade.56,57 Ligands for the inhibitory killer-cell immunoglobulin-like receptors (KIR) are HLA-Bw4, HLA-C alleles with a Lys at position 80 (HLA-C1), or an Arg at position 80 (HLA-C2). Natural killer cells are defined as alloreactive when they express inhibitory KIR specific for HLA class I epitopes (Bw4, C1, C2) not expressed on the patient’s cells,56,57 and/or activating KIR that recognize ligands expressed by the recipient’s cells, for example KIR2DS1-positive natural killer cells and an HLA-C2-positive recipient.58 In its simplest form, the search for a haploidentical donor is based on incompatibility between the donor’s and recipient’s HLA ligands. For example, when a patient has a C1/C1 genotype and the donor a C1/C2 genotype, the inhibitory signal provided by the C2 epitope is lacking in the patient. The absence on patient’s cells of one HLA ligand (Bw4, C1 or C2) recognized by inhibitory KIR can thus lead to potential alloreactivity mediated by ‘licensed’ natural killer cells from donors who are positive for this ligand. The role of activating KIR might also be considered in donor selection, i.e. by identifying the presence of the KIR2DS1 locus58 or by identifying KIR B haplotype-positive donors.59,60 An even more sophisticated approach would take into account the KIR2DL1 allelic polymorphism affecting the strength of the inhibitory receptor.61

Influence of next-generation sequencing methods on unrelated donor selection The recent developments in next-generation sequencing technologies based on single molecule sequencing enable haematologica | 2016; 101(6)

high-quality resolution of full-length HLA sequences up to the fourth-field level (intron and untranslated sequences). Next-generation sequencing is rapidly entering clinical HLA laboratories because it provides powerful and efficient HLA typing that also meets the turn-around-time requirements of the HSCT field. The advantages of nextgeneration sequencing technology in the selection of unrelated donors are the following: (i) HLA typing results will be available without any ambiguity on the level of resolution, i.e. there will be no confusion on the concept of high resolution/allele level typing, especially with the identification of the null alleles; (ii) complete, or almost complete sequence information will be available for all loci, meaning that all loci can be taken into account simultaneously for donor selection. This could be particularly relevant when only donors with multiple mismatches are available. For example donors with single class I mismatches could be sorted out on the basis of additional HLA-DRB3/DRB4/DPB1 compatibility; (iii) not only will typing by next-generation sequencing enable matching at the protein level (corresponding to second-field typing, or to the former ‘four-digit typing’), but it will also provide information on non-coding variations that can potentially affect the level of expression of a given HLA antigen.49 The potential impact of non-coding polymorphisms is presently only speculative and will eventually be determined by large, retrospective collaborative studies. Some polymorphisms may be surrogate markers of HLA haplotypes associated with higher risks of posttransplant complications. A promising example is the rs9277534 single nucleotide polymorphism in the HLADPB1 regulatory region which has been shown to be correlated with acute GVHD risk.50 Patient/donor matching algorithms should still be based on second-field level allele typing, with the exception of the null alleles defined by fourth-field level typing. It is, however, highly recommended that the full length HLA sequence information is recorded for each transplant pair.

Conclusion When no HLA-identical sibling donor is available, an estimate of the probability of finding a fully matched unrelated donor, based on the frequency of the patient’s haplotypes, will help the transplant center in taking a decision on whether to search for an unrelated donor or look for an alternative source of hematopoietic stem cells (haploidentical donor or cord blood unit). Various software programs based on high resolution HLA haplotype frequencies, such as HapLogic (NMDP), Optimatch (Germany) and Easymatch (France), can predict the number of potentially matched donors. When no 10/10matched unrelated donor is found, prioritization of a 9/10 matched unrelated donor or an alternative donor remains difficult in the absence of randomized trials. Clearly the urgency of the transplantation and the transplant center’s expertise will influence the algorithm. On the basis of published studies, some considerations and practical recommendations for the selection of optimally matched unrelated donors can be made, as summarized below: (i) patient/donor HLA typing is mandatory for all loci taken into consideration by the transplant protocol: the minimal level is HLA-A, -B, -C, -DRB1, -DQB1 high resolution typing, i.e. exons 2 + 3 for class I and exon 2 for 685

J.-M. Tiercy

class II, but second-field level typing (i.e. including polymorphisms outside the peptide binding site) is recommended; (ii) single mismatches (first- or second-field level typing) at any of the four HLA-A, -B, -C, -DRB1 loci are associated with an increased risk of acute GVHD and mortality; (iii) more than one mismatch among the HLA-A, -B, -C, -DRB1, and -DQB1 loci should be avoided; (iv) before considering a mismatch, donor-specific antibodies must be identified in allosensitized patients; (v) HLA-A, -B, -C, and -DRB1 mismatches involving residues located outside the peptide binding site (e.g. A*02:01/02:09), or residues that only fine tune the set of peptides bound to the HLA molecule (e.g. DRB1*11:01/11:04), or residues that are not seen by the Tcell receptor (e.g. C*03:03/03:04) could possibly be considered as weakly or non-immunogenic; (vi) there is no evidence that allele mismatches should be preferred to antigen mismatches; (vii) when no potentially HLA-A, -B, and -DRB1 compatible donor is available in the BMDW file, a potential 9/10-matched donor may be identified after selecting for HLA-A antigen mismatched donors; (viii) HLA-DQB1 and -DRB3/4/5 mismatches should be preferred to other mismatches; (ix) whenever two or more 10/10-matched donor are available, donor age14 ABO blood group and HLA-DPB1 matching should be prioritized. Permissive HLA-DPB1 mismatches can be defined either by the T-cell epitope matching algorithm,31 or by taking into account the level of DP expression tagged by the rs9277534 polymor-

References 1. Petersdorf EW. Genetics of graft-versus-host disease: the major histocompatibility complex. Blood Rev. 2013;27(1):1-12. 2. Shaw BE, Arguello R, Garcia-Sepulveda CA, Madrigal JA. The impact of HLA genotyping on survival following unrelated donor haematopoietic stem cell transplantation. Br J Haematol. 2010;150(3):251-258. 3. Cereb N, Kim HR, Ryu J, Yang SY. Advances in DNA sequencing technologies for high resolution HLA typing. Hum Immunol. 2015;76(12):923-927. 4. Passweg JR, Baldomero H, Bader P, et al. Hematopoietic SCT in Europe 2013: recent trends in the use of alternative donors showing more haploidentical donors but fewer cord blood transplants. Bone Marrow Transplant. 2015;50(4):476-482. 5. Furst D, Muller C, Vucinic V, et al. High-resolution HLA matching in hematopoietic stem cell transplantation: a retrospective collaborative analysis. Blood. 2013;122(18): 3220-3229. 6. Shaw BE, Mayor NP, Russell NH, et al. Diverging effects of HLA-DPB1 matching status on outcome following unrelated donor transplantation depending on disease stage and the degree of matching for other HLA alleles. Leukemia. 2010;24(1):58-65. 7. Tiercy JM. Unrelated hematopoietic stem cell donor matching probability and search algorithm. Bone Marrow Res. 2012;2012: 695018.

686

phism.50 On a simple, practical basis, mismatches among low expression HLA-DPB1 alleles (DPB1*02, 04, 17) should be prioritized over mismatches among high expression HLA-DPB1 alleles (DPB1*01, 03, 05, 06, 10, 11, 13, 15, 16, 19). In the case that HLA DPB1 incompatibilities are present, the patient should be tested for potential anti-DP donor-specific antibodies; (x) the impact of mismatches may vary depending on the type and state of the underlying disease, the GVHD prophylaxis (T-cell depletion) used, and the conditioning regimen. Although it is extremely difficult to predict the impact of any single HLA mismatch reliably, our current understanding of the immunogenetics of HSCT allows selection of mismatched donors whose cells are likely to induce a minimal alloresponse. Presently the choice between a mismatched unrelated donor, mismatched cord blood, or a haploidentical donor seems to depend on the transplant centerâ&#x20AC;&#x2122;s expertise whereas clear information on the optimal strategy awaits the results of randomized trials. Retrospective studies have nevertheless shown that it is possible to overcome the HLA barrier, to prioritize specific HLA disparities with potentially lower immunogenicity, and thus to increase the number of patients who can have access to HSCT. Acknowledgments This work was supported by grant 310030-146306 from the Swiss National Science Foundation. The exellent collaboration with the SBSC (Swiss Blood Stem Cells) registry and the Swiss transplant centers is acknowledged.

8. Nunes E, Heslop H, Fernandez-Vina M, et al. Definitions of histocompatibility typing terms. Blood. 2011;118(23):e180-183. 9. Elsner HA, Blasczyk R. Immunogenetics of HLA null alleles: implications for blood stem cell transplantation. Tissue Antigens. 2004;64(6):687-695. 10. Zanone-Ramseier R, Gratwohl A, Gmur J, Roosnek E, Tiercy JM. Sequencing of two HLA-A blank alleles: implications in unrelated bone marrow donor matching. Transplantation. 1999;67(10):1336-1341. 11. Lee SJ, Klein J, Haagenson M, et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110 (13):4576-4583. 12. 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. 13. Fleischhauer K, Shaw BE, Gooley T, et al. Effect of T-cell-epitope matching at HLADPB1 in recipients of unrelated-donor haemopoietic-cell transplantation: a retrospective study. Lancet Oncol. 2012;13(4): 366-374. 14. Kollman C, Spellman SR, Zhang MJ, et al. The effect of donor characteristics on survival after unrelated donor transplantation for hematologic malignancy. Blood. 2016;127(2):260-267. 15. Morishima Y, Kashiwase K, Matsuo K, et al. Biological significance of HLA locus match-

16.

17.

18.

19.

20.

21. 22.

ing in unrelated donor bone marrow transplantation. Blood. 2015;125(7):1189-1197. Passweg JR, Schanz U, Chalandon Y, et al. High-resolution HLA matching in unrelated donor transplantation in Switzerland: differential impact of class I and class II mismatches may reflect selection of nonimmunogenic or weakly immunogenic DRB1/DQB1 disparities. Bone Marrow Transplant. 2015; 50(9):1201-1205. Hirv K, Bloch K, Fischer M, Einsiedler B, Schrezenmeier H, Mytilineos J. Prediction of duration and success rate of unrelated hematopoietic stem cell donor searches based on the patient's HLA-DRB1 allele and DRB1-DQB1 haplotype frequencies. Bone Marrow Transplant. 2009;44(7):433-440. Gragert L, Eapen M, Williams E, et al. HLA match likelihoods for hematopoietic stemcell grafts in the U.S. registry. N Engl J Med. 2014;371(4):339-348. Horan J, Wang T, Haagenson M, et al. Evaluation of HLA matching in unrelated hematopoietic stem cell transplantation for nonmalignant disorders. Blood. 2012;120 (14):2918-2924. Woolfrey A, Klein JP, Haagenson M, et al. HLA-C antigen mismatch is associated with worse outcome in unrelated donor peripheral blood stem cell transplantation. Biol Blood Marrow Transplant. 2011;17(6):885-892. Tiercy JM. HLA-C incompatibilities in allogeneic unrelated hematopoietic stem cell transplantation. Front Immunol. 2014; 5:216. Arora M, Weisdorf DJ, Spellman SR, et al. HLA-identical sibling compared with 8/8

haematologica | 2016; 101(6)

How to select the best available donor

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

matched and mismatched unrelated donor bone marrow transplant for chronic phase chronic myeloid leukemia. J Clin Oncol. 2009;27(10):1644-1652. Bettens F, Passweg J, Schanz U, et al. Impact of HLA-DPB1 haplotypes on outcome of 10/10 matched unrelated hematopoietic stem cell donor transplants depends on MHC-linked microsatellite polymorphisms. Biol Blood Marrow Transplant. 2012;18(4): 608-616. Ludajic K, Balavarca Y, Bickeboller H, et al. Impact of HLA-DPB1 allelic and single amino acid mismatches on HSCT. Br J Haematol. 2008;142(3):436-443. Shaw BE, Gooley TA, Malkki M, et al. The importance of HLA-DPB1 in unrelated donor hematopoietic cell transplantation. Blood. 2007;110(13):4560-4566. Detrait M, Morisset S, Chalandon Y, et al. Suggestive evidence of a role of HLA-DRB4 mismatches in the outcome of allogeneic hematopoietic stem cell transplantation with HLA-10/10-matched unrelated donors: a French-Swiss retrospective study. Bone Marrow Transplant. 2015;50(10):1316-1320. Fernandez-Vina MA, Klein JP, Haagenson M, et al. Multiple mismatches at the low expression HLA loci DP, DQ, and DRB3/4/5 associate with adverse outcomes in hematopoietic stem cell transplantation. Blood. 2013;121(22):4603-4610. Kanda Y, Kanda J, Atsuta Y, et al. Impact of a single human leucocyte antigen (HLA) allele mismatch on the outcome of unrelated bone marrow transplantation over two time periods. A retrospective analysis of 3003 patients from the HLA Working Group of the Japan Society for Blood and Marrow Transplantation. Br J Haematol. 2013;161(4): 566-577. Fernandez-Vina MA, Wang T, Lee SJ, et al. Identification of a permissible HLA mismatch in hematopoietic stem cell transplantation. Blood. 2014;123(8):1270-1278. Verneris MR, Lee SJ, Ahn KW, et al. HLA Mismatch is associated with worse outcomes after unrelated donor reduced-intensity conditioning hematopoietic cell transplantation: an analysis from the Center for International Blood and Marrow Transplant Research. Biol Blood Marrow Transplant. 2015;21(10):1783-1789. Zino E, Frumento G, Marktel S, et al. A Tcell epitope encoded by a subset of HLADPB1 alleles determines nonpermissive mismatches for hematologic stem cell transplantation. Blood. 2004;103(4):1417-1424. Amir AL, D'Orsogna LJ, Roelen DL, et al. Allo-HLA reactivity of virus-specific memory T cells is common. Blood. 2010;115(15): 3146-3157. Thus KA, Ruizendaal MT, de Hoop TA, et al. Refinement of the definition of permissible HLA-DPB1 mismatches with predicted indirectly recognizable HLA-DPB1 epitopes. Biol Blood Marrow Transplant. 2014;20(11): 1705-1710. Thus KA, Te Boome L, Kuball J, Spierings E. Indirectly recognized HLA-C mismatches and their potential role in transplant outcome. Front Immunol. 2014;5:210. Oudshoorn M, Doxiadis, II, van den BergLoonen PM, Voorter CE, Verduyn W, Claas FH. Functional versus structural matching: can the CTLp test be replaced by HLA allele typing? Hum Immunol. 2002;63(3):176-184. Roosnek E, Tiercy JM. Search for an unrelated HLA-compatible stem cell donor. Curr

haematologica | 2016; 101(6)

Opin Hematol. 1999;6(6):365-370. 37. van der Meer A, Joosten I, Schattenberg AV, de Witte TJ, Allebes WA. Cytotoxic T-lymphocyte precursor frequency (CTLp-f) as a tool for distinguishing permissible from nonpermissible class I mismatches in T-celldepleted allogeneic bone marrow transplantation. Br J Haematol. 2000;111(2):685-694. 38. Joris MM, Lankester AC, von dem Borne PA, et al. Translating in vitro prediction of cytotoxic T cell alloreactivity to hematopoietic stem cell transplantation outcome. Transpl Immunol. 2014;30(2-3):59-64. 39. Bettens F, Schanz U, Tiercy JM. Lack of recognition of HLA class I mismatches outside alpha1/alpha2 domains by CD8+ alloreactive T lymphocytes: the HLA-B44 paradigm. Tissue Antigens. 2013;81(6):414418. 40. Crocchiolo R, Zino E, Vago L, et al. Nonpermissive HLA-DPB1 disparity is a significant independent risk factor for mortality after unrelated hematopoietic stem cell transplantation. Blood. 2009;114(7):14371444. 41. Kawase T, Matsuo K, Kashiwase K, et al. HLA mismatch combinations associated with decreased risk of relapse: implications for the molecular mechanism. Blood. 2009;113(12):2851-2858. 42. Pidala J, Wang T, Haagenson M, et al. Amino acid substitution at peptide-binding pockets of HLA class I molecules increases risk of severe acute GVHD and mortality. Blood. 2013;122(22):3651-3658. 43. Ferrara GB, Bacigalupo A, Lamparelli T, et al. Bone marrow transplantation from unrelated donors: the impact of mismatches with substitutions at position 116 of the human leukocyte antigen class I heavy chain. Blood. 2001;98(10):3150-3155. 44. Morishima S, Kashiwase K, Matsuo K, et al. High risk HLA alleles for severe acute graftversus-host disease and mortality in unrelated donor bone marrow transplantation. Haematologica. 2016;101(4):491-498. 45. Petersdorf EW, Gooley TA, Malkki M, et al. HLA-C expression levels define permissible mismatches in hematopoietic cell transplantation. Blood. 2014;124(26):3996-4003. 46. Joris MM, Lankester AC, von dem Borne PA, et al. The impact of frequent HLA haplotypes in high linkage disequilibrium on donor search and clinical outcome after unrelated haematopoietic SCT. Bone Marrow Transplant. 2013;48(4):483-490. 47. Morishima S, Ogawa S, Matsubara A, et al. Impact of highly conserved HLA haplotype on acute graft-versus-host disease. Blood. 2010;115(23):4664-4670. 48. Petersdorf EW, Malkki M, Gooley TA, et al. MHC-resident variation affects risks after unrelated donor hematopoietic cell transplantation. Sci Transl Med. 2012;4(144): 144ra101. 49. Petersdorf EW, Malkki M, Horowitz MM, Spellman SR, Haagenson MD, Wang T. Mapping MHC haplotype effects in unrelated donor hematopoietic cell transplantation. Blood. 2013;121(10):1896-1905. 50. Petersdorf EW, Malkki M, O'HUigin C, et al. High HLA-DP expression and graft-versushost disease. N Engl J Med. 2015;373(7): 599609. 51. Ballen KK, Koreth J, Chen YB, Dey BR, Spitzer TR. Selection of optimal alternative graft source: mismatched unrelated donor, umbilical cord blood, or haploidentical transplant. Blood. 2012;119(9):1972-1980.

52. Luznik L, O'Donnell PV, Symons HJ, et al. HLA-haploidentical bone marrow transplantation for hematologic malignancies using nonmyeloablative conditioning and highdose, posttransplantation cyclophosphamide. Biol Blood Marrow Transplant. 2008;14(6):641-650. 53. Fuchs EJ. Human leukocyte antigen-haploidentical stem cell transplantation using Tcell-replete bone marrow grafts. Curr Opin Hematol. 2012;19(6):440-447. 54. Ciurea SO, Thall PF, Milton DR, et al. Complement-binding donor-specific antiHLA antibodies and risk of primary graft failure in hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2015;21(8):1392-1398. 55. Yoshihara S, Taniguchi K, Ogawa H, Saji H. The role of HLA antibodies in allogeneic SCT: is the 'type-and-screen' strategy necessary not only for blood type but also for HLA? Bone Marrow Transplant. 2012;47(12):1499-1506. 56. Locatelli F, Moretta F, Brescia L, Merli P. Natural killer cells in the treatment of highrisk acute leukaemia. Semin Immunol. 2014;26(2):173-179. 57. Ruggeri L, Parisi S, Urbani E, Curti A. Alloreactive natural killer cells for the treatment of acute myeloid leukemia: from stem cell transplantation to adoptive immunotherapy. Front Immunol. 2015;6: 479. 58. Venstrom JM, Pittari G, Gooley TA, et al. HLA-C-dependent prevention of leukemia relapse by donor activating KIR2DS1. N Engl J Med. 2012;367(9):805-816. 59. Cooley S, Trachtenberg E, Bergemann TL, et al. Donors with group B KIR haplotypes improve relapse-free survival after unrelated hematopoietic cell transplantation for acute myelogenous leukemia. Blood. 2009;113(3): 726-732. 60. Mancusi A, Ruggeri L, Urbani E, et al. Haploidentical hematopoietic transplantation from KIR ligand-mismatched donors with activating KIRs reduces nonrelapse mortality. Blood. 2015;125(20):3173-3182. 61. Bari R, Rujkijyanont P, Sullivan E, et al. Effect of donor KIR2DL1 allelic polymorphism on the outcome of pediatric allogeneic hematopoietic stem-cell transplantation. J Clin Oncol. 2013;31(30):3782-3790. 62. Heemskerk MB, van Walraven SM, Cornelissen JJ, et al. How to improve the search for an unrelated haematopoietic stem cell donor. Faster is better than more! Bone Marrow Transplant. 2005;35(7):645-652. 63. Querol S, Mufti GJ, Marsh SG, et al. Cord blood stem cells for hematopoietic stem cell transplantation in the UK: how big should the bank be? Haematologica. 2009;94(4): 536-541. 64. Rosenmayr A, Pointner-Prager M, Winkler M, et al. The austrian bone marrow donor registry: providing patients in austria with unrelated donors for transplant - a worldwide cooperation. Transfus med hemother. 2011;38(5):292-299. 65. Testi M, Andreani M, Locatelli F, et al. Influence of the HLA characteristics of Italian patients on donor search outcome in unrelated hematopoietic stem cell transplantation. Tissue Antigens. 2014;84(2):198-205. 66. Grubic Z, Jankovic KS, Maskalan M, et al. HLA allele and haplotype polymorphisms among Croatian patients in an unrelated hematopoietic stem cell donor search program. Transpl Immunol. 2014;31(3):119124.

687

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Red Cell Biology & its Disorders

Ferrata Storti Foundation

The LSD1 inhibitor RN-1 recapitulates the fetal pattern of hemoglobin synthesis in baboons (P. anubis) Angela Rivers,1,2 Kestis Vaitkus,2,3 Vinzon Ibanez,2,3 Maria Armila Ruiz,2,3 Ramasamy Jagadeeswaran,1,2 Yogen Saunthararajah,4 Shuaiying Cui,5 James D. Engel,5 Joseph DeSimone,2,3 and Donald Lavelle2,3

Haematologica 2016 Volume 101(6):688-697

1 Department of Pediatrics, University of Illinois at Chicago, Chicago, IL; 2Jesse Brown VA Medical Center, Chicago, IL; 3Department of Medicine, University of Illinois at Chicago, Chicago, IL; 4Department of Hematology and Oncology, Taussig Cancer Institute, Cleveland Clinic, Cleveland, OH; and 5Department of Cell and Developmental Biology, University of Michigan Medical School, Ann Arbor, MI, USA

ABSTRACT

Correspondence: dlavelle@uic.edu

Received: December 11, 2015. Accepted: February 3, 2016. Pre-published: February 8, 2016. doi:10.3324/haematol.2015.140749

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/688

©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to the Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. Permission in writing from the publisher is required for any other use.

688

ncreased fetal hemoglobin levels lessen the severity of symptoms and increase the lifespan of patients with sickle cell disease. Hydroxyurea, the only drug currently approved for the treatment of sickle cell disease, is not effective in a large proportion of patients and therefore new pharmacological agents that increase fetal hemoglobin levels have long been sought. Recent studies identifying LSD-1 as a repressor of γ-globin expression led to experiments demonstrating that the LSD-1 inhibitor RN-1 increased γ-globin expression in the sickle cell mouse model. Because the arrangement and developmental stage-specific expression pattern of the β-like globin genes is highly conserved between man and baboon, the baboon model remains the best predictor of activity of fetal hemoglobin-inducing agents in man. In this report, we demonstrate that RN-1 increases γ-globin synthesis, fetal hemoglobin, and F cells to high levels in both anemic and non-anemic baboons with activity comparable to decitabine, the most potent fetal hemoglobin-inducing agent known. RN-1 not only restores high levels of fetal hemoglobin but causes the individual 5’ Iγ- and 3’ Vγ -globin chains to be synthesized in the ratio characteristic of fetal development. Increased fetal hemoglobin was associated with increased levels of acetylated Histone H3, H3K4Me2, H3K4Me3, and RNA polymerase II at the γ-globin gene, and diminished γ-globin promoter DNA methylation. RN-1 is likely to induce clinically relevant levels of fetal hemoglobin in patients with sickle cell disease, although careful titration of the dose may be required to minimize myelotoxicity.

Introduction The term “haemoglobin switching” describes the sequential, highly regulated pattern of expression of the α- and β-like globin genes during development.1 In humans, the e-globin gene is expressed during the first eight weeks of gestation, followed by high level expression of the duplicated γ-globin genes during the fetal period. The γ-globin genes are expressed at very low levels (<1%) in the adult stage when expression of the β-globin gene predominates. A secondary level of developmental regulation characterizes expression of the γ-globin genes. The duplicated γglobin genes can be distinguished by an amino acid difference at aa136 where the 5’ Gγ-globin gene contains glycine while the 3’ Aγ-globin gene contains alanine (Figure 1). During the fetal period the Gγ- and Aγ genes are expressed in a ratio of 7:3, but in adult life this ratio is 2:3. The baboon (P. anubis) is an important animal model for studies of fetal hemoglobin regulation because the structure and developmental regulation of the β-globin gene complex is nearly identical to man.2,3 In haematologica | 2016; 101(6)

The LSD-1 inhibitor RN-1 increases HbF in baboons the baboon the duplicated γ-globin genes are designated Iγ and Vγ because their amino acid sequence differs by the presence of an Ile or Val at aa75.4 The ratio of Iγ /Vγ-globin expression is more than 9:5 in early gestational age (5060dpc) fetuses, 3:2 at birth and 1:2 in the adult.4,5 Therefore, the developmental pattern of expression of the baboon Iγ- and Vγ-globin genes is very similar to the human Gγ- and Aγ-globin genes (Figure 1). In sickle cell disease (SCD), the substitution of glutamic acid for valine at the sixth amino acid of the β-globin protein leads to the formation of abnormal hemoglobin S (HbS).6 Following deoxygenation in red blood cells (RBCs), HbS forms polymers causing the RBCs to become deformed and adherent leading to painful vaso-occlusive events and organ injury. In vitro studies have shown that both fetal hemoglobin (HbF) (α2γ2) tetramers and (α2βSγ) tetramers inhibit HbS polymerization.7 Because increased levels of HbF lessen the severity of symptoms and increase the life expectancy of patients with SCD, therapeutic approaches to increase HbF levels would be highly desirable.8,9 While many drugs increase HbF in cultured erythroid cells, only a few have been shown to increase HbF in vivo. Studies in simian primate animal models that exhibit conservation of developmental globin gene regulation,10 such as baboons, are considered one of the best tests of the activity of HbF-inducing drugs for therapies in patients designed to alleviate the symptoms of SCD and β-thalassemia.11 Early studies in anemic baboons showed that treatment with the DNA methyltransferase inhibitor 5-azacytidine increased HbF to levels predicted to be therapeutic in patients.12 Later studies confirmed that 5-azacytidine and its closely related analog decitabine did indeed

increase HbF to therapeutic levels in patients with SCD and β-thalassemia.13-18 Based on experiments conducted in baboons, a phase I trial is currently in progress to test the safety and HbF-inducing activity of oral administration of decitabine in combination with the cytidine deaminase inhibitor tetrahydrouridine in SCD patients.19 Hydroxurea (HU), the first drug approved for the treatment of SCD, was initially shown to increase HbF levels in monkeys20 and subsequent clinical trials also confirmed that HU increased HbF levels in patients with SCD.21 Because a large percentage of patients do not respond to HU, developing new drugs that increase HbF levels remains a priority. The TR2/TR4 heterodimer and Bcl11A have been identified as γ-globin gene repressors that act by recruiting corepressor complexes that establish repressive epigenetic modifications at specific sites to inactivate gene expression.22 Recognition of the role of Direct Repeat (DR) elements in the e- and γ-globin gene promoters in repression of these respective genes led to the identification of the TR2 and TR4 non-steroidal nuclear receptor proteins that specifically bind these elements and subsequently recruit a multi-protein co-repressor complex that includes DNA methyltransferase 1 (DNMT1), lysine specific demethylase 1 (LSD1), and histone deacetylases (HDACs) to repress gene expression.23-25 Both DNMT1 and LSD1 have also been identified as components of co-repressor complexes also recruited by Bcl11a.26,27 LSD1, the first histone demethylase identified, demethylates lysine 4 (H3K4) and lysine 9 (H3K9) residues of histone H3 and represses gene expression during hematopoietic differentiation.28-30 Tranylcypromine (TCP), an LSD-1 inhibitor, induced HbF production in cultured human hematopoietic erythroid

Figure 1. Amino acid substitutions in the baboon and human duplicated γ-globin genes and the difference in their expression at birth and in adults. haematologica | 2016; 101(6)

689

A. Rivers et al.

progenitors and transgenic mice containing the entire human β-globin gene complex in the context of a yeast artificial chromosome (β-YAC mice).31 Two recent studies have shown that RN-1, a more potent and specific LSD-1 inhibitor than TCP, induced γ-globin expression in human and baboon erythroid progenitors and a sickle cell mouse model.32,33 In this study we have tested the effect of RN-1 in baboons and confirm that it is a highly potent in vivo inducer of HbF synthesis.

subcutaneous injection. All procedures were approved by the Animal Care Committee of the University of Illinois at Chicago.

F cells and F reticulocytes Levels of F cells and F reticulocytes were analyzed by flow cytometry using a Cytomics FC500 (Beckman Coulter) after staining with thiazole orange and PE-conjugated anti-HbF (BD Bioscience).

HbF and globin chain synthesis Methods Baboons Baboons were housed at the University of Illinois at Chicago Biologic Resources Laboratory (UIC BRL) under conditions that meet the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) standards. Bone marrow aspirations were performed from the hips of animals under ketamine/xylazine anesthesia (10 mg/kg; 1 mg/kg). Prior to bone marrow sampling, Buprenex (0.01 mg/kg IM) was given and later in the afternoon a second dose of Buprenex (0.01 mg/kg IM) was administered to alleviate potential pain and suffering. The effect of RN-1 was tested in both anemic and non-anemic baboons. Anemia was induced by repeated phlebotomies during a 2-week period prior to administration of drug to attain a hematocrit (HCT) of 20 and animals were maintained at this HCT during the course of the experiment by periodic phlebotomies. Baboons were subjected to ketamine/midazolam (10 mg/kg; 3-5 mg/kg) anesthesia to allow removal of sufficient volumes of blood (15% of body weight) to induce and maintain anemia. RN-1 was dissolved in phosphate-buffered saline and passed through a 0.45 micron filter prior to administration by

HbF levels in peripheral blood were determined by alkali denaturation.34 Measurement of globin chain synthesis in peripheral blood reticulocytes was performed by biosynthetic radiolabeling of globin chains in the presence of [3H] leucine.35 For non-anemic animals, a fraction enriched in reticulocytes was obtained by centrifugation of 7 mL whole blood on Percoll step gradients as previously described.36 Globin chain separation was achieved by highperformance liquid chromatography (HPLC) as previously described (Online Supplementary Methods).37

RNA analysis Real-time PCR assays were performed as previously described38 to determine levels of globin transcripts (Online Supplementary Methods).

Bisulfite sequence analysis For bisulfite sequence analysis, DNA was purified from washed cell pellets using QIAmp Blood DNA Minikits. Bisulfite modification was performed using Epitect Bisulfite kits (Qiagen #59104) according to the manufactuer's instructions. Amplification of bisulfite modified DNA was performed as previously described (Online Supplementary Methods).35

Figure 2. Effect of RN-1 on hemoglobin F in anemic baboons. Three individual anemic baboons were treated with varying doses of RN-1 as indicated by subcutaneous injection (arrows correspond to treatment days). HbF levels were determined by alkali denaturation on the indicated days. 690

haematologica | 2016; 101(6)

The LSD-1 inhibitor RN-1 increases HbF in baboons

Chromatin immunoprecipitation analysis Chromatin immunoprecipitation analysis was performed as previously described39 with modifications (Online Supplementary Appendix).

Statistics analysis Student's t-test was used for all the statistical analyses. Error bars represent standard deviation.

Results Effect of RN-1 in anemic baboons A series of dose-response experiments were conducted in anemic baboons to investigate the effect of RN-1 on HbF levels, F cells, and F reticulocytes. Anemic baboons were treated with five different doses (0.125, 0.20, 0.25, 0.50, and 2.5 mg/kg/d) of RN-1 administered subcutaneously for 4-5d (Online Supplementary Table S2, Exp. 1-5). The highest dose of RN-1 (2.5 mg/kg) increased HbF from 5.8% on the first day of drug treatment to 27.3% 7d after the last injection (Figure 2 and Online Supplementary Table S2, Exp. 1). Peak γ-globin chain synthesis (0.78 γ/γ+β) was observed 3d following the last day of drug administration demonstrating a near-complete reversion to fetal-stage synthesis (Figure 3). At this time the ratio of chain synthesis of the 5’ Iγ gene and the 3’ Vγ gene (Iγ/Vγ ratio) was more than 2, characteristic of the fetal stage. F reticulocytes increased from 34.6% to 92.3% while F cells increased from 28.5% to 59.2% (Figure 4 and Online Supplementary Table S2). Neutropenia was observed with an absolute neutrophil count (ANC) nadir of 29x10/9L 10d after the last RN-1 dose (Figure 5 and Online Supplementary Table S2). A second animal (PA8549) was treated with a reduced dose of RN-1 (0.5 mg/kg/d). The HbF level

increased from 2.4% on the first day of drug treatment to 29.4% 6d after the last injection (Figure 2 and Online Supplementary Table S2, Exp. 2). Peak γ-globin chain synthesis (0.68 γ/γ+β) was attained three days after the last RN-1 treatment (Figure 3). The Iγ/Vγ synthesis ratio at this time was more than 2, characteristic of fetal-stage synthesis. F cells increased from 19.1% to 60.8% (Figure 4 and Online Supplementary Table S2) and F reticulocytes from 33.9% to 97.5% (Figure 4 and Online Supplementary Table S2). Neutropenia was observed with an ANC nadir of 0.57x10/9L 10d after the last RN-1 dose (Figure 5 and Online Supplementary Table S2). A third animal (PA8000) was treated with a reduced RN-1 dose (0.25 mg/kg/d). The HbF increased from 4.1% on the first day of treatment to 20.6% 6d after the last injection (Figure 2 and Online Supplementary Table S2, Exp. 3). Peak γ-globin chain synthesis (0.68 γ/γ+β) was 0.49 and the Iγ/Vγ chain ratio was 1.18 (Figure 3). F cells increased from 20.3% to 47.0% and F reticulocytes from 45.7% to 80.7% (Figure 4 and Online Supplementary Table S2). The decrease in ANC (nadir=0.87x10/9L) was less severe than at the higher doses (Figure 4 and Online Supplementary Table S2). No changes in total hemoglobin production (α/γ+β=1.06+0.02) were observed in these 3 animals treated with the highest doses of RN-1. In a fourth experiment, PA8548 was re-treated with a reduced RN-1 dose (0.2 mg/kg/d). HbF increased from 3.7% on the first day of treatment to 20.5% six days post treatment (Figure 2 and Online Supplementary Table S2, Exp. 4). F-cells increased from 36.9% to 54.8% and F reticulocytes from 36.9% to 89.2% (Figure 4 and Online Supplementary Table S2). Peak γ-globin chain synthesis (γ/γ+β) was 0.52 (Online Supplementary Table S2). Neutropenia (ANC nadir=0.8x10/9L) was observed (Figure 5 and Online Supplementary Table S2). Another animal

Figure 3. Effect of RN-1 on globin chain synthesis in anemic baboons. Globin chain synthesis in reticulocytes of 3 baboons pre and post treatment with varying doses of RN-1 was measured by biosynthetic radiolabeling with [3H] leucine followed by HPLC separation of globin chains. haematologica | 2016; 101(6)

691

A. Rivers et al.

(PA8001) was treated with a reduced RN-1 dose (0.125 mg/kg/d) for 5d. HbF increased from 1.7% pre treatment to 4.8% post treatment, F-cells increased from 13.5% to 21.8% and F reticulocytes from 22.9% to 31.7% (Figure 4 and Online Supplementary Table S2). Neutropenia was not observed in this animal (Figure 4 and Online Supplementary Table S2). These experiments defined a linear dose relationship between HbF, F cells, and F reticulocytes and RN-1 dose within the 0.125 to 0.5 mg/kg range (Figure 4). Dose-related decreases in ANC were also observed (Figure 5 and Online Supplementary Figure S1). Dose-dependent decreases in platelets 8-9d following the initial RN-1 treatment were followed by thrombocytosis peaking 15-20d after the initial dose (Figure 5 and Online Supplementary Figure S1). In addition, dose-dependent increases in monocytes were observed that peaked 15-17d following the initial drug dose (Figure 5 and Online Supplementary Figure S1). Increases in mean corpuscular volume (MCV) were observed at the higher RN-1 doses coinciding with the day of maximal HbF levels (Online Supplementary Figure S2A) and mean corpuscular hemoglobin concentration (MCHC) levels were decreased at all doses (Online Supplementary Figure S2B). Significant increases in HbF were observed at lower doses of RN-1 that were not associated with large increases in MCV (Online Supplementary Figure S2C). In an effort to achieve significant increases in HbF in the absence of neutropenia, PA8549 was re-treated with a dose of 0.125 mg/kg for 10d (Online Supplementary Table S2, Exp. 6). The HbF increased from 3.6% on the first day of treatment to 16% 5d after the last injection.

F cells increased from 17.6% to 42.7% and observed, however, F reticulocytes from 21% to 70%. Peak γ-globin chain synthesis (γ/γ+β) was 0.22. Decreases in ANC were still observed however, with an ANC nadir of 0.87x10/9L. Two other experiments were performed to test the effects of a 2d per week dose schedule. When PA8000 was treated with 0.25 mg/kg (Online Supplementary Table S2, Exp. 7), HbF increased from 4.7% to 10.4%, F cells from 18.2% to 28.1%, and F reticulocytes from 29.2 to 52.2% associated with an ANC nadir of 2.36x10/9L, indicating that significant increases in HbF could be attained with this schedule and dose in the absence of neutrophil toxicity. When the dose was increased to 0.5 mg/kg at the 2d per week dose schedule in another animal (PA8549) HbF increased from 6.2% to 22.1%, F cells from 22.2% to 53.6%, and F reticulocytes from 24.1 to 86.4%, however an ANC nadir of 0.57x10/9L was observed (Online Supplementary Table S2, Exp. 8). The effect of RN-1 on globin mRNAs was determined in pre-and post-treatment peripheral blood samples in 3 animals treated with different doses (Online Supplementary Table S3). RN-1 treatment increased the level of γ-globin mRNA nearly 5-fold (P<0.05) in post-treatment samples (0.39+0.12 γ/e+γ+β) compared to pre-treatment samples (0.08+0.01 γ/e+γ+β), while levels of e-globin mRNA remained low.

Effect of RN-1 on histone modifications To investigate the mechanism of action of RN-1, chromatin immunoprecipitation (ChIP) assays were performed to determine the effect of the drug on histone acetylation

Figure 4. Effect of RN-1 on HbF, F cells, and F reticulocytes. Five baboons were treated with varying doses of RN-1 (0.125-2.5 mg/kg/d) for five days (days 1-5). (Left) Changes in HbF, F cells, and F reticulocytes in treated baboons. (Right) Relationship between the change in HbF, F cells, and F reticulocytes and RN-1 dose (maximal value post-treatment - pre-treatment value). 692

haematologica | 2016; 101(6)

The LSD-1 inhibitor RN-1 increases HbF in baboons and methylation associated with the e-, γ-, and β-globin genes. For these experiments, anemic baboons (n=4) were treated for 2d with RN-1 (0.5 or 0.25 mg/kg/d). Bone marrow (BM) erythroid cells were purified by immunomagnetic column selection of bone marrow samples obtained pre-treatment from bled baboons and post-treatment 4872 h after the last dose of drug. Levels of pol II, histone H3K4Me2, H3K4Me3, H3K9Ac, and H3K9Me2 associated with the e-, γ-, and β-globin promoter and IVS II regions were measured by real-time PCR in DNA samples obtained from formaldehyde fixed, immunoprecipitated chromatin of purified BM erythroid cells (Figure 6A). Increased levels of pol II associated with the γ-globin promoter (4.7-fold; P<0.02) and γ-globin IVS II region (3.2fold; P<0.05) were observed. Levels of H3K9Ac were also increased at the γ-globin promoter (6.2-fold; P<0.01) and γ-globin IVS II regions (4.2-fold; P<0.02). Histone H3K4Me2 levels were increased at the γ-globin IVS II (2.1fold; P<0.05) but not at the γ-globin promoter while H3K4Me3 levels were increased at both the γ-globin promoter (4.7-fold; P<0.02) and at the γ-globin IVS II region (4.6-fold; P<0.02). No significant differences in the levels of these modifications at the e- and β-globin promoter and IVS regions were observed between pre- and post-treatment samples. Levels of Histone H3K9Me2 were increased at the β-globin promoter (2.8-fold; P<0.02) and IVS II region (5.3-fold; P<0.01) in post-treatment samples.

Effect of RN-1 on DNA methylation

Bisulfite sequence analysis of the baboon 5’ γ-globin promoter region in DNA isolated from erythroid precursors purified from BM aspirates obtained pre-and posttreatment from 7 baboons treated with varying doses of

RN-1 (0.125-0.5 mg/kg/d) was performed to investigate whether increased γ-globin expression in RN-1 treated baboons was associated with loss of DNA methylation. The level of DNA methylation of five CpG residues within this region located at -54, -51, +5, +16 and +48 with respect to the transcription start site in purified erythroid precursors of post-treatment BM samples (0.80+0.05% dmC) was reduced compared to pre-treatment samples (0.73+0.09% dmC; P=0.053). CpG residues at -51 and +16 are polymorphic in the baboon with G to A transitions eliminating the CpG site. Further analysis of only the 3 non-polymorphic CpG sites within the 5’ γ-globin promoter region at -54, +5, and -48 confirmed a small but significant (P<0.05) decrease in the level of DNA methylation in post-treatment (0.79+0.05% dmC) compared to pretreatment samples (0.71+0.09% dmC; P<0.05); (Figure 6B). Measurement of globin chain synthesis in peripheral blood reticulocytes in 6 of these baboons showed increased γ-globin synthesis in post-treatment (0.032+0.15 γ/γ+β) compared to pre-treatment samples (0.09+0.04 γ/γ+β; P<0.01); (Figure 6B).

Effect of RN-1 in non-anemic baboons

To investigate whether the effect of RN-1 on γ-globin expression required erythroid expansion, additional experiments were conducted in normal, non-anemic baboons. In the first experiment, treatment of PA8549 with RN-1 (0.5 mg/kg/d; 5d) increased HbF synthesis in peripheral blood reticulocytes from 0.03 γ/γ+β pre treatment to 0.71 post treatment (Figure 7 and Online Supplementary Table S4, Exp. 9). Neutropenia was not observed. In the second experiment a reduced dose of RN-1 (0.25 mg/kg/d; 5d) administered to PA8549 increased γ-globin synthesis in

Figure 5. Effect of RN-1 on hematologic parameters. Five baboons were treated with varying doses of RN-1 (0.125-2.5 mg/kg/d) for five days (days 1-5). Changes in neutrophils (ANC), monocytes, platelets (Plt), and reticulocytes are shown. haematologica | 2016; 101(6)

693

A. Rivers et al. peripheral blood reticulocytes from 0.015 γ/γ+β pre-treatment to 0.26 post-treatment (Figure 7 and Online Supplementary Table S4, Exp. 10) while again neutropenia was not observed. In the third experiment this same RN-1 dose (0.25 mg/kg/d; 5d/wk) given for two weeks to PA8698 increased γ-globin synthesis to 0.24 γ/γ+β (Figure 7 and Online Supplementary Table S4, Exp. 11). F cells increased more than 4-fold from 2.0% pre treatment to 9.2% post treatment. A transient decrease in ANC to less than 1.5x109/L (1.21) was observed. To investigate the effect of a longer term treatment schedule, PA8696 was treated with RN-1 (0.2-0.25 mg/kg/d) for a period of ten weeks (Online Supplementary Table S4, Exp. 12) while 2 other animals (PA8695 and P8698, Online Supplementary Table S4, Exp. 13 and 14) were treated with 0.25 mg/kg/d RN-1 for six weeks. Increased F cells were observed in all animals and ranged from 5- to 8-fold (Figure 8 and Online Supplementary Table S4). HbF levels increased 3-4-fold in all animals (Figure 8 and Online Supplementary Table S4). Within the F-cell populations of the 3 baboons, the HbF levels rose more than 3fold (Online Supplementary Figure S3). Importantly, these increases in HbF levels and F cells were associated with minimal neutropenia that resolved quickly (Online Supplementary Table S4 and Figure S4). In all 3 non-anemic

animals treated for either ten or six weeks, increases in monocytes and decreases in platelets were also observed (Online Supplementary Figure S4). Treatment of these nonanemic animals was not associated with changes in MCV or in MCHC (Online Supplementary Figure S5).

Discussion Treatment of baboons with the LSD-1 inhibitor RN-1 substantially increased HbF, F cells and γ-globin synthesis in peripheral blood reticulocytes. The results confirm and extend our previous studies in the sickle cell mouse model showing that RN-1 increased F cells, F reticulocytes, and γ-globin mRNA to levels similar to decitabine, the most potent HbF-inducer known.32,33 However, both these drugs elicit only small increases of γ-globin expression in the SCD mouse model compared to the magnitude observed in baboons, emphasizing the importance of using nonhuman primate, such as baboons, to test the effect of HbFinducing drugs. Comparison of the level of γ-globin synthesis in RN-1treated baboons with previous results from our laboratory of baboons treated with DNMT inhibitors 3-5,12,38,40-42

Figure 6. Effect of RN-1 on histone modifications and γ-globin promoter DNA methylation. (A) ChIP analysis of levels of pol II, Histone H3K4Me2, H3K4Me3, H3K9Ac, and H3K9Me2 associated with the e-, γ-, and β-globin promoters and IVS regions in purified bone marrow (BM) erythroid cells from RN-1 treated baboon pre and post treatment. Anemic baboons were treated with 0.5 mg/kg/d RN-1 for 2d and BM aspirates were obtained 48 h following the last RN-1 treatment. Asterisks (*) designate significant differences (P<0.05) between pre- and post-treatment samples. (B) (Left) Levels of cytosine methylation at three non-polymorphic CPG residues in the 5’ γ-globin promoter region in purified BM erythroid cells from RN-1 treated baboons pre and post treatment. Paired pre and post treatment samples from individual baboons are shown. The horizontal bar shows the mean % dmC value of all samples. Anemic baboons were treated with either 0.5 mg/kg/d (filled triangles, filled squares, filled circles, filled diamonds, lined squares) or 0.25 mg/kg/d (open triangles, open circles). (Right) Changes in γ-globin synthesis in peripheral blood reticulocytes of baboons pre and post treatment. Paired pre and post treatment samples from each individual are shown. The horizontal bar shows the mean level of γ-globin synthesis (γ/γ+β) of all samples. 694

haematologica | 2016; 101(6)

The LSD-1 inhibitor RN-1 increases HbF in baboons

shows that the HbF-inducing activity of RN-1 in the baboon model is comparable to DNMT inhibitors. In addition, RN-1 treatment of baboons resulted in the complete reversion of the pattern of synthesis of the individual 5’ Iγ - and 3’ Vγ-globin genes (Iγ/Vγ ratio) to that characteristic of fetal development in anemic baboons, an effect only partially achieved by DNMT1 inhibitors.4 The preferential effect of RN-1 on expression of the 5’ Iγ-globin gene was even greater in the non-anemic animals. This effect has not been previously observed and appears to be a property unique to RN-1. The induction of HbF in the absence of changes in total β-like globin production (α/β ratio) and MCV in the treated animals suggests that increased γ-globin expression results from a direct effect of the drug on the repressive chromatin environment of the γ-globin gene, consistent with previous mechanistic studies, demonstrating a critical role for LSD1 in mediating γ-globin repression,31 although effects elicited through alteration of erythroid differentiation kinetics cannot be excluded. Further experiments will be required to determine the mechanism of action of RN-1 in vivo. ChIP experiments showing RN-1 increased levels of H3K4Me2, H3K4Me3 and H3K9ac associated with the γ-globin but not the e-globin gene suggest a specific effect of the drug on γ-globin expression and are consistent with previous studies in AML cell lines showing that LSD1 inhibition did not induce genome-wide changes in H3K4Me2 but rather acted only at a specific subset of genes.43 The lack of effect of RN-1 on e-globin expression in the baboon contrasts with increases in Ey and βh1 observed in the

SCD mouse model. It is possible that this difference is dose-related and that the e-globin gene is less sensitive to de-repression by RN-1 because the e-globin promoter contains two DR elements that bind DRED with higher affinity than the single DR1 element in the γ-globin promoter.2224 LSD1 demethylates other proteins including DNMT1. Demethylation of DNMT1 by LSD1 increases DNMT1 stability while targeted deletion of LSD1 results in loss of DNMT1 and DNA demethylation in ES cells.44 In this regard, we observed small effects on γ-globin promoter DNA methylation in BM erythroid cells from RN-1-treated baboons that were significantly less than changes previously observed in decitabine-treated baboons. DNA methylation changes are, therefore, not likely to be the major mechanism responsible for increased HbF levels in RN-1 treated baboons, although these small changes in DNA methylation could have contributed, in part, to the magnitude of the HbF response. Because RN-1 is not covalently incorporated into DNA (as are decitabine and 5-azacytidine), it is potentially a less genotoxic and safer drug. The clinical usefulness of RN-1 will likely depend on the design of a dose and schedule that limit its undesirable effects on other hematopoietic lineages. The anemic baboon was initially developed and has been traditionally used to model the anemia and reticulocytosis observed in SCD.12 In this model, anemia produced by repeated phlebotomies results in a near-complete replacement of the erythrocyte population within a 2-week period, removal of all formed elements, and alterations in levels of multiple cytokines in addition to ery-

Figure 7. Effect of RN-1 on globin chain synthesis in non-anemic baboons. HPLC analysis of globin chain synthesis in peripheral blood reticulocytes of normal, non-anemic baboons treated with varying doses of RN-1 pre treatment and post treatment. (Top) Effect of RN-1 on globin chain synthesis in peripheral blood reticulocytes of PA8549. This baboon was treated in two separate experiments for five days with RN-1 (0.5 mg/kg/d or 0.25 mg/kg/d). Pre-treatment and post-treatment values at each dose are shown. (Bottom) Effect of RN-1 on globin chain synthesis of peripheral blood reticulocytes of PA8698. The baboon was treated with RN-1 (0.25 mg/kg/d; 10d). Analysis of chain synthesis pre treatment and following five days and ten days of treatment are shown. haematologica | 2016; 101(6)

695

A. Rivers et al.

Figure 8. Effect of RN-1 on HbF and F cells in non-anemic baboons treated with RN-1. (A) Effect of RN-1 on HbF (left) and F cells (right) in 3 non-anemic baboons. PA8696 was treated for ten weeks (0.2-0.25 mg/kg/d; 5d/wk; sc). PA8698 and PA8695 were treated for six weeks (0.25 mg/kg/d; 5d/week/ sc). (B) Flow cytometric analysis of F cells and F reticulocytes on d1 and d70 following RN-1 treatment in a non-anemic baboon (PA8696).

thropoietin.45 To complement studies in anemic baboons, experiments were also performed in non-anemic animals where base-line hematopoiesis was not perturbed. The neutropenia observed in phlebotomized baboons was lessened in non-anemic animals where ANC over 1.5x109/L were generally maintained. The greater decreases in ANC in anemic animals may be due to depletion of neutrophils by repeated phlebotomies or the effect of these phlebotomies on growth factor levels. Decreased platelets and increased monocyte counts were observed in both the anemic and non-anemic animals, and are likely due to effects of the drug on hematopoietic differentiation. In an LSD1 conditional knockdown mouse model, expansion of granulocytic, erythroid, and megakaryocytic progenitors was observed, while granulopoiesis, erythropoiesis, and platelet production was inhibited, suggesting that the similar effects observed in RN-1-treated baboons are due to inhibition of LSD1 activity.30 RN-1 has been reported to affect long-term memory in mice and, therefore, this may be an additional possible adverse effect of the drug.46 Our results show that the LSD1 inhibitor RN-1 induces high levels of HbF in baboons and are consistent with pre-

References 1. Stamatoyannopoulos G. Molecular and cellular basis of hemoglobin switching. In: Steinberg MH, Forget BG, Higgs DR, Nagel RL, eds. Disorders of Hemoglobin: Genetics, Pathophysiology, and Clinical Management. Cambridge: Cambridge University Press; 2001, p. 131-145. 2. Barrie PA, Jeffreys AJ, Scott AF. Evolution

696

vious studies showing that LSD1 plays a critical role in γ-globin silencing as a component of the DRED complex.25,31 However, the possibility of a role for perturbation of erythroid differentiation in the mechanism of HbF reactivation cannot be excluded, and the in vivo mechanism of action will be investigated in future studies. The level of the HbF response varies between different baboon species,47 and, therefore, the HbF response in man and baboons may differ somewhat due to differences between species and/or physiological differences between the experimental animal model and patients with hemoglobinopathies. However, results in P. anubis are generally predictive of HbF-inducing effects in man and have been successfully translated in clinical studies in SCD and β-thalassemia patients.13-18 Therefore, we suggest that RN1 and/or other LSD1 inhibitors are excellent candidates for clinical evaluation as therapeutic agents in β-thalassemia and sickle cell disease. Funding This work was supported by NIH U01 HL117658 and NIH R01 HL114561.

of the beta-globin gene cluster in man and primates. J Mol Biol. 1981;149(3):319-336. 3. DeSimone J, Mueller AL. Fetal hemoglobin synthesis in baboons. (Papio cynocephalus). J Lab Clin Med. 1979;91(6):862871. 4. Schroeder WA, DeSimone J, Shelton JB, et al. Changes in the gamma chain heterogeneity of hemoglobin F of the baboon (Papio cynocephalus) postnatally and after

partial switching to hemoglobin F production by various stimuli. J Biol Chem. 1983; 258(5):3121-3125. 5. Chin J, Singh M, Banzon V, et al. Transcriptional activation of the gammaglobin gene in baboons treated with decitabine and in cultured erythroid progenitor cells involves different mechanisms. Exp Hematol. 2009;37(10):1131-1142. 6. Ingram VM. Gene mutations in human

haematologica | 2016; 101(6)

The LSD-1 inhibitor RN-1 increases HbF in baboons

8. 9.

10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

hemoglobin: the chemical difference between normal and sickle hemoglobin. Nature. 1957;275(4581):238-240. Sunshine HR, Hofrichter J, Eaton WA. Requirement for therapeutic inhibition of sickle haemoglobin gelation. Nature. 1978;275(5677):238-240. Platt OS, Thorington BD, Brambilla DJ, et al. Pain in sickle cell disease. Rates and risk factors. N Engl J Med. 1991;325(1):11-16. 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. 1984;330(23):1639-1644. Tagle DA, Koop BF, Goodman M, et al. Embryonic epsilon and gamma globin genes of a prosimian primate (Galago crassicaudatus) Nucleotide and amino acid sequences, developmental regulation and phylogenetic footprints. J Mol Biol. 203(2):439-455. Schechter AN. Hemoglobin research and the origins of molecular biology. Blood. 2008;112(10):3927-3938. DeSimone J, Heller PH, Hall L, Zwiers D. 5azacytidine stimulates fetal hemoglobin synthesis in anemic baboons. Proc Natl Acad Sci USA. 1982;79(14):4428-4431. Ley TJ, DeSimone J, Noguchi CT, et al. 5-azacytidine increases gamma-globin synthesis and reduces the proportion of dense cells in patients with sickle cell anemia. Blood. 1983;62(2):370-380. Ley TJ, DeSimone J, Anagnou NP, et al. 5azacytidine selectively increases synthesis in a patient with beta+ thalassemia. N Engl J Med. 1982;307(24):1469-1475. Lowrey CH, Nienhuis AW. Brief report; treatment with azacitidine of patients with end-stage beta-thalassemia. N Engl J Med. 1993;329(12):845-848. Koshy M, Dorn L, Bressler L, et al. 2-deoxy 5-azacytidine and fetal hemoglobin induction in sickle cell anemia. Blood. 2000;96(7):2379-2384. DeSimone J, Koshy M, Dorn L, et al. Maintenance of elevated fetal hemoglobin levels by decitabine during dose interval treatment of sickle cell anemia. Blood. 2002;99(11):3905-3908. Saunthararajah Y, Hillery CA, Lavelle D, et al. Effects of 5-aza-2â&#x20AC;&#x2122;-deoxycytidine on fetal hemoglobin levels, red cell adhesion, and hematopoietic differentiation. Blood. 2003;102(12):3865-3870. Lavelle D, Vaitkus K, Ling Y, et al. Effects of tetrahydrouridine on pharmacokinetics and pharmacodynamics of oral decitabine. Blood. 2012;119(5): 1240-1247. Letvin NL, Linch DC, Beardsley GP, et al. Augmentation of fetal-hemoglobin production in anemic monkeys by hydroxyurea. N Engl J Med. 1984;310(14):869-873. Charache S, Terrin ML, Moore RD, et al. Effect of hydroxyurea on the frequency of painful crises in sickle cell anemia.

haematologica | 2016; 101(6)

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Investigators of the Multicenter Study of Hydroxyurea in Sickle Cell Anemia. N Engl J Med. 1995;332(20):1317-1322. Suzuki M, Yamamoto M, Engel JD. Fetal globin gene repressors as drug targets for molecular therapies to treat the beta-globinopathies. Mol Cell Biol. 2014; 34(19):3560-3569. Tanabe O, Katsuoka F, Campbell AD, et al. An embryonic/fetal beta-type globin gene repressor contains a nuclear receptor TR2/TR4 heterodimer. EMBO J. 2002; 21(13):3434-3442. Tanabe O, McPhee D, Kobayashi S, et al. Embryonic and fetal beta-globin gene repression by the orphan nuclear receptors, TR2 and TR4. EMBO J. 2007;26(9):22952306. Cui S, Kolodziej KE, Obara N, et al. Nuclear receptors TR2 and TR4 recruit multiple epigenetic transcriptional corepressors that associate specifically with the embryonic beta-type globin promoters in differentiated adult erythroid cells. Mol Cell Biol. 2011;31(16):3298-3311. Sankaran VG, Menne TF, Xu J, et al. Human fetal hemoglobin expression is regulated by the developmental stage-specific repressor BCL11A. Science. 2008;322 (5909):1839-1842. Xu J, Bauer DE, Kerenyi MA, et al. Corepressor-dependent silencing of fetal hemoglobin expression by BCL11A. Proc Natl Acad Sci USA. 2013;110(16):6518-6523. Amente S, Lania L, Majello B. The histone LSD1 demethylase in stemness and cancer transcription programs. Biochim Biophys Acta. 2013;1829(10):981-986. Kereni MA, Shao Z, Hsu YJ, et al. Histone demethylase Lsd1 represses hematopoietic stem and progenitor cell signatures during blood cell maturation. Elife. 2013;2:e00633. Sprussel A, Schulte JH, Weber S, et al. Lysine-specific demethylase 1 restricts hematopoietic stem and progenitor proliferation and is essential for terminal differentiation. Leukemia. 2012;26(9):20392051. Shi L, Cui S, Engel JD, Tanabe O. Lysinespecific demethylase 1 is a therapeutic target for fetal hemoglobin induction. Nat Med. 2013;19(3):291-304. Rivers A, Vaitkus K, Ruiz MA, et al. RN-1, a potent and selective lysine-specific demethylase 1 inhibitor, increases -globin expression, F reticulocytes, and F cells in a sickle cell disease mouse model. Exp Hematol. 2015;43(7):546-553. Cui S, Lim KC, Shi L, et al. The LSD1 inhibitor RN-1 induces fetal hemoglobin synthesis and reduces disease pathology in sickle cell mice. Blood. 2015;26(3):386-396. Singer K, Chernoff AI, Singer L. Studies on abnormal hemoglobins. I. Their demonstration in sickle cell anemia and other hematologic disorders by means of alkali

denaturation. Blood. 1951;6(5):413-428. 35. Banzon V, Ibanez V, Vaitkus K, et al. siDNMT1 increases gamma-globin expression in chemical inducer of dimerization (CID)-dependent mouse betaYAC bone marrow cells and in baboon erythroid progenitor cell cultures. Exp Hematol. 2011;39(1):26-36. 36. Alderman EM, Fudenberg HH, Lovins RE. Binding of immunoglobulin classes to subpopulations of human red blood cells separated by density-gradient centrifugation. Blood. 1980;55(5):817-822. 37. Leone L, Monteleone M, Gabiutti V, Amione C. Reversed-phase high-performance liquid chromatography of human haemoglobin chains. J Chromatogr. 1985; 321(2):407-419. 38. Akpan I, Banzon V, Ibanez V, Vaitkus K, DeSimone J, Lavelle D. Decitabine increases fetal hemoglobin in Papio Anubis by increasing gamma-globin gene transcription. Exp Hematol. 2010;38(11):989-993. 39. Lavelle D, Vaitkus K, Hankewych M, Singh M, DeSimone J. Developmental changes in DNA methylation and covalent histone modifications of chromatin associated with the epsilon-, gamma-, and beta-globin promoters in Papio anubis. Blood Cells Mol Dis. 2006;36(2):269-278. 40. Lavelle D, DeSimone J, Heller P, Zwiers D, Hall L. On the mechanism of HbF elevations in the baboon by erythropoietic stress and pharmacologic manipulation. Blood. 1986;67(4):1083-1089. 41. Lavelle D, Chin J, Vaitkus K, et al. Oral decitabine reactivates expression of the methylated gamma-globin gene in Papio anubis. Am J Hematol. 2007;82(11):981985. 42. Lavelle D, Saunthararajah Y, Vaitkus K, et al. S110, a novel decitabine dinucleotide, increases fetal hemoglobin levels in baboons (P. Anubis). J Transl Med. 2010; 8:92. 43. Schenk T, Chen WC, Gollner S, et al. Inhibition of the LSD1 (KDM1A) demethylase reactivates the all-trans-retinoic acid differentiation pathway in acute myeloid leukemia. Nat Med. 2012;18(4):605-611. 44. Wang J, Hevi S, Kurash JK, et al. The lysine demethylase LSD1 (KDM1) is required for maintenance of global DNA methylation. Nat Genet. 2009;41(1):125-129. 45. Abraham E. Effects of stress on cytokine production. Methods Achiev Exp Pathol. 1991;14:45-62. 46. Neelamegam R, Ricq EL, Malvaez M, et al. Brain-penetrant LSD1 inhibitors can block memory consolidation. ACS Chem Neurosci. 2012;3(2):120-128. 47. DeSimone J, Schroeder WA, Shelton JB, et al. Speciation in the baboon and its relation to gamma-chain heterogeneity and to the response to induction of HbF by 5-azacytidine. Blood. 1984;63(5):1088-1095.

697

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Platelet Biology & its Disorders

Ferrata Storti Foundation

A distinct plasmablast and naïve B-cell phenotype in primary immune thrombocytopenia Shaun M. Flint,1,2 Adele Gibson,1 Geoff Lucas,3 Raghava Nandigam,4 Louise Taylor,4 Drew Provan,4 Adrian C. Newland,5 Caroline O. Savage,1 and Robert B. Henderson1

Haematologica 2016 Volume 101(6):698-706

1 Immunoinflammation TAU, GSK, Stevenage; 2Department of Medicine, University of Cambridge; 3Histocompatibility and Immunogenetics Laboratory, NHS Blood & Transplant; 4Department of Haematology, Royal London Hospital, London; 5Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

ABSTRACT

Correspondence: robbie.b.henderson@gsk.com

Received: October 6, 2015. Accepted: March 8, 2016. Pre-published: March 11, 2016. doi:10.3324/haematol.2015.137273

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/698

©2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is requested for any other use.

698

rimary immune thrombocytopenia is an autoimmune disorder in which platelet destruction is a consequence of both B- and T-cell dysregulation. Flow cytometry was used to further characterize the B- and T-cell compartments in a cross-sectional cohort of 26 immune thrombocytopenia patients including antiplatelet antibody positive (n=14) and negative (n=12) patients exposed to a range of therapies, and a cohort of matched healthy volunteers. Markers for B-cell activating factor and its receptors, relevant B-cell activation markers (CD95 and CD21) and markers for CD4+ T-cell subsets, including circulating T-follicular helper-like cells, were included. Our results indicate that an expanded population of CD95+ naïve B cells correlated with disease activity in immune thrombocytopenia patients regardless of treatment status. A population of CD21- naïve B cells was specifically expanded in autoantibody-positive immune thrombocytopenia patients. Furthermore, the B-cell maturation antigen, a receptor for B-cell activating factor, was consistently and strongly up-regulated on plasmablasts from immune thrombocytopenia patients. These observations have parallels in other autoantibody-mediated diseases and suggest that loss of peripheral tolerance in naïve B cells may be an important component of immune thrombocytopenia pathogenesis. Moreover, the B-cell maturation antigen represents a potential target for plasma cell directed therapies in immune thrombocytopenia.

Introduction Primary immune thrombocytopenia (ITP) is a clinical diagnosis given to patients with an unexplained, prolonged isolated thrombocytopenia. ITP is a rare but chronic condition in adults and is associated with significant bleeding-related morbidity and mortality.1 The condition is characterized by both platelet destruction and impaired platelet production. A role for platelet-directed antibodies was established in the 1960s with transfer experiments showing that thrombocytopenia could be induced by transfer of the gamma-globulin fraction of ITP patient serum.2 Using the most sensitive assays, antibodies binding platelet membrane glycoproteins are present in approximately 50% of patients.3 The mechanism by which B-cell tolerance is lost is a subject for debate, but an elevated serum level of B-cell Activating Factor (BAFF) is likely to be an important contributing factor.4 BAFF drives B-cell maturation, promotes B-cell survival and augments immunoglobulin production by binding three surface B-cell receptors: BAFF receptor (BAFF-R), transmembrane activator and calcium modulator and cyclophilin ligand interactor (TACI), and B-cell maturation antigen (BCMA).5 An expanded CD95 (Fas receptor) positive population of haematologica | 2016; 101(6)

Immunophenotype of primary ITP

B cells has also been described in ITP and there are reports of fewer regulatory B cells, defined both as CD24hiCD38hi B cells and by IL-10 production.6,7 A modern view of ITP pathogenesis places these B-cell abnormalities within a complex network of abnormalities affecting multiple immune cell lineages. T cells, in particular, contribute to platelet destruction both by facilitating the production of class-switched, high affinity autoantibody and through B-cell independent mechanisms such as cell-mediated cytotoxicity directed against platelets.8 The latter may be the primary mechanism of disease in a subset of patients with no detectable anti-platelet antibodies.9 High-affinity autoantibody production is facilitated by T follicular helper cells (TFH), a subset recently reported to be expanded proportional to germinal center and plasma cell numbers within the spleens of ITP patients.10 This study sought to extend existing knowledge of immune dysregulation in ITP by performing detailed flow cytometry-based immunophenotyping of the B- and T-cell compartments. An interest in the therapeutic potential of belimumab, an anti-BAFF humanized monoclonal antibody, led us to focus on BAFF and its receptors in B cells. While recent studies of immune populations in splenectomy specimens from patients with ITP have by their nature enrolled patients with refractory disease receiving significant immunodulatory therapy, we chose to enroll a crosssection of ITP patients in order to ensure the broadest possible applicability of our findings. Therefore, autoantibody-positive and -negative ITP patients were recruited across a range of platelet counts and prior treatments including rituximab and splenectomy, despite the known effects of these therapies on B cells with the intention of identifying candidate biomarkers of relevance to future clinical trials. An initial analysis was performed comparing splenectomy- and rituximab-naïve ITP patients with healthy volunteers, and significant results were evaluated in the larger cohort.

Methods Patients and healthy volunteers A cross-sectional cohort of adult patients with a clinical diagnosis of chronic ITP was recruited from patients in the UK ITP registry visiting the outpatient clinic of the Royal London Hospital Department of Haematology (Table 1 and Online Supplementary Table S1). All patients able to give informed consent were considered for inclusion; the only exclusion criterion was ongoing immunosuppressive or cytotoxic therapy for a non-ITP diagnosis (one renal transplant recipient). Recruitment was stratified to give approximately equal numbers of patients by anti-platelet antibody status. All participants provided one venous blood sample; a subset of patients provided a second sample at a later time point. None of the patients had received a platelet transfusion in the ten days prior to venesection or intravenous immunoglobulin in the 21 days prior to venesection. Age- (within 10 years) and sex-matched healthy volunteers (HV) were recruited locally from within the GSK donor pool in parallel with the ITP patients. Ethical approval was obtained from the National Research Ethics Service, London, UK, REC, Ref. 07/H0718/57 (ITP patients) and National Research Ethics Service, Hertfordshire, UK, REC, Ref. 07/H0311/103 (GSK donor pool). The human biological samples were sourced ethically and their research use was in accordance with the terms of the informed consent procedures. haematologica | 2016; 101(6)

Anti-platelet autoantibodies EDTA-anticoagulated venous blood was tested within three days of venesection. Direct tests (for platelet bound IgG and IgM) were determined using the platelet immunofluorescence test (PIFT)11 and by the monoclonal antibody-specific immobilization of platelet antigens (MAIPA) assay12 to determine whether there was platelet bound IgG localized to platelet glycoproteins (GP)IIb/IIIa, GPIa/IIa or GPIb/IX. A patient was considered to have platelet autoantibodies if the results of the direct PIFT or the direct MAIPA assay were greater than the mean +3 standard deviations of the values obtained for platelets obtained from at least 3 normal blood donors on the same day. These investigations were performed at the Histocompatibility and Immunogenetics Laboratory, NHS Blood & Transplant, Filton. UK.

Flow cytometry Venous blood was collected into lithium heparin tubes (Fisher Scientific) and arrived at the laboratory within three hours. Peripheral blood mononuclear cells (PBMC) were prepared by density gradient centrifugation using Ficoll-Paque Plus (GE Healthcare). For transitional B cells, PBMC were pre-incubated with MitoTracker green (Invitrogen) at 10 nM concentration for 20 min at 37°C, then washed before adding antibodies.13 For B-cell immunophenotyping, a common antibody panel consisting of anti-CD38 (PE-Cy7, eBioscience 25-0389), anti-CD27 (APC, eBioscience 17-0279), anti-CD19 (APC-Cy7, BD Biosciences 557791), anti-IgD (biotin, BD Biosciences cat 555777), and antiCD3 (Pac Orange, Invitrogen CD0330) was used, with the following antibodies added to individual tubes as required: anti-CD10 (PE, BD Bioscience 555375), anti-IgG (FITC, BD Biosciences 555786), anti-IgM (PerCP Cy5.5, BioLegend 314512), anti-CD95 (PE, BioLegend 305608), anti-CD21 (PE, BD Biosciences 555422), anti-CD24 (PerCP eFluor710, eBioscience 46-0247), anti-BCMA (PerCP eFluor710, custom GSK conjugate), anti-BAFFR (FITC, BioLegend 316904), and anti-TACI (PE, BioLegend 311906). For T-cell immunophenotyping, a common skeleton consisting of anti-CD3 (APC-Cy7, BioLegend 300318), anti-CD4 (Pacific Blue, BioLegend 317429), anti-CD45RA (PerCP Cy5.5, eBioscience 45-0458), and anti-CXCR5 (PE, R&D Systems FAB190P) was supplemented with anti-CCR6 (PE-Cy7, BioLegend 353418), anti-PD1 (APC, BD Biosciences 558694), and anti-CXCR3 (FITC, R&D Systems FAB160F) for general T-cell subsets/T-follicular helper cells and with anti-CD25 (PE-Cy7, BD Biosciences 557741) and anti-CD127 (AF647, BD Biosciences 558598) for regulatory T cells. Normal rat serum and normal mouse serum was added to all tubes to minimize non-specific binding (Online Supplementary Table S2). In each case, 1x106 cells were stained in 100 uL at room temperature in the dark for 20 min. Tubes containing anti-IgD were then stained with streptavidin eFluor 450 (eBiosciences 48-4317) for a further 20 min. Cells were resuspended in 200 uL FACS buffer and acquired immediately on a BD Canto II (BD Biosciences). A compensation matrix was calculated using BD CompBeads for all stains except Mitotracker, where positive and negative live cells were used. Gating was performed in FlowJo vX (Miltenyi). B-cell gating is shown (Figure 1A). CD4+ T cells are gated as CD4+CD3+ cells in the lymphocyte gate. Memory CD4+ T cells are gated as CD45RA–. Tregs are gated as a discrete CD25high, CD127low population. For quality assurance, each ITP sample was run in parallel with an HV sample using the same antibody mixes; only technically adequate samples were included in the analysis.

BAFF enzyme linked immunosorbent assay Serum was extracted from venous blood collected into serum tubes and centrifuged at 2000 rpm for 15 min. This was stored at 699

S.M. Flint et al.

-80°C until the enzyme linked immunosorbent assay (ELISA) could be performed, in three overlapping batches. The Quantikine Human BAFF/BLyS/TNFSF13B Immunoassay (R&D Systems, cat. n. DBLYS0) was used according to the manufacturer’s instructions.

Statistical analysis The main analysis was performed using a 'patient-level' cohort, comprising one sample per ITP patient and a matched HV sample, stratified by prior splenectomy or rituximab use. Where patients provided multiple samples, the time point with the lowest platelet count was chosen. Twenty-six ITP patients fulfilled these criteria for the B-cell analysis and 18 patients for the T-cell analysis, with HV matched to each analysis. Two patients had FACS data without a platelet count; these were only excluded from analyses requiring a platelet count. Four patients in the ITP cohort had platelet count and FACS data from multiple time points, allowing a repeated measures analysis.

All data analysis was performed using R (www.r-project.org). Between group comparisons were performed using pairwise Wilcoxon signed-rank tests, except for a repeated measures analysis of the association between platelet count and CD95+ naïve B cells where a linear mixed effects model was implemented using the nlme package in R. An alpha level of 0.05 was considered significant.

Results Major B-cell populations A cross-sectional cohort of 26 ITP patients, 12 of whom had active disease (i.e. platelet count < 50x109/L) was matched to 26 HV (Table 1). Nine ITP patients were on no current treatment and had not received prior B-cell modulating therapies; 12 patients had previously received rituximab or a splenectomy, and a number were also receiving

700

Figure 1. Analysis of major B-cell subsets: differences by diagnosis. (A) Gating strategy for B-cell subsets. (B) Major B-cell subsets expressed as a percentage of CD19+ cells and stratified by diagnosis (i.e. healthy volunteers or splenectomy- and rituximabnaïve immune thrombocytopenia patients). (C) Gating strategy for transitional B cells. Mitotracker positive gate defined using expression on switched memory B cells. Transitional B cells are Mitotracker+ IgD+CD27–. T1 transitional population is CD10+; T2 and 3 populations are distinguished by CD38 expression. (D) Transitional B-cell populations, expressed as a percentage of B cells and stratified by diagnosis as in (B). Significant pairwise Pvalues (Wilcoxon signed-rank test) are shown. ritux/splen: rituximab or splenectomy; mem: memory. Plasmablasts are gated out of other B-cell subsets. haematologica | 2016; 101(6)

Immunophenotype of primary ITP

other agents including mycophenolate, azathioprine and romiplostim (Online Supplementary Table S2). Fourteen ITP patients had detectable platelet-bound IgG or IgM antibodies, and 7 of these had antibodies with the following specificities: GPIIb/IIIa (4), GPIb/IX (1) or both GPIIb/IIIa and GPIb/IX (2).

differences were observed in these headline populations between ITP patients and HV (Figure 1B). There were also no differences between untreated ITP patients and HV in the most immature population of peripheral blood B cells, transitional B cells, despite phenotyping these in detail (Figure 1C and D).

The major B-cell subsets [i.e. B cells overall, naïve (CD27–IgD+) B cells, CD27+IgD+ memory B cells, CD27+IgD– memory B cells, ‘double-negative’ (CD27–IgD– ) B cells and circulating plasmablasts gated as shown in Figure 1A) were compared between splenectomy- and rituximab-naïve (‘untreated’) ITP patients (n=14) and HV. No

B-cell expression of CD95 (Fas receptor) and CD21 (Complement receptor 2)

B-cells

There were, however, differences in the B-cell surface expression of CD95 and CD21, both linked previously to autoimmune disease.14,15 Overall, CD95 was expressed in a bimodal fashion on the B-cell surface (Figure 2A) with the

CD95+ B-cells

Years post rituximab

Disease activity

Figure 2. CD95+ (Fas receptor+) naïve B cells are more frequent in immune thrombocytopenia (ITP) patients. (A) Representative gating, showing distribution of CD95+ cells across the major B-cell subsets. (B) Proportions of CD95+ cells across B-cell subsets, stratified by diagnosis (i.e. healthy volunteers or splenectomy- and rituximab-naïve ITP patients). (C) Proportions of CD95+ naïve and IgD+CD27+ B cells stratified by diagnosis and prior rituximab or splenectomy. (D) Proportions of CD95+ naïve and IgD+CD27+ B cells by timing of rituximab, for those ITP patients who had received prior rituximab. (E) CD95+ cells as a proportion of naïve B cells, stratified by platelet count. (F) CD95+ naïve B cells by platelet count for ITP patients where a second time point is available (Δ Splenectomy; Rituximab; o Neither). Significant pairwise P-values (Wilcoxon signed-rank test) are shown in (B, C and E). (F) P-value is generated using a repeated measures linear mixed effects model. PB: plasmablast; DN: IgD-CD27- B cells; plt: platelet count; ritux/splen: rituximab or splenectomy. Plasmablasts are gated out of other B-cell subsets.

Platelet count (x109/L)

haematologica | 2016; 101(6)

701

S.M. Flint et al.

proportion of CD95+ cells increasing stepwise along the Bcell differentiation pathway (i.e. median proportion in HV of CD95+ naïve B cells = 1%, IgD+ CD27+ memory B cells = 16%, IgG+ IgD–CD27+ memory B cells = 43% and circulating plasmablasts = 98%) (Figure 2B). In naïve and IgD+ CD27+ memory B cells, there was a small but highly statistically significant expansion of the CD95+ population in untreated ITP patients compared to HV (Figure 2B). ITP patients with prior splenectomy or rituximab also had a higher median proportion of CD95+ IgD+ CD27+ memory B cells than their untreated counterparts and a positive association was observed with time since rituximab therapy (Figure 2C and D), suggesting an additional effect of B-cell depletion. This is in contrast to the proportion of CD95+ naïve B cells, which was increased similarly in all ITP patients regardless of prior rituximab and splenectomy exposure, and which was not associated with time post rituximab. The size of the CD95+ naïve population was correlated with disease activity, with the largest expansion observed in those patients with platelet counts less than 50x109/L (Figure 2E). Using the 4 ITP patients who had been bled at a second time point (range 56-455 days between samples), including one who had previously had a splenectomy and one who had received rituximab 2.6 years prior to the first blood draw, we found that within individual patients, an improving platelet count was associated with a reduction in the proportion of CD95+ naïve B cells (P=0.05, repeated measures using a linear mixed effects model) (Figure 2F). Prior rituximab or splenectomy was also associated with an expansion of the proportion of CD95+ cells compared to HV in IgD-CD27+ memory B cells (data not shown). CD21, on the other hand, was present in HV on a median 98% of naïve B cells, 87% of IgG+ IgD-CD27+ memory B cells and 36% of plasmablasts, confirming other reports that this marker is lost stepwise during B-cell terminal differentiation (Figure 3A and B). There was a significant reduction in the proportion of naïve B cells expressing CD21 in untreated ITP patients compared with matched HV (Figure 3B). This was unrelated to disease activity (platelets < 50x109/L vs. platelets ≥ 50x109/L; P=0.79). Instead, this loss of CD21 on naïve B cells was restricted to patients with detectable platelet autoantibodies and appeared to be ameliorated by treatment with rituximab or splenectomy (Figure 3C). Median platelet counts were similar between antibody positive and negative patients (51.5x109/L vs. 56.5x109/L; P=0.93).

BAFF and its receptors Serum BAFF levels were increased in our cohort of ITP patients overall, consistent with other studies. Moreover, in untreated patients, there was a trend toward higher serum BAFF levels in active disease (i.e. platelet count < 50x109/L) compared with patients in remission (Figure 4A). This trend was not seen in the treated group, most likely due to increases in serum BAFF in the setting of rituximab and splenectomy. This is well-described and a consequence of the effect of B-cell depletion on BAFF release and synthesis.16 There were no differences in serum BAFF levels according to autoantibody status (data not shown). In the absence of published data on the expression of BAFF receptors following rituximab, we analyzed both treated and untreated ITP patients. BAFFR was detectable on all populations of B cells except for plasmablasts (Figure 702

Table 1. Baseline demographics, treatment received and autoantibody status for immune thrombocytopenia patients and healthy volunteers used in the B-cell analysis. Demographics Patients in cohort, number Median age, years (range) Gender, number (%) Female Male ITP duration, years (range) Prior rituximab or splenectomy Rituximab alone, number (%) Splenectomy alone, number (%) Both, number (%) Current thrombopoetin receptor agonist, number (%) Anti-platelet antibodies, number (%) Positive by direct MAIPA (IgG) Positive by direct PIFT (IgG & IgM) Positive by direct PIFT & direct MAIPA Median platelet count, ×109/L (IQR)*

ITP

26 36 (22-63)

26 37 (19-62)

16 (62%) 10 (38%) 7.3 (1.1-37.4)

16 (62%) 10 (38%)

7 (27%) 1 (3.8%) 4 (15%) 6 (24%)

7 (28%) 13 (50%) 14 (54%) 52 (36-119)

*Missing, n=2. MAIPA: monoclonal antibody immobilization of platelet antigens; PIFT: platelet immunofluorescence test; IQR: interquartile range; HV: healthy volunteer; ITP: immune thrombocytopenia.

4B). TACI was broadly expressed on memory B cells and plasmablasts but present on only a very small population of naïve B cells (Figure 4C). There were no significant differences in the B-cell expression of either of these markers between HV and untreated ITP patients, but we observed a significant decrease in BAFF receptor expression on a number of B-cell populations and increase in TACI expression on naïve B cells in samples from ITP patients after rituximab or splenectomy (Figure 4C). Neither of these showed a correlation with time since last rituximab dose (data not shown). BCMA was primarily present on plasmablasts, but was also detected on a small number of double-negative (CD27-IgD-) B cells. BCMA was markedly and consistently up-regulated on plasmablasts in ITP (Figure 4D), irrespective of platelet count, prior treatment or autoantibody status (Figure 4D-F).

T-cell phenotype Data on T-cell populations were also available for 18 ITP patients. We found a trend to proportionally fewer CD4+ T cells in splenectomy- and rituximab-naïve ITP patients compared to HV, consistent with previous reports of a reduced CD4:CD8 ratio in ITP (Wilcoxon P-value = 0.053) (Figure 5A).17 We found few other differences within the memory T cell, CXCR5+ memory T cell and Treg populations when these were analyzed as a proportion of their parent populations. Reduced numbers of CD4+ T cells and memory CD4+ T cells overall in our cohort of splenectomyand rituximab-naïve ITP patients resulted in a number of differences in subpopulations in absolute terms (Figure 5B).

Discussion Reduced CD21 and increased CD95 expression on naïve B cells and marked BCMA upregulation on plasmablasts are identified as important components of the immune haematologica | 2016; 101(6)

Immunophenotype of primary ITP

Figure 3. CD21– (complement receptor 2–) naïve B cells are more frequent in immune thrombocytopenia (ITP) patients. (A) Representative gating, showing distribution of CD21– cells across the major B-cell subsets. (B) Proportions of CD21– cells across B-cell subsets, stratified by diagnosis (i.e. healthy volunteers or splenectomy- and rituximab-naïve ITP patients). (C) CD21– cells as a proportion of naïve B cells, stratified by anti-platelet antibody status. Significant pairwise P-values (Wilcoxon signed-rank test) are shown. sw mem + PB: switched memory B cells and plasmablasts i.e. CD27+IgD– B cells; DN: double-negative B cells, i.e. CD27–IgD– B cells; nonsw memory: non-switched memory, i.e. CD27+IgD+ B cells; Ab: antiplatelet antibody; ritux/splen: rituximab or splenectomy.

phenotype of ITP. CD95 is up-regulated rapidly upon Bcell activation,18 and the ITP-specific expansion of CD95+ and of CD21– naïve B cells observed in this study may represent a population of autoreactive B cells activated in the presence of circulating platelet autoantigens. It is known that self-reactive B cells emerge from the bone marrow in reasonable quantities,19 and that peripheral mechanisms of tolerance can be overcome in the presence of high BAFF levels.20 Mechanistic data in support of this hypothesis comes from a recent study of B cells from patients with flaring systemic lupus erythematosus (SLE).21 Using B-cell receptor repertoire analysis by next generation sequencing, the authors demonstrated that a distinct subset of naïve B cells constituted an important source of autoreactive antibody secreting cells. This activated subset was characterized by increased CD95 expression and reduced expression of CD21. Although our flow panels were not designed to study the co-expression of these markers, these findings mirror the expanded CD95+ and CD21– naïve B-cell populations we describe here. In SLE, this subset was expanded further during flaring disease, consistent with our observation that expansion of CD95+ naïve B cells in ITP correlated with disease activity. The loss of CD21, a receptor for C3d that interacts with CD19 to lower the threshold for signaling through the Bcell receptor,22 has been linked to a subset of naïve B cells in which autoreactive cells are over-represented. Expansion of this CD21– naïve B-cell population has been described in other immune-mediated conditions as well as ITP and SLE, including Sjogren’s syndrome, common variable immunodeficiency and rheumatoid arthritis.23-25 It is of interest that, in our study, CD21– naïve B cells were found predominantly in antiplatelet antibody positive patients, regardless of platelet count. This would be consistent with the hypothesis that, even though the size of the CD95+ CD21– naïve B-cell population was small, it plays an important role in the generation of autoantibody. Given this hypothesis, it would have been informative to study the expression of CD95 and CD21 in a cohort with non-immune thrombocytopenia. This population haematologica | 2016; 101(6)

was not available to this study, but is a potential avenue of future research. The lack of differences in the transitional B-cell compartment is in apparent contradiction to an earlier report of fewer CD24hiCD38hi B cells in non-splenectomized patients with active ITP.6 However, the changes reported were subtle and only found in patients with a platelet count less than 50x109/L. Comparatively small numbers restricted our ability to perform detailed subanalyses, and as such, it is possible that our cross-sectional cohort was not powered to replicate this observation. Alternatively, this may be because we used the gating strategy of Palanichamy et al.13 to distinguish three populations of transitional B cells. The CD24hiCD38hi population of the earlier report would represent the more immature T1-2 populations in our study. However, we were able to demonstrate a trend toward expansion of these early B-cell populations in patients post rituximab, consistent with a previous report.26 The dominant finding of our T-cell analysis was a CD4+ T-cell lymphopenia, most prominent in the memory compartment. Such a lymphopenia has not been described previously in ITP and should be interpreted with caution. Lymphopenia is a recognized consequence of therapy and 4 of the 16 ITP patients analyzed were receiving antiproliferative agents (i.e. azathioprine, mycophenolate).27 The lack of observed differences in other subsets, especially Tregs, may reflect the cross-sectional nature of the patient cohort, which was recruited across a range of disease activity. Finally, there was a strong upregulation of BCMA on plasmablasts in ITP. The role of BCMA in the pathogenesis of autoimmune disease is complex and incompletely understood. While the sole B-cell phenotype of BCMA–/– mice appears to be impaired survival of long-lasting plasma cells,28 when crossed onto a lupus prone background, BCMA–/– mice exhibited a range of pathologies, including increased plasma cell number, elevated systemic BAFF, and an increased titer of anti-nuclear antibody compared to their BCMA sufficient counterparts.29 Similarly, little is known about the regulation of BCMA transcription and membrane expression. Certainly Blimp-1 and IRF4 are 703

S.M. Flint et al. important positive regulators,30 consistent with BCMAs predominant expression on plasma cells and plasmablasts; however, whether BCMA regulates, or is regulated by, its ligands BAFF and APRIL is unknown. Additional weight for a role for BCMA in autoimmune disease comes from studies in SLE, where B-cell expression of BCMA has been shown to be elevated.31 Current therapies, including rituximab, do not target plasma cells well. The restricted tissue

expression pattern of BCMA, its important role in plasma cell survival, and its increased expression in ITP and other autoimmune diseases together make it an attractive target for novel plasma cell-directed therapies. Our study adds significant detail to an emerging B-cell phenotype that is shared between a number of antibodymediated autoimmune diseases. This phenotype is characterized by expanded populations of CD95+ and CD21â&#x20AC;&#x201C;

Figure 4. Analysis of serum BAFF and its receptors: BCMA is markedly up-regulated on the surface of plasmablasts in immune thrombocytopenia, regardless of activity or autoantibody status. (A) Serum BAFF, stratified by past treatment and diagnosis. BAFFR (B) and TACI (C) median fluorescence intensity (MFI) across the major B-cell subpopulations, stratified by diagnosis and prior treatment. (D) Plasmablast BCMA MFI, stratified by diagnosis and prior treatment. (E and F) Plasmablast BCMA MFI, stratified by platelet count and autoantibody status. Significant pairwise P-values (Wilcoxon signed-rank test) are shown. Plt: platelet count; ritux/splen: rituximab or splenectomy.

704

haematologica | 2016; 101(6)

Immunophenotype of primary ITP

Figure 5. CD4+ T cells in immune thrombocytopenia (ITP) and healthy volunteer samples by diagnosis. (A) T-cell populations expressed as a proportion of their parent population, stratified by diagnosis [i.e. healthy volunteers (white, n=18) or splenectomy- and rituximab-naïve ITP patients (dark gray, n=9)]. (B) The same T-cell populations, expressed in absolute numbers [healthy volunteers (white, n=9); splenectomy- and rituximab-naïve ITP patients (dark gray, n=8)]. Significant pairwise P-values (Wilcoxon signed-rank test) are shown. ritux/splen: rituximab or splenectomy; mem: memory, i.e. CD45RA–.

naïve B cells, elevated levels of serum BAFF and elevated BCMA expression on plasmablasts. As well as ITP, aspects of this phenotype have also been observed in SLE, common variable immunodeficiency, and rheumatoid arthritis.23,32 Individually, these autoimmune diseases are rare, but a common immune phenotype may streamline the development of effective therapies targeting multiple diseases. The CD95+ population of naïve B cells in particular is identified as warranting further study as a potential biomarker for this phenotype, being both correlated with disease activity within and between individuals, and robust to commonly used B-cell modulating therapies. Acknowledgments The authors would like to thank patients and healthy volunhaematologica | 2016; 101(6)

teers who have contributed samples for this study. We are grateful for the advice that Prof. Ken Smith, Iñaki Sanz and Chungwen Wei provided during the study design and to Andrea Itano, who initiated the program. Funding Primary funding for this study was from GSK. SMF was funded by a Translational Medicine and Therapeutics PhD studentship jointly funded by the Wellcome Trust and GSK. The UK ITP Registry (www.ukitpregistry.com) is supported through unrestricted educational grants from GSK and Amgen. We thank the staff of the serology section of Histocompatibility and Immunogenetics Laboratory, NHS Blood & Transplant – Filton for performing the platelet antibody investigations. 705

S.M. Flint et al.

References 1. Nørgaard M, Jensen A, Engebjerg MC, et al. Long-term clinical outcomes of patients with primary chronic immune thrombocytopenia: A Danish population-based cohort study. Blood. 2011;117(13):3514-3520. 2. Shulman NR, Marder VJ, Weinrach RS. Similarities between known antiplatelet antibodies and the factor responsible for thrombocytopenia in idiopathic purpura. Physiologic, serologic and isotopic studies. Ann N Y Acad Sci. 1965;124(2):499-542. 3. McMillan R. Antiplatelet Antibodies in Chronic Immune Thrombocytopenia and Their Role in Platelet Destruction and Defective Platelet Production. Hematol Oncol Clin North Am. 2009;23(6):11631175. 4. Zhu XJ, Shi Y, Peng J, et al. The effects of BAFF and BAFF-R-Fc fusion protein in immune thrombocytopenia. Blood. 2009;114(26):5362-5367. 5. Emmerich F, Bal G, Barakat A, et al. Highlevel serum B-cell activating factor and promoter polymorphisms in patients with idiopathic thrombocytopenic purpura. Br J Haematol. 2007;136(2):309-314. 6. Li X, Zhong H, Bao W, et al. Defective regulatory B-cell compartment in patients with immune thrombocytopenia. Blood. 2012;120(16):3318-3325. 7. Martinez-Gamboa L, Mei H, Loddenkemper C, et al. Role of the spleen in peripheral memory B-cell homeostasis in patients with autoimmune thrombocytopenia purpura. Clin Immunol. 2009;130(2):199-212. 8. Olsson B, Andersson P-O, Jernås M, et al. T-cell-mediated cytotoxicity toward platelets in chronic idiopathic thrombocytopenic purpura. Nat Med. 2003;9(9):11231124. 9. Najaoui A, Bakchoul T, Stoy J, et al. Autoantibody-mediated complement activation on platelets is a common finding in patients with immune thrombocytopenic purpura (ITP). Eur J Haematol. 2012;88(2):167-174. 10. Audia S, Rossato M, Santegoets K, et al. Splenic TFH expansion participates in B-cell differentiation and antiplatelet-antibody production during immune thrombocytopenia. Blood. 2014;124(18):2858-2866.

706

11. Von Dem Borne AE, Verheugt FW, Oosterhof F, Von Riesz E, De La Rivière AB, Engelfriet CP. A simple immunofluorescence test for the detection of platelet antibodies. Br J Haematol. 1978;39:195207. 12. Campbell K, Rishi K, Howkins G, et al. A modified rapid monoclonal antibody-specific immobilization of platelet antigen assay for the detection of human platelet antigen (HPA) antibodies: A multicentre evaluation. Vox Sang. 2007;93:289-297. 13. Palanichamy A, Barnard J, Zheng B, et al. Novel human transitional B cell populations revealed by B cell depletion therapy. J Immunol. 2009;182(10):5982-5993. 14. Siegel RM, Chan FK, Chun HJ, Lenardo MJ. The multifaceted role of Fas signaling in immune cell homeostasis and autoimmunity. Nat Immunol. 2000;1(6):469-474. 15. Haas KM, Tedder TF. Role of the CD19 and CD21/35 receptor complex in innate immunity, host defense and autoimmunity. Adv Exp Med Biol. 2005;560:125-139. 16. Lavie F, Miceli-Richard C, Ittah M, Sellam J, Gottenberg J-E, Mariette X. Increase of B cell-activating factor of the TNF family (BAFF) after rituximab treatment: insights into a new regulating system of BAFF production. Ann Rheum Dis. 2007;66(5):700703. 17. Hu Y, Ma DX, Shan NN, et al. Increased number of Tc17 and correlation with Th17 cells in patients with immune thrombocytopenia. PLoS One. 2011;6(10):e26522. 18. Daniel PT, Krammer PH. Activation induces sensitivity toward APO-1 (CD95)mediated apoptosis in human B cells. J Immunol. 1994;152(12):5624-5632. 19. Wardemann H, Yurasov S, Schaefer A, Young JW, Meffre E, Nussenzweig MC. Predominant autoantibody production by early human B cell precursors. Science. 2003;301(5638):1374-1377. 20. Thien M, Phan TG, Gardam S, et al. Excess BAFF rescues self-reactive B cells from peripheral deletion and allows them to enter forbidden follicular and marginal zone niches. Immunity. 2004;20(6):785798. 21. Tipton CM, Fucile CF, Darce J, et al. Diversity, cellular origin and autoreactivity of antibody-secreting cell population expansions in acute systemic lupus erythemato-

sus. Nat Immunol. 2015;16(7):755-765. 22. Fearon DT, Carroll MC. Regulation of B lymphocyte responses to foreign and selfantigens by the CD19/CD21 complex. Annu Rev Immunol. 2000;18:393-422. 23. Quách TD, Manjarrez-Orduño N, Adlowitz DG, et al. Anergic responses characterize a large fraction of human autoreactive naïve B cells expressing low levels of surface IgM. J Immunol. 2011;186(8):4640-4648. 24. Isnardi I, Ng YS, Menard L, et al. Complement receptor 2/CD21- human naïve B cells contain mostly autoreactive unresponsive clones. Blood. 2010; 115(24):5026-5036. 25. Saadoun D, Terrier B, Bannock J, et al. Expansion of autoreactive unresponsive CD21-/low B cells in Sjogren’s syndromeassociated lymphoproliferation. Arthritis Rheum. 2013;65(4):1085-1096. 26. Anolik JH, Friedberg JW, Zheng B, et al. B cell reconstitution after rituximab treatment of lymphoma recapitulates B cell ontogeny. Clin Immunol. 2007;122(2):139145. 27. Ritter ML, Pirofski L. Mycophenolate mofetil: Effects on cellular immune subsets, infectious complications, and antimicrobial activity: Review article. Transpl Infect Dis. 2009;11(4):290-297. 28. O’Connor BP, Raman VS, Erickson LD, et al. BCMA is essential for the survival of long-lived bone marrow plasma cells. J Exp Med. 2004;199(1):91-98. 29. Jiang C, Loo WM, Greenley EJ, Tung KS, Erickson LD. B cell maturation antigen deficiency exacerbates lymphoproliferation and autoimmunity in murine lupus. J Immunol. 2011;186(11):6136-6147. 30. Deng S, Yuan T, Cheng X, Jian R, Jiang J. Blymphocyte-induced maturation protein1 up-regulates the expression of B-cell maturation antigen in mouse plasma cells. Mol Biol Rep. 2010;37(8):3747-3755. 31. Zhao LD, Li Y, Smith MF, et al. Expressions of BAFF/BAFF receptors and their correlation with disease activity in Chinese SLE patients. Lupus. 2010;19(13):1534-1549. 32. Rakhmanov M, Keller B, Gutenberger S, et al. Circulating CD21low B cells in common variable immunodeficiency resemble tissue homing, innate-like B cells. Proc Natl Acad Sci USA. 2009;106(32):13451-13456.

haematologica | 2016; 101(6)

ARTICLE

Myelodysplastic Syndromes

Loss of B cells and their precursors is the most constant feature of GATA-2 deficiency in childhood myelodysplastic syndrome

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Michaela Nováková,1 Markéta Žaliová,1 Martina Suková,2 Marcin Wlodarski,3 Aleš Janda,3 Eva Froňková,1,2 Vít Campr,4 Kateřina Lejhancová,5 Ondřej Zapletal,6 Dagmar Pospíšilová,7 Zdeňka Černá,8 Tomáš Kuhn,9 Peter Švec,10 Vendula Pelková,1 Zuzana Zemanová,11 Gitte Kerndrup,12 Marry van den Heuvel-Eibrink,13 Vincent van der Velden,14 Charlotte Niemeyer,3 Tomáš Kalina,1 Jan Trka,1,2 Jan Starý,2 Ondřej Hrušák,1,2 and Ester Mejstříková1,2

CLIP-Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University and University Hospital Motol, Prague, Czech Republic; 2Department of Pediatric Hematology and Oncology, 2nd Faculty of Medicine, Charles University and University Hospital Motol, Prague, Czech Republic; 3Center for Pediatrics and Adolescent Medicine, University Medical Center, Freiburg, Germany; 4Department of Pathology and Molecular Medicine, University Hospital Motol, Prague, Czech Republic; 5Department of Pediatrics, Charles University, University Hospital Hradec Králové, Czech Republic; 6 Department of Pediatric Hematology, Children's University Hospital, Brno, Czech Republic; 7Department of Pediatrics, Palacky University and University Hospital Olomouc, Czech Republic; 8Department of Pediatrics, University Hospital Pilsen, Czech Republic; 9Department of Pediatrics Ostrava, University Hospital Ostrava, Czech Republic; 10Department of Pediatric Hematology and Oncology, University Hospital Bratislava, Slovakia; 11Centre of Oncocytogenetics, Institute of Clinical Biochemistry and Laboratory Diagnostics, 1st Faculty of Medicine and General University Hospital and Charles University, Prague, Czech Republic; 12Department of Pathology, Aarhus University Hospital, Denmark; 13Department of Pediatric Oncology, Princess Máxima Centre for Paediatric Oncology, Utrecht, the Netherlands; and 14Department of Immunology, Erasmus MC, Rotterdam, the Netherlands

Haematologica 2016 Volume 101(6):707-716

ABSTRACT

Correspondence:

ester.mejstrikova@lfmotol.cuni.cz

ATA-2 deficiency was recently described as common cause of overlapping syndromes of immunodeficiency, lymphedema, familiar myelodysplastic syndrome or acute myeloid leukemia. The aim of our study was to analyze bone marrow and peripheral blood samples of children with myelodysplastic syndrome or aplastic anemia to define prevalence of the GATA2 mutation and to assess whether mutations in GATA-2 transcription factor exhibit specific immunophenotypic features. The prevalence of a GATA2 mutation in a consecutively diagnosed cohort of children was 14% in advanced forms of myelodysplastic syndrome (refractory anemia with excess blasts, refractory anemia with excess blasts in transformation, and myelodysplasiarelated acute myeloid leukemia), 17% in refractory cytopenia of childhood, and 0% in aplastic anemia. In GATA-2-deficient cases, we found the most profound B-cell lymphopenia, including its progenitors in blood and bone marrow, which correlated with significantly diminished intronRSS-Kde recombination excision circles in comparison to other myelodysplastic syndrome/aplastic anemia cases. The other typical features of GATA-2 deficiency (monocytopenia and natural killer cell lymphopenia) were less discriminative. In conclusion, we suggest screening for GATA2 mutations in pediatric myelodysplastic syndrome, preferentially in patients with impaired B-cell homeostasis in bone marrow and peripheral blood (low number of progenitors, intronRSS-Kde recombination excision circles and naïve cells) . haematologica | 2016; 101(6)

Received: October 15, 2015. Accepted: March 18, 2016. Pre-published: March 24, 2016. doi:10.3324/haematol.2015.137711

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/707

707

M. Novakova et al.

Introduction Myelodysplastic syndrome (MDS) is a rare disease of childhood with an approximate frequency of 0.8 to 1.8 per million children.1 The most common subtype of MDS is refractory cytopenia of childhood (RCC), which represents a distinct category that was introduced as a provisional entity in the 2008 WHO classification.2 Aplastic anemia (AA) shares several clinical and laboratory features with RCC, and nowadays histopathological assessment is a key method to distinguish between the two diseases.3 Advanced MDS in children can be separated into three categories: 1) refractory anemia with excess blasts (RAEB); 2) RAEB in transformation (RAEB-t); or 3) myelodysplasiarelated acute myeloid leukemia (MDR-AML).4 In some children, MDS or hypoplastic bone marrow failure is associated with an underlying genetic predisposition (e.g. Fanconi anemia, dyskeratosis congenita or ShwachmanDiamond syndrome).5 A mutation in the GATA2 gene, which encodes the transcription factor GATA-2, was recently found by whole genome sequencing6,7 or by candidate approaches8,9 as a common cause of several overlapping syndromes: familial MDS/acute myeloid leukemia (AML), dendritic cell, monocyte, B- and NK-lymphoid (DCML) deficiency, mycobacterial infections and monocytopenia (MonoMAC), and hereditary lymphedema (Emberger syndrome).6,7 Several abnormalities, identifiable by flow cytometry (FC) in peripheral blood (PB), are known to be present in patients with GATA2 mutations: a decreased number of B cells, NK cells, monocytes and dendritic cells;10,11 plasma cells with an aberrant immunophenotype in bone marrow (BM); clonal T-large granular lymphocyte (LGL) proliferation; and aberrant maturation patterns of granulocytic lineage.10,12 MDS manifests in GATA-2-deficient patients earlier than in the general population.11 A GATA2 mutation in pediatric non-familial MDS patients was found in 16% of patients with aberrant karyotype (monosomy 7).13 Flow cytometry is recognized to be an important diagnostic method especially in adult forms of MDS.14-16 In children, the amount of FC abnormalities in comparison to adults is often limited, especially in RCC.17 The myeloid compartment is severely reduced in both RCC and AA in comparison to healthy controls, but in AA the reduction is more pronounced.18 All Czech patients with suspected

MDS and AA have undergone trephine biopsy analysis by one of 2 expert pathologists since 2005, and BM aspirates are always analyzed in parallel using FC when material is available. We also analyzed the level of intronRSS-Kde recombination excision circles (KRECs) in PB and BM to assess B-cell production in children with MDS and AA. The aims of our study were 2-fold. Our first aim was to define prevalence of GATA2 mutation in a nation-wide pediatric cohort of MDS/AA patients. Our second aim was to identify FC profile characteristics for GATA-2-deficient patients.

Methods Patients Patients entered the study after their parents or guardians signed informed consent and the institutional ethics committee approved the study. Patients with RCC and AA were analyzed between 2005 and 2014, and all samples underwent histopathological analysis. Non-RCC (RAEB n=12, RAEB-t n=5, MDR-AML n=3) patients were analyzed in the period 1998-2014. Only those patients with available material for screening of the GATA2 mutation entered the study (3 additional AA patients were analyzed using FC during the study period. No material was available for GATA2 mutation screening, neither FC nor GATA2 screening was performed in one RCC patient, and no material for GATA2 mutation screening was available in 3 non-RCC patients). The prevalence of GATA2 mutation was analyzed among Czech pediatric primary MDS/AA patients: RCC n=30, AA n=38, nonRCC n=22. The flow chart describing the patient cohort is available in the Online Supplementary Appendix. We used residual material from grafts for stem cell transplantation and samples taken for infiltration assessments of nonhematopoietic tumors as control BM samples (n=35). All control samples were obtained from individuals under 20 years of age (median 4.6 years, range 0.01-19.3). Only control samples with no tumor cell infiltration as assessed by morphology entered the study.

Diagnostic criteria Diagnosis was established according to WHO classification (2008).2 The distinction between RCC and AA was based on histopathological criteria and cytogenetic findings (both summarized in the Online Supplementary Appendix). Patients with cytogenetic aberration before start of the treatment were classified

Table 1. Disease group characteristics. Monosomy 7 or trisomy 8 was categorized as positive if present at any time point during follow up.

GATA-2 Non-RCC RCC AA Controls flow cytometry

M/F ratio

Age, years median (range)

Follow up since first symptoms to therapy, weeks median (range)

Monosomy 7

Trisomy 8

12 20 27 39 35

5.0 2.3 1.1 1.3 1.2

16.3 (4.3 - 21.3) 8.3 (0.79 - 17.8) 7.4(3.7-19) 9.9 (1.1 - 17.9) 4.6 (0.01 - 19.3)

125.4 (3.1 -1046.1) 7.1 (0.43-66.1) 9.9 (0.71-422.9) 3.9 (0.71-74)

8/12 4/20 6/27 0/39

2/12 1/20 4/27 0/39

One patient (UPN2) with trisomy 8 and one patient (UPN7) developed cytogenetic abnormality and refractory cytopenia in childhood (RCC) in adulthood. M: male; F: female; AA: aplastic anemia.

708

haematologica | 2016; 101(6)

GATA-2 deficiency in childhood MDS

Table 2. Overview of clinical and laboratory findings in GATA-2-deficient patients. Exons of GATA-2 are numbered according to their 5’ to 3’ order within the GATA-2 gene (NCBI RefSeqGene NG_029334.1).

Unique patient number (UPN)

Familiar/ Primary Age Mutation Mutation Mutation sporadic manifestation (MDS exon/intron cDNA consequence diagnosis) (years)

UPN1 RCC

Sporadic

MDS-RCC

UPN2 ID/RCC Sporadic Immunodefiency 21 in adulthood

UPN3 RCC

Sporadic

MDS RCC

UPN4 RCC

Familiar

MDS RCC

UPN5 ID/RCC Familiar Immunodefiency 17

UPN6 RCC

Familiar

MDS RCC

UPN7 ID/ Familiar Immunodefiency 17 RCMD in adulthood

Exon 4

Clinical symptoms

Cytogenetics Treatment Follow up (anytime during follow up)

c.222_229+6 Frameshift del14ins21 c.1187G>A Substitution

Café-au-lait, monosomy 7, hydrocele trisomy 8 Exon 8 Recurrent trisomy 8 urinary tract infections, skin herpetic infections, bronchial asthma, aphthous stomatitis, cholecystolithiasis Exon 5 c.391_395del7 Frameshift Vesicouretheral reflux, monosomy 7 congenital hydrops, hearing disorder, lymphedema

MUD SCT MSD SCT

Intron 6 c.1017+572C>T Intronic Recurrent orthostatic monosomy 7 mutation collapses Exon 7 c.1081C>T Substitution Decreased physical monosomy 7 performance, 17 years bilateral pneumonia, interstitial lung disease, chronic active EBV infection, recurrent pain in testes, cholecystolithiasis, HLA-B27 Exon 7 c.1081C>T Substitution Asymptomatic carrier, Not brother of UPN7, found leukopenia and thrombocytopenia Intron 6 c.1017+ Intronic Orthostatic monosomy 7 572C>T mutation collapses

MSD SCT

15 years, healthy 30 years, healthy

2 relapses, died of progressive disease (age 14 years) Intron 6 c.1017+572C>T Intronic Chronic active EBV monosomy 7 MUD 26 years, mutation infection SCT healthy Exon 7 c.1084C>T Stop gain Chronic bronchitis, Not found Corticosteroids Died sinusitis, interstitial of lung lung disease, multiple warts insufficiency, CMV pneumonitis (age 21 years)

UPN8 FS

Familiar

Identified through family search

UPN9 RAEB/AML

Sporadic

MDS RAEB/AML

UPN10 RCC/RAEB-t

Sporadic

MDS RCC/RAEB-t

Exon 7

UPN11 RAEB Sporadic

MDS RAEB

4.4

Exon 7

UPN12 ID/RCC Sporadic

MDS RCC

Exon 7

c.1066_1095 del30

Deletion

Recurrent urinary monosomy 7 tract infections, 15 years weight loss, decreased physical performance, attention deficiency and hyperactive disorder c.1035_1038 Frameshift Recurrent otitis, monosomy 7 dupCGGC bronchitis, aphthous stomatitis, bronchial asthma, hypospadia, vesicouretheral reflux, speech disorder, Asperger syndrome, nail dysplasia c.1128C>G Stop gain Lymphedema, Not found deafness, recurrent infections

MUD SCT

Recently ongoing MUD SCT

Watch and wait

26 years, healthy 21 years, alive

17 years, healthy

AML Died of MDS therapy progression, mycotic lung infection (age 18 years) MUD 28 years, SCT healthy

MSD SCT

14 years, healthy

MUD SCT

15 years, healthy

MDS: myelodysplastic syndromes; FS: family search; ID: immunodeficiency; SCT: stem cell transplantation; MUD: matched unrelated donor; MSD: matched sibling donor; CMV: cytomegalovirus; EBV: Epstein Barr virus.

haematologica | 2016; 101(6)

709

M. Novakova et al. Table 3. Summary of the analyzed bone marrow (BM) parameters. The values are defined as medians with range in brackets.

Bone marrow populations

GATA-2

Non-RCC

RCC

Cell populations Granulocytes (%) 49 (26-72) 35 (7-79)*o 48 (1.3-82) Lymphocytes (%) 26 (7.3-48) 14 (2.8-76) 33 (9.1-78) Monocytes CD14pos (%) 3.3 (0.27-23) 3.6 (0.03-26) 4.7 (0.05-9.6) Erythroid cells CD45neg71bright (%) 4.5 (1.5-9.5) 3.6 (0.58-22) 2.4 (0.39-15) Progenitors Progenitors CD34pos (%) 0.49 (0.1-19)* 4.9 (0.13-32)o 0.49 (0.08-2.9)* pos Progenitors CD117 (%) 1.4 (0.12-19) 8.3 (1.1-25)*o 0.55 (0.13-2.3)*o B lineage CD19pos total B cells (%) 1.4 (0.19-5.2)* 3.7 (0.1-21)*o 6.6 (1.4-17)*o pos pos pos o Progenitors CD34 10 19 (%) 0.01 (0-0.04)* 0.04 (0-1.7)* 0.07 (0-1.7)*o CD10pos out of CD19pos (%) 4.7 (1.2-46)* 13 (1.3-70)*o 16 (1.1-97)*o pos neg pos CD20 10 out of CD19 (%) 75 (40-93)* 69 (8.5-90)* 79 (1.5-96)* Plasma cells CD10neg20neg out of B cells (%) 11 (3.1-48)* 8.4 (0.76-69)* 3.6 (0.02-22)*o Plasma cells CD10neg20neg out of all cells (%) 0.11 (0.02-1.4) 0.34 (0.03-1.9) 0.19 (0.04-1.3) KREC (copies/mg DNA) 25 (25-118)* 371 (25-26071)*o 7149 (23-123732)*o T lineage CD3pos total T cells (%) 23 (7.3-42)* 10 (1.9-54) 23 (2.9-60)* CD3pos4pos T cells (%) 9.8 (1.1-20)* 5.4 (1.3-35) 9.4 (0.31-35)* CD3pos8pos T tells (%) 10 (3.9-24)* 4.3 (0.7-19)o 9.7 (1-37)* CD4/8 ratio 0.75 (0.2-1.5)* 1.5 (0.44-2.4)o 0.97 (0.11-2.1)* CD3pos16,56pos cells (%) 2.5 (0.5-8.4)* 1.3 (0.17-3.7)* 0.8 (0.06-5.1)*o HLA DRpos out of CD3pos (%) 3.2 (1.4-8.9)* 5.3 (2.8-40)o 8.1 (0.79-87)o HLA DRpos out of CD3pos4pos (%) 2.6 (1.5-4.1)* 13 (2.4-35)o 5.1 (2-77)o HLA DRpos out of CD3pos8pos (%) 3.5 (1.5-9.2)* 30 (3.4-43) 29 (4.5-89)*o TREC (copies/mg DNA) 2248 (241-15385) 4083 (230-11108) 6257 (25-61522) Other lymphoid populations NK cells (%) 1 (0.1-7.9) 3 (0.78-6.9) 1.3 (0.24-9.1) DNA analysis DNA G01 (%) 85 (74-94) 89 (67-98) 91 (77-96)*o DNA SG2M (%) 9 (4.5-13) 9.7 (0.5-18) 6.7 (2.7-19)*o

Controls

26 (1.3-85)*o 51 (6.5-90)*o 3.5 (0.3-11)* 1.3 (0.1-9.5)*o

51 (19-83) 27 (8.5-66) 4.8 (2.9-12) 3.6 (0.93-11)

0.2 (0.03-1.3)*o 0.23 (0.02-1.6)*o

2.1 (1-8.6) 1.4 (0.77-4.1)

10 (0.9-40)o 0.03 (0-0.99)*o 10 (0.58-96)* 80 (3.1-97)* 2.3 (0.13-48)*o 0.35 (0.02-5)* 6655 (107-157298)o

11 (3-47) 1 (0.22-5.3) 60 (30-95) 36 (4.7-65) 1.6 (0.24-4.6) 0.17 (0.03-0.45) 15875 (10226-81858)

32 (2-70)* 12 (0.22-38)* 14 (1.4-37)* 1 (0.12-2.3) 0.81 (0.01-4.4)*o 11 (1.1-67)o 7.5 (1.7-41)o 30 (5.8-82)*o 12736 (278-75757)*o

11 (4.4-20) 5.4 (1.4-13) 4.2 (1.2-10) 1.2 (0.52-3.4) 0.14 (0.03-2) 8.1 (2.1-20) 6.2 (1.7-14) 12 (1.7-28) 3123 (1599-8779)

2.4 (0.18-10)*

1.5 (0.34-6.8)

93 (70-98)*o 4.2 (1-15)*o

89 (75-92) 8.1 (5.1-13)

RCC: refractory cytopenia in childhood; AA; aplastic anemia. *Indicates significantly different level (P<0.05, Mann-Whitney test) of the respective parameter compared to control samples. oIn non RCC, RCC and AA means significantly different from GATA-2 by Mann-Whitney test (P<0.05).

as RCC regardless of the histopathological picture. Two patients were classified as RCC without histopathological dysplasia: one RCC GATA2 wild-type patient with familiar history of MDS had simultaneous monosomy 7 and trisomy 8, and one GATA-2-deficient patient had monosomy 7. Further details on diagnostics including flow cytometry, cytogenetics and DNA isolation may be found in the Online Supplementary Appendix.

copies per 1 mg of DNA. Thus, the final KREC (TREC) levels in unsorted populations serve as a surrogate marker of developing B (T) lymphocytes, irrespective of their proliferation history.22 The detection limit was 25 copies/mg DNA. The controls for KREC and TREC analyses in PB consisted of 87 samples (median age 8 years; range 0-18 years) and 5 samples from BM (median age 10.1 years; range 5.2-14.2 years). The control groups for PB and dried blood spot had been included in a previous study.21

GATA2 sequencing

Statistical analysis

GATA2 mutation status was investigated in all MDS/AA patients with available material. Genomic DNA was extracted from BM or PB samples. The entire coding region of GATA2 and an intronic enhancer region 3â&#x20AC;&#x2122; to exon 6 were amplified using genomic PCR. Further details may be found in the Online Supplementary Appendix.

KREC/TREC detection Albumin gene level was quantitatively detected in isolated samples using qPCR, and a standard dilution series was derived from human genomic DNA with a known starting concentration (Roche, Basel, Switzerland). The levels of T-cell receptor excision circle (TREC) and KREC signal joints were assessed separately using serial dilutions of cloned plasmid standards, as previously described.19-21 The results were subsequently recalculated to albumin gene levels and expressed as the number of TREC (KREC) 710

Details may be found in the Online Supplementary Appendix.

Results Prevalence of GATA-2 deficiency in pediatric MDS/AA We investigated the prevalence of GATA-2 deficiency in samples taken from Czech pediatric patients under 18 years of age who developed MDS or AA. Eight of 90 pediatric MDS/AA patients had a GATA2 mutation. Three of these patients were diagnosed with non-RCC and 5 patients were diagnosed with RCC (Online Supplementary Figure S1). The prevalence of a GATA2 mutation was 17%, 14% and 0% within RCC, non-RCC, and AA groups, respectively. The prevalence of a GATA2 mutation in patients with cytogenetic aberrations was 41% and 17% in patients with monosomy 7 and trisomy 8, respectively. haematologica | 2016; 101(6)

GATA-2 deficiency in childhood MDS

Figure 1. Cell populations in bone marrow (BM). (A) B-cell subpopulations and kappa-deleting recombination excision circles (KRECs). (B) T-cell subpopulations and T-cell receptor excision circles (TRECs). (C) Natural killer cells. (D) Monocytes. (E) Progenitors. Braces indicate significant difference between the parameters using the non-parametric Mann-Whitney test (P<0.05). Black lines represent medians. A gray area indicates the range of control samples.

Bone marrow histology of GATA-2-deficient patients Histopathology analysis showed no difference between GATA-2-deficient RCC patients and other RCC patients; similarly, no difference was observed between GATA2 mutated patients with advanced MDS and other advanced patients. In our study, all BM samples from RCC patients, including GATA-2-deficient patients, were hypocellular. There was only one patient (UPN3 with Emberger syndrome) in the GATA-2-deficient group with a higher degree of fibrosis (MF-2).

B-cell compartment composition and production of B cells exhibit distinct features, especially in GATA-2deficient patients The proportion of B cells in BM in our control group was inversely correlated with age (Online Supplementary Figure S2), which is in line with previously published data.23,24 The lowest proportion of B cells was present in GATA-2deficient patients (Table 3 and Figure 1). The highest proportion of B cells was present in the AA group, which may be explained by a severe reduction in the myeloid compartment and relative lymphocytosis. The B-cell compartment in AA is composed primarily of mature B cells. B-cell progenitors (defined as haematologica | 2016; 101(6)

CD19pos10pos34pos) out of all cells were significantly lower in all disease groups compared with controls, and the proportion was lowest in GATA-2-deficient patients (Table 3 and Figure 1). There was no significant difference between AA and RCC in the percentage of B-cell progenitors out of all cells. The highest proportion of plasma cells (CD19pos10neg20neg) in B cells was observed in GATA-2-deficient patients (Table 3 and Figure 1). B-cell subsets were analyzed in PB using the lineagedefining marker CD19 in combination with CD27, IgM and IgD. A decrease in na誰ve B cells was observed in 7 of 9 patients with GATA2 mutations (range 15%-64%; median 32%; normal range 47.3%-82.5% for age >5 years25) (Figure 3A). A decreased percentage of PB B cells was present in 10 of 12 patients (Figure 3A). A normal percentage of B cells was observed in one RCC patient (UPN3), and one value out of three was below the normal range in another RCC patient (UPN6) (Figure 3A). The level of KREC signal joints, which correlates with de novo production of B cells, was examined in BM and PB.22 The lowest KREC levels were observed in GATA-2-deficient patients in BM and PB (range BM 25-118; median 25; range PB 25-146, median 25) (Table 3 and Figures 1 and 2). Very low levels of KRECs, together with a decrease in B711

M. Novakova et al. A

x109/L

Absolute count B cells

Absolute count T cells

x109/L

Figure 2. Cell populations in peripheral blood (PB). (A) B cells and kappa-deleting recombination excision circles (KRECs). (B) T cells and T-cell receptor excision circles (TRECs). (C) Natural killer cells. (D) Monocytes. Braces indicate significant differences between parameters using the non-parametric Mann-Whitney test (P<0.05). A gray area indicates the range of control samples. Absolute counts are shown. In relative counts, similar results were found except for relative T cells, which were increased in GATA-2-deficient patients compared to controls.

cell progenitors and a proportional increase in plasma cells in BM, indicate a defect in B-cell production in GATA-2deficient patients. We identified 3 RCC patients with almost no KRECs in BM and PB and no GATA2 mutation (Figures 1 and 2). Notably, one of these patients had a family history of MDS; her mother had undergone SCT for MDS RAEB. We also analyzed available newborn blood spots (Guthrie cards) from 4 GATA-2-deficient patients, which were used for the neonatal screening of metabolic disorders. We observed normal KREC levels at birth in 3 patients (UPN1, UPN6, UPN8), but there were no KRECs at birth in another patient (UPN11) who exhibited the earliest MDS RAEB manifestation in our GATA-2-deficient cohort (Table 2 and Figure 2). Patient UPN9 with a GATA2 intronic mutation pro712

gressed rapidly to AML within three months after initial MDS RAEB diagnosis. AML blasts were immunophenotypically characterized by the co-expression of progenitor markers CD34 and CD117 and myeloid marker CD33. Simultaneously, a subpopulation of CD34pos blasts demonstrated clear B-cell differentiation by CD19, CD10 and CD20 markers (Online Supplementary Figure S3). This uncommon simultaneous presence of AML and B-cell precursor acute lymphoblastic leukemia blasts has not yet been reported in GATA-2-deficient patients.

T cells in GATA-2-deficient patients are proportionally increased in bone marrow and peripheral blood There was no correlation between the percentage of CD3pos T cells in BM and age (Online Supplementary Figure haematologica | 2016; 101(6)

GATA-2 deficiency in childhood MDS

% out of lymphocytes

A Relative count B cells

Absolute count B cells

Relative monocyte count

Absolute monocyte count

Relative neutrophil count

Absolute neutrophil count

Naive B cells CD19pos27negIgDpos

Class switched memory B cells CD19pos27posIgMnegIgDneg

10 8 6 4 2 0

% out of leukocytes

B Relative count NK cells

Absolute count NK cells

100 10 1

% out of leukocytes

0.1

100 80 60 40 20 0

Figure 3. Follow-up peripheral blood (PB) samples of GATA-2-deficient patients. Each column represents one patient. Black lines represent medians. A gray area indicates a normal range for the age category of most of the patients [either 12-18 years for B cells, natural killer (NK) cells; >15 years for monocytes and neutrophils; or >16 years for class switched memory B cells and na誰ve B cells]. (A) B cells and B-cell subpopulations. (B) Monocytes, NK cells, neutrophils.

S2). We found a significantly higher percentage of CD3pos, CD4pos and CD8pos T cells in BM of GATA-2-deficient patients compared with healthy controls (Table 3). Activation of T cells measured by HLA DR expression was significantly lower in GATA-2-deficient patients compared to healthy controls. All disease groups, including GATA-2-deficient patients, exhibited significantly higher amounts of CD3pos16,56pos cells compared with healthy controls (Table 3). We observed an increased percentage of T cells also in PB in GATA-2-deficient patients.

The progenitor compartment is severely reduced in AA The percentage of CD34pos, CD117pos and CD34pos19pos10pos progenitors in healthy pediatric controls was inversely correlated with age (P<0.05) (Online Supplementary Figure S2), as previously published.23,24 As expected, AA presented with a lower total progenitor frequency (assessed as CD34pos or CD117pos) than RCC or GATA-2 deficiency. In contrast, immature B-cell progenitors CD34pos19pos10pos were lower in GATA-2 deficiency compared to AA or RCC subsets. All progenitors (assessed as CD34pos, CD117pos or CD34pos19pos10pos cells) were decreased in all three conditions (AA, RCC and GATA-2 deficiency) compared with controls. However, neither CD34 nor CD117 alone can be used for diagnostic purposes to discriminate between RCC and AA because of the substantial overlap. haematologica | 2016; 101(6)

Myeloid populations and NK cells in GATA-2 deficiency Unexpectedly, the analysis of BM monocytes revealed that the only group that differed from controls was AA presenting with BM monocytopenia (Table 3 and Figure 1). In PB, monocytopenia is often regarded as one of the hallmarks of GATA-2 deficiency.10 Although we did observe absolute monocytopenia at least in some specimens of 10 of 11 GATA-2-deficient patients, the majority of the patients show monocytopenia only in less than half of the investigated periods (Online Supplementary Figure S4). A stable decrease in the percentage and absolute count of monocytes was present in only 2 patients, who both suffered from severe lung disease (Figure 3 and Online Supplementary Figure S4). One patient (UPN7) was recently described as exhibiting diffuse parenchymal lung disease as the first manifestation of GATA-2 deficiency.26 The granulocytic compartment in BM generally contains neutrophils; eosinophils and basophils are only minor subpopulations in normal BM. We focused on the evaluation of the total percentage of granulocytes in BM. There was no difference in percentage of granulocytes between GATA-2-deficient patients and controls (Table 3), but we frequently found aberrancies in maturation as detected by CD16 and CD13 expression in this group. We observed a complete absence of CD16 on all forms of granulocytes in one of 6 patients with GATA-2 deficiency analyzed. We observed a disturbed maturation profile with an accumulation of CD16neg13neg granulocytes (43%) and a reduction 713

M. Novakova et al.

of the mature forms CD16pos13pos (16%) in one patient. This result is consistent with an earlier report.10 Non-RCC and AA patients presented with fewer granulocytes than the remaining cohorts, including controls (Table 3). In PB, neutropenia was frequently present in GATA-2-deficient patients (Figure 3B), and at least some of the absolute neutrophil count (ANC) values were below 1x109/L in all but one patient (UPN7). NK lymphopenia was present in half of the GATA-2-deficient cases (Figure 3B).

B lymphopenia is a more specific and sensitive parameter for discriminating pediatric patients with GATA-2 deficiency To determine which parameters best identify GATA-2deficient patients among patients with MDS and AA, we performed receiver operating characteristic (ROC) curves analysis comparing peripheral blood monocytes, B cells and NK cells (absolute and relative counts). Higher sensitivity and specificity can be reached using either relative or absolute counts of B cells compared to both monocytes and NK cells (Figure 4). Lower-than-physiological counts of B cells were found in 10 of 12 GATA-2-deficient patients (Figure 3A), and decreased KREC levels were found in all GATA-2-deficient patients (Figures 1A and 2A).

Discussion Mutations in the transcription factor GATA-2 leading to haploinsufficiency is a frequent germline genetic aberration found in pediatric MDS. We compared flow cytometry results, KREC and TREC levels in pediatric patients with MDS and AA to pediatric control samples with a focus on GATA2 mutation. The most typical feature in GATA-2-deficient patients is the profound reduction of B cells and their progenitors in BM and PB. A decreased production of B cells was also documented by the low levels of KRECs in BM and PB. KREC levels on Guthrie cards taken for the neonatal screening of inherited disorders of metabolism revealed normal levels in some of the patients, which indicates the normal production of immature B cells prenatally. Three RCC patients without GATA2 mutation and mostly absent KRECs were highly suspected of having an unknown underlying genetic aberration that was responsible for MDS development, but we could not identify any common genetic aberration. Nevertheless, significant decrease of KRECs among GATA-2-deficient patients in BM and in PB (Figures 1A and 2A) indicates that KREC is useful in the diagnostic workup, possibly as a genetic prescreening procedure. Peripheral B-cell subpopulations in GATA-2-deficient patients shift towards mature memory subsets. The production of B cells was defective in GATA2-deficient patients, but immunoglobulin levels were largely normal in most of our patients (data not shown). Immunoglobulin substitution is rarely required in GATA-2deficient patients.27 B-cell progenitors defined as CD19pos10pos34pos were also significantly reduced in all other disease groups (RCC, non-RCC and AA without GATA2 mutations) compared with controls, but the reduction of these cells in GATA-2 deficiency was even more profound (P<0.05). In our cohort, the other features known to be associated with GATA-2 deficiency (monocytopenia and NK lymphopenia) were less discriminative. The most profound BM monocytopenia in the AA group is in contrast to a recently 714

Figure 4. Receiver operating characteristic curves for peripheral blood monocytes, B cells and natural killer (NK) cells in GATA-2-deficient patients in comparison with all other patients with aplastic anemia and myelodysplastic syndromes. Light green: relative B-cell count; dark green: absolute B-cell count; light red: relative monocyte count; dark red: absolute monocyte count; light blue: relative NK-cell count; dark blue: absolute NK-cell count.

published study by Ganapathi et al.10 who found the lowest amount of monocytes in GATA-2-deficient MDS patients, with levels lower than AA. In our GATA-2-deficient cohort, the most profound monocytopenia in BM and PB was found in 2 patients with immunodeficiency and severe lung problems. Two patients with advanced form of MDS presented with monocytosis in BM. This result is consistent with previous observations that GATA-2-deficient patients whose disease progresses into advanced MDS may exhibit monocytosis.28 A recently published study by Wlodarski et al. also reported a tendency towards higher monocytes in pediatric MDS cases with GATA2 mutation (the patients partially overlap with our study). The difference in results between our cohort and those published by Ganapathi et al. might be explained by a lower incidence of advanced MDS cases (RAEB/RAEB-t) (3 of 52 vs. 3 of 12 in our study; Ď&#x2021;2 P=0.04) and by lower incidence of monosomy 7 (4 of 48 vs. 4 of 12; Ď&#x2021;2 P=0.01). A study by Pasqet et al. found significant monocytopenia and analyzed blood counts before the MDS/AML phase. A study by Spinner et al. also identified significant monocytopenia in GATA-2deficient patients and included predominantly immunodeficient cases (this study overlaps with the study by Ganapathi et al., which selected patients with MDS/AML). The median age of our study was lower in comparison to Ganapathi et al. In our pediatric cohort, both PB relative and absolute B-cell lymphopenia were more specific and sensitive parameters in comparison with absolute and relative monocytopenia (Figure 4). Neutropenia is a frequent finding in GATA-2-deficient patients, and it contributes to their immunodeficiency symptoms.29,30 A study by Pasquet et al. identified GATA-2-deficient patients in the cohort based on neutropenia.30 Nearly normal ANC values were present only in UPN5 and UPN7, which was likely related to longterm use of corticosteroids due to lung disease. Relative Tcell counts were increased in GATA-2-deficient patients in PB and BM. Low levels of TRECs in GATA-2 may be explained by the decreased production and/or by expansion of mature T cells during infections. In contrast to prehaematologica | 2016; 101(6)

GATA-2 deficiency in childhood MDS

viously published GATA-2-deficient cohorts, we did not observe an expansion of T-cell large granular lymphocytes in any of our patients. T-cell deficiency, namely CD4 lymphopenia, contributes to immunodeficiency in GATA2mutated patients.28 We observed CD4 lymphopenia below 0.4x109/L in 3 patients; 2 of them were followed for severe lung disease. We assessed the prevalence of GATA-2 deficiency in Czech children with MDS or AA. A GATA2 mutation was exclusively identified in patients with RCC (17%) or advanced MDS (non-RCC; 14%). Aplastic anemia and RCC generally exhibit similar clinical and laboratory features. Some of the differences in flow cytometry between RCC and AA that we had observed in the past had been driven by the GATA-2-deficient group (namely B-cell lymphopenia).31,32 Differences between the overlapping categories RCC and AA are a frequent subject of discussion, especially in patients with hypocellular BM and without adverse cytogenetics. The separation of patients into two categories seems to be less relevant because immunosuppressive therapy is indicated for both disease groups,33-35 and there is no difference in prognosis nor in the probability of progression into advanced MDS.35 We observed significant differences between RCC and AA in a limited number of parameters in BM (decreased in AA: CD34pos, CD117pos, granulocytes and erythroid precursors, increased in AA: CD19pos and lymphocytes). In summary, we found that the disturbances in the B-cell compartment were the strongest distinguishing biological feature of GATA-2 deficiency in childhood MDS, in contrast to other recently published factors, such as monocytopenia, which were less common and unspecific in our

References 1. Hasle H, Wadsworth LD, Massing BG, McBride M, Schultz KR. A populationbased study of childhood myelodysplastic syndrome in British Columbia, Canada. Br J Haematol. 1999;106(4):1027-1032. 2. Baumann I, Niemeyer CM, Bennett JM, Shannon K. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. In: Swerdlow SH, Campo E, Harris NL, Jaffe ES, Pileri SA, Stein H, et al., eds. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon: International Agency of Research on Cancer (IARC); 2008. p 104-107. 3. Baumann I, FĂźhrer M, Behrendt S, et al. Morphological differentiation of severe aplastic anaemia from hypocellular refractory cytopenia of childhood: reproducibility of histopathological diagnostic criteria. Histopathology. 2012;61(1):10-17. 4. Strahm B, NĂśllke P, Zecca M, et al. Hematopoietic stem cell transplantation for advanced myelodysplastic syndrome in children: results of the EWOG-MDS 98 study. Leukemia. 2011;25(3):455-462. 5. Shimamura A, Alter BP. Pathophysiology and management of inherited bone marrow failure syndromes. Blood Rev. 2010; 24(3):101-122. 6. Ostergaard P, Simpson MA, Connell FC, et al. Mutations in GATA2 cause primary lymphedema associated with a predisposition to acute myeloid leukemia (Emberger

haematologica | 2016; 101(6)

10.

11. 12.

13.

study. The finding of decreased B-cell numbers in BM and PB, and most specifically, low levels of B-cell progenitors in BM, together with very low or absent KRECs in BM and PB, can identify appropriate candidates for GATA2 mutation testing in pediatric MDS patients. Information on mutational status in the family is of importance not only when matched family donor is considered for the transplantation. GATA-2 deficiency does not only predispose to cancer, but also to an immunodeficient condition in which close immunological monitoring with careful treatment of infections might prevent organ damage, as we observed in patient UPN7. Funding This work was a main result of grant NT14534-3. Z.Z. was supported by RVO-VFN64165/2012, JT was supported by P302/12/G101, EM was supported by Ministry of Health of the Czech Republic grant nr. 15-28525A, MN was supported by GAUK 802214 and UNCE 204012. This work was also supported by the project for conceptual development of research organization 00064203 (University Hospital Motol, Prague, Czech Republic) and Ministry of Education, Youth and Sports NPU I nr. LO1604, infrastructure was supported by EU-Prague project CZ.2.16/3.1.00/24505. Acknowledgments The authors would like to thank Iveta Janotova for data management, Pavel Semerak, Pavla Luknarova and Daniel Thurner for processing flow cytometry samples, Jan Stuchly for consulting statistics and the Czech Pediatric Hematology Group for their collaboration (Doctors Sterba, Timr, Mihal, Prochazkova, Blazek and Hak).

syndrome). Nat Genet. 2011;43(10):929931. Dickinson RE, Griffin H, Bigley V, et al. Exome sequencing identifies GATA-2 mutation as the cause of dendritic cell, monocyte, B and NK lymphoid deficiency. Blood. 2011;118(10):2656-2658. Hahn CN, Chong C-E, Carmichael CL, et al. Heritable GATA2 mutations associated with familial myelodysplastic syndrome and acute myeloid leukemia. Nat Genet. 2011;43(10):1012-1017. Hsu AP, Johnson KD, Falcone EL, et al. GATA2 haploinsufficiency caused by mutations in a conserved intronic element leads to MonoMAC syndrome. Blood. 2013;121(19):3830-3837, S1-S7. Ganapathi KA, Townsley DM, Hsu AP, et al. GATA2 deficiency-associated bone marrow disorder differs from idiopathic aplastic anemia. Blood. 2015;125(1):56-70. Dickinson RE, Milne P, Jardine L, et al. The evolution of cellular deficiency in GATA2 mutation. Blood. 2014;123(6):863-874. Calvo KR, Vinh DC, Maric I, et al. Myelodysplasia in autosomal dominant and sporadic monocytopenia immunodeficiency syndrome: diagnostic features and clinical implications. Haematologica. 2011; 96(8):1221-1225. Hirabayashi S, Strahm B, Urbaniak S, et al. Unexpected High Frequency of GATA2 Mutations in Children with Non-Familial MDS and Monosomy 7. ASH Annual Meeting Abstracts. 2012;120(21):1699.

14. Westers TM, Ireland R, Kern W, et al. Standardization of flow cytometry in myelodysplastic syndromes: a report from an international consortium and the European LeukemiaNet Working Group. Leukemia. 2012;26(7):1730-1741. 15. Della Porta MG, Picone C, Pascutto C, et al. Multicenter validation of a reproducible flow cytometric score for the diagnosis of low-grade myelodysplastic syndromes: results of a European LeukemiaNET study. Haematologica. 2012;97(8):1209-1217. 16. van de Loosdrecht AA, Westers TM, Westra AH, et al. Identification of distinct prognostic subgroups in low- and intermediate-1-risk myelodysplastic syndromes by flow cytometry. Blood. 2008;111(3):10671077. 17. Aalbers AM, van den Heuvel-Eibrink MM, de Haas V, et al. Applicability of a reproducible flow cytometry scoring system in the diagnosis of refractory cytopenia of childhood. Leukemia. 2013;27(9):19231925. 18. Aalbers AM, van den Heuvel-Eibrink MM, Baumann I, et al. Bone marrow immunophenotyping by flow cytometry in refractory cytopenia of childhood. Haematologica. 2015;100(3):315-323. 19. van Zelm MC, Szczepanski T, van der Burg M, van Dongen JJM. Replication history of B lymphocytes reveals homeostatic proliferation and extensive antigen-induced B cell expansion. J Exp Med. 2007;204(3):645655.

715

M. Novakova et al.

20. Weinberg K, Blazar BR, Wagner JE, et al. Factors affecting thymic function after allogeneic hematopoietic stem cell transplantation. Blood. 2001;97(5):1458-1466. 21. Fronková E, Klocperk A, Svatoñ M, et al. The TREC/KREC assay for the diagnosis and monitoring of patients with DiGeorge syndrome. PLoS One. 2014;9(12):e114514. 22. Fronkova E, Muzikova K, Mejstrikova E, et al. B-cell reconstitution after allogeneic SCT impairs minimal residual disease monitoring in children with ALL. Bone Marrow Transplant. 2008;42(3):187-196. 23. Veltroni M, Sainati L, Zecca M, et al. Advanced pediatric myelodysplastic syndromes: can immunophenotypic characterization of blast cells be a diagnostic and prognostic tool? Pediatr Blood Cancer. 2009;52(3):357-363. 24. Bras AE, van den Heuvel-Eibrink MM, van der Sluijs-Gelling AJ, et al. No significant prognostic value of normal precursor B-cell regeneration in paediatric acute myeloid leukaemia after induction treatment. Br J Haematol. 2013;161(6):861-864. 25. Piatosa B, Wolska-Kušnierz B, Pac M, et al. B cell subsets in healthy children: reference values for evaluation of B cell maturation

716

26.

27.

28.

29.

30.

31.

process in peripheral blood. Cytometry B Clin Cytom. 2010;78(6):372-381. Svobodova T, Mejstrikova E, Salzer U, et al. Diffuse parenchymal lung disease as first clinical manifestation of GATA-2 deficiency in childhood. BMC Pulm Med. 2015;15:8. Chou J, Lutskiy M, Tsitsikov E, et al. Presence of hypogammaglobulinemia and abnormal antibody responses in GATA2 deficiency. J Allergy Clin Immunol. 2014; 134(1):223-226. Spinner MA, Sanchez LA, Hsu AP, et al. GATA2 deficiency: a protean disorder of hematopoiesis, lymphatics, and immunity. Blood. 2014;123(6):809-821. Dotta L, Badolato R. Primary immunodeficiencies appearing as combined lymphopenia, neutropenia, and monocytopenia. Immunol Lett. 2014;161(2):222-225. Pasquet M, Bellanné-Chantelot C, Tavitian S, et al. High frequency of GATA2 mutations in patients with mild chronic neutropenia evolving to MonoMac syndrome, myelodysplasia, and acute myeloid leukemia. Blood. 2013;121(5):822-829. Mejstrikova E, Pelkova V, Reiterova M, et al. Composition of Cellular Subsets by Flow Cytometry Identifies Differences

32.

33.

34.

35.

Between MDS Subtypes and Aplastic Anemia but No Differences Are Identified Between Cases with and without Monosomy 7. ASH Annual Meeting Abstracts. 2009;114(22):3802. Reiterova M, Kramarzova K, Sukova M, et al. Changes Identified by Flow Cytometry and WT1 Expression in Consecutive Bone Marrow Samples in Refractory Cytopenia of Childhood and Aplastic Anemia Before Start of the Therapy. ASH Annual Meeting Abstracts. 2011;118(21):1342. Yoshimi A, van den Heuvel-Eibrink MM, Baumann I, et al. Comparison of horse and rabbit antithymocyte globulin in immunosuppressive therapy for refractory cytopenia of childhood. Haematologica. 2014; 99(4):656-663. Führer M, Rampf U, Baumann I, et al. Immunosuppressive therapy for aplastic anemia in children: a more severe disease predicts better survival. Blood. 2005; 106(6):2102-2104. Forester C, Sartain S, Guo D, et al. Pediatric aplastic anemia and refractory cytopenia: A retrospective analysis assessing outcomes and histomorphologic predictors. Am J Hematol. 2015;90(4):320-326.

haematologica | 2016; 101(6)

ARTICLE

Chronic Myeloid Leukemia

Imatinib withdrawal syndrome and longer duration of imatinib have a close association with a lower molecular relapse after treatment discontinuation: the KID study Sung-Eun Lee,1 Soo Young Choi,1 Hye-Young Song,1 Soo-Hyun Kim, 1 Mi-Yeon Choi, 1 Joon Seong Park,2 Hyeoung-Joon Kim,3 Sung-Hyun Kim,4 Dae Young Zang,5 Sukjoong Oh,6 Hawk Kim,7 Young Rok Do,8 Jae-Yong Kwak,9 Jeong-A Kim,10 Dae-Young Kim,11 Yeung-Chul Mun,12 Won Sik Lee,13 Myung Hee Chang,14 Jinny Park,15 Ji Hyun Kwon,16 and Dong-Wook Kim1,17

Department of Hematology, Seoul St. Mary's Hospital, The Catholic University of Korea, Seoul; 2Department of Hematology-Oncology, Ajou University School of Medicine, Suwon; 3Department of Hematology-Oncology, Chonnam National University Hwasun, Hospital; 4Department of Internal Medicine, Dong-A University College of Medicine, Busan; 5Department of Internal Medicine, Hallym University College of Medicine, Anyang; 6Division of Hematology-Oncology, Department of Internal Medicine, Kangbuk Samsung Hospital, School of Medicine, Sungkyunkwan University, Seoul; 7Division of Hematology and Cellular Therapy, Ulsan University Hospital, University of Ulsan College of Medicine; 8Division of Hematology-Oncology, School of Medicine, Keimyung University, Daegu; 9Division of Hematology-Oncology, Chonbuk National University Medical School, Jeonju; 10Department of Hematology, St. Vincent’s Hospital, The Catholic University of Korea, Suwon; 11Department of Hematology, Asan Medical Center, University of Ulsan College of Medicine, Seoul; 12Department of Hematology, School of Medicine, Ewha Womans University, Seoul; 13Department of Internal Medicine, Inje University College of Medicine, Inje University Busan Paik Hospital; 14Department of Hematology-Oncology, National Health Insurance Service Ilsan Hospital, Ilsan; 15 Department of Hematology, Gachon University Gil Hospital, Incheon; 16Department of Hematology-Oncology, Chungbuk National University Hospital, Cheongju; and 17Catholic Leukemia Research Institute, The Catholic University of Korea, Seoul, Korea

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2016 Volume 101(6):717-723

Correspondence: ABSTRACT

he aim of the Korean Imatinib Discontinuation Study was to identify predictors for safe and successful imatinib discontinuation. A total of 90 patients with a follow-up of ≥12 months were analyzed. After a median follow-up of 26.6 months after imatinib discontinuation, 37 patients lost the major molecular response. The probability of sustained major molecular response at 12 months and 24 months was 62.2% and 58.5%, respectively. All 37 patients who lost major molecular response were retreated with imatinib therapy for a median of 16.9 months, and all achieved major molecular response again at a median of 3.9 months after resuming imatinib therapy. We observed newly developed or worsened musculoskeletal pain and pruritus in 27 (30%) patients after imatinib discontinuation. Imatinib withdrawal syndrome was associated with a higher probability of sustained major molecular response (P=0.003) and showed a trend for a longer time to major molecular response loss (P=0.098). Positivity (defined as ≥ 17 positive chambers) of digital polymerase chain reaction at screening and longer imatinib duration before imatinib discontinuation were associated with a higher probability of sustained major molecular response. Our data demonstrated that the occurrence of imatinib withdrawal syndrome after imatinib discontinuation and longer duration of imatinib were associated with a lower rate of molecular relapse. In addition, minimal residual leukemia measured by digital polymerase chain reaction had a trend for a higher molecular relapse. (Trial registered at ClinicalTrials.gov: NCT01564836). haematologica | 2016; 101(6)

dwkim@catholic.ac.kr

Received: November 23, 2015. Accepted: February 17, 2016. Pre-published: February 17,2016. doi:10.3324/haematol.2015.139899

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/06/717

717

S.-E. Lee et al.

Introduction

Table 1. Patient characteristics.

Parameters Since the publication of several recent reports showing that imatinib (IM) discontinuation can be performed in patients who have had sufficient IM therapy and have undetectable molecular residual disease (UMRD), treatment-free remission (TFR) has emerged as a goal in the treatment of chronic myeloid leukemia (CML).1-5 Although UMRD with a sensitivity of 4.5-log for at least 2 years is typically required and some factors associated with a higher rate of sustained UMRD after IM discontinuation have been reported, the precise identification of the minimum requirement for a safe and successful TFR and the underlying mechanism of sustained UMRD remain elusive. The IM discontinuation studies have presented several predictors for TFR.1,3,4,6,7 Recently, Horn et al. provided a predictive mathematical model to prognosticate the patient-specific risk of molecular relapse on treatment discontinuation,8 and the TWISTER study tested the utility of DNA analysis by polymerase chain reaction (PCR).4 Currently, the ISAV and EURO-SKI studies are exploring the additional predictors of age, digital polymerase chain reaction (dPCR) results, and withdrawal syndromes such as musculoskeletal pain and pruritus.9,10 Our previous study of health-related profiles including quality of life (HRQOL) also demonstrated that IM withdrawal syndrome can develop in some patients.11 This Korean multicenter prospective study (Korean Imatinib Discontinuation Study; KID Study) has been performed to identify predictors for safer and more successful IM discontinuation and to explore additional contributing factors for a sustained molecular response (MR) in specific patient cohorts. Herein, we analyzed the results of patients with a sufficient follow-up to confirm the significant predictors of molecular relapse using preliminary identified factors.

Patients with follow-up â&#x2030;Ľ12 months (N=90)

Age (yr), median (range) Sex, Female, N(%) Transcript type,N(%) b3a2*/b2a2/NA Previous interferon therapy, N(%) No/Yes Sokal risk, N(%) Low/intermediate High/ NA Reason for IM discontinuation, N(%) Low-grade adverse events Inconvenience of daily doses Medication cost Pregnancy/planning pregnancy Patient preference Time from IM therapy to UMRD (mo), median (range) UMRD duration before IM cessation (mo), median (range) IM therapy duration (mo), median (range)

56.2 (26-82) 52 (58) 66 (73)/19 (21)/5 (6) 82 (91) / 8(9) 29 (32) / 23 (26) 15 (17) / 23 (26) 45 (50) 3 (3) 1 (1) 3 (3) 38 (42) 25.7 (4.8-114.1) 39.9 (22.1-130.7) 80.8 (38.2-141.3)

IM: imatinib; NA: not available; UMRD: undetectable molecular residual disease; mo: months.*One patient had a type of b3a2+b2a2.

tion).11 As newly developed or worsened musculoskeletal pain and pruritus were observed in a significant proportion of patients, all patients were asked to complete an additional questionnaire for IM withdrawal syndrome (IMWS).

Evaluation of molecular response Methods Study patients Patients with chronic-phase CML who were treated with firstline IM for more than 3 years and had undetectable BCR-ABL1 transcript by quantitative reverse transcription PCR (qRT-PCR) for at least 2 years were eligible for the KID study. Previous reports from the KID study included 48 patients who were enrolled from 9 October 2010 to 7 November 2012; 28 patients who had IM without transplantation, and 20 patients who received IM for post-transplant relapse.3 By the data cut-off date of 31 July 2015 for this analysis, a total of 156 patients were enrolled. Among them, 90 non-transplant patients with at least 12 months of follow-up were included in this analysis. Written informed consent was obtained from each patient before enrollment. This study was approved by the Institutional Review Board of each participating institution and conducted in accordance with the Declaration of Helsinki. The study protocol was registered with the National Institutes of Health clinical trial registry as #NCT01564836.

To confirm undetectable BCR-ABL1 transcript levels by qRTPCR for at least 2 years, duplicated qRT-PCR analyses were performed at more than six time points. For all screening and subsequent follow-up samples, duplicated qRT-PCR and nested RTPCR with at least 4.5-log sensitivity was performed in the central laboratory (Cancer Research Institute, The Catholic University of Korea, Seoul, Korea) and only qRT-PCR results with more than 50,000 ABL1 transcripts were analyzed. Major molecular response (MMR) was defined as a BCR-ABL1 transcript level of 0.1% or lower on the international scale (IS). UMRD was defined as negative PCR results in duplicated qRT-PCR assays with 5-log sensitivity. After IM discontinuation, the molecular response was monitored every month for the first 6 months, every 2 months up to 12 months, and every 3 months thereafter. Loss of MMR and UMRD was defined after 2 consecutive assessments within 4 weeks, and in cases of MMR loss (molecular relapse) IM treatment was reintroduced. After IM resumption for MMR loss, the molecular response was evaluated every month until MMR was re-achieved and every 3 months thereafter.

Questionnaire relating to withdrawal syndrome Patients were provided with a questionnaire composed of 43 parameters in 3 sections (IM-related adverse events, physical health, and mental health). Each parameter was rated using a fivepoint score at each time point and changes were calculated as differences in score from the baseline (at the time of IM discontinua718

Digital PCR For more sensitive quantitation of minimal residual disease, all screening samples were tested with 10-7 sensitivity by digital PCR (dPCR) as previously reported.9,12 A brief method is provided in the Online Supplementary Methods. To identify the optimal cut-off haematologica | 2016; 101(6)

Predictors of imatinib free remission in CML

in the 765 chambers analyzed by dPCR for predicting loss of MMR, we performed a receiver operating characteristic curve analysis and applied different cut-off levels iteratively in steps of 1 positive chamber. A cut-off of more than 17 positive chambers was chosen for a positive result of dPCR with a significantly lower probability of sustained MMR.

Statistical analysis Time to loss of UMRD and MMR was calculated from the date of IM discontinuation to the date of the first detection of BCR-ABL1 in two consecutive analyses or, for patients who did not lose MR, was determined by default on the date of the last molecular examination. The probability of sustained MR was plotted using the Kaplan-Meier method and compared using the log-rank test.

Results Patient characteristics A total of 90 patients with UMRD (38 men and 52 women) with a median age of 56.2 years (range, 26-82 years) were analyzed. The percentage of patients with low, intermediate, and high Sokal risk scores was 32%, 26%, and 17%, respectively, and 26% had an unknown risk. Eight patients had a previous history of interferon treatment. The reasons for participating in this study included low-grade adverse events (n=45), inconvenience of daily doses (n=3), medication cost (n=1), pregnancy/planning pregnancy (n=3), and patientâ&#x20AC;&#x2122;s wishes (n=38). Prior to discontinuation, the median duration of IM therapy was 80.8 months (range, 38.2-141.3 months) and the duration of sustained UMRD was 39.9 months (range, 22.1-130.7 months). Patient characteristics are summarized in Table 1.

Outcomes after discontinuation of IM therapy After a median follow-up of 26.6 months (range, 12.658.1 months) after IM discontinuation, 45 patients (50%) lost UMRD. Eight patients who lost UMRD but not MMR exhibited different patterns of BCR-ABL1 kinetics: 7 patients spontaneously re-achieved UMRD after a median time of 1.2 months (range, 0.7-5.4 months) and 1 patient showed fluctuation of BCR-ABL1 transcript under the level of 0.1% on IS for 20.6 months. The other 37 patients lost MMR in 2 consecutive analyses. The overall 12month and 24-month probability of sustained MMR was 62.2% and 58.5%, respectively (Figure 1A). The overall 12-month and 24-month probability of sustained UMRD was 50.0% and 50.0%, respectively. Among the 37 patients with molecular relapse, the median time to MMR loss was 3.3 months (range, 0.9-20.8 months) after IM discontinuation (Figure 1B) and the median time from loss of UMRD to loss of MMR was 1.0 month (range, 0.0-13.6 months) (Table 2).

Molecular kinetics after IM resumption All 37 patients who lost MMR were retreated with IM therapy for a median of 16.9 months (range, 4.6-47.4 months). All patients re-achieved MMR at a median of 3.9 months (range, 0.5-11.1 months) after resuming IM therapy, and 32 re-achieved UMRD at a median of 7.2 months (range, 3.3-14.7 months). The other five patients are still on MMR for a median of 6.3 months (range 4.6-23.2) of follow-up duration. haematologica | 2016; 101(6)

Figure 1. Probability of sustained MMR (A) and time to MMR loss (B) in patients with follow-up â&#x2030;Ľ 12 months (N=90).

Association of baseline minimal residual disease with molecular kinetics Among 88 patients with available dPCR data at screening, 16 (18%) patients had a positive result of dPCR based on a cut-off of 17 positive chambers. Ten out of 16 (63%) patients with positivity of dPCR and 26 out of 72 (36%) patients with negativity of dPCR had molecular relapse. Patients in the dPCR-negative group had a higher probability of sustained MMR than patients in the dPCR-positive group (63.8% vs. 37.5%, P=0.021) (Figure 2A). Among 37 patients with molecular relapse, the median time to MMR loss was not different according to positivity of dPCR at baseline (3.6 months in the dPCR negative group vs. 2.8 months in the dPCR positive group, P=0.989) (Figure 2B).

Association of withdrawal syndrome with molecular kinetics Changes in musculoskeletal pain and pruritus after IM discontinuation were carefully evaluated for all 90 patients. Interestingly, aggravation or new development of musculoskeletal pain and/or pruritus after IM discontinuation was presented in 27 (30%) patients, which was defined as IMWS. Online Supplementary Tables S1 and S2 present the characteristics of patients with aggravation or new development of musculoskeletal pain and pruritus after IM discontinuation, respectively. Musculoskeletal pain seemed to continue for quite a while, with a median duration of 6 months (range 1-36 months), and pruritus continued for a median of 3 months (range 2-54 months). In addition, there was no patient who restarted IM due to IMWS. Five patients in the IMWS group (18%) lost MMR compared with 32 (51%) patients in the no IMWS group. Patients with IMWS had a higher probability of sustained 719

S.-E. Lee et al. Table 2. Outcomes after imatinib discontinuation.

Patients with follow-up â&#x2030;Ľ 12 months (N=90) Follow-up duration (mo), median (range) Loss of UMRD, N (%) Loss of MMR, N (%) Time to loss of UMRD (mo) median (range) Time to loss of MMR (mo), median (range) Time to from loss of UMRD to loss of MMR (mo), median (range)

26.6 (12.6-58.1) 45* (50) 37 (41) 2.0 (0.9-9.5) 3.3 (0.9-20.8) 1.0 (0-13.6)

mo: months; UMRD: undetectable molecular residual disease; MMR: major molecular response. *Among 8 patients who lost UMRD but not MMR, 7 patients spontaneously reachieved UMRD and 1 patient showed fluctuation of BCR-ABL1 transcript under the level of 0.1% on IS.

MMR than patients without IMWS (79.5% vs. 49.2%, P=0.003) (Figure 3A). Among 37 patients with molecular relapse, the median time to MMR loss was longer in the patients with IMWS than in those without IMWS (3.0 months vs. 7.6 months, P=0.098) (Figure 3B). In addition creatine phosphokinase (CPK) was not noted in association with the IMWS (Online Supplementary Figure S1). The elevation of erythrocyte sedimentation rate (ESR) and Creactive protein (CRP) were noted in 5 and 4 of 8 evaluable patients with IMWS, respectively, whereas both were normal in 9 evaluable patients without IMWS (data not shown).

Overall predictive factors affecting sustained MMR Univariate and multivariate analyses for predictive factors of sustained MMR were performed (Table 3). In addition to IMWS (P=0.003), negativity of digital PCR at screening (P=0.021) and longer duration of IM therapy (P=0.013) were identified as potential factors for sustained MMR. Longer duration of UMRD (P=0.084) showed a trend for a higher rate of sustained MMR. Multivariate analyses showed that the presence of IMWS (P=0.021) and longer duration of IM therapy (P=0.033) were independent factors for sustained MMR, and negativity of dPCR at screening (P=0.055) maintained a trend for sustained MMR. Combining models according to the number of predictive factors for sustained MMR allows better discrimination (Figure 4).

Discussion Although the results of prospective clinical trials indicated the possibility of TFR in 30-50% of patients after IM discontinuation, because the variables for a sustained molecular response have changed and relapse rates have increased with longer follow-up, many clinical questions regarding safe and successful IM discontinuation remain.2, 13-15 We previously reported results from 48 patients showing that the 12-month probability of sustained MMR and UMRD was 79.9% and 80.8%, respectively (64.4% and 66.3% among 28 non-transplant patients).3 In this further analysis, the 12-month probability of sustained MMR and UMRD was 62.2% and 50.0%, respectively. This rate was similar to that reported by the A-STIM trial (median IM duration 79 months; median UMRD duration 41 months)2 and slightly higher than that reported by the TWISTER trial (median IM duration 70 months; median UMRD duration 36 months).4 In our study, the median duration of 720

Figure 2. Probability of sustained MMR (A) and time to MMR loss (B) according to positivity of dPCR (defined as > 17 positive chambers) at baseline.

IM therapy and UMRD before IM discontinuation was 80.8 months and 39.9 months, respectively. We employed stringent PCR sensitivity criteria for accurate measurement of BCR-ABL1 transcript levels prior to discontinuation and examined BCR-ABL1 kinetics in response to IM resumption and after IM cessation in detail. The confirmation of predictive factors for durable TFR is a key issue. Increased IM therapy duration was reported to be strongly associated with a higher probability of sustained UMRD in previous studies.1,3,6,7 Sokal risk score,1,4,6 sex,1 prior interferon treatment,4,7 UMRD duration before IM discontinuation,3,7 time to UMRD,6 age,9 and dPCR9 were also reported as potential predictive factors, albeit with conflicting data. To identify additional predictors and to validate predictors explored in our previous study,3 we performed further analysis with more patients who had at least 12 months of follow-up. This analysis was based on additional clinical observations: first, we observed an interesting phenomenon in some patients who developed musculoskeletal pain with or without pruritus after discontinuation of IM, implying a possible link to autoimmunity;11 and second, detectable minimal residual leukemia by the more sensitive approach of dPCR may have a significant positive predictive value for molecular relapse.9 In this study, we observed that most IM-related adverse events such as nausea, indigestion, peripheral edema, and skin whitening and fragility were rapidly resolved, whereas musculoskeletal pain and pruritus were newly developed or worsened in some patients after IM discontinuation. These observations are similar to the findings of the haematologica | 2016; 101(6)

Predictors of imatinib free remission in CML

EURO-SKI study,10 in which 15 (30%) out of 50 patients reported musculoskeletal pain evolving gradually from 1 to 6 weeks after TKI discontinuation, and may suggest that the newly developed or worsened musculoskeletal

pain and pruritus were due to IMWS. Interestingly, our study found an association between IMWS and sustained MMR: the patients with IMWS showed a higher rate of sustained MMR and a longer time to MMR loss.

Table 3. Univariate and multivariate analyses of variables affecting the probability of sustained MMR in the patients with follow-up ≥12 months (N=90).

Variables

Univariate analyses Probability of sustained MMR,%

Age of patient, years Sex of patient Male Female Transcript type b3a2 b2a2 (NA=5) Sokal risk Low Intermediate High (NA=23) IM withdrawal syndrome Yes No Nested RT-PCR at IM cessation Positive Negative Digital PCR at IM cessation Positive* Negative (NA = 2) Time from IM therapy to UMRD <24 months ≥24 months UMRD duration before IM cessation <36 months ≥36 months IM therapy duration <62 months ≥62 months

(continuous)

38 52

55.1±8.1 60.7±6.9

66 19

58.4±6.2 47.4±11.5

29 23 15

62.1±9.0 60.6±10.3 46.7±12.9

27 63

79.5±8.4 49.2±6.3

14 76

64.3±12.8 57.6±5.7

16 72

37.5±12.1 63.8±5.7

35 55

48.6±8.4 64.8±6.6

36 54

48.4±8.7 64.8±6.5

25 65

40.0±9.8 65.7±6.0

Multivariate analyses RR (95% CI)

0.404 0.643

0.424

0.300

1 3.09 (1.19-8.03)

0.021

1 0.48 (0.23-1.01)

0.055

0.152

0.084

1 0.48 (0.24-0.94)

0.033

0.003

0.756 0.021

0.013

*Positivity of digital PCR was defined as >17 positive chambers of the 765 chambers.

Figure 3. Probability of sustained MMR (A) and time to MMR loss (B) according to the presence of IM withdrawal syndrome. haematologica | 2016; 101(6)

721

S.-E. Lee et al.

Figure 4. Probability of sustained MMR according to the number of predictive factors (negative dPCR, presence of IMWS, and longer than 62 months of IM therapy duration).

Considering the role of inflammatory cytokines in musculoskeletal pain,16 whether IMWS is linked to immunologic background warrants further investigation. Briefly, we observed that ESR and CRP were increased in some patients with IMWS (data not shown) but CPK were not noted in association with the musculoskeletal symptoms. The importance of immunologic status for successful TFR is illustrated by the beneficial effect of natural killer (NK) cell activity17,18 and interferon treatment on a higher TFR rate,4,7 indicating that the re-activation of autologous immunity may contribute to the elimination of MRD after IM discontinuation via the emergence of NK cells or cytotoxic T lymphocytes (CTLs) against CML-associated antigens.19-21 Although we analyzed the frequency and function of NK cells after IM discontinuation in some patients, no significant association was found. In our previous study, we showed that 24 out of 32 PCR-negative samples as assayed by conventional qRTPCR showed detectable BCR-ABL1 in dPCR. More importantly, dPCR with preamplification showed a greater sensitivity with 2-3-log improvement and a continuous decrease in BCR-ABL1 upon IM treatment was evident in the follow-up samples analyzed by dPCR.12 These results suggest that the duration of UMRD on continued IM treatment will be an important predictor of sustained molecular remission and that the detection of very low levels of BCR-ABL1 by dPCR may help to select candidates for a trial of TFR. In this study, positivity of dPCR at baseline was confirmed as an important determinant of molecular relapse. However, there are conflicting data regarding the association of the depth of MRD at baseline. In A-STIM, which included patients with intermittent low-level positive qRT-PCR tests during the 2 years prior to IM discontinuation, there was no difference in the rate of loss of MMR according to MRD.2 In

References 1. Mahon FX, Rea D, Guilhot J, et al. Discontinuation of imatinib in patients with chronic myeloid leukaemia who have maintained complete molecular remission for at least 2 years: the prospective, multicentre Stop Imatinib (STIM) trial. Lancet Oncol. 2010;11(11):1029-1035.

722

the TWISTER study, which used a highly sensitive (6-log below baseline) patient-specific assay for BCR-ABL1 DNA, the presence or absence of MRD was not informative with regard to the subsequent loss of MMR, although this approach did enable earlier detection of MMR loss than qRT-PCR.4 In addition to the persistence of BCR-ABL1 genomic DNA in non-cycling CML stem cell clones, this discrepancy may be caused by the diverse factors affecting the sustained MR, including the biologic nature of high-risk disease and host immune response. Therefore, further studies are warranted to validate the utility of dPCR in the TFR setting and to determine concomitant immunobiologic factors. In terms of BCR-ABL1 transcript kinetics after IM discontinuation, in our study, among the 37 patients who lost MMR, 20 lost MMR within 3 months of discontinuing IM but the latest loss of MMR was at 20.8 months. This implies that qRT-PCR monitoring should be continued to safeguard patients with delayed relapse. However, the fact that most patients re-achieved MMR at approximately 4 months after resuming IM therapy supported the safety of IM discontinuation under intensive follow-up. Our data demonstrated that the occurrence of IMWS after discontinuation of long-term IM therapy was closely associated with a lower molecular relapse, and longer duration of IM provided a predictive factor for successful IM discontinuation. In addition, minimal residual leukemia measured by dPCR had a trend for a higher molecular relapse. Funding This study was supported by a grant from the National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea (1020400) and the Korea Leukemia Bank (NRF-2013M3A9B8031236).

2. Rousselot P, Charbonnier A, ConyMakhoul P, et al. Loss of major molecular response as a trigger for restarting tyrosine kinase inhibitor therapy in patients with chronic-phase chronic myelogenous leukemia who have stopped imatinib after durable undetectable disease. J Clin Oncol. 2014;32(5):424-430. 3. Lee SE, Choi SY, Bang JH, et al. Predictive factors for successful imatinib cessation in

chronic myeloid leukemia patients treated with imatinib. Am J Hematol. 2013;88(6): 449-454. 4. Ross DM, Branford S, Seymour JF, et al. Safety and efficacy of imatinib cessation for CML patients with stable undetectable minimal residual disease: results from the TWISTER study. Blood. 2013;122(4):515522. 5. Thielen N, van der Holt B, Cornelissen JJ, et

haematologica | 2016; 101(6)

Predictors of imatinib free remission in CML

10.

al. Imatinib discontinuation in chronic phase myeloid leukaemia patients in sustained complete molecular response: a randomised trial of the Dutch-Belgian Cooperative Trial for Haemato-Oncology (HOVON). Eur J Cancer. 2013;49(15):32423246. Yhim HY, Lee NR, Song EK, et al. Imatinib mesylate discontinuation in patients with chronic myeloid leukemia who have received front-line imatinib mesylate therapy and achieved complete molecular response. Leuk Res. 2012;36(6):689-693. Takahashi N, Kyo T, Maeda Y, et al. Discontinuation of imatinib in Japanese patients with chronic myeloid leukemia. Haematologica. 2012;97(6):903-906. Horn M, Glauche I, Muller MC, et al. Model-based decision rules reduce the risk of molecular relapse after cessation of tyrosine kinase inhibitor therapy in chronic myeloid leukemia. Blood. 2013;121(2):378384. Mori S, Vagge E, le Coutre P, et al. Age and dPCR can predict relapse in CML patients who discontinued imatinib: The ISAV study. Am J Hematol. 2015;90(10):910-914. Richter J, Soderlund S, Lubking A, et al. Musculoskeletal pain in patients with chronic myeloid leukemia after discontinu-

haematologica | 2016; 101(6)

11.

12.

13.

14.

15.

ation of imatinib: a tyrosine kinase inhibitor withdrawal syndrome? J Clin Oncol. 2014;32(25):2821-2823. Park JS, Lee SE, Jeong SH, et al. Change of health-related profiles after Imatinib cessation in chronic phase chronic myeloid leukemia patients. Leuk Lymphoma. 2015 Jun 12:1-7. [Epub ahead of print]. Goh HG, Lin M, Fukushima T, et al. Sensitive quantitation of minimal residual disease in chronic myeloid leukemia using nanofluidic digital polymerase chain reaction assay. Leuk Lymphoma. 2011;52(5):896-904. Ross DM, Hughes TP. How I determine if and when to recommend stopping tyrosine kinase inhibitor treatment for chronic myeloid leukaemia. Br J Haematol. 2014;166(1):3-11. Breccia M, Alimena G. Discontinuation of tyrosine kinase inhibitors and new approaches to target leukemic stem cells: treatment-free remission as a new goal in chronic myeloid leukemia. Cancer Lett. 2014; 347(1):22-28. Mahon F-X, Rea D, Guilhot J, et al. Discontinuation of Imatinib in Patients with Chronic Myeloid Leukemia Who Have Maintained Complete Molecular Response: Update Results of the STIM Study. ASH

Annual Meeting Abstracts 2011;118(21):603. 16. Marchand F, Perretti M, McMahon SB. Role of the immune system in chronic pain. Nat Rev Neurosci. 2005;6(7):521-532. 17. Mizoguchi I, Yoshimoto T, Katagiri S, et al. Sustained upregulation of effector natural killer cells in chronic myeloid leukemia after discontinuation of imatinib. Cancer Sci. 2013;104(9):1146-1153. 18. Ohyashiki K, Katagiri S, Tauchi T, et al. Increased natural killer cells and decreased CD3(+)CD8(+)CD62L(+) T cells in CML patients who sustained complete molecular remission after discontinuation of imatinib. Br J Haematol. 2012;157(2):254-256. 19. Molldrem JJ, Lee PP, Wang C, et al. Evidence that specific T lymphocytes may participate in the elimination of chronic myelogenous leukemia. Nat Med. 2000;6(9):1018-1023. 20. Kanodia S, Wieder E, Lu S, et al. PR1-specific T cells are associated with unmaintained cytogenetic remission of chronic myelogenous leukemia after interferon withdrawal. PLoS One. 2010;5(7):e11770. 21. Burchert A, Wolfl S, Schmidt M, et al. Interferon-alpha, but not the ABL-kinase inhibitor imatinib (STI571), induces expression of myeloblastin and a specific T-cell response in chronic myeloid leukemia. Blood. 2003;101(1):259-264.

723

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Myeloid Leukemia

Ferrata Storti Foundation

Defining the dose of gemtuzumab ozogamicin in combination with induction chemotherapy in acute myeloid leukemia: a comparison of 3 mg/m2 with 6 mg/m2 in the NCRI AML17 Trial

Alan Burnett,1 Jamie Cavenagh,2 Nigel Russell,3 Robert Hills,1 Jonathan Kell,4 Gail Jones,5 Ove Juul Nielsen,6 Asim Khwaja,7 Ian Thomas,1 and Richard Clark8 on behalf of the UK NCRI AML Study Group

Department of Haematology, Cardiff University School of Medicine, UK; 2Department of Haematology, St Bartholomew's Hospital, West Smithfield, London UK; 3Department of Haematology, Nottingham University Hospital NHS Trust, UK; 4Department of Haematology, University Hospital of Wales, Cardiff, UK; 5Department of Haematology, Newcastle Teaching Hospitals NHS Trust, UK; 6Department of Haematology, Rigshospitalet, Copenhagen, Denmark; 7Department of Haematology, University College, London Cancer Institute, UK; and 8Department of Haematology, Royal Liverpool University Hospital, UK 1

Haematologica 2016 Volume 101(6):724-731

ABSTRACT

Correspondence: akburnett719@gmail.com

Received: January 5, 2016. Accepted: February 23, 2016. Pre-published: February 26, 2016. doi:10.3324/haematol.2016.141937

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/724

724

recent source data meta-analysis of randomized trials in adults assessing the immunoconjugate gemtuzumab ozogamicin combined with standard chemotherapy in acute myeloid leukemia showed a significant survival benefit in patients without an adverse karyotype. It is not clear whether the optimal dose should be 3 mg/m2 or 6 mg/m2. In this study, we randomized 788 patients to a single dose of gemtuzumab ozogamicin 3 mg/m2 or 6 mg/m2 with the first course of induction therapy. We found that the rate of complete remission was higher with 3 mg/m2 [82% vs. 76%; odds ratio 1.46 (1.04-2.06); P=0.03], but this was balanced by a higher rate of complete remission with incomplete peripheral blood count recovery in the 6 mg/m2 treatment (10% vs. 7%) resulting in similar overall response rate [89% vs. 86%; hazard ratio 1.34 (0.88-2.04); P=0.17]. There was no overall difference in relapse or survival at four years between the arms: 46% vs. 54%; hazard ratio 1.17 (0.94-1.45), P=0.5, and 50% versus 47%; hazard ratio 1.10 (0.90-1.34), P=0.3, respectively. The 30- and 60-day mortality was significantly higher in the 6 mg/m2 recipients: 7% versus 3%; hazard ratio 2.07 (1.11-3.87), P=0.02, and 9% versus 5%; hazard ratio 1.99 (1.17-3.39), P=0.01, respectively, which in addition was associated with a higher rate of veno-occlusive disease (5.6% vs. 0.5%; P<0.0001). Our conclusion from this trial is that there is no advantage in using a single dose of 6 mg/m2 of gemtuzumab ozogamicin in combination with induction chemotherapy when compared with a 3 mg/m2 dose, with respect to response, disease-free and overall survival, either overall, or in any disease subgroup. (AML17 was registered as ISRCTN55675535.)

Introduction Gemtuzumab ozogamicin (GO) was the first antibody directed chemotherapy in cancer, but has had a chequered development in acute myeloid leukemia (AML) where it has had clinical exposure in a number of settings. None of the five randomized trials in adults (involving 3325 patients) or the trial in children (n=1022)1-6 has shown that GO improves the rate of remission. In all but one there was no excess in induction mortality, and even in the trial where this was seen, the explanation was not that the mortality in the combination arm was high, but that mortality in haematologica | 2016; 101(6)

GO + induction chemotherapy in AML

the control arm was unusually low.3 Five of the six trials showed a significant reduction in relapse risk. No trial showed a benefit for patients with adverse cytogenetics. An up-dated individual patient data-based meta-analysis of the adult patients confirmed that there was an overall survival benefit due to the reduction in relapse risk.7 Similar data have since emerged in children.6 In these trials, the dose and schedule of GO differed, being 6 mg/m2 on day 4 in two trials,3,5 3 mg/m2 on day 1 in two trials,1,4 and a fractionated schedule of 3 mg/m2 (to a maximum dose of 5 mg per dose) on days 1, 4 and 7.2 The latter schedule was derived from an encouraging study in relapsed disease.8 The meta-analysis further suggested that a single dose of 3 mg/m2 was as effective at preventing relapse as a 6 mg/m2 dose, while having less toxicity, and, therefore, may be the optimal dose, with the question of dose frequency remaining to be confirmed. The majority of patients in these studies received 3 mg/m2 as a single dose which was given on day 1 of course 1. This was based on a pilot study which suggested that toxicity was minimal and efficacy was encouraging at this dose,9 in contrast to the experience of combining the licensed dose of 9 mg/m2 with chemotherapy.10 This assumption turned out to be the case in our large multicenter setting trial.1 Since the toxicity found in the 3 mg/m2 single dose experience was modest, we postulated that there was a rationale to compare the higher dose (6 mg/m2) with the 3 mg/m2 dose in the hope that efficacy could be increased, particularly in patients with adverse risk cytogenetics. Therefore, as part of the UK NCRI AML17 trial, we prospectively compared two doses of GO, 3 mg/m2 versus 6 mg/m2, to be given on day 1 of the first course of induction treatment.

Methods The NCRI AML17 trial (ISRCTN55675535) was open to patients (the majority under 60 years of age, and also including children) who had any form of de novo or secondary AML and high-risk MDS (defined as marrow blasts >10%). Acute promyelocytic leukemia was excluded. As induction treatment, 759 adult patients were randomized between DA (3+10) (n=380) or ADE (10+3+5) (n=379) (Figures 1 and 2); 29 children (patients <16 years) received only ADE. GO was administered on day 1 of the induction chemotherapy except when the white blood cell (WBC) count was more than 30x109/L when cyto-reduction treatment with hydroxyurea could be given to reduce the count to less than 30x109/L, or the GO delayed until day 4 of chemotherapy. Liver function biochemistry was required to be less than 2 x ULN (upper limit of normal). Toxicity was defined as in NCI CTCAE v.3.0. Veno-occlusive disease (VOD) of the liver was defined by published criteria.11 After the first induction course, patients with a FLT-3 mutation were eligible to enter a randomization to receive the experimental FLT3 inhibitor lestaurtinib or placebo in a 2:1 ratio after each course of chemotherapy. For other patients who completed the first induction course, a previously reported validated score was used to assess the risk of relapse.12 Factors used were age, presenting WBC count, secondary disease, cytogenetics and the morphological response of the bone marrow (% blasts) after the first course. Patients with good or standard risk disease received the second daunorubicin/cytosine arabinoside course (with or without etoposide), and were then randomized to receive either one or two courses of high-dose cytosine arabinoside or MACE/MidAC as consolidation (Figure 1). High-risk patients haematologica | 2016; 101(6)

were allocated to a randomization between FLAG-Ida (fludarabine/ara-C/G-CSF/idarubicin) or daunorubicin/clofarabine for up to three courses with the intention to undergo allogeneic trans-

Table 1. Patientsâ&#x20AC;&#x2122; characteristics.

Characteristic Chemotherapy ADE DA Age 0-15 16-29 30-39 40-49 50-59 60+ Median Range Sex Female Male Diagnosis De novo Secondary MDS WHO PS 0 1 2 3 4 Not reported* WBC 0-9.9 10-49.9 50-99.9 100+ Median Range Cytogenetics Favorable Intermediate Adverse Unknown FLT3 ITD WT Mutant Unknown NPM1c WT Mutant Unknown ITD/NPM1c ITD WT, NPM1c WT ITD WT, NPM1c Mutant ITD Mutant, NPM1c WT ITD Mutant, NPM1c Mutant Unknown Post course 1 risk score Good risk Standard risk Poor risk Not assessable**

GO 6 mg/m2 (n=395)

GO 3 mg/m2 (n=393)

205 190

203 190

15 50 40 86 149 55 50 0-81

14 51 41 86 148 53 50 0-81

190 205

178 215

336 37 22

337 36 20

279 91 11 4 0 10

280 91 8 4 1 9

215 126 25 29 8.8 0.4-386.5

196 126 39 32 10.0 0.5-291.8

44 272 60 19

52 249 73 19

304 77 14

321 57 15

267 107 21

271 102 20

233 66 34 41 21

244 73 27 29 20

62 180 115 38

77 165 132 19

*Children under the age of 10 completed the WHO play performance score. **Post course 1 validated risk score12 is not available for patients who suffer induction death, have missing cytogenetics or in whom a response to course 1 is not available.

725

A. Burnett et al.

plantation. Patients who were not good risk, not FLT3 positive or adverse risk progressed with the core chemotherapy, but could be randomized or not to the mTOR inhibitor, everolimus, which was given between chemotherapy courses. The results of these randomizations will be reported elsewhere, but are taken into account when assessing the GO dose question in this trial. Diagnosis was confirmed locally and immunophenotyping and cytogenetics (20 metaphases) were performed in regional accredited laboratories and classified as previously published. 13 Molecular characterization was undertaken in two reference labs. Supportive care was determined by the policy of each center. Stem cell transplantation was undertaken in regional transplant centers. The trial was sponsored by Cardiff University and approved by Wales Research Ethics Committee 3 on behalf of all UK investigators, by the Danish Medicines Agency for sites in Denmark, and by MEDSAFE for sites in New Zealand. The trial was conducted in accordance with the Declaration of Helsinki, and received research funding from Cancer Research UK. GO was provided by Pfizer Inc. who had no role in the design or management of the trial.

Statistical analysis Response end point definitions are as described by Cheson.14 All analyses are by intention-to-treat. Categorical end points [e.g. complete remission (CR) rates] were compared using MantelHaenszel tests, to give Peto odds ratios and confidence intervals. Continuous/scale variables were analyzed by non-parametric (Wilcoxon rank sum) tests. Time-to-event outcomes were analyzed using the log rank test, with Kaplan-Meier survival curves. Odds/hazard ratios (OR/HR) less than 1 indicate benefit for the investigational therapy GO 6 mg/m2 versus GO 3 mg/m2. All survival percentages are at four years unless otherwise stated. In addition to overall analyses, exploratory analyses were performed stratified by the randomization stratification parameters and other important variables, with suitable tests for interaction.

Because of the well-known dangers of subgroup analysis, these were interpreted with caution. The randomization was originally planned to run over the course of the entire AML17 trial (5 years recruitment); however, the supply of GO enabled 788 patients to be randomized. This gave 80% power to detect a 10% absolute improvement in survival at five years from 45% to 55%, requiring 382 events. Follow up is complete as at 1st March 2015, with a median follow up for survival of 50 months (range 26.8-67.8 months) and 386 events.

Results Between June 2009 and October 2011 788 patients were randomized: their median age was 50 years (range 0-81 years, with 29 aged <16 years). Eighty-five percent had de novo AML, 9% secondary AML and 5% high-risk MDS; 53% were male. Thirteen percent were favorable, 69% intermediate, and 18% adverse cytogenetic risk; the median presenting WBC was 9.2 (0.4-386.5); 18% had a FLT3 ITD and 28% had an NPM1c mutation (Table 1). Four hundred and eight patients received ADE (including all 29 children recruited who were allocated ADE therapy) and 380 received DA (Figure 2) as induction treatment.

Remission induction Eighty-seven percent of all patients entered CR/CRi (79% CR, 8% complete remission with incomplete blood count (CRi)]. CR rates were higher in the GO 3 mg/m2 arm [82% vs. 76%, OR 1.46 (1.04-2.06); P=0.03], but this was balanced by more CRi with 6 mg/m2 (7% vs. 10%), leading to no significant difference in overall response rate (ORR) [9% vs. 86%; OR 1.34 (0.88-2.04); P=0.17] (Table 2).

Figure 1. Trial design of AML17. ADE: course 1, daunorubicin 50 mg/m2 d 1,3,5; ara-C 100 mg/m2 every 12 hours d 1-10, etoposide 100 mg/m2 d1-5; course 2, daunorubicin 50 mg/m2 d 1,3,5; ara-C 100 mg/m2 every 12 hours d 1-8, etoposide 100 mg/m2 d1-5; DA (no children): course 1 daunorubicin 50 mg/m2 d 1,3,5; ara-C 100 mg/m2 every 12 hours d 1-10, course 2 daunorubicin 50 mg/m2 d 1,3,5; ara-C 100 mg/m2 every 12 hours d 1-8; GO3: gemtuzumab ozogamicin 3 mg/m2 given on day 1 of course 1 of chemotherapy; GO6: gemtuzumab ozogamicin 6 mg/m2 given on day 1 of course 1 of chemotherapy; Lestaurtinib: lestaurtinib (CEP-701) 40-80 mg bd (depending on azole antifungals) from 2 days post chemo to 2 days pre subsequent course, up to a maximum of 28 days; mTOR (everolimus, available post October 2009): everolimus 5-10 mg/day, from 2 days post chemo to 2 days pre subsequent course, up to a maximum of 28 days; D Clofarabine (available post November 2009): daunorubicin 50 mg/m2 d 1,3,5; clofarabine 20 mg/m2 d 1-5; FLAG-Ida: fludarabine 30 mg/m2 (d 2-6); ara-C 2 g/m2 (4 h post fludarabine), d 26; G-CSF 263 g s.c. d 1-7; ara-C (post July 2010): ara-C 3 g/m2 12-hourly, d 1, 3, 5. Patients allocated either CEP-701 or everolimus post course 1 carried this allocation forward into subsequent courses. * Prior to July 2010 patients in the 3 versus 4 course randomization were randomized between MACE (amsacrine 100 mg/m2 d 1-5, ara-C 200 mg/m2 d 1-5, etoposide 100mg/m2 d 1-5) and MACE/MidAC (course 3 as above, MidAC: mitoxantrone 10 mg/m2 d1-5; ara-C 1 g/m2 twice daily d 1-3). 726

haematologica | 2016; 101(6)

GO + induction chemotherapy in AML

There was no significant difference in remission rates achieved with the first induction course between the arms.

Toxicity and supportive care The 30-day mortality [3% vs. 7%; OR 2.07 (1.11-3.87); P=0.02] and 60-day mortality [5% vs. 9%; OR 1.99 (1.173.39); P=0.01] were both significantly increased in the 6

mg/m2 arm (Table 2). There were 18 versus 36 deaths within 60 days: the causes of which were infection (10 vs. 11); infection+hemorrhage (0 vs. 1); hemorrhage (3 vs. 4); resistant disease (2 vs. 6); veno-occlusive disease (0 vs. 5); cardiac (1 vs. 3); pulmonary (2 vs. 1); renal (0 vs. 3); or multiple causes (0 vs. 2). Survival beyond 60 days was the same in both arms [53% vs. 52%, HR 1.00 (0.81-1.24); P=1.0]. When grade 3 or 4 toxicities were compared

Table 2. Trial outcomes and results of dose comparisons. CR CRi CR/CRi Induction death Resistant disease CR/CRi post course 1 30-day mortality 60-day mortality 4 year OS 4 year RFS 4 year cumulative incidence of relapse 4 year cumulative incidence of death in CR 4 year OS from CR 4 year OS censored at SCT

GO 6 mg/m2

GO 3 mg/m2

OR/HR & CI

76% 10% 86% 7% 7% 73% 7% 9% 47% 38% 54% 9% 53% 53%

82% 7% 89% 3% 8% 78% 3% 5% 50% 44% 46% 10% 56% 60%

1.46 (1.04-2.06)

0.03

1.34 (0.88-2.04)

0.17

1.29 (0.93-1.78) 2.07 (1.11-3.87) 1.99 (1.17-3.39) 1.10 (0.90-1.34) 1.11 (0.91-1.35) 1.17 (0.94-1.45) 0.83 (0.51-1.36) 1.07 (0.85-1.35) 1.20 (0.93-1.54)

0.13 0.02 0.01 0.3 0.3 0.15 0.5 0.6 0.16

OR: odds ratio; HR: hazard ratio; CI: confidence interval; CR: complete remission; CRi: complete remission with incomplete blood count; OS: overall survival; RFS: relapse-free survival; SCT: stem cell transplantation.

Figure 2. CONSORT diagram. haematologica | 2016; 101(6)

727

A. Burnett et al.

between the 3 mg/m2 and 6 mg/m2 doses, alanine transaminase (ALT), creatinine and hematuria in course 1 (7% vs. 17%; 1% vs. 2%; 1% vs. 2%, respectively) were the only significant differences during courses 1 and 2 (Figure 3). Although the kinetics of peripheral blood recovery were similar, with the exception of slower platelet recovery, there was increased platelet transfusion requirement and

increased days on antibiotics during course 1 in the 6 mg/m2 arm, but there were no differences after course 2 (Online Supplementary Table S1). Central assessment of VOD was confirmed as definite (n=17) or possible (n=5) in 22 of 395 (5.6%) patients on 6 mg/m2 compared with 2 definite and 0 possible in 2 of 393 on the 3 mg/m2 arm (0.5%) (P<0.0001).

Figure 3. Grade 3-4 toxicities following (A) course 1, (B) course 2.

Figure 4. Outcomes for gemtuzumab ozogamicin dose randomization: (A) cumulative incidence of relapse; (B) cumulative incidence of death in remission; (C) relapse-free survival; (D) overall survival. 728

haematologica | 2016; 101(6)

GO + induction chemotherapy in AML

Figure 5. Stratified analysis of survival. haematologica | 2016; 101(6)

729

A. Burnett et al.

Relapse and survival There was no significant difference in relapse [46% vs. 54%, HR 1.17 (0.94-1.45), P=0.15] or death in remission [(10% vs. 9%, HR 0.83 (0.51-1.36) P=0.5)] leading to no significant differences in relapse-free survival which was 41% overall [44% vs. 38%; HR 1.11 (0.91-1.35); P=0.3] (Table 2 and Figure 4). Four-year overall survival was 49% [50% vs. 47%; HR 1.10 (0.90-1.34); P=0.3] (Table 2 and Figure 4D). There was no difference in mortality after day 60. A total of 329 patients received transplant; of the 273 allografts, 158 were performed in first remission. The distribution of transplants was similar between the arms, and censoring survival at stem cell transplant did not alter the overall treatment effect. Neither the 3 mg/m2 or the 6 mg/m2 dose caused excessive post-transplant toxicity; in particular no liver toxicity was observed.

Exploration of subgroups No subgroup showed any suggestion of a significant response or survival benefit from the 6 mg/m2 dose, but there was a trend for benefit for both response rate (test for heterogeneity P=0.02) (Online Supplementary Figure S1) and overall survival (test for heterogeneity P=0.04) (Figure 5) in the adverse cytogenetics patients (n=133).

Discussion In the early studies of GO, it was observed that there was an association with the then licensed dose (9 mg/m2) and liver toxicity if combined with chemotherapy or stem cell transplant.15 The risk was related to the time interval to or from transplant.16 The pathological diagnosis was referred to as sinusoidal obstructive syndrome (SOS) and thought likely to be caused by cytotoxicity to leukemia cells which had accumulated in the liver sinusoids.11 However, the concept of augmenting treatment with targeted chemotherapy remained attractive, but establishing the dose which could safely be combined with standard intensive chemotherapy required a dose-finding trial. To do this, we tested the DAT chemotherapy schedule (daunorubicin, Ara-C, thioguanine). It emerged that the observed liver toxicity was associated with the use of thioguanine, which, somewhat fortunately, became unavailable.9 Since the efficacy in that pilot study was encouraging at the 3 mg/m2 dose, this was adopted as the study dose for the subsequent large trials conducted by our group. Since minimal toxicity was seen, the question remained as to whether we were under-dosing at this level, even although the dose was effective in patients who did not have adverse risk disease. Subsequently a 6 mg/m2 was adopted by others,3,5 so we thought it relevant to examine

References 1. Burnett AK, Hills RK, Milligan D, et al. Identification of patients with Acute Myeloblastic Leukemia who benefit from the addition of Gemtuzumab Ozogamicin: Results of the MRC AML15 Trial. J Clin Oncol. 2011;29(4):369-377. 2. Castaigne S, Pautas C, Terre C, et al. Effect

730

this dose to better define the dose level required. In this large randomized study we saw no significant benefit of using GO at the 6 mg/m2 dose, although there was a possible trend for benefit in the adverse risk patients who have not been shown to benefit from GO irrespective of dose or schedule in other trials. The 6 mg/m2 dose did have a detrimental effect with respect to liver toxicity and platelet count recovery, and significantly increased the day 30 and 60 mortality; this suggests that in cases in which a single dose schedule is used, the 3 mg/m2 dose is adequate. The remaining issue is whether a single dose is inferior to the fractionated dosing schedule developed by the French group. The rationale for this approach is that, following exposure to antibody, there is re-expression of CD33 within hours, and, therefore, it is logical to expect a repeated challenge.17 This was endorsed by the initial study in relapsed disease8 where 57 patients were treated with the 1-, 4- and 7-day schedule and achieved an overall response rate of 33% [26% CR and 7% complete remission with incomplete platelet recovery (CRp)] with minimal toxicity. The French ALFA group took this forward to front-line treatment in a randomized study involving 280 non-favorable risk patients who were randomized to receive the addition of GO in a dose of 3 mg/m2 in a day 1, 4 and 7 schedule with the first induction course, followed by a single dose each of two post induction courses. This produced a very encouraging initial report with a significant survival benefit,2 but one that longer follow up has shown to be a little less clear.18 However, in that study, each dose was capped at 5 mg and the schedule given was on day 1, 4 and 7, and also as a single dose in consolidation. It is plausible that at least some of the benefit was due to dosing in consolidation. There was some hematologic toxicity, particularly to platelets. So while a 3 mg/m2 dose appears adequate, an optimal schedule still has to be defined. However, the MRC AML15 trial did not show any additional benefit of adding GO to consolidation irrespective of whether it had been given with the first induction course,1 so the pressing issue to be resolved is whether the single or fractionated schedule is to become the standard approach. To resolve this, we have initiated a direct comparison of a 3 mg/m2 dose on day 1 versus days 1 and 4 in our ongoing trials (registered as ISRCTN31682779 and ISRCTN78449203). Acknowledgments We are grateful to Cancer Research UK for research funding for this trial and to Pfizer Japan for supplying gemtuzumab ozogamicin at a discounted price, to the staff of the Haematology Trials Office and Cardiff Experimental Cancer Medicine Centre for supporting the trial and to our co-investigators.

of gemtuzumab ozogamicin on survival of adult patients with de-novo acute myeloid leukaemia (ALFA-0701): a randomised, open-label, phase 3 study. Lancet. 2012; 379(9825):1508-1516. 3. Petersdorf SH, Kopecky KJ, Slovak M, et al. A phase 3 study of gemtuzumab ozogamicin during induction and postconsolidation therapy in younger patients with acute myeloid leukemia. Blood. 2013;

121(24):4854-4860. 4. Burnett AK, Russell NH, Hills RK, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy improves survival in older patients with acute myeloid leukemia. J Clin Oncol. 2012;30(32):39243931. 5. Delaunay J, Recher C, Pigneux A, et al. Addition of Gentuzumab ozogamycin to chemotherapy improves event-free surivval

haematologica | 2016; 101(6)

GO + induction chemotherapy in AML

but not overall survival of AML patients with intermediate cytogenetics not eligible for allogeneic transplantation. Results of the GOELAMS AML2006 IR study. Blood. 2011;118:79(Abstract). Gamis AS, Alonzo TA, Meshinchi S, st al. Gemtuzumab ozogamicin in children and adolescents with de novo acute myeloid leukemia improves event-free survival by reducing relapse risk: results from the randomized phase III Childrenâ&#x20AC;&#x2122;s Oncology Group trial AAML0531. J Clin Oncol 2014: 32(27):3021-3032. Hills RK, Castaigne S, Appelbaum FR, et al. Addition of gemtuzumab ozogamicin to induction chemotherapy in adult patients with acute myeloid leukaemia: a metaanalysis of individual patient data from randomised controlled trials. Lancet Oncol. 2014;15(9):986-996. Taskin AL, Legrand O, Raffoux E, et al. High efficacy and safety profile of fractionated doses of Mylotarg as induction therapy in patients with relapsed acute myeloblastic leukemia: a prospective study of the alfa group. Leukemia. 2007;21(1):6671. Kell WJ, Burnett AK, Chopra R, et al. A fea-

haematologica | 2016; 101(6)

10.

11.

12.

13.

sibility study of simultaneous administration of gemtuzumab ozogamicin with intensive chemotherapy in induction and consolidation in younger patients with acute myeloid leukemia. Blood. 2003; 102:4277-4283. Giles FJ, Kantarjian HM, Kornblau SM, et al. Mylotarg (gemtuzumab ozogamicin) therapy is associated with hepatic venoocclusive disease in patients who have not received stem cell transplantation. Cancer. 2001;92(2):406-413. Rajvanshi P, Schulman HM, Sievers EL, McDonald GB. Hepatic sinusoidal Obstruction after Gemtuzumab Ozogamicin (mylotatg) therapy. Blood. 2002;99(7):2310-2314. Burnett AK, Hills RK, Wheatley K, et al. A sensitive risk score for directing treatment in younger patients with AML. Blood. 2006;108(11):10a. Grimwade D, Hills RK, Moorman AV, et al. Refinement of cytogenetic classification in acute myeloid leukemia: determination of prognostic significance of rare recurring chromosomal abnormalities among 5876 younger adult patients treated in the United Kingdom Medical Research Council trials.

Blood. 2010;116(3):354-365. 14. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendation of the International Working Group for diagnosis standardisation, of response criteria treatment outcomes and reporting standards for therapeutic trials in acute myeloid leukaemia. J Clin Oncol. 2003;21(24):4642-4649. 15. Larson RA, Sievers EL, Stadtmauer EA, et al. Final report of the efficacy and safety of gemtuzumab ozogamicin (Mylotarg) in patients with CD33-positive acute myeloid leukemia in first recurrence. Cancer. 2001; 104(7):1442-1452. 16. Wadleigh M, Richardson PG, Zahrieh D, et al. Prior gemtuzamab ozogamicin exposure significantly increases the risk of venoocclusive disease in patients who undergo myeloablative allogeneic stem cell transplantation. Blood. 2003;102(5):1578-1582. 17. van Der Velden VH, te Marvelde JG, Hoogeveen PG, et al. Targeting of the CD33-calicheamicin immunoconjugate Mylotarg (CMA-676) in acute myeloid leukemia: in vivo and in vitro saturation and internalization by leukemic and normal myeloid cells. Blood. 2001; 97(10):31973204.

731

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Lymphoblastic Leukemia

Ferrata Storti Foundation

An early thymic precursor phenotype predicts outcome exclusively in HOXA-overexpressing adult T-cell acute lymphoblastic leukemia: a Group for Research in Adult Acute Lymphoblastic Leukemia study

Jonathan Bond,1* Tony Marchand,2* Aurore Touzart,1 Agata Cieslak,1 Amélie Trinquand,1 Laurent Sutton,3 Isabelle Radford-Weiss,4 Ludovic Lhermitte,1 Salvatore Spicuglia,5 Hervé Dombret,6 Elizabeth Macintyre,1 Norbert Ifrah,7 Jean-François Hamel,7§ and Vahid Asnafi1§

Université Paris Descartes Sorbonne Cité, Institut Necker-Enfants Malades (INEM), Institut National de Recherche Médicale (INSERM) U1151, and Laboratory of OncoHematology, Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Necker EnfantsMalades; 2Department of Hematology, University Hospital and INSERM UMR 917, Rennes 1 University; 3Department of Hematology, Centre Hospitalier Argenteuil; 4 Université Paris 5 Descartes, Department of Cytogenetics, Assistance PubliqueHôpitaux de Paris, Hôpital Necker-Enfants Malades; 5Technological Advances for Genomics and Clinics (TAGC), INSERM U1090, Aix-Marseille University UMR-S 1090; 6 Université Paris Diderot, Institut Universitaire d’Hématologie, EA-3518, Assistance Publique-Hôpitaux de Paris, University Hospital Saint-Louis; and 7PRES LUNAM, CHU Angers Service des Maladies du Sang et INSERM U 892, Angers, France 1

Haematologica 2016 Volume 101(6):732-740

*These authors contributed equally to this work. §Co-corresponding authors

ABSTRACT

Correspondence: vahid.asnafi@nck.aphp.fr

Received: December 18, 2015. Accepted: February 26, 2016. Pre-published: March 4, 2016. doi:10.3324/haematol.2015.141218

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/732

732

ene expression studies have consistently identified a HOXA-overexpressing cluster of T-cell acute lymphoblastic leukemias, but it is unclear whether these constitute a homogeneous clinical entity, and the biological consequences of HOXA overexpression have not been systematically examined. We characterized the biology and outcome of 55 HOXApositive cases among 209 patients with adult T-cell acute lymphoblastic leukemia uniformly treated during the Group for Research on Adult Acute Lymphoblastic Leukemia (GRAALL)-2003 and -2005 studies. HOXA-positive patients had markedly higher rates of an early thymic precursor-like immunophenotype (40.8% versus 14.5%, P=0.0004), chemoresistance (59.3% versus 40.8%, P=0.026) and positivity for minimal residual disease (48.5% versus 23.5%, P=0.01) than the HOXA-negative group. These differences were due to particularly high frequencies of chemoresistant early thymic precursorlike acute lymphoblastic leukemia in HOXA-positive cases harboring fusion oncoproteins that transactivate HOXA. Strikingly, the presence of an early thymic precursor-like immunophenotype was associated with marked outcome differences within the HOXA-positive group (5-year overall survival 31.2% in HOXA-positive early thymic precursor versus 66.7% in HOXA-positive non-early thymic precursor, P=0.03), but not in HOXA-negative cases (5year overall survival 74.2% in HOXA-negative early thymic precursor versus 57.2% in HOXA-negative non-early thymic precursor, P=0.44). Multivariate analysis further revealed that HOXA positivity independently affected eventfree survival (P=0.053) and relapse risk (P=0.039) of chemoresistant T-cell acute lymphoblastic leukemia. These results show that the underlying mechanism of HOXA deregulation dictates the clinico-biological phenotype, and that the negative prognosis of early thymic precursor acute lymphoblastic leukemia is exclusive to HOXA-positive patients, suggesting that early treatment intensification is currently suboptimal for therapeutic rescue of HOXA-positive chemoresistant adult early thymic precursor acute lymphoblastic leukemia. Trial Registration: The GRAALL-2003 and -2005 studies were registered at http://www.clinicaltrials.gov as #NCT00222027 and #NCT00327678, respectively. haematologica | 2016; 101(6)

ETP predicts outcome in HOXA-positive adult T-ALL.

Introduction Modern management of acute leukemia is predicated upon the identification of biologically distinct subgroups whose prognosis might benefit from timely alterations in treatment intensity.1 T-cell acute lymphoblastic leukemia (T-ALL) is associated with a wide range of acquired genetic abnormalities that contribute to developmental arrest and abnormal proliferation of malignant lymphoid progenitors.2,3 Despite the diversity of observed mutations and deletions, transcriptional microarray studies have consistently shown that T-ALL can be classified by five recurrent patterns of gene expression, namely the Immature/LYL1, TAL1, TLX1, TLX3 and HOXA clusters.46 The last subgroup is characterized by aberrant activation of the HOXA gene locus on chromosome 7. Homeobox (HOX) factors normally regulate the transcription of genes that are critical for development and proliferation.7,8 In murine models, Hoxa overexpression induces a hematopoietic differentiation block and leukemic transformation of normal progenitor cells,9-11 suggesting that HOXA overexpression may directly affect the biology of human T-ALL. HOXA-positive (HOXAPos) T-ALL is associated with a number of recurrent chromosomal translocations. Juxtaposition with TCRB regulatory elements via translocation (7;7)(p15;q34) or inversion(7)(p15q34) directly activates HOXA by a cis-like mechanism;12,13 however, the majority of HOXA locus deregulation has been described to occur in trans. Fusion proteins that arise from rearrangements involving the Mixed Lineage Leukemia gene (MLL)4, MLLT10 (formerly AF10)14-17 and the SETNUP214 translocation18 have been shown to recruit DOT1 Ligand (DOT1L), which stimulates HOXA expression through aberrant methylation of Lys79 of Histone H3.19,20 DOT1L is additionally known to methylate a range of target genes that are also likely to contribute to the leukemic phenotype,21 and it is therefore probable that the molecular mechanisms of leukemogenesis within the HOXAPos subgroup are heterogeneous. In support of this, HOXA dysregulation does not necessarily predict inclusion in the HOXA gene expression cluster, as a proportion of these cases segregate preferentially with the Immature/ LYL1 subgroup.5,6 This immature cluster shows a high level of enrichment of transcripts that are associated with early thymic precursor (ETP)-ALL,22 a subgroup of T-ALL that exhibit a stem cell/immature myeloid-like immunophenotype, resistance to treatment and poor outcome.15,23-26 Genomic analysis of ETP-ALL has revealed high rates of mutations in factors involved in cytokine receptor and RAS signaling, hematopoiesis and epigenetic modification,15 but the precise molecular basis of these patients’ adverse prognosis remains unclear. We analyzed the biological and clinical characteristics of a cohort of HOXAPos adult T-ALL patients who were treated as part of the Group for Research on Adult Acute Lymphoblastic Leukemia (GRAALL)-2003 and -2005 studies. Notably, we found that the underlying mechanism of HOXA deregulation is highly predictive of phenotypic immaturity and early treatment resistance. Survival analyses revealed that the HOXAPos group did not have an inferior overall outcome, and that poor prognosis was restricted to a subset of patients who had an ETP-like immunophenotype, chemoresistance and activation of haematologica | 2016; 101(6)

the HOXA locus in trans. Strikingly, these parameters did not predict survival in the HOXANeg group, indicating that the adverse prognosis of ETP-ALL in adults is exclusive to HOXAPos patients.

Methods The GRAALL-2003 and GRAALL-2005 studies The GRAALL-2003 study was a phase II trial that enrolled 77 adults with T-ALL between November 2003 and November 2005.27 The GRAALL-2005 study was the subsequent phase III trial that included randomized evaluation of hyper-fractionated cyclophosphamide in induction and late intensification. Twohundred and sixty-one adults with T-ALL were enrolled between May 2006 and September 2011. Informed consent was obtained from all patients at trial entry. Studies were conducted in accordance with the Declaration of Helsinki and approved by local and multicenter research ethical committees. The complete study protocols are detailed in the Online Supplementary File ‘GRAALL_2003_2005 protocol’. At a point date on March 1st 2013, the median follow-up was 2.9 years (5.5 and 2.7 years for GRAALL-2003 and GRAALL-2005, respectively). The sole criteria for inclusion in the current project were a diagnosis of T-ALL and availability of diagnostic material for HOXA9 measurement. Survival outcomes of the 209 patients (42 GRAALL-2003 and 167 GRAALL-2005) who fulfilled these criteria did not differ from those of the remaining 129 T-ALL patients of the study cohorts. A full comparison of the clinical features of each group is shown in Online Supplementary Table S1.

Statistical analysis The considered cut-off level for dichotomizing white blood cell count was 100x109/L. The considered cut-off ratio for dichotomizing HOXA status was defined as the lowest HOXA ratio associated with a genetic abnormality known to activate HOXA. Categorical data are presented as percentages and compared using Fisher exact tests. Continuous data are presented as medians and inter-quartile ranges and compared using MannWhitney tests. Censored data (i.e. overall survival, event-free survival and disease-free survival) were analyzed using Cox models. Competing risk events (i.e. cumulative incidence of relapse) were analyzed using Fine & Gray models. Overall survival and event-free survival were calculated from the date of pre-phase initiation. Events considered for event-free survival were induction failure, first hematologic relapse and death from any cause in first complete remission. The cumulative incidence of relapse and disease-free survival were calculated from the date of achieving complete remission. The chosen adjustment covariates were defined based on their clinical relevance, in order to minimize the risk of over-adjustment. The adjustment covariates were white blood count, stem cell transplantation, risk classifier, ETP status and chemosensitivity status. Stem cell transplantation was analyzed as a time-dependent covariate using the Mantel-Byar approach. Interactions were assessed by introducing interaction terms in the multivariate models. Specific hazards of relapse and hazard ratios are given with 95% confidence intervals. All tests were two-sided with a significance level of 0.05, except for interactions for which a significance level of 0.1 was considered. Statistical analyses were performed using Stata/mp 13.1 (Stata Corporation, College Station, TX, USA). Additional details are provided in the Online Supplementary Methods. 733

J. Bond et al.

Results Definition of HOXA-positive adult T-cell acute lymphoblastic leukemia In order to characterize the spectrum of HOXA deregulation in adult T-ALL, we measured the levels of HOXA9 in the T-ALL cohort of the GRAALL-2003 and -2005 studies. Diagnostic material was available for 209 of 328 patients. HOXA9 levels were normalized to a reference gene and expressed as a ‘HOXA ratio’ (see Online Supplementary Methods). This ratio varied greatly among samples, ranging from 0 to 66.2 (Figure 1A). Most patients had low HOXA ratios, and the median was 0.06. As HOXAPos T-ALL comprises cases that express leukemic fusion proteins which have been shown to upregulate HOXA transcription directly, we defined the cut-off for positivity as the lowest HOXA ratio associated with a genetic abnormality known to activate the HOXA locus. This threshold of 0.66, as defined by the lowest ratio in a PICALM-MLLT10 T-ALL, classified 55/209 cases as HOXAPos. Of note, 52 of these cases corresponded to the highest quartile of HOXA ratio in the entire study cohort. Thirteen HOXAPos cases had sufficient diagnostic material available for evaluation of the global pattern of HOX locus transcription. As expected, the entire HOXA gene cluster was deregulated in 100% of these samples (Figure

1B). We then tested eight HOXANeg samples from the third quartile of expression, and found that 0% exhibited activation of HOXA gene transcription. We additionally quantified the levels of HOXA5 by quantitative real-time polymerase chain reaction in 180 patients, and found that this measurement was strongly correlated with HOXA9 (Online Supplementary Figure S1). We therefore conclude that global deregulation of the HOXA locus in adult T-ALL could be predicted by a HOXA ratio of 0.66, corresponding to about 25% of patients overall.

Molecular mechanism of HOXA deregulation in adult T-cell acute lymphoblastic leukemia The 55 HOXAPos patients were extensively investigated for anomalies known to cause HOXA overexpression in TALL: translocations involving MLL, PICALM-MLLT10, SET-NUP214 and TCRB-HOXA (‘primary screen’ in Figure 2A). Comprehensive assessment was completed in 52/55 cases. The three remaining cases (including one from the third quartile of HOXA ratio) did not undergo TCRBHOXA testing due to a lack of sample availability. One of these three cases also lacked sufficient material for MLL fluorescence in situ hybridization, although the absence of any chromosome 11q abnormality by karyotyping made the possibility of MLL translocation unlikely. This initial screen identified an explicatory translocation in 33 HOXAPos patients (Figure 2A,B and Table 1). These

Figure 1. Definition of HOXAPos adult T-ALL. (A) HOXA ratios (HOXA9/ABL) were calculated for 209 T-ALL patients treated as part of the GRAALL-2003 and -2005 studies. Each point represents an individual measurement. The threshold of HOXA positivity (0.66) was defined by the lowest HOXA ratio associated with a known HOXAderegulating abnormality (PICALM-MLLT10). For HOXAPos cases for which the etiology of HOXA locus activation remained undefined, the measurement point is indicated by a diamond. Cases with an ETP-like phenotype are shown in red. (B) Taqman low density array (TLDA) analysis of HOX gene expression in HOXANeg (n = 8) and HOXAPos (n = 13) patients, compared with normal thymus (Thy) (n = 3) controls. HOXAPos cases exhibit specific activation of the HOXA locus, while HOXANeg samples have uniformly low expression of all HOX genes.

734

haematologica | 2016; 101(6)

ETP predicts outcome in HOXA-positive adult T-ALL.

comprised 11 TCRB-HOXA, eight PICALM-MLLT10, nine SET-NUP214 and six MLL rearrangements. One patient had co-existing translocations of both PICALM-MLLT10 and TCRB-HOXA. Two further patients were found to have the NUP98-RAP1GDS1 translocation, while the presence of a visible t(10;12) (p12;q14) translocation during conventional karyotyping led us to investigate and confirm a NAP1L1-MLLT10 translocation in another case (Figure 2C,D). For five patients from whom material was available, we also performed RNA-sequencing in an effort to identify cryptic HOXA-activating fusions. This revealed two additional MLLT10 translocations, one involving the XPO1 locus and the other involving the DDX3X locus (Figure 2E). We additionally performed fluorescence in situ hybridization screening for MLLT10 in ten patients, identifying one further MLLT10 rearrangement with an unknown partner. After completion of these investigations, the etiology of HOXA activation remained undefined in 16/55 HOXAPos patients. Where possible (12/16 cases), we performed polymerase chain reaction testing for NUP98-RAP1GDS1, XPO1-MLLT10, DDX3X-MLLT10 and NAP1L1-MLLT10, but all evaluated cases were negative for each translocation.

Clinico-biological characterization of HOXA-positive adult T-cell acute lymphoblastic leukemia Initial comparison of the clinico-biological characteristics of HOXAPos and HOXANeg T-ALL revealed substantial heterogeneity within the HOXAPos group, whereby the patients with cis-activated TCRB-HOXA differed markedly from those with trans-activated MLL, MLLT10, SETNUP214 or NUP98-RAP1GDS1. For the ensuing analyses, the HOXAPos patients were therefore grouped according to the underlying mechanism of locus deregulation (Table 1). A comprehensive breakdown of the characteristics of all HOXAPos genetic subgroups is shown in Online Supplementary Table S4. HOXAPos and HOXANeg cases did not differ significantly with regard to sex (male:female ratio 1.9 versus 3.8; P=0.22), median age (30.4 versus 31.9 years; P=0.4), white blood cell count (33.0 versus 37.7x109/L; P=0.3) or central nervous system involvement (18.5% versus 11.3%; P=0.23). There were similar rates of NOTCH1/FBXW7 mutations (65.5% versus 72.1%; P=0.39). In addition, analysis using our recently described oncogenetic risk predictor28 that includes classification by mutations in NRAS, KRAS and PTEN did not reveal any differences between the two groups (49% HOXAPos high risk versus 41.3% HOXANeg high risk; P=0.41).

haematologica | 2016; 101(6)

Figure 2. Molecular mechanism of HOXA deregulation in adult T-ALL. (A) Flowchart of investigation of the etiology of HOXA positivity. Numbers of patients in each diagnostic subgroup are shown. *One patient had co-existing diagnoses of both TCRβHOXA and PICALM-MLLT10. RNAseq = RNA-sequencing. FISH = Fluorescence in situ hybridization. Representative results are depicted in panels (B) – (E). (B) Positive FISH for TCRβ-HOXA. (C) Direct (Sanger) sequencing confirms the presence of the NAP1L1-MLLT10 translocation. Exon numbers are indicated. (D) Reverse transcriptase polymerase chain reaction amplification of NUP98-RAP1GDS1 using fusion-specific primers. Patients 2 and 3 are positive for NUP98RAP1GDS1, while patients 1 and 4 are negative. NTC = no template control. (E) Diagnosis of DDX3XMLLT10 by RNA-sequencing. A schematic representation of pairedend and fusion-spanning reads is shown. Plain lines indicate split reads spanning two exons, and dotted lines indicate two reads of the same fragment. Exon numbers are indicated.

735

J. Bond et al.

Notably, HOXAPos leukemias were more likely to be both genotypically and phenotypically immature than their HOXANeg counterparts. Genotypic immaturity, as defined by lack of detectable rearrangement of the TCRB locus in the leukemic blasts,29 was considerably more common in HOXAPos cases (56.9% versus 16.5%; P<0.0001). Furthermore, HOXAPos samples had significantly higher rates of an ETP-like immunophenotype,25 as defined by low expression of CD5, lack of CD1a/CD8, and expression of at least one stem cell or myeloid antigen (CD34, CD13, CD33, CD117). Rates of ETP-like immunophenotype were 40.8% for HOXAPos patients, compared with 14.5% for HOXANeg cases (P=0.0004). Strikingly, this immaturity was not observed amongst the TCRB-HOXA subgroup, which presented a more mature cortical profile of developmental arrest. As the TCRB-HOXA rearrangement is the only subgroup of HOXAPos T-ALL that presents a cis-activation of the HOXA locus, this suggests that the stage of differentiation block of T-ALL blasts correlates with the mechanism of HOXA deregulation. There were also major differences between the groups with regard to initial treatment response. The HOXAPos subgroup had significantly lower proportions of both early corticosteroid response (36.4% versus 59.7%; P=0.0045) and early bone marrow chemosensitivity (40.7% versus 59.2%; P=0.026) in comparison with the HOXANeg cases. These differences were not seen when the patients with cis-activated TCRB-HOXA were analyzed separately, as these had comparatively high rates of both corticosensitivity (72.7%) and chemosensitivity (90.9%). Assessment of

minimal residual disease (MRD) response gave similar results, as HOXAPos cases were more likely than HOXANeg cases to have positive (>10-4) MRD1 after induction (48.5% versus 23.5%; P=0.01). Again, these differences were confined to the trans-activated HOXAPos subgroup, as all TCRB-HOXA patients who were assessed were negative for MRD1. Taken together with the observed heterogeneity of developmental arrest between cis- and transactivated cases, these results suggest that the underlying mode of HOXA activation affects the biological phenotype of HOXAPos T-ALL.

HOXA positivity is not directly linked to altered clinical outcome in adult T-cell acute lymphoblastic leukemia In order to determine whether HOXA positivity correlates with prognosis in adult T-ALL, we performed global survival comparisons of HOXAPos and HOXANeg cases. There were very similar outcomes in the two cohorts for 5-year overall survival (55.0% for HOXAPos versus 58.1% for HOXANeg; P=0.91), event-free survival (45.9% versus 48.9%; P=0.95) and disease-free survival (50% versus 51.1%; P=0.92) (Figure 3A,B and Online Supplementary Figure S2A). Additional analysis according to HOXA ratio revealed no differences in survival between quartile groups (Online Supplementary Figure S2B-D), further indicating a lack of direct correlation between the degree of HOXA locus activation and patient outcome. The limited size of the HOXAPos subgroups precluded satisfactory analysis of the survival risks associated with individual translocations (Online Supplementary Figures S3A-C).

Table 1. Clinico-biological characteristics of HOXAPos and HOXANeg adult T-ALL.

HOXANeg HOXA activation mechanism N. (%) TCR molecular status Immature (IM0. Imd. Img) αβ lineage γδ lineage ETP immunophenotype EGIL 1-2 EGIL 3 EGIL 4 NOTCH1/FBXW7mutated High risk classifier* Clinical subsets analyzed Age median, years Age >35 years Sex ratio M/F White blood cell count median, x109/L White blood cell count >100x109/L Central nervous system involvement Corticosensitivity Chemosensitivity Complete remission Relapse Death MRD ≥ 10-4

HOXAPos

Total

P value† Neg vs. Pos

P value† Trans vs. cis

All HOXAPos

Trans

Cis

Unknown

154 (74)

55 (100)

29 (53)

11 (20)

16 (29)

209

17% 72% 11% 15% 23% 63% 14% 72% 41%

57% 27% 16% 41% 60% 28% 11% 65% 49%

70% 11% 19% 44% 74% 11% 15% 45% 67%

10% 70% 20% 11% 27% 55% 18% 100% 11%

60% 27% 13% 53% 56% 38% 6% 81% 38%

28% 60% 13% 21% 33% 54% 13% 70% 43%

<0.0001*** <0.0001*** 0.46 0.0004*** <0.0001*** <0.0001*** 0.81 0.39 0.41

0.002** 0.0011** 1 0.11 0.001* 0.009* 1 0.001** 0.006**

30.4 40% 3.8 37.7 30% 11% 60% 59% 94% 32% 32% 24%

31.9 40% 1.9 33 16% 19% 36% 41% 93% 26% 35% 48%

30.2 38% 2.2 33 21% 18% 17% 21% 97% 29% 38% 68%

35 46% 1.8 56.1 0% 18% 73% 91% 100% 9% 27% 0%

31.2 41% 1.8 22.2 24% 24% 47% 44% 82% 29% 35% 33%

30.5 40% 3.1 37 26% 13% 54% 54% 94% 30% 33% 31%

0.48 1 0.22 0.3 0.052 0.23 0.004** 0.026* 0.75 0.48 0.74 0.01*

0.68 0,73 0.7 0.8 0.16 1 0.0017** <0.0001*** 1 0.4 0.72 0.0052**

HOXAPos cases are grouped according to the underlying mechanism of HOXA locus activation: trans- (MLL, MLLT10, SET-NUP214, NUP98-RAP1GDS1), cis- (TCR -HOXA) or unknown. TCR: T-cell receptor; EGIL: European Group for the Immunological classification of Leukaemia; ETP: early thymic precursor. *High risk classifier incorporating the effects of RAS and PTEN mutations25. †t test (Mann-Whitney test) or Fisher exact test were used where appropriate. Statistically significant differences are highlighted in bold.

736

haematologica | 2016; 101(6)

ETP predicts outcome in HOXA-positive adult T-ALL.

Overall, these results suggest that despite the high associated rates of early treatment resistance, HOXA positivity does not influence patient outcomes directly.

HOXA. These analyses revealed a significant interaction between HOXA positivity and chemosensitivity, whereby HOXA positivity conferred significant decreases in both the event-free survival and cumulative incidence of relapse of chemoresistant patients (P=0.053 and P=0.039, respectively). Strikingly, this effect was independent of white blood cell count, stem cell transplantation, EGIL classification, and our recently reported risk classifier that integrates the prognostic effects of mutations of NOTCH1, FBXW7, RAS and PTEN25 (Online Supplementary Table S5). Taken together, these analyses indicate that the prognostic value of an ETP-like chemoresistant phenotype in adult TALL is specific to the HOXAPos cohort.

An early thymic precursor-like immunophenotype is associated with an inferior prognosis in HOXA-positive, but not HOXA-negative adult T-cell acute lymphoblastic leukemia Patients with HOXAPos or HOXANeg T-ALL had markedly different profiles of both genotypic and phenotypic maturity (Table 1). In particular, cases with trans-activation of the HOXA locus had very high rates of an ETP-like immunophenotype. This led us to speculate that HOXAPos ETP-ALL may constitute a distinct subgroup of adult TALL, and that HOXA overexpression might modulate the biology of ETP-ALL. We initially performed univariate survival analyses after division of the HOXAPos and HOXANeg patients into ETP and non-ETP cohorts. We found that the presence of an ETP-like immunophenotype correlated with marked differences in outcome within the HOXAPos group for overall survival (31.2% in HOXAPos ETP versus 66.7% in HOXAPos non-ETP; P=0.03), event-free survival (25% versus 52.8%; P=0.02), disease-free survival (28.6% versus 53.6%; P=0.02) and cumulative incidence of relapse (53.7% versus 25.4%; P=0.0095) at 5 years (Figure 4). In contrast, these survival differences were not seen in HOXANeg patients, among whom ETP and non-ETP cases had similar 5-year overall survival (74.2% in HOXANeg ETP versus 57.2% in HOXANeg non-ETP; P=0.44), event-free survival (60.8% versus 50.7%; P=0.72), disease-free survival (64.7% versus 52.2%; P=0.9) and cumulative incidence of relapse (29.2% versus 39.2%; P=0.57) (Figure 4). Multivariate analysis revealed that the statistical interaction between HOXA positivity and ETP-like phenotype did not reach independent significance when other prognostic factors were included in the model (Online Supplementary Table S5). As an ETP-like phenotype was usually associated with a profile of additional biological characteristics that might also influence patient outcome (Table 1 and Online Supplementary Table S4), we used multivariate statistical models to determine which clinical covariates were specifically modulated by the presence of

Discussion We have characterized the clinico-biological consequences of HOXA positivity in a large cohort of adult TALL patients uniformly treated as part of the GRAALL2003 and -2005 studies. We found that HOXA9 transcript levels robustly predicted global HOXA locus activation, thereby justifying this measurement as a proxy for definition of HOXA positivity. T-ALL cases exhibited a wide and continuous range of HOXA ratios, making rigid categorization of HOXAPos T-ALL difficult. In order to arrive at a practical cut-off, we chose the lowest ratio associated with a known HOXA-activating translocation. The legitimacy of this approach was supported by the finding that the HOXA locus was globally activated exclusively in HOXAPos patients, while diagnostic screening revealed no evidence of HOXA-activating translocations in patients with borderline ratios. We nevertheless cannot exclude the possibility that some cases classified as HOXANeg may have lesser degrees of HOXA activation which might affect disease biology. In addition, HOXA positivity was unexplained in 16 patients, despite extensive investigation. These cases had similar clinico-biological profiles to those of the trans-activated HOXAPos cohort, suggesting the presence of similar mechanisms of HOXA deregulation which remain to be discovered. Although HOXA overexpression has been linked to adverse prognosis in acute myeloid leukemia,30-33 the clini-

Figure 3. HOXAPos and HOXANeg adult T-ALL patients have similar survival outcomes. (A) Overall survival (OS) and (B) event-free survival (EFS) of HOXAPos and HOXANeg patients are shown. Five-year OS was estimated to be 55% (95% CI, 38.46% - 68.79%) in the HOXAPos group, as compared with 58.1% (95% CI, 48.50% - 66.49%) in HOXANeg cases. The corresponding figures for 5-year EFS were 45.9% (95% CI, 30.58% - 59.91%) for HOXAPos and 48.9% (95% CI, 39.73% - 57.53%) for HOXANeg.

haematologica | 2016; 101(6)

737

J. Bond et al.

cal impact of HOXA positivity in T-ALL has not previously been examined. We found that survival did not differ between the HOXAPos and HOXANeg patients when the groups were compared as a whole, but that significant disparities in outcome within the HOXAPos cohort were intimately linked to clinico-biological phenotype. Notably, cases that harbored an activation of the HOXA locus in trans had a high rate of an ETP-like immunophenotype that was typically associated with early treatment resistance and inferior survival. In keeping with these findings, HOXA positivity has recently been reported to be associated with an ETP-like gene expression profile and induction failure in pediatric T-ALL.34 Of note, the incidence of HOXA overexpression in that study was 25%, which is very similar to what we observed in this adult T-ALL cohort, and in excess of previous estimates of HOXA positivity based purely on transcriptomic clustering. Immunophenotypic classification of ETP-ALL in adults is complicated by a higher frequency of cases with over-

lapping patterns of antigen expression than that which is seen in children.35 This variability in the estimation of the incidence of ETP-ALL has in turn hindered identification of prognostic factors that may help predict the outcome of the disease.36 We found that the negative outcome of adults with ETP-ALL in this study was exclusive to the HOXAPos cohort, despite similarly high rates of chemoresistance in HOXAPos and HOXANeg ETP-ALL. Our results therefore identify HOXA positivity as a novel prognostic variable in adults with ETP-ALL, and the results of multivariate statistical analysis suggest that HOXA overexpression is directly correlated with the outcome of chemoresistant ETP-ALL. These results are also consistent with previous reports of poor outcome in HOXAPos PICALMMLLT10 T-ALL, which seems to be confined to cases with an immature immunophenotype.37,38 The negative prognosis that was originally described in pediatric ETP-ALL appears to have improved with the implementation of targeted treatment intensification

Figure 4. An ETP-like immunophenotype is associated with an inferior prognosis in HOXAPos, but not HOXANeg adult T-ALL. The results of survival analyses after separation of the HOXAPos and HOXANeg groups according to the presence or absence of an ETP-like immunophenotype are shown. The 5-year survival figures were as follows: (A) overall survival (OS) for HOXAPos ETP 31.3% (95% CI, 11.4% - 53.7%), HOXAPos non-ETP 66.7% (95% CI, 42.5% - 82.5%), HOXANeg ETP 74.2% (95% CI, 45% 89.4%), HOXANeg non-ETP 57.2% (95% CI, 46.1% - 66.9%). (B) Event-free survival (EFS) for HOXAPos ETP 25% (95% CI, 10.2% - 43.1%), HOXAPos non-ETP 52.7% (95% CI, 35.5% - 67.4%), HOXANeg ETP 60.8% (95% CI, 39.4% - 76.6%), HOXANeg non-ETP 50.7% (95% CI, 41.1% - 59.5%). (C) Disease-free survival (DFS) for HOXAPos ETP 28.6% (95% CI, 11.7% - 48.2%), HOXAPos non-ETP 53.6% (95% CI, 35.9% - 68.5%), HOXANeg ETP 64.7% (95% CI, 43.1% - 79.8%), HOXANeg non-ETP 52.2% (95%CI, 42.1% - 61.3%). (D) Cumulative incidence of relapse (CIR) for HOXAPos ETP 53.7% (95% CI, 34.3% - 75.6%), HOXAPos non-ETP 25.4% (95% CI, 13.6% - 44.4%), HOXANeg ETP 29.2% (95% CI, 15.1% - 51.6%), HOXANeg non-ETP 39.2% (95% CI, 30.6% - 49.3%).

738

haematologica | 2016; 101(6)

ETP predicts outcome in HOXA-positive adult T-ALL.

based on early MRD assessment.25,26,39,40 It remains to be seen whether similar strategies will improve the outcome of HOXAPos adult ETP-ALL patients. The introduction of more intensive pediatric-based regimens has been central to recent improvements in the outcome of hitherto resistant adult ALL.41-43 Patients included in the GRAALL-2003 and -2005 studies received enhanced induction and/or salvage therapy in the event of poor early treatment response; however, systematic early MRD monitoring was not performed for all patients. Nevertheless, our results suggest that this approach offered significant survival benefits for the HOXANeg cohort, as ETP-like and non-ETP-like cases ultimately had comparable outcomes. Conversely, we found that these treatment modifications were inadequate for therapeutic rescue of the majority of chemoresistant HOXAPos ETP-ALL cases, and that the outlook for these patients remains poor. We propose that the dramatically inferior prognosis of this group mandates consideration of alternative treatments in the context of future clinical trials. Previous data revealing the requirement for DOT1L activity in HOXA-overexpressing acute myeloid leukemia21 suggest that pharmacological DOT1L inhibition might also have therapeutic benefit in T-ALL. In addition, recent evidence from in vitro and animal models

References 1. Alexander S. Clinically defining and managing high-risk pediatric patients with acute lymphoblastic leukemia. Hematology Am Soc Hematol Educ Program. 2014(1):181189. 2. Van Vlierberghe P, Pieters R, Beverloo HB, Meijerink JP. Molecular-genetic insights in paediatric T-cell acute lymphoblastic leukaemia. Br J Haematol. 2008;143(2):153168. 3. Teitell MA, Pandolfi PP. Molecular genetics of acute lymphoblastic leukemia. Annu Rev Pathol. 2009;4:175-198. 4. Ferrando AA, Neuberg DS, Staunton J, et al. Gene expression signatures define novel oncogenic pathways in T cell acute lymphoblastic leukemia. Cancer Cell. 2002;1(1):75-87. 5. Soulier J, Clappier E, Cayuela JM, et al. HOXA genes are included in genetic and biologic networks defining human acute Tcell leukemia (T-ALL). Blood. 2005;106(1):274-286. 6. Homminga I, Pieters R, Langerak AW, et al. Integrated transcript and genome analyses reveal NKX2-1 and MEF2C as potential oncogenes in T cell acute lymphoblastic leukemia. Cancer Cell. 2011;19(4):484-497. 7. 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. 8. Shah N, Sukumar S. The Hox genes and their roles in oncogenesis. Nat Rev Cancer. 2010;10(5):361-371. 9. Bach C, Buhl S, Mueller D, Garcia-Cuellar MP, Maethner E, Slany RK. Leukemogenic transformation by HOXA cluster genes. Blood. 2010;115(14):2910-2918. 10. Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg AM, Sauvageau G. Hoxa9

haematologica | 2016; 101(6)

11.

12.

13.

14.

15.

16.

17.

18.

19.

suggests that combined inhibition of glycogen synthase kinase and poly(ADP-ribose) polymerase effectively suppresses growth of chemoresistant HOXA-overexpressing acute myeloid leukemia,44 suggesting a similar potential avenue of investigation in HOXAPos T-ALL. Acknowledgments The authors would like to thank all participants in the GRAALL-2003 and GRAALL-2005 study groups for collecting and providing data and patients’ samples, and V. Lheritier for collecting clinical data. The GRAALL-2003 study was sponsored by the Hôpitaux de Toulouse, and the GRAALL-2005 study by the Assistance Publique-Hôpitaux de Paris. The authors also thank Nicolas Boissel, Marina Lafage and MarieChristine Béné for their constructive and critical appraisal of the manuscript. JB was supported by a Kay Kendall Leukaemia Fund Intermediate Research Fellowship. The Necker laboratory is supported by the Association Laurette Fugain, La Ligue Contre le Cancer and the INCa CARAMELE Translational Research and PhD programs. The IBiSA ‘Transcriptomics and Genomics Marseille-Luminy (TGML)' platform was supported by the France Genomique National infrastructure, funded as part of the ‘Investissements d’Avenir’ programme (contract ANR-10-INBS-09).

transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. Embo J. 1998;17(13):37143725. Thorsteinsdottir U, Kroon E, Jerome L, Blasi F, Sauvageau G. Defining roles for HOX and MEIS1 genes in induction of acute myeloid leukemia. Mol Cell Biol. 2001;21(1):224-234. Cauwelier B, Dastugue N, Cools J, et al. Molecular cytogenetic study of 126 unselected T-ALL cases reveals high incidence of TCRbeta locus rearrangements and putative new T-cell oncogenes. Leukemia. 2006;20(7):1238-1244. Le Noir S, Ben Abdelali R, Lelorch M, et al. Extensive molecular mapping of TCRalpha/delta- and TCRbeta-involved chromosomal translocations reveals distinct mechanisms of oncogene activation in TALL. Blood. 2012;120(16):3298-3309. Bohlander SK, Muschinsky V, Schrader K, et al. Molecular analysis of the CALM/AF10 fusion: identical rearrangements in acute myeloid leukemia, acute lymphoblastic leukemia and malignant lymphoma patients. Leukemia. 2000;14(1):93-99. Zhang J, Ding L, Holmfeldt L, et al. The genetic basis of early T-cell precursor acute lymphoblastic leukaemia. Nature. 2012;481(7380):157-163. Brandimarte L, Pierini V, Di Giacomo D, et al. New MLLT10 gene recombinations in pediatric T-acute lymphoblastic leukemia. Blood. 2013;121(25):5064-5067. Bond J, Bergon A, Durand A, et al. Cryptic XPO1-MLLT10 translocation is associated with HOXA locus deregulation in T-ALL. Blood. 2014;124(19):3023-3025. Van Vlierberghe P, van Grotel M, Tchinda J, et al. The recurrent SET-NUP214 fusion as a new HOXA activation mechanism in pediatric T-cell acute lymphoblastic leukemia. Blood. 2008;111(9):4668-4680. Okada Y, Jiang Q, Lemieux M, Jeannotte L,

20. 21.

22.

23.

24.

25.

26.

27.

28.

Su L, Zhang Y. Leukaemic transformation by CALM-AF10 involves upregulation of Hoxa5 by hDOT1L. Nat Cell Biol. 2006;8(9):1017-1024. Okada Y, Feng Q, Lin Y, et al. hDOT1L links histone methylation to leukemogenesis. Cell. 2005;121(2):167-178. Bernt KM, Zhu N, Sinha AU, et al. MLLrearranged leukemia is dependent on aberrant H3K79 methylation by DOT1L. Cancer Cell. 2011;20(1):66-78. Van Vlierberghe P, Ambesi-Impiombato A, Perez-Garcia A, et al. ETV6 mutations in early immature human T cell leukemias. J Exp Med. 2011;208(13):2571-2579. Neumann M, Heesch S, Gokbuget N, et al. Clinical and molecular characterization of early T-cell precursor leukemia: a high-risk subgroup in adult T-ALL with a high frequency of FLT3 mutations. Blood Cancer J. 2012;2(1):e55. Neumann M, Heesch S, Schlee C, et al. Whole-exome sequencing in adult ETP-ALL reveals a high rate of DNMT3A mutations. Blood. 2013;121(23):4749-4752. Coustan-Smith E, Mullighan CG, Onciu M, et al. Early T-cell precursor leukaemia: a subtype of very high-risk acute lymphoblastic leukaemia. Lancet Oncol. 2009;10(2):147156. Inukai T, Kiyokawa N, Campana D, et al. Clinical significance of early T-cell precursor acute lymphoblastic leukaemia: results of the Tokyo Children's Cancer Study Group Study L99-15. Br J Haematol. 2012;156(3):358-365. Huguet F, Leguay T, Raffoux E, et al. Pediatric-inspired therapy in adults with Philadelphia chromosome-negative acute lymphoblastic leukemia: the GRAALL-2003 study. J Clin Oncol. 2009;27(6):911-918. Trinquand A, Tanguy-Schmidt A, Ben Abdelali R, et al. Toward a NOTCH1/FBXW7/RAS/PTEN-based onco-

739

J. Bond et al.

29.

30.

31.

32.

33.

34.

740

genetic risk classification of adult T-cell acute lymphoblastic leukemia: a Group for Research in Adult Acute Lymphoblastic Leukemia study. J Clin Oncol. 2013;31(34):4333-4342. Asnafi V, Beldjord K, Boulanger E, et al. Analysis of TCR, pT alpha, and RAG-1 in Tacute lymphoblastic leukemias improves understanding of early human T-lymphoid lineage commitment. Blood. 2003;101(7):2693-2703. Zangenberg M, Grubach L, Aggerholm A, et al. The combined expression of HOXA4 and MEIS1 is an independent prognostic factor in patients with AML. Eur J Haematol. 2009;83(5):439-448. Andreeff M, Ruvolo V, Gadgil S, et al. HOX expression patterns identify a common signature for favorable AML. Leukemia. 2008;22(11):2041-2047. Li Z, Huang H, Li Y, et al. Up-regulation of a HOXA-PBX3 homeobox-gene signature following down-regulation of miR-181 is associated with adverse prognosis in patients with cytogenetically abnormal AML. Blood. 2012;119(10):2314-2324. Golub TR, Slonim DK, Tamayo P, et al. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science. 1999;286(5439): 531-537. Matlawska-Wasowska K, Kang H, Devidas,

35.

36. 37.

38.

39.

M, et al. Mixed lineage leukemia rearrangements (MLL-R) are determinants of high risk disease in homeobox a (HOXA)-deregulated T-lineage acute lymphoblastic leukemia: a Children's Oncology Group Study. Am Soc Hematol Annual Meeting 2015 Abstract 694. Van Vlierberghe P, Ambesi-Impiombato A, De Keersmaecker K, et al. Prognostic relevance of integrated genetic profiling in adult T-cell acute lymphoblastic leukemia. Blood. 2013;122(1):74-82. Litzow MR, Ferrando AA. How I treat T-cell acute lymphoblastic leukemia in adults. Blood. 2015;126(7):833-841. van Grotel M, Meijerink JP, Beverloo HB, et al. The outcome of molecular-cytogenetic subgroups in pediatric T-cell acute lymphoblastic leukemia: a retrospective study of patients treated according to DCOG or COALL protocols. Haematologica. 2006;91 (9):1212-1221. Ben Abdelali R, Asnafi V, Petit A, et al. The prognosis of CALM-AF10-positive adult Tcell acute lymphoblastic leukemias depends on the stage of maturation arrest. Haematologica. 2013;98(11):1711-1717. Patrick K, Wade R, Goulden N, et al. Outcome for children and young people with early T-cell precursor acute lymphoblastic leukaemia treated on a contemporary protocol, UKALL 2003. Br J

Haematol. 2014;166(3):421-424. 40. Wood B, Winter S, Dunsmore K, et al. Tlymphoblastic leukemia (T-ALL) shows excellent outcome, lack of significance of the early thymic precursor (ETP) immunophenotype, and validation of the prognostic value of end-induction minimal residual disease (MRD) in Childrenâ&#x20AC;&#x2122;s Oncology Group (COG) Study AALL0434. Am Soc Hematol Annual Meeting 2014 Abstract 1. 41. Boissel N, Auclerc MF, Lheritier V, et al. Should adolescents with acute lymphoblastic leukemia be treated as old children or young adults? Comparison of the French FRALLE-93 and LALA-94 trials. J Clin Oncol. 2003;21(5):774-780. 42. Haiat S, Marjanovic Z, Lapusan S, et al. Outcome of 40 adults aged from 18 to 55 years with acute lymphoblastic leukemia treated with double-delayed intensification pediatric protocol. Leuk Res. 2011;35(1):6672. 43. Rijneveld AW, van der Holt B, Daenen SM, et al. Intensified chemotherapy inspired by a pediatric regimen combined with allogeneic transplantation in adult patients with acute lymphoblastic leukemia up to the age of 40. Leukemia. 2011;25(11):1697-1703. 44. Esposito MT, Zhao L, Fung TK, et al. Synthetic lethal targeting of oncogenic transcription factors in acute leukemia by PARP inhibitors. Nat Med. 2015;21(12):1481-1490.

haematologica | 2016; 101(6)

ARTICLE

Acute Lymphoblastic Leukemia

Risk assessment of relapse by lineage-specific monitoring of chimerism in children undergoing allogeneic stem cell transplantation for acute lymphoblastic leukemia

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Sandra Preuner,1 Christina Peters,2 Ulrike Pötschger,1 Helga Daxberger,1 Gerhard Fritsch,1 Rene Geyeregger,1 André Schrauder,3 Arend von Stackelberg,4 Martin Schrappe,3 Peter Bader,5 Wolfram Ebell,4 Cornelia Eckert,4 Peter Lang,6 Karl-Walter Sykora,7 Johanna Schrum,8 Bernhard Kremens,9 Karoline Ehlert,10 Michael H. Albert,11 Roland Meisel,12 Anita Lawitschka,2 Georg Mann,2 Renate Panzer-Grümayer,1 Tayfun Güngör,13 Wolfgang Holter,14,2,19 Brigitte Strahm,15 Bernd Gruhn,16 Ansgar Schulz,17 Wilhelm Woessmann,18 and Thomas Lion1,19

Children’s Cancer Research Institute, Vienna, Austria; 2St. Anna Children’s Hospital, Vienna, Austria; 3University Medical Center Schleswig-Holstein and Christian-AlbrechtsUniversity Kiel, Department of Pediatrics, Kiel, Germany; 4Charité, University Medicine Berlin, Germany; 5Johann Wolfgang Goethe University, Frankfurt, Germany; 6University Hospital Tübingen, Germany; 7Hannover Medical School, Hannover, Germany; 8University Hospital UKE, Hamburg, Germany; 9University Hospital Essen, Germany; 10University Children’s Hospital Münster, Germany (current address: Medical University Greifswald, Germany); 11Dr. von Hauner University Children’s Hospital, München, Germany; 12 University Hospital Düsseldorf, Germany; 13University Children’s Hospital Zürich, Division of Stem Cell Transplantation, Switzerland; 14Children’s University Hospital Erlangen, Germany; 15Pediatric Hematology and Oncology, University Medical Center Freiburg, Germany; 16University Hospital Jena, Germany; 17University Hospital Ulm, Germany; 18University Children’s Hospital Giessen, Germany; and 19Department of Pediatrics, Medical University Vienna, Austria 1

ABSTRACT

Haematologica 2016 Volume 101(6):741-746

Correspondence: thomas.lion@ccri.at

llogeneic hematopoietic stem cell transplantation is required as rescue therapy in about 20% of pediatric patients with acute lymphoblastic leukemia. However, the relapse rates are considerable, and relapse confers a poor outcome. Early assessment of the risk of relapse is therefore of paramount importance for the development of appropriate measures. We used the EuroChimerism approach to investigate the potential impact of lineage-specific chimerism testing for relapserisk analysis in 162 pediatric patients with acute lymphoblastic leukemia after allogeneic stem cell transplantation in a multicenter study based on standardized transplantation protocols. Within a median observation time of 4.5 years, relapses have occurred in 41/162 patients at a median of 0.6 years after transplantation (range, 0.13-5.7 years). Prospective screening at defined consecutive time points revealed that reappearance of recipient-derived cells within the CD34+ and CD8+ cell subsets display the most significant association with the occurrence of relapses with hazard ratios of 5.2 (P=0.003) and 2.8 (P=0.008), respectively. The appearance of recipient cells after a period of pure donor chimerism in the CD34+ and CD8+ leukocyte subsets revealed dynamics indicative of a significantly elevated risk of relapse or imminent disease recurrence. Assessment of chimerism within these lineages can therefore provide complementary information for further diagnostic and, potentially, therapeutic purposes aiming at the prevention of overt relapse. This study was registered at clinical.trials.gov with the number NC01423747.

haematologica | 2016; 101(6)

Received: August 11, 2015. Accepted: February 2, 2016. Pre-published: February 11, 2016. doi:10.3324/haematol.2015.135137

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/741

741

S. Preuner et al.

Introduction Treatment of acute lymphoblastic leukemia (ALL) according to current Berlin-Frankfurt-Münster (BFM)-ALL or similar intensive protocols results in cure rates of approximately 80% with chemotherapy alone.1,2 Nevertheless, a significant proportion of patients with resistant or relapsed disease require allogeneic hematopoietic stem cell transplantation (HSCT) as rescue therapy. Across all subtypes of pediatric ALL, about 20% of patients in industrialized countries currently undergo allogeneic HSCT from related or unrelated donors.3 Disease relapse, with an overall incidence of approximately 25%, is the dominant cause of mortality in this setting.4 Clone-specific markers for the detection of minimal residual disease (MRD) are available in most instances, and the current detection limit of these approaches is in the range of one in ten thousand cells (10-4).5,6 Potentially more sensitive detection of MRD can be achieved by real-time polymerase chain reaction (PCR) analysis of leukemia-specific fusion gene transcripts, but such markers are available only in a limited proportion of ALL patients.7 In patients undergoing allogeneic HSCT for treatment of various types of leukemia, persistence or recurrence of autologous cells detectable in either whole peripheral blood (PB) samples or within specific leukocyte subsets expected to harbor the malignant cells, if present, was shown to be indicative of imminent disease relapse.8,9 The identification of recipient-derived cells in whole PB specimens is hampered by the limited sensitivity offered by the most common approaches to chimerism analysis based on PCR amplification of microsatellite/short tandem repeat markers.10 These techniques are highly variable among different centers, and usually do not permit detection of recipient cells below the level of 10-2, thus lacking the sensitivity required for the assessment of residual leukemia.11 We and others have shown that it is readily possible to isolate individual leukocyte subsets by immunophenotype-based flow sorting, even if they account for as little as 1% of the total white blood cell count.12 The performance of chimerism analysis within specifically enriched leukocyte populations also has a detection limit in the range of 10-2, thereby permitting the identification of autologous cells in PB with an overall sensitivity of up to 10-4.10 Lineage-specific analysis of chimerism therefore offers a limit of detection for autologous and potentially leukemic cells in the range of sensitivity achievable by the commonly used methods for monitoring MRD. We have recently demonstrated that the assessment of lineage-specific chimerism within the first weeks after allogeneic HSCT facilitates prediction of the risk of graft rejection in transplant recipients, including children with ALL.13 In the present prospective multicenter study performed in a large cohort of pediatric patients with highrisk ALL over a period of 10 years, we have addressed the possibility of exploiting lineage-specific monitoring of chimerism for timely assessment of the risk of relapse after allogeneic HSCT. The study was performed in a blinded fashion to prevent the lineage-specific chimerism test results from having any influence on clinical decisions.

Methods Patients The present study was an ancillary research project of the international multicenter ALL-SCT-BFM 2003 trial,14 and was per742

formed with the approval of the local institutional review board at each participating site in accordance with the Declaration of Helsinki. Patients and/or their legal guardians provided written informed consent before enrollment. During the recruitment period between September 2003 and December 2008, a total of 162 high-risk ALL patients with a median age of 10.5 years (range, 0.623.1) who underwent allogeneic HSCT at 16 pre-selected centers were included in our study. The indication for allogeneic HSCT was established according to the risk-adapted stratification criteria of the BFM Study Group,15-17 and included children in first, second or subsequent complete morphological remission of ALL. Of 41 documented relapses in the entire cohort of patients studied, 28 occurred only in the bone marrow, eight were extramedullary, and five combined. Among these, six relapses were observed in patients with T-ALL, including three extramedullary and three bone marrow relapses, while the majority of bone marrow, extramedullary and combined relapses occurred in patients with Bcell precursor ALL. The blast cells in all relapses occurring in patients with B-cell precursor ALL displayed the CD19+/CD34+ marker profile.

Sample collection and handling PB specimens (≥3 mL EDTA-PB) were investigated at predefined time points including weekly intervals between days +7 and +49, followed by 2-week intervals until day +100, and four additional time points on days +120, +150, +180 and +360, whenever possible. Eight or more consecutive PB specimens were available for chimerism analysis in the majority of transplant recipients studied, with a mean of 8.6 samples per patient (median 8; range, 2-20).

Isolation of specific leukocyte subsets by flow-sorting The post-transplant PB samples were analyzed by four-color quantitative fluorescence activated cell sorting (FACS) on a FACSCalibur (BD Biosciences) as described elsewhere.18 All cell types forming ≥1% of nucleated cells were targets for cell sorting. Routinely sorted cells comprised CD33+ monocytes, CD15+ granulocytes, CD3+/CD4+ helper T cells, CD3+/CD8+ cytotoxic T cells, CD19+/CD34- mature B cells, CD3-/CD56+ natural killer cells and CD34+ early hematopoietic progenitor cells (also including CD34+/CD19+ cells). The targeted number of cells sorted for subsequent analysis of chimerism19,20 was 4 000 per cell population, and ranged between 1 000 and 15 000. The purity of the sorted leukocyte fractions was >98%.

Analysis of donor/recipient chimerism and minimal residual disease DNA was extracted from nucleated cells using the QIAamp Blood kit (Qiagen, Hilden, Germany). Chimerism was evaluated using the EuroChimerism approach, an internationally validated assay for quantitative analysis of donor and recipient cells after transplantation.11 Data from PCR-based MRD analysis21 were available at similar time points to chimerism testing in ten patients, thereby permitting a limited direct comparison between the methodological approaches.

Definitions Complete donor chimerism (DC) was defined as the exclusive presence of donor-derived cells, as determined by PCR analysis of an informative microsatellite (short tandem repeat) marker in the specimen investigated. Mixed chimerism (MC) was defined as the presence of both donor and recipient cells at any ratio in the sample tested. ALL relapse following allogeneic HSCT was defined as the presence of ≥5% blasts in representative bone marrow smears. haematologica | 2016; 101(6)

Relapse risk in pediatric ALL

Statistics The analysis was performed using the Statistical Analysis System (SAS Institute, Cary, NC, USA). Since mortality not attributable to relapse was low (9%) in the cohort of patients investigated, the statistical analysis was restricted to the assessment of chimerism data with regard to relapse occurrence. Cumulative incidences of relapse were estimated considering the competing risk of death. Starting from the time of transplantation, the proportional cause-specific hazard model with time-dependent co-variates considering the time points of first achievement of DC, and first recurrence of MC after a period of DC, was used to determine the correlation of lineage-specific chimerism with relapse. The hazard ratios (HR) for patients with recurrent MC in comparison to patients who achieved persistent DC are indicated with 95% confidence intervals (CI). The comparison of patients who had recurrent MC with those who had persistent DC was restricted to individuals who had achieved DC during the post-transplant course, and the first observation of DC was the time of entry into the risk evaluation set. In patients with reappearance of MC after previous documentation of DC, the cumulative incidence of relapse was estimated from the time of MC recurrence as the starting point. Cumulative incidences of the residual relapse risk in patients who achieved DC were estimated for illustrative and exploratory purposes by censoring patients who subsequently switched to MC. The interval for this evaluation started at the median time to MC recurrence, and only patients who had reached DC at this time point were included. The correlation of recurring MC in the CD34+ and CD8+ cell fractions with ensuing relapse was assessed by multivariate analysis including other known risk factors such as the remission status before transplantation and the type of donor (Table 3). All indicated P-values are twosided, and P-values ≤0.05 are considered statistically significant.

Results In the present study, lineage-specific and overall chimerism were monitored at defined time points at a single reference center during the first year after allogeneic HSCT in 162 children with high-risk ALL. The patients’ characteristics are shown in Table 1. The risk of relapse was assessed in relation to the presence and dynamics of chimerism in whole blood and within individual leukocyte subsets. The 5-year cumulative incidence of relapse in the cohort of patients presented was 26%, and the correlation of relapse occurrence with overall and lineage-specific chimerism was evaluated in patients who achieved persistent DC, in comparison with those who had continuous MC or reappearance of autologous cells after a phase of DC (Table 2). Significant associations with relapse were observed for recurrence of MC at any level in the CD34+ and CD8+ compartments, and to a lesser extent in whole PB, in patients with a preceding period of DC in the indicated cell compartments (Figure 1). The detection of reappearing autologous cells in one or more consecutive specimens within the CD34+ and CD8+ cell compartments enriched from PB revealed markedly elevated hazard ratios for ensuing disease relapse of 5.2 (P=0.003) and 2.8 (P=0.008), respectively (Table 2). Recurrence of MC detectable in whole PB specimens was also associated with an elevated risk of relapse (HR 2.1), albeit with lower statistical significance (P=0.024) (Table 2). No significant correlations were found for the remaining leukocyte lineages investigated. The cumulative incidence of relapse for patients with documented reappearance of MC in the CD34+ cell fraction haematologica | 2016; 101(6)

was 67% (±16). The switch from DC to MC in this compartment occurred a mean time of 6 months after transplantation when the intervals of investigation were rather large, and the time span between detection of MC and hematologic relapse was less than 1 month in most instances (Figure 1). Since a proportion of relapses occurred already during the first weeks and months after transplantation, the time-dependent residual relapse risk in the subset of patients with detectable CD34+ cells in PB and the presence of DC within this cell fraction at 6 months after transplantation was 31% (±10). By contrast, in patients in whom CD34+ cells in PB were undetectable by FACS analysis or insufficiently abundant for isolation by flow sorting (usually <1% of total leukocytes), the cumulative incidence of relapse was 8% (±3). Hence, the presence of circulating CD34+ cells in PB at levels readily amenable to isolation by flow sorting (≥1% of total leukocytes) also correlated with an elevated risk of relapse (HR 4.3, P<0.001, 95% CI 2.1168.735), irrespectively of the chimerism status within this cell fraction. The median time span between first detection of circulating CD34+ cells in PB after transplantation and onset of relapse was 183 days (lower quartile: 75 days, upper quartile: 315 days). Monitoring by flow cytometry revealed the presence of CD34+ cells in PB amenable to sorting and molecular testing in a limited proportion of patients (Table 2), while successful chimerism analysis within total leukocyte populations was possible in virtually all specimens tested. Given the commonly observed delayed reconstitution of T-lymphocytes, CD8+ cells could not be readily isolated for chimerism testing during the first weeks after transplantation. However, from 6 to 7 weeks after HSCT onward, chimerism analysis within the CD8+ cell subset was possible in ≥80% of all cases with available PB specimens.

Table 1. Patients’ baseline characteristics.

Patients (n=162) Sex Female Male Age Median (min-max), years Remission First complete remission Second complete remission Beyond second complete remission No complete remission Immunophenotype1 B-cell precursor-ALL T-cell ALL Other Donor Matched sibling donor Matched donor Mismatched donor Stem cell source2 Bone marrow Peripheral blood Cord blood Combined

58 (36%) 104 (64%) 10.2 (0.5-22.7) 80 (49%) 62 (38%) 15 (9%) 5 (3%) 113 (74%) 34 (22%) 5 (3%) 49 (30%) 96 (59%) 17 (10%) 109 (69%) 45 (28%) 2 (1%) 2 (1%)

Unknown in ten patients; 2Unknown in four patients.

743

S. Preuner et al.

Among the children who had a recurrence of MC either in the CD8+ cell compartment or within whole PB, the cumulative incidence of relapse was 33% (±7). The switch from DC to MC in whole PB and within the CD8+ cell lineage was detected a median period of 3 months after allografting, at stages when the time-dependent residual relapse risk was 18% (±4) and 15%(±4), respectively. The median time to relapse after detection of recurrent MC in whole PB or CD8+ cells was 5 months (range, 0.5-24 months), as displayed in Figure 1. Rising levels of autologous cells were identified by quantitative monitoring of MC in a proportion of patients who subsequently progressed to overt relapse, but the observation of expanding recipient cells did not further increase the statistical significance of risk assessment for ensuing relapse when compared to the recurrence of MC at any level. In contrast to reappearance of MC after a period of DC, the constant presence and persistence of MC did not indicate a significantly increased risk of relapse, irre-

spectively of the cell populations affected (CD34+: HR 1.3, P=0.504; CD8+: HR 1.9, P=0.185; PB: HR 1.2, P=0.697). The CD19+/CD34- cell compartment was the only leukocyte subset analyzed in which the observation of MC revealed hazard ratios for relapse below 1.0 in all statistical analyses performed (Table 2). Although the indication of decreased risk of relapse (HR 0.7) failed to reach statistical significance (P=0.441), the relapse rate in patients with reappearance or persistence of MC within CD19+/CD34cells was 12.5% in cases of B-cell precursor ALL, and 0% in cases with T-ALL, as compared to 28% and 11% in patients with DC in this cell compartment.

Discussion In this prospective, multicenter study, we have investigated the potential role of lineage-specific chimerism analysis

Table 2. Post-transplant evolution of chimerism in whole blood and selected cell lineages. Within the entire cohort of 162 pediatric patients with ALL undergoing rescue therapy by allogeneic HSCT, chimerism was successfully analyzed within total leukocytes from peripheral blood (total PB), and specific cell fractions isolated by flow sorting from PB in the indicated number of instances. Enrichment of CD34+ cells from PB by flow sorting was achievable in 91/162 patients, but the cell numbers were sufficient for subsequent analysis of chimerism in 62 cases only. The leukocyte lineages statistically correlating in univariate analysis with the risk of relapse, including the CD8+ and the CD34+ cell subsets, are displayed, and the respective numbers of patients who reached complete donor chimerism (DC) or switched to MC after previous documentation of DC (DC→MC) are indicated. The 5-year cumulative risk of relapse (CIR) and the hazard ratios (HR) derived from a Cox-regression analysis with timedependent covariates are indicated with the corresponding 95% confidence intervals (CI). The CD19+/CD34- cell lineage in which the presence of MC correlated with decreased CIR (statistically not significant) is shown for comparison.

Parameter

Patients with evaluable chimerism

Total PB CD8+ cells CD34+ cells CD19+/CD34- cells

162 156 62 151

DC → MC

DC Patients Relapses* 145 139 34 146

36 28 15 32

CIR*

CIR

95% CI

P values

0.33±0.07 0.31±0.07 0.67±0.16 0.12±0.06

2.1 2.8 5.2 0.7

1.1-4.2 1.3-5.9 1.8-15.1 0.2-1.9

0.024 0.008 0.003 0.441

Patients Relapses

0.25±0.04 0.20±0.03 0.44±0.09 0.22±0.03

52** 39 9 33

17 12 6 4

*The indicated numbers include relapses observed after recurrent MC following a period of DC. **A single patient who experienced a relapse a few days before detection of recurrent MC has been excluded.

Table 3. Multivariate analysis of risk factors for relapse after allogeneic HCST. In addition to the reappearance of mixed chimerism (MC) in the leukocyte subsets CD34+, CD8+ and total peripheral blood (indicated as →MC in the Table), the influence of additional factors including the remission status before HSCT and the type of donor, which indicated a correlation with relapse risk in univariate analysis, were subjected to multivariate analysis. While the donor type had no influence on the relapse risk in this setting, the achievement of first or second complete remission prior to HSCT was significantly correlated with a decreased risk of relapse. Immunophenotype was not associated with higher relapse hazards and the results remained essentially unchanged when immunophenotype was added to the model. Reappearing MC in the indicated cell populations remained an independent risk factor for ensuing leukemia relapse even after adjustment for all other variables tested. The 95% confidence intervals (95%CI) of the calculated hazard ratios are indicated.

Parameter +

CD34

CD8+

Factor

→MC CR(vs. CR2) Donor (vs. MSD) →MC CR (no CR)

Donor (vs. MSD) Total PB

→MC CR (no CR)

Donor (vs. MSD)

CR1 MD MMD CR1 CR2 MD MMD CR1 CR2 MD MMD

P value

Hazard ratio

0.001 0.065 0.228 0.026 0.005 0.001 0.003 0.865 0.991 0.022 0.004 0.031 0.297 0.700

8.36 0.34 2.08 18.27 3.06 0.06 0.09 0.93 0.00 2.21 0.10 0.19 0.68 0.80

95% CI 2.44 0.11 0.63 1.42 1.41 0.01 0.02 0.42 0.00 1.12 0.02 0.04 0.34 0.26

28.60 1.07 6.87 234.91 6.64 0.32 0.45 2.07 4.37 0.49 0.86 1.40 2.50

CR1/CR2,: first/second complete remission; MSD: matched sibling donor; MD: matched (unrelated) donor; MMD: mismatched donor.

744

haematologica | 2016; 101(6)

Relapse risk in pediatric ALL

for risk assessment of relapse in a large cohort of pediatric patients with high-risk ALL who underwent allogeneic HSCT according to uniform protocols. The significant correlation of recurring MC in the CD34+ and CD8+ cell fractions with ensuing relapse was confirmed by multivariate analysis including other known risk factors such as the remission status before transplantation and the type of donor. The observation that recurrent but not persistent MC was significantly associated with an increased risk of relapse may suggest that reappearing autologous cells mediate or indicate hitherto unknown immune processes facilitating disease recurrence. It is of note that flow cytometrybased detection of CD34+ cells circulating in the PB also indicated a significantly elevated risk of relapse, irrespectively of the chimerism status. One may speculate that increased mobilization of CD34+ cells from the bone marrow could be related to perturbation of the stem cell niche or cytokine-mediated release attributable to leukemic progenitor cells. The detection of residual or reappearing recipient cells by chimerism analysis in whole PB or within specific cell lineages isolated according to the immunophenotype of the original leukemia may reflect the presence of malignant cells.10 The ability to identify cells belonging to the leukemic clone by chimerism analysis would be particularly useful in transplant settings in which neither diseasespecific nor clone-specific markers are available for the detection of minimal residual or recurrent leukemia.12,22 However, in patients with ALL, specific markers permitting sensitive detection of leukemic cells are usually available, and can also be exploited for the monitoring of MRD after allogeneic HSCT.23 The identification of autologous cells in the CD34+ compartment may indicate incomplete myeloablation by the conditioning regimen. Alternatively, it may reflect persistence or recurrence of the malignant clone, but this finding is inherently less specific than surveillance of clone- or disease-specific molecular MRD markers. In the present study, identification of recurrent MC in the CD34+ cell fraction correlated in most instances with available results of MRD analysis based on PCR detection of clone-specific immunoglobulin or T-cell receptor gene rearrangements at levels of â&#x2030;Ľ10-4. The limited amount of MRD data obtained at time points similar to those of lineage-specific chimerism analyses did not permit systematic comparison of the diagnostic approaches. Nevertheless, in one patient who ultimately progressed to overt relapse, reversion from DC to MC in the CD34+ compartment was observed at a time at which MRD analysis did not provide any evidence of the presence of leukemic

cells. This anecdotal observation highlights the potential complementarity of different diagnostic approaches usually yielding similar information.24,25 Although the analysis of chimerism within total leukocyte preparations from PB is not sensitive enough to reveal small numbers of leukemic cells, it was shown that persistence or recurrence of low-level MC may nevertheless correlate with an increased risk of relapse.9 These observations indicate that chimerism testing and MRD analysis assess different aspects contributing to relapse occurrence. The monitoring of chimerism provides insights beyond mere detection of potentially leukemic cells, including particularly immunological processes contributing to graft rejection and relapse which are not amenable to detection by MRD analysis. Findings indicative of impending allograft rejection also herald the loss of the graft-versusleukemia effect which may be associated with a high risk of disease recurrence. It is also conceivable that additional hitherto incompletely understood immunological interactions occurring between donor and recipient play a role in the ensuing leukemia relapse. The observed correlation between the recurrence of MC at any level within the CD8+ lymphocyte fraction and the risk of relapse, both in patients with T-cell and B-cell precursor leukemia, may serve as evidence supporting this notion. Although most patients with recurrent MC in the CD8+ cell fraction reverted again to DC, the risk of relapse remained elevated, similar to that of patients who retained MC, possibly indicating that the effect mediated by the recurrence of autologous cells within this compartment may not be reversible. Lineage-specific analysis of chimerism provides a sensitive approach to the investigation of such interactions, and the findings presented highlight the potential of this methodology to contribute to early assessment of the risk of relapse. The surveillance of chimerism can yield important information complementary to direct molecular detection of residual leukemia, and could provide a rational basis for timely considerations aiming at the prevention of overt relapse. In addition to close diagnostic monitoring of patients identified as having an increased risk of relapse, the measures could include immunomodulatory treatment to stabilize the graft and to enhance the graft-versusleukemia effect. Recurrence of recipient CD34+ cells in PB after transplantation, which could be assessed in a limited proportion (38%) of the patients studied, revealed the highest hazard ratio for ensuing disease relapse in the setting investigated, albeit with a short time to detection of hematologic relapse in most cases. In order to better exploit the informativeness

Figure 1. Cumulative risk of relapse (CIR) in relation to post-transplant chimerism. In patients showing recurrence of MC within specific cell fractions including CD34+ and CD8+ cells or within total leukocytes isolated from peripheral blood (PB), a significantly increased risk of relapse was observed, with hazard ratios (HR) and CIR values specified in Table 2. The median time between recurrence of MC and relapse was different for the cell subsets analyzed: 5 months for CD8+ and total PB, and less than 1 month for CD34+ cells.

haematologica | 2016; 101(6)

745

S. Preuner et al.

of CD34+ cells for relapse prediction, it would be of interest to investigate bone marrow specimens. The CD34+ cells are present in greater abundance in this compartment and isolation of cell numbers adequate for chimerism analysis is, therefore, easier. Bone marrow aspirations are routinely performed in most centers around days +28 and +100 after transplantation for assessment of hematologic remission and engraftment status. This material could serve as a readily available source for analysis of lineage-specific chimerism in the CD34+ or other cell fractions, and could contribute to improving the predictive value of chimerism testing for impending relapse. Based on the observations presented, prospective evaluation of this concept would be warranted. The observed association of persistent or reappearing MC in the mature B-cell compartment with decreased risk of relapse is intriguing, despite the fact that this finding did not reach statistical significance. The majority of ALL patients studied had a B-cell precursor phenotype, and it is very unlikely that the autologous CD19+/CD34- cells observed belonged to the leukemic clone. Moreover, the same observation was also made in patients with T-ALL.

References 1. Schrappe M, Hunger SP, Pui CH, et al. Outcomes after induction failure in childhood acute lymphoblastic leukemia. N Engl J Med. 2012;366(15):1371-1381. 2. Pui CH, Pei D, Campana D, et al. A revised definition for cure of childhood acute lymphoblastic leukemia. Leukemia. 2014;28(12): 2336-2343. 3. Balduzzi A, Valsecchi MG, Uderzo C, et al. Chemotherapy versus allogeneic transplantation for very-high-risk childhood acute lymphoblastic leukaemia in first complete remission: comparison by genetic randomisation in an international prospective study. Lancet. 2005;366(9486):635-642. 4. Fagioli F, Quarello P, Zecca M, et al. Hematopoietic stem cell transplantation for children with high-risk acute lymphoblastic leukemia in first complete remission: a report from the AIEOP registry. Haematologica. 2013;98(8):1273-1281. 5. Gandemer V, Pochon C, Oger E, et al. Clinical value of pre-transplant minimal residual disease in childhood lymphoblastic leukaemia: the results of the French minimal residual disease-guided protocol. Br J Haematol. 2014;165(3):392-401. 6. Balduzzi A, Di Maio L, Silvestri D, et al. Minimal residual disease before and after transplantation for childhood acute lymphoblastic leukaemia: is there any room for intervention? Br J Haematol. 2014;164(3): 396-408. 7. Pui CH, Mullighan CG, Evans WE, Relling MV. Pediatric acute lymphoblastic leukemia: where are we going and how do we get there? Blood. 2012;120(6):1165-1174. 8. Lion T, Daxberger H, Dubovsky J, et al. Analysis of chimerism within specific leukocyte subsets for detection of residual or recurrent leukemia in pediatric patients after allogeneic stem cell transplantation. Leukemia. 2001;15(2):307-310. 9. Bader P, Kreyenberg H, Hoelle W, et al. Increasing mixed chimerism is an important

746

10. 11.

12.

13.

14.

15.

16.

17.

This observation cannot be readily explained and may, therefore, provide the grounds for a future, appropriately designed study to address this phenomenon. Lineage-specific monitoring of leukocytes isolated from the PB of patients transplanted for treatment of ALL provides a basis for close diagnostic surveillance in individuals with increased risk of relapse, and for timely therapeutic considerations aiming at the prevention of overt relapse. The observations presented suggest that more frequent analyses of chimerism within the CD34+ compartment could enable earlier identification of patients with a particularly high risk of disease recurrence. Furthermore, the data provide an impetus for future studies addressing the potential benefit of lineage-specific chimerism analyses of bone marrow specimens. These efforts could increase the sensitivity and timeliness of identifying patients at high risk of leukemia relapse and contribute to improved outcomes in ALL patients undergoing allogeneic HSCT. Acknowledgments The study was supported by an institutional grant from the Childrenâ&#x20AC;&#x2122;s Cancer Research Institute (CCRI).

prognostic factor for unfavorable outcome in children with acute lymphoblastic leukemia after allogeneic stem-cell transplantation: possible role for pre-emptive immunotherapy? J Clin Oncol. 2004;22 (9):1696-1705. Preuner S, Lion T. Post-transplant monitoring of chimerism by lineage-specific analysis. Methods Mol Biol. 2014;1109:271-291. Lion T, Watzinger F, Preuner S, et al. The EuroChimerism concept for a standardized approach to chimerism analysis after allogeneic stem cell transplantation. Leukemia. 2012;26(8):1821-1828. Lion T. Detection of impending graft rejection and relapse by lineage-specific chimerism analysis. Methods Mol Med. 2007;134:197-216. Breuer S, Preuner S, Fritsch G, et al. Early recipient chimerism testing in the T- and NK-cell lineages for risk assessment of graft rejection in pediatric patients undergoing allogeneic stem cell transplantation. Leukemia. 2012;26(3):509-519. Peters C, Schrappe M, von Stackelberg A, et al. Stem-Cell Transplantation in children with acute lymphoblastic leukemia: a prospective international multicenter trial comparing sibling donors with matched unrelated donors-the ALL-SCT-BFM-2003 trial. J Clin Oncol. 2015;33(11):1265-1274. Schrauder A, Reiter A, Gadner H, et al. Superiority of allogeneic hematopoietic stem-cell transplantation compared with chemotherapy alone in high-risk childhood T-cell acute lymphoblastic leukemia: results from ALL-BFM 90 and 95. J Clin Oncol. 2006;24(36):5742-5749. Schrauder A, von Stackelberg A, Schrappe M, et al. Allogeneic hematopoietic SCT in children with ALL: current concepts of ongoing prospective SCT trials. Bone Marrow Transplant. 2008;41 (Suppl 2):S71-74. von Stackelberg A, Volzke E, Kuhl JS, et al. Outcome of children and adolescents with relapsed acute lymphoblastic leukaemia and non-response to salvage protocol therapy: a retrospective analysis of the ALL-REZ BFM

Study Group. Eur J Cancer. 2011;47(1):90-97. 18. Fritsch G, Witt V, Dubovsky J, et al. Flow cytometric monitoring of hematopoietic reconstitution in myeloablated patients following allogeneic transplantation. Cytotherapy. 1999;1(4):295-309. 19. Schraml E, Daxberger H, Watzinger F, Lion T. Quantitative analysis of chimerism after allogeneic stem cell transplantation by PCR amplification of microsatellite markers and capillary electrophoresis with fluorescence detection: the Vienna experience. Leukemia. 2003;17(1):224-227. 20. Schraml E, Lion T. Interference of dye-associated fluorescence signals with quantitative analysis of chimerism by capillary electrophoresis. Leukemia. 2003;17(1):221-223. 21. van der Velden VH, Panzer-Grumayer ER, Cazzaniga G, et al. Optimization of PCRbased minimal residual disease diagnostics for childhood acute lymphoblastic leukemia in a multi-center setting. Leukemia. 2007;21 (4):706-713. 22. Lion T. Chimerism testing after allogeneic stem cell transplantation: importance of timing and optimal technique for testing in different clinical-biological situations. Leukemia. 2001;15(2):292. 23. Bruggemann M, Schrauder A, Raff T, et al. Standardized MRD quantification in European ALL trials: proceedings of the Second International Symposium on MRD assessment in Kiel, Germany, 18-20 September 2008. Leukemia. 2010;24(3):521535. 24. Bader P, Willasch A, Klingebiel T. Monitoring of post-transplant remission of childhood malignancies: is there a standard? Bone Marrow Transplant. 2008;42 (Suppl 2):S31-34. 25. Bader P, Kreyenberg H, von Stackelberg A, et al. Monitoring of minimal residual disease after allogeneic stem-cell transplantation in relapsed childhood acute lymphoblastic leukemia allows for the identification of impending relapse: results of the ALL-BFMSCT 2003 trial. J Clin Oncol. 2015;33(11): 1275-1284.

haematologica | 2016; 101(6)

ARTICLE

Complications in Hematology

Late thyroid complications in survivors of childhood acute leukemia. An L.E.A. study

Claire Oudin,1,2 Pascal Auquier,2 Yves Bertrand,3 Philippe Chastagner,4 Justyna Kanold,5 Maryline Poirée,6 Sandrine Thouvenin,7 Stephane Ducassou,8 Dominique Plantaz,9 Marie-Dominique Tabone,10 Jean-Hugues Dalle,11 Virginie Gandemer,12 Patrick Lutz,13 Anne Sirvent,14 Virginie Villes,2 Vincent Barlogis,1,2 André Baruchel,11 Guy Leverger,10 Julie Berbis,2 and Gérard Michel1,2

Department of Pediatric Hematology and Oncology, Timone Enfants Hospital and AixMarseille University; 2Research Unit EA 3279 and Department of Public Health, AixMarseille University and Timone Hospital Marseille; 3Department of Pediatric Hematology and Oncology, University Hospital of Lyon; 4Department of Pediatric OncoHaematology, Children’s Hospital of Brabois, Vandoeuvre Les Nancy; 5Department of Pediatric Hematology and Oncology, CIC Inserm 501, University Hospital of ClermontFerrand; 6Pediatric Hematology and Oncology Department, University Hospital L’Archet, Nice; 7Pediatric Hematology, University Hospital, Saint Etienne; 8Department of Pediatric Hematology and Oncology, University Hospital of Bordeaux; 9Department of Pediatric Hematology-Oncology, University Hospital of Grenoble; 10Pediatric Hematology Department, Trousseau Hospital, Paris; 11Pediatric Hematology Department, Robert Debré Hospital, Paris; 12Department of Pediatric Hematology and Oncology, University Hospital of Rennes; 13Department of Pediatric Hematology-Oncology, University Hospital, Strasbourg; and 14Pediatric Hematology and Oncology Department, University Hospital, Montpellier, France

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2016 Volume 101(6):747-756

ABSTRACT

hyroid complications are known side effects of irradiation. However, the risk of such complications in childhood acute leukemia survivors who received either central nervous system irradiation or hematopoietic stem cell transplantation is less described. We prospectively evaluated the incidence and risk factors for thyroid dysfunction and tumors in survivors of childhood acute myeloid or lymphoid leukemia. A total of 588 patients were evaluated for thyroid function, and 502 individuals were assessed for thyroid tumors (median follow-up duration: 12.6 and 12.5 years, respectively). The cumulative incidence of hypothyroidism was 17.3% (95% CI: 14.1-21.1) and 24.6% (95% CI: 20.4-29.6) at 10 and 20 years from leukemia diagnosis, respectively. Patients who received total body irradiation (with or without prior central nervous system irradiation) were at higher risk of hypothyroidism (adjusted HR: 2.87; P=0.04 and 2.79, P=0.01, respectively) as compared with transplanted patients who never received any irradiation. Patients transplanted without total body irradiation who received central nervous system irradiation were also at higher risk (adjusted HR: 3.39; P=0.02). Patients irradiated or transplanted at older than 10 years of age had a lower risk (adjusted HR: 0.61; P=0.02). Thyroid malignancy was found in 26 patients (5.2%). Among them, two patients had never received any type of irradiation: alkylating agents could also promote thyroid cancer. The cumulative incidence of thyroid malignancy was 9.6% (95% CI: 6.015.0) at 20 years. Women were at higher risk than men (adjusted HR: 4.74; P=0.002). In conclusion, thyroid complications are frequent among patients who undergo transplantation after total body irradiation and those who received prior central nervous system irradiation. Close monitoring is thus warranted for these patients. Clinicaltrials.gov identifier: NCT 01756599.

haematologica | 2016; 101(6)

Correspondence: claire.oudin@ap-hm.fr

Received: December 2, 2015. Accepted: March 1, 2016. Pre-published: March 11, 2016. doi:10.3324/haematol.2015.140053

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/747

747

C. Oudin et al.

Introduction As high cure rates have led to an increase in long-term survivors treated for acute childhood leukemia, a mounting number of studies have focused on the long-term health status of such patients.1 It has long been known that the side effects which can occur include late thyroid complications, such as thyroid dysfunction2-9 and thyroid cancer.10-13 Thyroid dysfunction is rather frequent and has been mostly reported as a complication of irradiation, particularly for solid tumors, such as brain cancers10,14 and Hodgkin lymphoma.15,16 In leukemia survivors, thyroid complications have been mainly described as a consequence of central nervous system or total body irradiation and hematopoietic stem cell transplantation. However, the conclusions of many published studies may have been biased by the small numbers of patients included, the heterogeneity of the study cohorts (e.g., different disease types and treatment modalities), and the methodology used (e.g., questionnaires, retrospective studies). Concerning patients who receive central nervous system radiation, the possible effect of scatter radiation on the thyroid is poorly understood. Moreover, little is known regarding the risk of thyroid complications following busulfan-containing conditioning regimens and hematopoietic stem cell transplantation.17,18 Lastly, few studies have assessed both thyroid dysfunction and thyroid malignancies simultaneously. We therefore aimed to describe the incidence of and risk factors for thyroid dysfunction and thyroid malignancy in a large cohort of long-term survivors of childhood acute leukemia. As chemotherapy alone usually does not lead to an increased risk of late thyroid complications,4,7,16,19 our study was focused on survivors of childhood acute leukemia who received either central nervous system irradiation (who could be vulnerable to the scatter effects of this type of radiation on the thyroid gland) or underwent hematopoietic stem cell transplantation (in whom the thyroid could be damaged both by the direct effect of total body irradiation and the possible effect of high-dose alkylating agents). The current study was based on data from the French Leucémie de l’Enfant et de l’Adolescent (L.E.A.) prospective cohort. All patients underwent systematic and repeated hormonal and ultrasound evaluations of any thyroid complications at predefined dates, thus minimizing potential bias while providing estimates of cumulative incidences over time.

National Program for Clinical Research and the National Cancer Institute and by our local institutional review boards. The data were collected during specific medical visits at predefined dates, initially every 2 years during a 10-year post-transplantation follow-up period and subsequently every 4 years (for details, see the Online Supplementary Methods). From 2004 to 2012, it was recommended that all patients undergoing a new L.E.A. evaluation who received either a hematopoietic stem cell transplant or central nervous system irradiation were systematically assessed for thyroid function and tumors. Participants in the thyroid function analysis had at least one assessment of hormone levels, while participants in the thyroid tumor evaluation had at least one ultrasound scan during their follow-up. Thyroid function was assessed by monitoring thyroid stimulating hormone (TSH) and free thyroxine (T4) by immunoassay with a high molecular weight ligand or labeled antibody.22 Uncompensated peripheral hypothyroidism was diagnosed in cases of elevated TSH levels (> 5 mIU/L) and low T4 levels (< 10 pmol/L), while compensated peripheral hypothyroidism was diagnosed in cases of elevated TSH levels and normal T4 levels. Central hypothyroidism was defined by low T4 (<10 pmol/L) and normal/low TSH levels. Thyroid tumors (nodules or micronodules) were evaluated using thyroid ultrasound scans, which were carried out to assess tumor size and characteristics. In the case of suspicious nodules, fine needle aspiration, nodule biopsy and/or thyroidectomy were performed. Malignancy and cancer histopathology subtype were defined based on the histopathology findings. Statistical analyses were performed using SPSS 20.0 (SPSS Inc., Chicago, IL, USA) and Intercooled Stata 9.0 for Windows. The χ2 and Fisher exact tests were used to compare qualitative variables. Quantitative variables were compared using the MannWhitney U-test. The prevalence rates of thyroid dysfunction and tumors were assessed with a 95% confidence interval (CI). Cumulative incidence rates of hypothyroidism and thyroid cancers over time were estimated using the Kaplan-Meier method, with 95% confidence intervals, and compared using the log rank test. The proportional hazards assumption was assessed via examination of the log-minus-log survival plot of each cofactor. For the analysis of cumulative incidence rates of hypothyroidism, patients were censored when they underwent thyroidectomy, regardless of the underlying cause. Hazard ratios (HR) and the associated 95% CI were estimated using Cox proportional hazard models; P values <0.05 are considered statistically significant.

Results Methods We assessed thyroid dysfunction and tumors in patients from the L.E.A. cohort. The L.E.A. program was implemented in 2004 to prospectively evaluate the long-term health status, quality of life and socio-economic status of childhood acute leukemia survivors enrolled in French treatment programs from 1980 to present, in 13 cancer centers. Further details of the program have been described elsewhere.20,21 The eligibility criteria for this study were the following: (i) provision of written informed consent to participation in the L.E.A. program between 2004 and 2012, and (ii) having received central nervous system irradiation and/or undergone hematopoietic stem cell transplantation with a myeloablative conditioning regimen as part of the treatment. All patients (or their parents) provided written informed consent to participation in the study, which was approved by the French 748

Patients’ characteristics: comparison between participating and eligible but not-participating patients Among 665 eligible patients from 13 French centers, 588 (88.4%) were evaluated for thyroid function during the years 2004 to 2012 and 502 (75.5%) for thyroid tumors (Online Supplementary Figure S1). The patients’ characteristics and treatment modalities are shown in Online Supplementary Table S1. The percentages of transplanted patients were 75.7% among the group in which thyroid function was evaluated (of whom 48.1% were transplanted after relapse) and 76.3% in the group in which thyroid tumors were evaluated (49.6% were transplanted after relapse). The median (interquartile range, IQR) duration of follow-up from diagnosis to last horhaematologica | 2016; 101(6)

Thyroid complications after childhood leukemia

haematologica | 2016; 101(6)

Figure 1. Cumulative incidences of hypothyroidism. (A) Cumulative incidence of hypothyroidism; impact of transplantation and irradiation type. (P<10-3). (B) Cumulative incidence of hypothyroidism in patients treated with hematopoietic stem cell transplantation without total body irradiation (n=135): comparison between patients who received prior central nervous system irradiation and patients who did not (P=0.005). (C) Cumulative incidence of hypothyroidism among patients treated with hematopoietic stem cell transplantation who received total body irradiation as part of their conditioning: comparison between patients who were exposed to central nervous system irradiation and those who were not (P=non significant). HSCT: hematopoietic stem cell transplantation; CNS: central nervous system. TBI: total body irradiation.

749

C. Oudin et al.

monal assessment and last thyroid ultrasound scan was 12.6 (IQR: 6.4-18.8) and 12.4 (IQR: 6.4-19.1) years, respectively. Patients in the participating groups had more frequently experienced leukemia relapse and were more likely to have received total body irradiation before transplantation; they were also less likely to have received central nervous system irradiation than eligible patients who did not participate in the study. Gender, leukemia subtype, age at diagnosis, at central nervous system irradiation and at hematopoietic stem cell transplantation, central nervous system irradiation doses and fields were not different between the participating and non-participating patients.

Thyroid dysfunction Prevalence and cumulative incidence The most common thyroid dysfunction was hypothyroidism, which was observed in 105/588 patients (17.9%). Primary hypothyroidism was reported in 99 cases (94.3%), while central hypothyroidism was observed in six cases (5.7%). Hypothyroidism was uncompensated in 34 cases and compensated in 62 patients. A total of 98/105 patients received L-thyroxine substitution. The cumulative incidence of hypothyroidism for all patients was 17.3% (95% CI: 14.1-21.1) at 10 years and 24.6% at 20 years (95% CI: 20.4-29.6). The median age at diagnosis of hypothyroidism was 12.6 (IQR: 7.8-15.2) years. The median delay between irradiation or transplantation and hypothyroidism diagnosis

was 3.3 (IQR 1.4-6.4) years. One patient developed hyperthyroidism (Grave disease).

Risk factors for hypothyroidism We found that the type of irradiation had a marked impact on the risk of hypothyroidism (Figure 1A), which was higher in cases of total body irradiation-based conditioning regimens than in cases of transplantation without total body irradiation or cases of central nervous system irradiation without transplantation [cumulative incidences at 20 years were 36.7% (95%CI: 30.3-44.2), 19.2% (95% CI: 10.9-32.5) and 5.0% (95% CI: 1.9-13.2) for each group, respectively, P<10-3]. Among the patients who underwent stem cell transplantation without total body irradiation, the risk was higher in cases of prior central nervous system irradiation than in cases without prior central nervous system irradiation [cumulative incidences at 20 years: 44.2% (95% CI: 22.6-73.4) versus 12.1% (95% CI: 4.9-28.4), respectively; P=0.005] (Figure 1B). In contrast, the incidence of hypothyroidism among patients who received a total body irradiation-based conditioning regimen was not affected by prior central nervous system irradiation (Figure 1C). In a univariate analysis, we did not find any impact of acute leukemia subtype (lymphoid versus myeloid), transplantation type (autologous versus allogeneic), or central nervous system irradiation field (cranial versus craniospinal) among patients treated with central nervous system irradiation (Table 1). Of note, the risk of hypothy-

Table 1. Univariate analysis: risk factors for hypothyroidism (n=588).

Gender Male Female Leukemia type Acute myeloid leukemia Acute lymphoid leukemia Age at transplantation/irradiation < 10 years â&#x2030;Ľ 10 years Hematopoietic stem cell transplantation type Autologous Allogeneic Treatment modality HSCT without TBI, without prior CNS irradiation CNS irradiation, no HSCT HSCT with TBI, with prior CNS irradiation HSCT with TBI, without prior CNS irradiation HSCT without TBI, with prior CNS irradiation CNS irradiation type Cranial irradiation Cerebrospinal irradiation CNS irradiation dose 18 Grays 24 Grays Other

Number (%)

P value

52 (15.3) 53 (21.4)

10-year cumulative incidence (%)

[95% CI]

0.16

15.7 19.3

[11.9-20.7] [14.4-25.7]

19.8 30.5

[15.1-25.7] [23.5-39.0]

16 (11.0) 89 (20.1)

0.08

12.5 18.6

[7.6-20.0] [14.9-23.1]

15.5 27.0

[9.0-25.9] [22.0-32.7]

74 (20.5) 31 (13.7)

0.19

19.2 14.2

[15.1-24.2] [9.9-20.7]

27.3 19.3

[22.0-33.6] [13.2-27.8]

26 (28.6) 73 (20.6)

0.71

23.2 23.0

[15.6-33.6] [18.1-28.9]

29.4 33.6

[20.6-40.8] [26.4-42.2]

7.0 10 24.4 29.1 21.0

[3.5-13.5] [0.1-6.8] [10.9-49.1] [23.4-35.9] [7.3-52.1]

12.1 5.0 42.9 35.4 44.2

[4.9-28.4] [1.9-13.2] [23.9-68.4] [28.8-42.9] [22.6-73.4]

0.49

4.7 6.6

[2.0-10.9] [2.2-19.2]

12.1 19.2

[6.6-21.9] [9.4-37.0]

0.07

4.5 6.8 25.0

[1.9-10.5] [1.7-25.0]] [6.9-68.5]

11.5 23.9 25.0

[6.2-20.8] [11.4-46.0] [6.9-68.5]

9 (7.5) 6 (4.2) 9 (42.9) 74 (25.6) 7 (46.7) 13 (10.7) 8 (15.1) 11 (8.2) 8 (22.2) 3 (37.5)

<10-3

20-year cumulative incidence (%)

[95% CI]

CNS: central nervous system; HSCT: hematopoietic stem cell transplantation; TBI: total body irradiation. In bold: statistically significant value.

750

haematologica | 2016; 101(6)

Thyroid complications after childhood leukemia

roidism tended to be higher among patients who received central nervous system irradiation with 24 Grays than in those who received only 18 Grays, although the difference was not statistically significant (P=0.07). Patients were then stratified into five groups according to the therapeutic modalities they had received: (i) transplantation without total body irradiation, without prior central nervous system irradiation [n=120 (20.4%)]; (ii) central nervous system irradiation, no transplantation [n=143 (24.3%)]; (iii) transplantation after total body irradiation with prior central nervous system irradiation [n=21 (3.6%)]; (iv) transplantation after total body irradiation, without prior central nervous system irradiation [n=289 (49.1%)]; and (v) transplantation without total body irradiation, with prior central nervous system irradiation [n=15 (2.6%)].

The multivariate analysis confirmed that transplantation and irradiation modalities played a role in the development of hypothyroidism. The adjusted hazard ratio was 2.87 (95% CI: 1.03-7.99) for patients who received total body irradiation and prior central nervous system irradiation (P=0.04), 2.79 (95% CI: 1.28-6.1) for those treated with total body irradiation without prior central nervous system irradiation (P=0.01), and 3.39 (95% CI: 1.23-9.35) for those who received a transplant without total body irradiation but who had had prior central nervous system irradiation (P=0.02), as compared with the reference population (patients transplanted without total body irradiation, without prior irradiation). In contrast, the adjusted hazard ratio was 0.27 for patients who received only central nervous system irradiation (and no transplantation) (P=0.02). The patientsâ&#x20AC;&#x2122; age at transplantation or irradiation also

Figure 2. Cumulative incidence of thyroid cancer. The figure shows the cumulative incidence of thyroid cancer in the whole cohort (green line), and a comparison of the incidence between females (blue line) and males (red line). P<0.001.

Table 2. Risk factors for hypothyroidism: multivariate analysis (n=588).

Variables Age at transplantation/irradiation < 10 years â&#x2030;Ľ 10 years Gender Male Female Leukemia type Acute lymphoid leukemia Acute myeloid leukemia Treatment modality HSCT without TBI, without prior CNS irradiation CNS irradiation, no HSCT HSCT with TBI, with prior CNS irradiation HSCT with TBI, without prior CNS irradiation HSCT without TBI, with prior CNS irradiation

95% CI

P value

1 0.61

0.40-0.93

0.02

57.8 42.2

1 1.43

0.97-2.11

0.07

442 146

75.2 24.8

1 0.62

0.33-1.14

0.13

120 143 21 289 15

20.4 24.3 3.6 49.1 2.6

1 0.27 2.87 2.79 3.39

0.09-0.81 1.03-7.99 1.28-6.10 1.23-9.35

0.02 0.04 0.01 0.02

361 227

61.4 38.6

340 248

Adjusted HR

CNS: central nervous system; HSCT: hematopoietic stem cell transplantation; TBI: total body irradiation. In bold: statistically significant values.

haematologica | 2016; 101(6)

751

C. Oudin et al.

had a marked impact on hypothyroidism. Patients irradiated or transplanted at older than 10 years of age had a lower risk of hypothyroidism (adjusted hazard ratio: 0.61; P=0.02). The results of the multivariate analyses are summarized in Table 2.

was 1.7% at 10 years (95% CI: 0.8-3.9) and reached 9.6% (95% CI: 6.0-15.0) at 20 years (Figure 2). The median age at diagnosis of thyroid cancer was 20.5 (IQR: 17.1-24.1) years. Patients with thyroid cancer were treated with total or subtotal thyroidectomy and radioactive iodine therapy. No patients died from thyroid cancer.

Risk factors for primary hypothyroidism We also performed the same univariate and multivariate analyses exclusively considering patients presenting with primary hypothyroidism, excluding six patients with central hypothyroidism for whom we could not diagnose a potential primary form of the disease. The results were very similar to those reported in the previous section (including both primary and central hypothyroidism). In the univariate analysis, patients who received total body irradiation and those who underwent transplantation without total body irradiation but with prior central nervous system irradiation had a higher risk of primary hypothyroidism (P<10-3). This was confirmed by the multivariate analysis (for both univariate and multivariate analyses, see Online Supplementary Tables S2 and S3).

Thyroid tumor Prevalence and cumulative incidence A total of 162 patients (32.3%) were found to have thyroid nodules or micronodules [median age at nodule diagnosis: 21.1 (IQR: 15.7-27.8) years]. Of those 162 patients, 56 (34.6%) underwent cytological examination: 26 (46.4%) had papillary carcinoma, 28 had benign adenoma and two had dystrophic lesions without evidence of malignancy. The cumulative incidence of thyroid cancer

Risk factors for thyroid cancer In the univariate analysis, the only significant variable that had an impact on the incidence of thyroid cancer was gender; women were at a higher risk of developing thyroid cancer than were men [cumulative incidence at 20 years: 2.3% (95% CI: 0.6-8.2) and 17.1% (95% CI: 10.627) for men and women, respectively, P<0.001] (Figure 2). Leukemia subtype, age at transplantation or irradiation, type of stem cell transplant (autologous versus allogeneic), central nervous irradiation type (cranial or craniospinal) or dose (18 versus 24 Grays) had no impact on thyroid cancer (Table 3). However, we did observe a trend towards a higher cumulative incidence among patients who received total body irradiation, the cumulative incidence of thyroid cancer at 15 years being 8.3% and 5.5% for patients who received total body irradiation with and without central nervous system irradiation, respectively, 3.8% for patients who underwent transplantation without any irradiation, and 2.1% for those who received only central nervous system irradiation without transplantation. After adjusting for other variables, the prognostic value of gender remained the only significant value (adjusted HR: 4.74 for women, P=0.002) (Table 4).

Table 3. Univariate analysis: risk factors for thyroid cancer (n=502).

Number (%)

P value

15-year cumulative incidence (%)

[95%CI]

20-year cumulative incidence (%)

[95% CI]

5 (1.8) 21 (9.6)

<10-3

1.0 7.5

[0.2-3.9] [3.9-14.2]

2.3 17.1

[0.6-8.2] [10.6-27.0]

6 (5.1) 20 (5.2)

0.40

1.6 4.7

[0.2-10.9] [2.5-8.7]

12.3 8.9

[4.6-30.7] [5.3-14.9]

18 (5.9) 8 (4.1)

0.92

4.2 3.6

[2.0-8.7] [1.3-9.8]

11.0 6.0

[6.5-18.3] [2.2-15.4]

7 (10.0) 14 (4.5)

0.39

7.1 3.7

[2.7-18.0] [1.5-9.0]

9.4 13.5

[4.0-21.3] [6.9-25.5]

2 (2.0) 5 (4.2) 2 (10.5) 16 (6.3) 1 (10.0)

0.08

3.9 2.1 8.3 5.5 0.0

[0.6-24.3] [0.5-8.2] [1.2-46.1] [2.6-11.5]

8.2 6.0 8.3 14.3 0.0

[2.1-29.2] [2.2-16.0] [1.2-46.1] [7.9-25.2]

6 (5.7) 2 (5.1)

0.83

2.9 2.7

[0.7-11.3] [0.4-17.7]

7.1 2.7

[2.7-18.3] [0.4-17.7]

7 (6.1) 0 (0.0) 1 (11.1)

0.14

3.7 0.0 0.0

[1.2-11.2]

8.1 0.0 0.0

[3.3-19.2]

CNS: central nervous system; HSCT: hematopoietic stem cell transplantation; TBI: total body irradiation. In bold: statistically significant value.

752

haematologica | 2016; 101(6)

Thyroid complications after childhood leukemia

Correlation between hypothyroidism and thyroid cancer To assess the relationship between hypothyroidism and thyroid cancer, we selected patients who were evaluated for both diseases (488 patients). We did not find any correlation between hypothyroidism and thyroid cancer either in the entire cohort of 488 patients (P=0.80) or in a selected population of patients who had received total body irradiation as part of their conditioning (P=0.78).

Discussion We assessed the incidence of and risk factors for late thyroid complications (thyroid dysfunction and thyroid tumors) that can occur after treatment for acute leukemia during childhood. Few studies have focused on both forms of thyroid pathology. The high number of patients included and the prospective nature of this study enabled us to determine precisely the cumulative risk of thyroid complications over time among a homogeneous cohort of long-term survivors of childhood acute leukemia following treatment with cranial irradiation and/or hematopoietic stem cell transplantation. Hypothyroidism was found to be frequent in our cohort (17.9%) and was mostly of the primary type. Compensated hypothyroidism was reported in the majority of cases. Of note, the prevalence of compensated hypothyroidism in the French general population is 1.9% and 3.3% for men and women, respectively; these prevalences increase with age.23 Our findings are consistent with those of previous comparable studies. A retrospective questionnaire-based study from the Childhood Cancer Survivor study including 3467 survivors of leukemia found hypothyroidism in 6.5% of cases, with prevalences ranging from 6.8% (patients treated with cranial irradiation) to 29.8% (patients receiving total body irradiation and prior central nervous system irradiation).24 However, that study did not mention the cumulative incidence of hypothyroidism, which was 24.6% at 20 years (95% CI: 20.429.6) in our cohort. This cumulative incidence is much higher than that reported by Chow et al. (15-year cumu-

lative incidence of hypothyroidism: 1.6%), probably because their study also included patients treated with chemotherapy alone, and because the methodology used may have undervalued the incidence of hypothyroidism (retrospective study, self-reported thyroid dysfunction).4 Primary hypothyroidism has been linked to exposure to thyroid radiation,25 primarily in solid tumors such as Hodgkin lymphoma15,16 and brain tumors,10,14 although it was also found to be associated with cranial or craniospinal irradiation in patients with childhood acute leukemia.4 Cranial irradiation in patients with acute lymphoblastic leukemia has long been suggested to play a role in hypothyroidism,3,4,26 although a few authors have argued that the impact of prophylactic cranial radiation is controversial.3,5,7,27 However, those earlier studies should be interpreted with caution because of the small numbers of patients included, the retrospective nature of the studies and/or the short follow-up periods. In the current study, our results support the hypothesis that the risk of hypothyroidism is significantly lower in patients who received only central nervous system irradiation (adjusted HR=0.27, P=0.02) compared with that in the reference population (i.e., patients treated with hematopoietic stem cell transplantation without any history of irradiation). This might be due to a low dose of radiation received by the thyroid gland (most patients who underwent central nervous system irradiation received a dose of 18 Grays). This also raises the question of the potential impact of high-dose alkylating agents (used in many preparative regimens, instead of total body irradiation) on the thyroid gland, as previously suggested.17 According to the literature, the risk of primary hypothyroidism seems to be strongly associated with the total irradiation dose received by the thyroid gland.25 It has been suggested that the risk of hypothyroidism is higher in patients exposed to craniospinal irradiation than in those who undergo cranial irradiation.4 We, however, did not find similar results, which may be due to the small number of patients who received craniospinal irradiation. Several studies have focused on thyroid dysfunction after hematopoietic stem cell transplantation.17,28-31 In a

Table 4. Multivariate analysis: risk factors for thyroid cancer.

95% CI

P value

1 0.89

0.37-2.12

0.79

56.6 43.4

1 4.74

1.76-12.82

0.002

384 118

76.5 23.5

1 1.47

0.52-4.18

0.47

99 119 19 255 10

19.7 23.7 3.8 50.8 2

1 0.74 2.19 3.13 0.99

0.12-4.66 0.24-19.68 0.62-15.81 0.08-12.12

0.74 0.49 0.17 1

306 196

61.0 39.0

284 218

Adjusted HR

CNS: central nervous system; HSCT: hematopoietic stem cell transplantation; TBI: total body irradiation. In bold: statistically significant value.

haematologica | 2016; 101(6)

753

C. Oudin et al.

large, prospective study from the Fred Hutchinson Cancer Research Center,17 which included 791 patients transplanted for malignant or non-malignant disease during childhood, the prevalence of hypothyroidism was as high as 30%. In the present study, the 10-year cumulative incidence of hypothyroidism among transplanted patients was 23%, which is lower than in the aforementioned study. This could be due to the focus on acute leukemia survivors in our study, whereas Sanders et al. studied patients with a variety of pathologies. They also studied patients who had been treated with conditioning regimens including both busulfan and total body irradiation, which markedly increased the risk of thyroid dysfunction. In other studies concerning transplanted patients, based on smaller cohorts with various durations of follow-up, the prevalence of hypothyroidism ranged from 10.8% to 40%.18,28-30 This varying prevalence may be due to the heterogeneity of the cohorts, in terms of both follow-up time and pathology (e.g., leukemia, lymphoma, and non-malignant). As expected, in our study total body irradiation was strongly associated with an increased risk of hypothyroidism, compared with the risk in transplanted patients who did not receive any type of irradiation (i.e., total body irradiation or central nervous system irradiation) and in those who received only central nervous system irradiation (multivariate analysis) (Table 2). This observation is consistent with previously reported results.14,24,29 More interestingly, and more rarely compared to other studies, our study also included a large number of patients treated with hematopoietic stem cell transplantation without the use of total body irradiation (n=135). We found that, in transplanted patients who did not receive total body irradiation, prior cerebral irradiation increased the risk of developing hypothyroidism (adjusted HR=3.39, P=0.02), which suggests that irradiation and alkylating agents may exert a cumulative effect. It is, therefore, important to monitor thyroid dysfunction closely in transplanted patients who receive a non-totalbody-irradiation-based conditioning regimen with prior central nervous system irradiation as well as in those treated with total body irradiation. To our knowledge, this is the first study describing such results. The patients included in our study were more likely to have received total body irradiation and less likely to have been treated with central nervous system irradiation than eligible non-included patients. This could have introduced a degree of bias in our results. In the L.E.A. program, only patients who either underwent stem cell transplantation or received cranial/craniospinal radiotherapy were screened for thyroid dysfunction, which could also be considered as a bias in the present study. However, this choice was based on numerous studies that have failed to demonstrate that standarddose chemotherapy alone plays a role in promoting thyroid dysfunction.4,16,19,32 We found that age (<10 years) at transplantation or irradiation was another risk factor for the development of hypothyroidism (adjusted HR=0.61 for patients over 10 years of age, P=0.02), which correlates with previously reported findings.17,29,30 Nearly all of the patients diagnosed with hypothyroidism received replacement therapy with levothyroxine, both in cases of uncompensated and compensated hypothyroidism. This is in accordance with general practice in 754

France, as high TSH levels could be associated with an increased risk of thyroid cancer in the general population.33 Nevertheless, systematic replacement therapy for subclinical compensated hypothyroidism remains debated.34 We found only six cases of central hypothyroidism, all of which were in patients who received either total body irradiation or cranial irradiation. This is consistent with the reports of most studies in the field, which describe peripheral hypothyroidism as the most common thyroid disorder following treatment in leukemia survivors.5,25,29,35 However, Sanders et al. reported 74 cases of central hypothyroidism among 791 patients,17 which may be due to a high number of single doses of total body irradiation. It was previously shown that single-dose total body irradiation is associated with increased toxicity and notably greater thyroid toxicity.29 We also evaluated the occurrence of and risk factors for thyroid cancer. In the general population, the incidence of thyroid cancer is 1.5/100,000 men and 4.7/100,000 women.36 Many previous studies have focused on thyroid cancer after treatment for childhood cancer.13,30,37-39 The reported prevalences varied, which may be due to several factors, e.g., the studies were frequently retrospective, based on questionnaires, or included small numbers of patients. In contrast, our study was prospective and included a large number of patients. In our cohort, thyroid cancer was found in 26 (5.2%) of 502 patients who underwent systematic ultrasound examination of the thyroid. In all 26 cases, the histological subtype was papillary carcinoma. This subtype is the most frequent one, not only in the general population, but also following treatment for childhood malignancy.11,13,37,38,40,41 Some authors found that being less than 10 years of age was associated with an increased risk of thyroid cancer,11,13,38 although we did not find similar results. In contrast, we confirmed that women are at an increased risk of developing thyroid cancer (HR=4.74; P=0.002), which has been suggested by other studies13 and observed in the general population.36 Notably, thyroid cancer occurred at a very young age in our cohort [median age at thyroid cancer diagnosis: 20.5 (IQR: 17.1-24.1) years, median period from leukemia diagnosis: 16.1 (IQR: 11.5-21.1) years], whereas thyroid cancer is diagnosed much later in the general population, usually at 45-50 years of age.42 As the median follow-up duration was only 12.4 (IQR: 6.4-19.1) years, we probably underestimate the prevalence of thyroid cancer in our population, with disease occurrence increasing with age. This underscores the need for long-term follow-up of this population. Radiation therapy is a well-known risk factor for thyroid cancer,13,43 even in cases of low-dose radiation. Notably, two patients among the 26 who developed a thyroid cancer in our cohort had never received any irradiation but had been transplanted after receiving a busulfan-based conditioning regimen. This is consistent with the results published by Cohen et al., who found that among 32 patients who developed thyroid cancer after transplantation, seven had never been exposed to any radiation therapy.13 The use of alkylating agents may, therefore, also promote the development of thyroid cancer, as suggested by previous studies.18,39 We, therefore, suggest that all transplanted patients be carefully monitored for thyroid malignancy, irrespectively of the conditioning regimen applied. We did not evaluate whether thyroid ultrasound scan was the best way to haematologica | 2016; 101(6)

Thyroid complications after childhood leukemia

screen for thyroid malignancies. Systematic thyroid ultrasound scanning is not cost-effective for the general population. However, the incidence of thyroid malignancies among survivors of childhood cancer is high, and neck palpation is often insufficient to detect small suspicious nodules in these high-risk patients.37 Furthermore, in the follow-up of irradiated individuals, several studies have shown that many large, ultrasounddetected thyroid nodules escape detection via palpation.37,44 Lastly, thyroid ultrasound scanning is a simple and non-invasive method. Taken together, these arguments lead us to consider that thyroid ultrasound scanning is a valuable method for screening for malignancies, regardless of a lack of definitive data comparing this method with other screening modalities. There was no central review of the thyroid ultrasound scans, which could be considered another weakness of this study. However, given the nature of this imaging method, it is difficult to perform a central review. We did not find any relationship between hypothyroidism and thyroid cancer. Some authors have suggested a link between high TSH levels and the occurrence of thyroid cancer in the general population,33,45 based on the understanding that TSH is a major growth factor for thyroid cells and that some animal models have indicated that TSH plays a role in the development of follicular cellderived thyroid cancer. Nevertheless, this has never been

References 1. Pui CH, Pei D, Sandlund JT, et al. Risk of adverse events after completion of therapy for childhood acute lymphoblastic leukemia. J Clin Oncol. 2005;23(31):79367941. 2. de Fine Licht S, Winther JF, Gudmundsdottir T, et al. Hospital contacts for endocrine disorders in Adult Life after Childhood Cancer in Scandinavia (ALiCCS): a populationbased cohort study. Lancet. 2014;383(9933): 1981-1989. 3. Carter EP, Leiper AD, Chessells JM, Hurst A. Thyroid function in children after treatment for acute lymphoblastic leukaemia. Arch Dis Child. 1989;64(4):631. 4. Chow EJ, Friedman DL, Stovall M, et al. Risk of thyroid dysfunction and subsequent thyroid cancer among survivors of acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor study. Pediatr Blood Cancer. 2009;53(3):432-437. 5. Lando A, Holm K, Nysom K, et al. Thyroid function in survivors of childhood acute lymphoblastic leukaemia: the significance of prophylactic cranial irradiation. Clin Endocrinol (Oxf). 2001;55(1):21-25. 6. Follin C, Link K, Wiebe T, Moell C, Bjork J, Erfurth EM. Prolactin insufficiency but normal thyroid hormone levels after cranial radiotherapy in long-term survivors of childhood leukaemia. Clin Endocrinol (Oxf). 2013;79(1):71-78. 7. Voorhess ML, Brecher ML, Glicksman AS, et al. Hypothalamic-pituitary function of children with acute lymphocytic leukemia after three forms of central nervous system prophylaxis. A retrospective study. Cancer. 1986;57(7):1287-1291.

haematologica | 2016; 101(6)

confirmed for survivors of childhood cancer, which may be because of frequent systematic replacement therapy with levothyroxine in cases of elevated TSH levels. In conclusion, our findings suggest that irradiated and transplanted long-term survivors of childhood acute leukemia should undergo careful and prolonged monitoring for thyroid dysfunction. Patients exposed to total body irradiation or central nervous system irradiation followed by transplantation are at the highest risk of thyroid complications. Clinicians should be aware that thyroid cancer is frequent among survivors of acute leukemia treated with central nervous system irradiation and/or hematopoietic stem cell transplantation and such patients therefore warrant early and prolonged systematic screening with ultrasound scans. Acknowledgments The authors would like to thank the nursing staff for their excellent care of patients. They would also like to thank the L.E.A staff (Supplementary Data) for their help with data collection. They are also grateful to the patients and their families. The study was supported by the French National Clinical Research Program, the French National Cancer Institute (InCA), the French National Research Agency (ANR), the Canceropole PACA, the Regional Council PACA, the Herault Departmental Comity of the Ligue Contre le Cancer and the French Institute for Public Health Research (IRESP).

8. Mohn A, Chiarelli F, Di Marzio A, Impicciatore P, Marsico S, Angrilli F. Thyroid function in children treated for acute lymphoblastic leukemia. J Endocrinol Invest. 1997;20(4):215-219. 9. Pasqualini T, Zantleifer D, Balzaretti M, et al. Evidence of hypothalamic-pituitary thyroid abnormalities in children with endstage renal disease. J Pediatr. 1991;118(6): 873-878. 10. Schmiegelow M, Feldt-Rasmussen U, Rasmussen AK, Poulsen HS, Muller J. A population-based study of thyroid function after radiotherapy and chemotherapy for a childhood brain tumor. J Clin Endocrinol Metab. 2003;88(1):136-140. 11. Sigurdson AJ, Ronckers CM, Mertens AC, et al. Primary thyroid cancer after a first tumour in childhood (the Childhood Cancer Survivor study): a nested case-control study. Lancet. 2005;365(9476):20142023. 12. Socie G, Curtis RE, Deeg HJ, et al. New malignant diseases after allogeneic marrow transplantation for childhood acute leukemia. J Clin Oncol. 2000;18(2):348-357. 13. Cohen A, Rovelli A, Merlo DF, et al. Risk for secondary thyroid carcinoma after hematopoietic stem-cell transplantation: an EBMT Late Effects Working Party Study. J Clin Oncol. 2007;25(17):2449-2454. 14. Madanat LM, Lahteenmaki PM, Alin J, Salmi TT. The natural history of thyroid function abnormalities after treatment for childhood cancer. Eur J Cancer. 2007;43(7):1161-1170. 15. Hancock SL, Cox RS, McDougall IR. Thyroid diseases after treatment of Hodgkinâ&#x20AC;&#x2122;s disease. N Engl J Med. 1991;325 (9):599-605. 16. Sklar C, Whitton J, Mertens A, et al.

17.

18.

19.

20.

21.

22.

23.

Abnormalities of the thyroid in survivors of Hodgkin's disease: data from the Childhood Cancer Survivor Study. J Clin Endocrinol Metab. 2000;85(9):3227-3232. Sanders JE, Hoffmeister PA, Woolfrey AE, et al. Thyroid function following hematopoietic cell transplantation in children: 30 years' experience. Blood. 2009;113(2):306308. Slatter MA, Gennery AR, Cheetham TD, et al. Thyroid dysfunction after bone marrow transplantation for primary immunodeficiency without the use of total body irradiation in conditioning. Bone Marrow Transplant. 2004;33(9):949-953. Nygaard R, Bjerve KS, Kolmannskog S, Moe PJ, Wesenberg F. Thyroid function in children after cytostatic treatment for acute leukemia. Pediatr Hematol Oncol. 1988;5(1): 35-38. Berbis J, Michel G, Baruchel A, et al. Cohort Profile: the French Childhood Cancer Survivor Study for leukaemia (LEA cohort). Int J Epidemiol. 2015;44(1):49-57. Michel G, Bordigoni P, Simeoni MC, et al. Health status and quality of life in long-term survivors of childhood leukaemia: the impact of haematopoietic stem cell transplantation. Bone Marrow Transplant. 2007;40(9):897-904. Sapin R, Schlienger JL. [Thyroxine (T4) and tri-iodothyronine (T3) determinations: techniques and value in the assessment of thyroid function]. Ann Biol Clin (Paris). 2003;61(4):411-420. Valeix P, Dos Santos C, Castetbon K, Bertrais S, Cousty C, Hercberg S. [Thyroid hormone levels and thyroid dysfunction of French adults participating in the SU.VI.MAX study]. Ann Endocrinol (Paris). 2004;65 (6):477-486.

755

C. Oudin et al. 24. Chow EJ, Liu W, Srivastava K, et al. Differential effects of radiotherapy on growth and endocrine function among acute leukemia survivors: a childhood cancer survivor study report. Pediatr Blood Cancer. 2013;60(1):110-115. 25. Chemaitilly W, Sklar CA. Endocrine complications in long-term survivors of childhood cancers. Endocr Relat Cancer. 2010;17(3): R141-159. 26. Robison L, Nesbit MEJ, Sather HN, Meadows AT, Ortega JA, Hammond GD. Thyroid abnormalities in long-term survivors of childhood acute leukemia (ALL). Pediatr Res. 1985;19:2566A. 27. Hata M, Ogino I, Aida N, et al. Prophylactic cranial irradiation of acute lymphoblastic leukemia in childhood: outcomes of late effects on pituitary function and growth in long-term survivors. Int J Cancer. 2001;96 (Suppl):117-124. 28. Baker KS, Ness KK, Weisdorf D, et al. Late effects in survivors of acute leukemia treated with hematopoietic cell transplantation: a report from the Bone Marrow Transplant Survivor Study. Leukemia. 2010;24(12): 2039-2047. 29. Berger C, Le-Gallo B, Donadieu J, et al. Late thyroid toxicity in 153 long-term survivors of allogeneic bone marrow transplantation for acute lymphoblastic leukaemia. Bone Marrow Transplant. 2005;35(10):991-995. 30. Ishiguro H, Yasuda Y, Tomita Y, et al. Longterm follow-up of thyroid function in patients who received bone marrow transplantation during childhood and adolescence. J Clin Endocrinol Metab. 2004;89 (12):5981-5986.

756

31. Vivanco M, Dalle JH, Alberti C, et al. Malignant and benign thyroid nodules after total body irradiation preceding hematopoietic cell transplantation during childhood. Eur J Endocrinol. 2012;167(2):225-233. 32. van Santen HM, Vulsma T, Dijkgraaf MG, et al. No damaging effect of chemotherapy in addition to radiotherapy on the thyroid axis in young adult survivors of childhood cancer. J Clin Endocrinol Metab. 2003;88(8): 3657-3663. 33. McLeod DS, Watters KF, Carpenter AD, Ladenson PW, Cooper DS, Ding EL. Thyrotropin and thyroid cancer diagnosis: a systematic review and dose-response metaanalysis. J Clin Endocrinol Metab. 2012;97 (8):2682-2692. 34. Surks MI, Ortiz E, Daniels GH, et al. Subclinical thyroid disease: scientific review and guidelines for diagnosis and management. JAMA. 2004;291(2):228-238. 35. Cohen A, Bekassy AN, Gaiero A, et al. Endocrinological late complications after hematopoietic SCT in children. Bone Marrow Transplant. 2008;41 (Suppl 2):S4348. 36. Schonfeld SJ, Lee C, Berrington de Gonzalez A. Medical exposure to radiation and thyroid cancer. Clin Oncol (R Coll Radiol). 2011;23(4):244-250. 37. Brignardello E, Corrias A, Isolato G, et al. Ultrasound screening for thyroid carcinoma in childhood cancer survivors: a case series. J Clin Endocrinol Metab. 2008;93(12):48404843. 38. Taylor AJ, Croft AP, Palace AM, et al. Risk of thyroid cancer in survivors of childhood cancer: results from the British Childhood

39.

40.

41.

42. 43.

44.

45.

Cancer Survivor Study. Int J Cancer. 2009;125(10):2400-2405. Veiga LH, Bhatti P, Ronckers CM, et al. Chemotherapy and thyroid cancer risk: a report from the Childhood Cancer Survivor Study. Cancer Epidemiol Biomarkers Prev. 2012;21(1):92-101. Acharya S, Sarafoglou K, LaQuaglia M, et al. Thyroid neoplasms after therapeutic radiation for malignancies during childhood or adolescence. Cancer. 2003;97(10):23972403. Gow KW, Lensing S, Hill DA, et al. Thyroid carcinoma presenting in childhood or after treatment of childhood malignancies: An institutional experience and review of the literature. J Pediatr Surg. 2003;38(11):15741580. Schlumberger MJ. Papillary and follicular thyroid carcinoma. N Engl J Med. 1998;338 (5):297-306. Ron E, Lubin JH, Shore RE, et al. Thyroid cancer after exposure to external radiation: a pooled analysis of seven studies. Radiat Res. 1995;141(3):259-277. Schneider AB, Bekerman C, Leland J, et al. Thyroid nodules in the follow-up of irradiated individuals: comparison of thyroid ultrasound with scanning and palpation. J Clin Endocrinol Metab. 1997;82(12):40204027. Haymart MR, Repplinger DJ, Leverson GE, et al. Higher serum thyroid stimulating hormone level in thyroid nodule patients is associated with greater risks of differentiated thyroid cancer and advanced tumor stage. J Clin Endocrinol Metab. 2008;93(3): 809-814.

haematologica | 2016; 101(6)

ARTICLE

Non-Hodgkin Lymphoma

Frequent CTLA4-CD28 gene fusion in diverse types of T-cell lymphoma

Hae Yong Yoo,1,2* Pora Kim,3* Won Seog Kim,2,4* Seung Ho Lee,1* Sangok Kim,3,5* So Young Kang,6 Hye Yoon Jang,7 Jong-Eun Lee,7 Jaesang Kim,8 Seok Jin Kim,2,4 Young Hyeh Ko,2,6† and Sanghyuk Lee3,5,8†

Department of Health Sciences and Technology, Samsung Advanced Institute for Health Sciences and Technology, Sungkyunkwan University; 2Samsung Biomedical Research Institute, Research Institute for Future Medicine, Samsung Medical Center; 3 Ewha Research Center for Systems Biology (ERCSB), Ewha Womans University; 4 Division of Hematology and Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine; 5Department of Bio-Information Science, Ewha Womans University; 6Department of Pathology, Samsung Medical Center, Sungkyunkwan University School of Medicine; 7DNA Link Inc.; and 8Department of Life Science, Ewha Womans University, Seoul, Korea

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

*These authors contributed equally to this work. † These authors are co-corresponding authors of this article.

Haematologica 2016 Volume 101(6):757-763

ABSTRACT

TLA4 and CD28 are co-regulatory receptors with opposite roles in Tcell signaling. By RNA sequencing, we identified a fusion between the two genes from partial gene duplication in a case of angioimmunoblastic T-cell lymphoma. The fusion gene, which codes for the extracellular domain of CTLA4 and the cytoplasmic region of CD28, is likely capable of transforming inhibitory signals into stimulatory signals for T-cell activation. Ectopic expression of the fusion transcript in Jurkat and H9 cells resulted in enhanced proliferation and AKT and ERK phosphorylation, indicating activation of downstream oncogenic pathways. To estimate the frequency of this gene fusion in mature T-cell lymphomas, we examined 115 T-cell lymphoma samples of diverse subtypes using reverse transcriptase polymerase chain reaction analysis and Sanger sequencing. We identified the fusion in 26 of 45 cases of angioimmunoblastic T-cell lymphomas (58%), nine of 39 peripheral T-cell lymphomas, not otherwise specified (23%), and nine of 31 extranodal NK/T cell lymphomas (29%). We further investigated the mutation status of 70 lymphoma-associated genes using ultra-deep targeted resequencing for 74 mature T-cell lymphoma samples. The mutational landscape we obtained suggests that T-cell lymphoma results from diverse combinations of multiple gene mutations. The CTLA4-CD28 gene fusion is likely a major contributor to the pathogenesis of T-cell lymphomas and represents a potential target for anti-CTLA4 cancer immunotherapy.

Correspondence: yhko310@skku.edu or sanghyuk@ewha.ac.kr

Received: November 10, 2015. Accepted: January 22, 2016. Pre-published: January 27, 2016. doi:10.3324/haematol.2015.139253

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/5/757

Introduction Peripheral T-cell lymphoma is a malignant neoplasm of mature T cells. Recent genomic studies have identified highly recurrent somatic mutations in TET2, DNMT3A, IDH2, and RHOA in diverse subtypes of mature T-cell lymphoma (TCL).1-5 However, the roles of these mutations in the regulation of T-cell signaling and oncogenesis have yet to be elucidated. Furthermore, none of the mutant genes has been clearly demonstrated to be a dominant oncogenic driver. It is, therefore, still necessary to identify additional driver mutations and dissect the interplay with T-cell signaling components. CTLA4 and CD28, members of the immunoglobulin superfamily that are expressed on the surfaces of T cells, are regulatory co-receptors of T-cell signaling. They play critical and opposite roles in maintaining balanced T-cell signaling and, haematologica | 2016; 101(6)

757

H.Y. Yoo et al.

thus, the proper level of immune activation.6,7 Perturbing this balance can result in a number of undesirable consequences such as autoimmunity, transplant rejection, or even malignant TCL.6 Accordingly, controlling T-cell signaling through these two co-receptors has been a key strategy for recent cancer immunotherapies including antiCTLA4 antibody therapy8,9 and chimeric antigen receptor T-cell therapy utilizing the intracellular signaling domain of CD28.10 In this study, we identified a fusion between CTLA4 and CD28 in a case of angioimmunoblastic TCL by whole transcriptome sequencing and analyzed the frequency of gene fusion in 117 cases of TCL. A functional study of the fusion gene indicated that fusion between CTLA4 and CD28 results in the activation of downstream oncogenic pathways. The mutational status of 70 lymphoma-associated genes was analyzed using ultra-deep targeted resequencing in 74 samples of mature TCL.

Methods Sample description The patients’ clinical information is summarized in Online Supplementary Table S2. Formalin-fixed, paraffin-embedded tumor tissue was used for Sanger sequencing of 115 TCL tumor samples and targeted deep sequencing of 74 TCL tumor samples. The QIAamp DNA Mini Kit (Qiagen) and RNeasy Mini Kit (Qiagen) were used for DNA and RNA extraction, respectively. All patients’ samples were obtained with informed consent in Samsung Medical Center, Seoul, Korea, and the study was approved by the Institutional Review Board in accordance with the Declaration of Helsinki.

Detection of the CTLA4-CD28 mutation For detection of the CTLA4-CD28 fusion transcript, we carried out reverse transcription (RT) with random hexamers and total RNA isolated from formalin-fixed, paraffin-embedded samples followed by polymerase chain reaction (PCR) amplification with the following oligonucleotide primers; Fusion_cRT1 F, Fusion_cRT1 R, Fusion_cRT2 F, Fusion_cRT2 R and Fusion_cRT3 F. For characterization of the genomic rearrangement, the CTLA4CD28 fusion gene was amplified by PCR using genomic DNA of the clinical specimen as the template. Amplification was performed with Herculase 2 Fusion DNA Polymerase (Agilent Technologies) and the primers; Fusion_G1 F, Fusion_G1 R and Fusion_G2 R. The resulting PCR products were analyzed by agarose gel electrophoresis and sequenced using two independent primer pairs.

Cell proliferation and cytokine assays Jurkat (human T-cell acute lymphoblastic leukemia) and H9 (human cutaneous T lymphocyte lymphoma) cells were transfected with a construct expressing the CTLA4-CD28 fusion protein. Cells expressing CTLA4 and the CTLA4-CD28 fusion were seeded in 96-well plates in triplicate at a density of 5×103 cells/well in 100 mL of RPMI-1640 medium containing 10% fetal bovine serum and antibiotics. For stimulation, a 96-well plate was coated with 5 mg/mL goat anti-mouse IgG (AbFrontier) or with 2 mg/mL antiCD3 (HIT3a, BD Pharmingen) or with the combination of 2 mg/mL anti-CD3 (HIT3a, BD Pharmingen), 2 mg/mL anti-CD28 (CD28.2 BD Pharmingen) or 5 mg/mL anti-CTLA4 (BNI3, BD Pharmingen) overnight or for 2 h at 37°C. After 48 h, cell proliferation was evaluated using Cell Counting Kit-8 (Dojindo), according to the manufacturer's instructions, and the absorbance value for each well was measured at 450 nm using a microplate reader 758

(Spectra Max 180, Molecular Devices). Each experiment was repeated three times. For the cytokine assay, cells were stimulated with 15 ng/mL phorbol myristate acetate and 290 ng/mL ionomycin (eBioscience) for 3 h. Cells were plated on 24-well plates coated with 5 mg/mL goat anti-mouse IgG (AbFrontier) or with 2 mg/mL anti-CD3 (HIT3a, BD Pharmingen) or with the combination of 2 mg/mL anti-CD3 (HIT3a, BD Pharmingen), 2 mg/mL antiCD28 (CD28.2 BD Pharmingen) or 5 mg/mL anti-CTLA4 (BNI3, BD Pharmingen). After 24 h, the supernatants were examined using Human IL-2 ELISA kits (Thermo Scientific) according to the manufacturer's instructions. Each experiment was repeated three times.

Selection of 70 frequently mutated genes in lymphoma and bioinformatic analysis for targeted sequencing Through ten genomics studies on B- and T-cell lymphomas, we have identified 62 genes that are frequently mutated in these neoplasms.1,3,5,11-18 We added three p53-related genes and five JAKSTAT signaling genes for targeted deep sequencing. The list of 70 target genes is provided in Online Supplementary Table S1. We further compiled somatic mutations in the target genes from the original references which collectively yielded 832 mutations in 62 genes (Online Supplementary File S1). Targeted sequencing was performed using 74 TCL samples. The bioinformatics analysis of the targeted deep sequencing data is described in the Online Supplementary Methods.

Results Identification and validation of the CTLA4-CD28 fusion gene The fusion transcript was initially predicted from analyses of RNA-Seq data from previously described TCL patients5 using the FusionScan program. Multiple reads that mapped to exon 3 of the CTLA4 gene in tandem with exon 4 of the CD28 gene were identified (Figure 1). The depth profile of the RNA-Seq data, reflecting the expression level of each exon, shows abrupt changes at the break points in both genes, which is consistent with the presence of fusion transcripts (Figure 1B). We verified the fusion via RT-PCR using two different primer sets in fusion-positive samples from patients and seven TCL cell lines (Figure 2A; Online Supplementary Figure S1). CTLA4-CD28 fusion cDNA was successfully amplified from the patients’ samples and two T lymphoblastoid cell lines, CEMC1-15 and CEMC7-14. Subsequent Sanger sequencing confirmed that the PCR products contained the fusion transcripts between exon 3 of the CTLA4 gene and exon 4 of the CD28 gene. To estimate the frequency of this gene fusion in TCL patients, we examined 115 TCL patients’ samples of diverse subtypes using RT-PCR and Sanger sequencing (Online Supplementary Table S2). We found that the CTLA4-CD28 fusion occurred in all the tested subtypes of TCL with an overall frequency of 38%. The fusion was observed more frequently in TCL of follicular helper T cell phenotype (TFH) than in non-TFH TCL. The fusion was identified in 26 of 45 (58%) patients with angioimmunoblastic TCL (AITL), nine of 39 (23%) patients with peripheral TCL not otherwise specified (PTCL-NOS), and nine of 31 (29%) patients with NK/T cell lymphoma (Online Supplementary Table S2; Online Supplementary Figure S2). Among PTCL-NOS, five of nine PTCL-NOS with a TFH phenotype (56%) and five of 16 non-TFH PTCL-NOS haematologica | 2016; 101(6)

CTLA4-CD28 gene fusion in TCL

Figure 1. Identification of the CTLA4-CD28 gene fusion. (A) Top, schematic diagram of the gene fusion; bottom, sequencing chromatogram. Numbers on the transcript indicate the nucleotide position of exons. (B) Alignment of RNA-Seq data from a fusion-positive patient. Read-depth plots indicate the depth coverage of aligned RNA-Seq reads. SP: signal peptide; TM: transmembrane region.

(31%) showed the CTLA4-CD28 fusion. The CTLA4CD28 fusion was not observed in blood samples from 50 healthy individuals.

Functional analyses of the CTLA4-CD28 fusion gene Notably, the predicted protein generated from this fusion gene features the extracellular and transmembrane domains of CTLA4 and the cytosolic signaling domain of CD28 (Figure 1A). A possible outcome is inappropriate activation of T-cell signaling. We, therefore, proceeded to analyze the effect of the CTLA4-CD28 fusion on cell proliferation and cytokine production in T cells. Jurkat (human T-cell acute lymphoblastic leukemia) and H9 (human cutaneous T lymphocyte lymphoma) cells transfected with a construct expressing the CTLA4-CD28 fusion protein proliferated at a rate approximately 30% higher than that of cells transfected with only the vector or the wild-type CTLA4 expression construct after stimulation with anti-CTLA4 antibody (Figure 2B; Online Supplementary Figure S3A). The surface expression levels of CTLA4 and the CTLA4-CD28 fusion were comparable in the Jurkat cell line (Online Supplementary Figure S4). Furthermore, the production of interleukin 2 (IL-2), the definitive marker of T-cell activation, was 6-fold greater (Figure 2C; Online Supplementary Figure S3B). These results indicate that the CTLA4-CD28 fusion protein likely mediates activating signals upon T-cell stimulation. haematologica | 2016; 101(6)

Next, we examined the phosphorylation of AKT and ERK1/2, which represents the activation of two critical pathways downstream of T-cell receptor signaling. As expected, CD28-mediated co-stimulation of T cells led to increased AKT and ERK phosphorylation (Figure 2D). Importantly, cells expressing the CTLA4-CD28 fusion also showed increased AKT and ERK phosphorylation relative to cells expressing wild-type CTLA4 upon stimulation with an anti-CTLA4 antibody. These results strongly suggest that the CTLA4-CD28 fusion could lead to constitutive T-cell activation by converting inhibitory signals into activating signals. Of note, Shin et al. developed the CTLA4-CD28 chimera for adoptive T-cell therapy of cancer, and studied its biological roles and therapeutic efficacy using mouse T cells.19 Importantly, they also demonstrated that the fusion gene delivered activating rather than inhibitory T-cell signals.

Genomic structure of the CTLA4-CD28 fusion gene The two genes are located in tandem 129 kbp apart on chromosome 2, with CD28 preceding CTLA4 on the same strand. The gene order is reversed in the CTLA4-CD28 fusion, raising a strong possibility that partial duplications have occurred (Figure 3A). Extrachromosomal amplification by episome formation, which was observed for the NUP214-ABL1 fusion in T-cell acute lymphoblastic leukemia,20 could not be ruled out in all cases, but was 759

H.Y. Yoo et al. A

Figure 2. Validation and functional analyses of the CTLA4-CD28 fusion gene. (A) Validation of the CTLA4-CD28 fusion transcript by RT-PCR using samples from patients and cell lines. Arrows indicate the approximate positions of oligonucleotide primers on the CTLA4-CD28 fusion transcript. PCR products were amplified from patient 1, CEMC1-15 cells and CEMC7-14 cells and validated by Sanger sequencing. β-actin was used as an internal control, and NTC indicates the no template control. Normal tonsil tissue and 293T cells were used as negative controls. (B) Jurkat cells expressing the CTLA4-CD28 fusion showed enhanced cell proliferation (*P<0.05 and **P<0.01 compared with cells expressing wild-type CTLA4) after co-stimulation with anti-CD3/CTLA4 antibodies. Each experiment was repeated three times with five replicates, and the data are expressed as the mean ± standard deviation. (C) Expression of the CTLA4-CD28 fusion enhanced interleukin 2 (IL-2) production after CTLA4 activation (**P<0.01 compared with cells expressing wild-type CTLA4). Jurkat cells were stimulated with phorbol myristate acetate/Ionomycin (P/I) without or with anti-CD3/CTLA4 antibodies to activate CTLA4. IL-2 was measured at three independent times, and the data are expressed as the mean ± standard deviation. (D) Expression of the CTLA4-CD28 fusion enhanced phosphorylation of AKT and ERK1/2 after CTLA4 activation with an anti-CTLA4 antibody.

shown not to have occurred in multiple cases studied by fluorescence in situ hybridization (Online Supplementary Figure S5). Consistent with our hypothesis, quantitative PCR analyses of genomic DNA demonstrated that the copy number gains for portions of the CD28 and CTLA4 genes represented in the fusion were significantly higher in the fusion-positive patients than those in the fusionnegative lymphoma patients (Figure 3B; Online Supplementary Figure S6). Next, we mapped the exact positions of the break points in the genomic DNA of fusion-positive patients. For the fusion-positive patient shown in Figures 1 and 2, we amplified a 2.5 kb genomic DNA fragment (Figure 3D). Subsequent Sanger sequencing revealed that the fragment contained 696 bp of CTLA4 intron 3, 1457 bp of CD28 intron 3, and 47 bp of intervening sequence from a LINE retrotransposon element that is normally located 104 kb downstream of the CTLA4 gene on chromosome 2 (Figure 3E; Online Supplementary Figure S7). We characterized another patient whose genome 426 bp of CTLA4 intron 3 was joined to 3185 bp of CD28 intron 3. Surprisingly, the most frequent arrangement was direct fusion of two 760

exons with no intron sequences included. We observed six other patients and two cell lines with such fusions. Despite the diversity in their genomic structures, all the genome arrangements we observed in patients’ samples would generate identical fusion transcripts and proteins. We also analyzed the mRNA expression levels of the CD28, CTLA4, and CTLA4-CD28 fusion genes in samples from patients with AITL using quantitative RT-PCR (Figure 3C). Fusion mRNA was expressed only in fusionpositive patients, and the expression level was comparable to that of CTLA4. The levels of CD28 and CTLA4 expression were approximately three and seven times higher in fusion-positive patients, respectively, which suggests the presence of a feed-forward circuit that amplifies the signaling from the CTLA4-CD28 fusion receptor. A stronger binding affinity of CTLA4, compared to CD28, to CD80 and CD86 ligands may have a role in amplifying signals in the CTLA4-CD28 fusion receptor.6 In brief, patients with the CTLA4-CD28 fusion not only have a copy number gain at the genomic level but also show elevated CD28 and CTLA4 expression, which is consistent with abnormal activation of the T-cell population. haematologica | 2016; 101(6)

CTLA4-CD28 gene fusion in TCL

Figure 3. Genomic structure of the CTLA4-CD28 fusion gene. (A) Schematic diagram of the gene duplication producing the fusion gene. (B) Copy-number analysis of CD28 and CTLA4 genes in samples from patients with AITL using quantitative -PCR (*P=0.009, **P=0.001). Copy number changes, estimated relative to that in peripheral blood cells from a normal individual, are shown in the box plot. The values are from two independent experiments. (C) Expression levels of CD28, CTLA4, and CTLA4-CD28 fusion transcripts in samples from patients with AITL using quantitative-PCR (*P=0.001, **P=0.0001, ***P=0.002). The values are from three independent experiments. (D) Structural analysis of genomic loci for a patient with the CTLA4-CD28 fusion. The PCR product (2.5 kb) amplified from genomic DNA of patient 1 was subsequently validated by Sanger sequencing. No product was amplified from the control normal tonsil cells and the no template control (NTC). (E) Schematic diagrams of the genomic structure of the CTLA4-CD28 fusion from eight patients and two cell lines. The arrows indicate the position of primers (Fusion_G1 F and Fusion_G1 R).

Mutational landscape of lymphoma-related genes from targeted deep sequencing Several somatic mutations have already been implicated as driver mutations in TCL, including the point mutations IDH2 R172K, DNMT3A R882H, RHOA G17V, and CD28 T195P and loss-of-function (stop-gain or frameshift) mutations in the TET2 gene.3-5 Determining the functional relationship between these mutations and the CTLA4-CD28 fusion would be of the utmost importance to understanding the molecular mechanisms underlying TCL. We selected 70 genes that had been reported to be frequently mutated in various types of T- and B-cell lymphoma (Online Supplementary Table S1).1,3-5,11-17 Targeted deep sequencing data for all exons of the 70 genes were produced with an average sequencing depth of 1,204X and 98.9% of target coverage using 2,965 primer pairs. The ultra-high depth of the sequencing was expected to reveal haematologica | 2016; 101(6)

variants of low frequency but with important functional roles. In total, 74 samples of tumor tissue were examined: these samples came from 29 patients with AITL, 15 with PTCL-NOS, 26 with extranodal NK/T cell lymphoma (ENKL), two with enteropathy-associated T-cell lymphoma, and two with anaplastic large cell lymphoma. The sequencing data were analyzed using our own computational pipeline, which was designed to identify somatic mutations in tumor cells without the normal control (Online Supplementary Figure S8). We identified 11 missense mutations in ten different genes and 24 TET2 loss-of-function mutations, including eight novel mutations (Figure 4; Online Supplementary Figure S8). Loss-offunction mutations in the TET2 gene occurred most frequently in all subtypes (47 patients, 64%). Excluding TET2 mutations with low allele frequency (<5%), which were observed mostly for R544X and R1404X (Online 761

H.Y. Yoo et al.

Figure 4. Mutational landscape of driver genes from targeted resequencing. Each column represents an independent patient. Novel mutations are indicated with asterisks. TCL subtypes are color-coded above the main window, and bars on the right side show the relative proportions of the indicated mutation among TCL subtypes. The CTLA4-CD28 row indicates the presence of fusion gene based on RT-PCR and Sanger sequencing. TET2 mutations of all locations are shown in top rows, with the color of the highest allele frequency for patients with multiple TET2 mutations (position-wise mutation plot provided in Online Supplementary Figure S9). The cumulative numbers of patients positive for each mutation are indicated on the right.

Supplementary Figure S9), the TET2 mutation rate was only 30% (22 patients), with a significant concentration in the AITL subtype (15 patients, P=0.008). Many mutations were observed at low frequencies, below 20%, demonstrating the power of ultra-deep sequencing. These low frequency mutations are likely due to the heterogeneity of the tumor cells as well as the presence of normal or stromal cells. The RHOA mutation was mostly observed in the AITL subtype (21 of 29 AITL patients; 72%), with only two mutations observed in the 15 PTCL-NOS patients (13%). We identified eight low frequency mutations (allele frequency <5%) in addition to 12 (41% of AITL patients) high frequency mutations, which agrees well with previous reports on AITL based on Sanger sequencing.3-5 Among other recurrent mutations, IDH2 R172K and DNMT3A R882H mutations were found in seven and five patients, respectively. These latter two mutations have been described as ageing-related initiating mutations that are associated with the clonal expansion of pre-lymphoma stem cells and can be detected in the blood of elderly individuals without apparent hematologic malignancies.21-23 The frequency of the CTLA4-CD28 fusion was 30% in this targeted deep-sequencing analysis (22 out of 74 patients), appearing in all subtypes tested. Sixteen of the 22 fusion-positive patients had additional mutations. In 11 AITL patients with the CTLA4-CD28 fusion, nine (82%) and ten (91%) patients also harbored the TET2 mutation and RHOA mutation, respectively. The CTLA4-CD28 fusion was not found in four patients with a CD28 T195P mutation which also led to up-regulation of T-cell receptor signaling.24 indicating the relevance of CD28 in the process of lymphomagenesis. In ENKL, most of the fusion cases were devoid of additional mutations. In fact, a substantially high proportion of cases did not have any mutations among the tested genes, and even TET2 mutation-positive cases mostly did not 762

have other mutations. It is in this regard that a novel recurrent TCF3 G302S is notable, especially for the ENKL subtype. The transcription factor TCF3 (E2A) is required for B and T lymphocyte development, and the significance of mutations in TCF3 and its negative regulator ID3 has recently been highlighted in Burkitt lymphoma.14,15,17 Genomic examination of additional ENKL patients should be carried out to substantiate the TCF3 G302S mutation as a marker and a potential driver of the ENKL subtype.

Discussion Here, we report that the CTLA4-CD28 fusion gene is a novel, high-frequency mutation for diverse types of TCL. Two other groups have recently reported the identification of the CTLA4-CD28 fusion gene in SĂŠzary syndrome, an aggressive rare variant of cutaneous T-cell lymphoma.25,26 That this mutation is not limited to SĂŠzary syndrome but is found in a broad range of TCL types with an overall frequency of 30% and typically in combination with other mutations should be of significance. We also provide the overall mutational landscape for TCL which indicates that the CTLA4-CD28 fusion represents one of recurrent genetic events for full blown neoplastic transformation. Targeted deep sequencing analyses for 70 genes implicated in TCL showed that a large majority of patients had more than one mutated gene. Although not all of the mutations have been mechanistically demonstrated to be oncogenic, we suggest that multiple mutational events, including those described in this study, are required for the full development of TCL. It has been proposed and partly demonstrated that TCL and myeloid leukemia feature the age-related accumulation of premalignant mutations in DNMT3A, TET2, JAK2, and GNAS which are associated with subsequent clonal hematopoietic expansion.21-23 Similarly in B-cell lymphoma, it has been shown that circulating B cells bearing haematologica | 2016; 101(6)

CTLA4-CD28 gene fusion in TCL

BCL2 translocations do not cause follicular lymphoma per se but can evolve into overt follicular lymphoma with additional mutations.27 Our data indicate that the CTLA4-CD28 fusion disrupts cellular homeostasis via inappropriate activation of T-cell signaling. Given that such dysregulation likely lies at the core of oncogenesis, it is possible that the CTLA4-CD28 fusion provides a target for potential immunotherapy. In fact, Sekulic et al. reported a one-patient trial in a female who had suffered from Sézary syndrome for 8 years.25 Administration of the CTLA4-blocking antibody ipilimumab8 produced dramatic initial responses for first 2 months, but the disease subsequently progressed rapidly resulting in death 3 months after the last dose. Notwithstanding the tragic outcome, it was shown that an immunotherapy targeting the fusion gene is viable in principle, and with improved dosage and timing, the response may be more durable. Targeted therapy and immunotherapy as parts of a combination therapy regimen is an emerging paradigm of cancer treatment. Thus, the identification of the frequent CTLA4-CD28 fusion gene will pro-

References 1. Cairns RA, Iqbal J, Lemonnier F, et al. IDH2 mutations are frequent in angioimmunoblastic T-cell lymphoma. Blood. 2012;119(8):1901-1903. 2. Lemonnier F, Couronné L, Parrens M, et al. Recurrent TET2 mutations in peripheral Tcell lymphomas correlate with TFH-like features and adverse clinical parameters. Blood. 2012;120(7):1466-1469. 3. Palomero T, Couronné L, Khiabanian H, et al. Recurrent mutations in epigenetic regulators, RHOA and FYN kinase in peripheral T cell lymphomas. Nat Genet. 2014;46(2):166-170. 4. Sakata-Yanagimoto M, Enami T, Yoshida K, et al. Somatic RHOA mutation in angioimmunoblastic T cell lymphoma. Nat Genet. 2014;46(2):171-175. 5. Yoo HY, Sung MK, Lee SH, et al. A recurrent inactivating mutation in RHOA GTPase in angioimmunoblastic T cell lymphoma. Nat Genet. 2014;46(4):371-375. 6. Rudd CE, Taylor A, Schneider H. CD28 and CTLA-4 coreceptor expression and signal transduction. Immunol Rev. 2009;229 (1):12-26. 7. Kuehn HS, Ouyang W, Lo B, et al. Immune dysregulation in human subjects with heterozygous germline mutations in CTLA4. Science. 2014;345(6204):162316-162327. 8 Leach DR, Krummel MF, Allison JP. Enhancement of antitumor immunity by CTLA-4 blockade. Science. 1996;271(5256): 1734-1736. 9. Hodi FS, O'Day SJ, McDermott DF, et al. Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med. 2010;363(8):711-723.

haematologica | 2016; 101(6)

vide a new therapeutic opportunity for TCL patients with this fusion, and elucidation of the exact mechanism by which the CTLA4-CD28 fusion interacts with other mutations should provide further insights into the molecular nature of TCL development, as well as new strategies for curbing this disease. Acknowledgments This work was supported by grants from the Technology Innovation Program of the Ministry of Trade, Industry and Energy, Republic of Korea (10050154 to SL), the National Research Foundation of Korea (NRF-2014M3C9A3065221, NRF2012M3A9D1054744 to SL, NRF-2015K1A4A3047851 to SL and JK), the Samsung grants (SM01132671 and SMO1150911 to YHK), and a grant from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (HI14C3414 to YHK and HYY, HI14C2331 to HYY). This study was also supported by the Ministry of Health & Welfare, Republic of Korea (A120262 to JK) and by the Ewha Womans University scholarship of 2015.

10. Johnson LA, Morgan RA, Dudley ME, et al. Gene therapy with human and mouse T-cell receptors mediates cancer regression and targets normal tissues expressing cognate antigen. Blood. 2009;114(3):535-546. 11. Gunawardana J, Gunawardana J, Chan FC, et al. Recurrent somatic mutations of PTPN1 in primary mediastinal B cell lymphoma and Hodgkin lymphoma. Nat Genet. 2014;46(4): 329-335. 12. Lohr JG, Stojanov P, Lawrence MS, et al. Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing. Proc Natl Acad Sci USA 2012;109(10):3879-3884. 13. 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. 14. Richter J, Schlesner M, Hoffmann S, et al. Recurrent mutation of the ID3 gene in Burkitt lymphoma identified by integrated genome, exome and transcriptome sequencing. Nat Genet. 2012;44(12):1316-1320. 15. Schmitz R, Young RM, Ceribelli M, et al. Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature. 2012;490(7418):116-120. 16. Zhang J, Grubor V, Love CL, et al. Genetic heterogeneity of diffuse large B-cell lymphoma. Proc Natl Acad Sci USA. 2013; 110(4):1398-1403. 17. Love C, Sun Z, Jima D, et al. The genetic landscape of mutations in Burkitt lymphoma. Nat Genet. 2012;44(12):1321-1325. 18. Couronné L, Bastard C, Bernard OA. TET2 and DNMT3A mutations in human T-cell lymphoma. N Engl J Med. 2012;366(1):9596.

19 Shin JH, Park HB, Oh YM, et al. Positive conversion of negative signaling of CTLA4 potentiates antitumor efficacy of adoptive Tcell therapy in murine tumor models. Blood. 2012;119(24):5678-5687. 20 Graux C, Cools J, Melotte C, et al. Fusion of NUP214 to ABL1 on amplified episomes in T-cell acute lymphoblastic leukemia. Nat Genet. 2004;36(10):1084-1089. 21. Genovese G, Kähler 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. 22. Jaiswal S, Fontanillas P, Flannick J, et al. Agerelated clonal hematopoiesis associated with adverse outcomes. N Engl J Med. 2014;371 (26):2488-2498. 23. Xie M, Lu C, Wang J, et al. Age-related mutations associated with clonal hematopoietic expansion and malignancies. Nat Med. 2014;20(12):1472-1478. 24. Lee SH, Kim JS, Kim J, et al. A highly recurrent novel missense mutation in CD28 among angioimmunoblastic T-cell lymphoma patients. Haematologica. 2015;100 (12):e505-507. 25. Sekulic A, Liang WS, Tembe W, et al. Personalized treatment of Sezary syndrome by targeting a novel CTLA4:CD28 fusion. Mol Genet Genomic Med. 2015;3(2):130-136. 26. Ungewickell A, Bhaduri A, Rios E, et al. Genomic analysis of mycosis fungoides and Sezary syndrome identifies recurrent alterations in TNFR2. Nat Genet. 2015;47(9): 1056-1060. 27. Correia C, Schneider PA, Dai H, et al. BCL2 mutations are associated with increased risk of transformation and shortened survival in follicular lymphoma. Blood. 2015;125(4): 658-667.

763

ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Stem Cell Transplantation

Ferrata Storti Foundation

Haematologica 2016 Volume 101(6):764-772

Improved graft-versus-host disease-free, relapse-free survival associated with bone marrow as the stem cell source in adults Rohtesh S. Mehta,1 Regis Peffault de Latour,2 Todd E DeFor,1 Marie Robin,2 Aleksandr Lazaryan,1 Ali茅nor Xhaard,2 Nelli Bejanyan,1 Flore Sicre de Fontbrune,2 Mukta Arora,1 Claudio G. Brunstein,1 Bruce R. Blazar,3 Daniel J. Weisdorf,1 Margaret L. MacMillan,3 Gerard Socie,2* and Shernan G. Holtan1*

Hematology, Oncology, Transplantation, Department of Medicine, University of Minnesota, Minneapolis, MN, USA; 2Haematology, H么pital Saint-Louis, Paris, France; and 3Blood and Marrow Transplant Program, Department of Pediatrics, University of Minnesota, Minneapolis, MN, USA 1

*GS and SGH contributed equally to this work.

ABSTRACT

Correspondence: rsmehta@umn.edu

Received: November 9, 2015. Accepted: March 22, 2016. Pre-published: April 1, 2016. doi:10.3324/haematol.2015.138990

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/764

漏2016 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights reserved to Ferrata Storti Foundation. Copies of articles are allowed for personal or internal use. A permission in writing by the publisher is required for any other use.

764

e previously reported that bone marrow grafts from matched sibling donors resulted in best graft-versus-host disease-free, relapse-free survival at 1-year post allogeneic hematopoietic cell transplantation. However, pediatric patients comprised the majority of bone marrow graft recipients in that study. To better define this outcome in adults and pediatric patients at 1- and 2-years post- allogeneic hematopoietic cell transplantation, we pooled data from the University of Minnesota and the H么pital Saint-Louis in Paris, France (n=1901). Graft-versus-host disease-free, relapse-free survival was defined as the absence of grade III-IV acute graft-versus-host disease, chronic graft-versus-host disease (requiring systemic therapy or extensive stage), relapse and death. In adults, bone marrow from matched sibling donors (n=123) had best graft-versus-host disease-free, relapse-free survival at 1- and 2years, compared with peripheral blood stem cell from matched sibling donors (n=540) or other graft/donor types. In multivariate analysis, peripheral blood stem cells from matched sibling donors resulted in a 50% increased risk of events contributing to graft-versus-host diseasefree, relapse-free survival at 1- and 2-years than bone marrow from matched sibling donors. With limited numbers of peripheral blood stem cell grafts in pediatric patients (n=12), graft-versus-host disease-free, relapse-free survival did not differ between bone marrow and peripheral blood stem cell graft from any donor. While not all patients have a matched sibling donor, graft-versus-host disease-free, relapse-free survival may be improved by the preferential use of bone marrow for adults with malignant diseases. Alternatively, novel graft-versus-host disease prophylaxis regimens are needed to substantially impact graft-versus-host disease-free, relapse-free survival with the use of peripheral blood stem cell. Introduction Disease relapse and graft-versus-host disease (GVHD) impact the length and quality of life after allogeneic hematopoietic cell transplantation (HCT). Yet, in clinical trials, outcomes of allogeneic HCT are often defined by isolated events of GVHD, relapse, and mortality. The Blood and Marrow Transplant Clinical Trials Network (BMT CTN) incorporated these major causes of morbidity and mortality into a composite endpoint named GVHD-free, relapse-free survival (GRFS) - defined as the absence of grade III-IV acute GVHD, chronic GVHD requiring systemic immunosuppression, relapse, or death. We previously reported 1-year GRFS of 31% in 907 haematologica | 2016; 101(6)

BM graft is associated with best GRFS after allogeneic HCT

consecutive patients who underwent allogeneic HCT at the University of Minnesota (UMN) from 2000-2012.1 In that study, bone marrow (BM) grafts from a human leucocyte antigen (HLA)-matched sibling donor (MSD) resulted in the best GRFS (51%) as compared with peripheral blood stem cell (PBSC) from MSD (25%), BM from matched unrelated donors (URD) (32%), or umbilical cord blood (UCB) transplants (31%). However, there were limited numbers of adult patients (age >21) who received BM from MSD, and thus specific conclusions regarding the impact of graft source in adults could not be drawn. To address that question in a larger dataset, we now report the outcomes of 1,901 patients pooled from the UMN and the H么pital Saint-Louis in Paris (Saint-Louis). Moreover, in our previous study, pediatric recipients of BM from MSD served as the reference group to which all other groups (including adult patients) were compared. In the current study, we report GRFS at 1- and 2-years post- transplantation in pediatric and adult patients independently, as their GRFS differs widely. With our pooled data, we show that allogeneic HCT with BM from a MSD results in the best GRFS in adult patients at both 1- and 2-years. In contrast, pediatric patients have similar GRFS with PBSC or BM graft from any donor, but the use of UCB was associated with inferior GRFS.

Methods Our primary objective was to compare GRFS among different donor/graft sources at 1- and 2-years. The secondary objectives were (a) to define the distribution of GRFS events at 1-year among different donor/graft sources and (b) to define disease free survival (DFS) and overall survival (OS) at 1- and 2-years post-HCT.

Figure 1. Kaplan-Meier estimates of GRFS, DFS and OS among (A) adults and (B) pediatric patients.

Patient population and definitions We included all consecutive patients who underwent first allogeneic HCT for hematological malignancy at the UMN from 2000 to 2013 (n=995) or Saint-Louis from 2000 to 2012 (n=906). Graft sources included PBSC, BM and UCB, while donors were MSD or matched URD (6/6-HLA matched), other related donors (5-6/6HLA matched), or mismatched URD (5/6-HLA matched). Patients with prior allogeneic HCT, recipients of syngeneic HCT, haploidentical HCT, experimental cellular therapies or graft manipulation techniques, and those with non-malignant diseases were excluded. GRFS events were defined at 1- and 2- years of HCT as the first occurrence of grade III-IV acute GVHD, extensive or systemic chronic GVHD requiring therapy, relapse, or death. Disease risk at the time of transplantation was classified into standard-risk or high-risk based on the American Society for Blood and Marrow Transplantation (ASBMT) 2006 risk scoring schema.2 Acute myeloid leukemia (AML) or acute lymphoblastic leukemia (ALL) in first or second complete remission, chronic myeloid leukemia (CML) in first chronic phase, Hodgkin or non-Hodgkin lymphoma in complete remission or chemotherapy-sensitive partial remission, chronic lymphocytic leukemia (CLL) in first remission were defined as standard-risk; all other diseases were classified as highrisk. DFS was defined as the time from transplantation to relapse of the underlying malignancy or death, and OS was defined as the time from transplantation to death. All HCT and data collection protocols were reviewed and approved by the University of Minnesota Institutional Review Board.

test was used to compare characteristics across centers for continuous factors and the Chi-square test was used for categorical variables. Kaplan-Meier curves were used to estimate the probability of GFRS at 1- and 2-years post-HCT.3 The log-rank test was used to complete the comparisons. Direct adjusted survival curves were also calculated based on a stratified Cox model.4 The Cox regression model was used to examine the independent effect of factors on GFRS.5 Proportional hazards were checked using martingale residuals.6 Transplant centers violated the proportional hazards assumption, and therefore models were stratified by center. Other factors which were examined included year of HCT (2000-2007 versus 2008-2013, based on natural cut-off point), age, gender (male versus female), diagnosis, recipient cytomegalovirus (CMV) serostatus, type of conditioning (myeloablative versus reduced intensity conditioning (RIC), with or without anti-thymocyte globulin (ATG)), GVHD prophylaxis, donor type (MSD versus other related donor versus matched URD versus mismatched URD versus UCB), graft type (BM versus PBSC versus UCB) and disease risk (standard versus high-risk). All analyses were performed using SAS version 9.3 (SAS Institute, Cary, NC, USA). All reported P-values are 2-sided.

Results Patient and treatment characteristics

Statistical methods Data were analyzed independently for pediatric patients (age <21 years) and adults (age>21 years). The Wilcoxon signed-rank haematologica | 2016; 101(6)

A total of 1901 patients, including 1466 adults were analyzed, of which 456 received BM grafts (Table 1). Overall, sixty percent of patients were males, 56% were CMV 765

R.S. Mehta et al. Table 1. Baseline patient and treatment characteristics.

Variable

Strata

N Age (Years) Median (range) Year of Transplant

243

Conditioning GvHD Prophylaxis

Recipient CMV+ Diagnosis

Disease Risk: High

192

University of Minnesota

Adults (Age ≥21 years) Hôpital Saint-Louis, Paris

752

714

50 (21-75) 437 (58%) 315 (42%) 441 (59%) 256 (34%) 292 (39%) 21 (3%) 69 (9%) 5 (1%) 365 (49%) 69 (9%) 317 (42%) 366 (49%) 310 (41%) 442 (59%) 529 (70%) 187 (25%) 36 (5%) 424 (56%) 122 (16%) 306 (41%) 180 (24%) 125 (17%) 19 (3%) 285 (38%)

44 (21-68) 379 (53%) 335 (47%) 417 (58%) 159 (22%) 371 (52%) 12 (2%) 155 (22%) 129 (18%) 47 (7%) 198 (28%) 469 (66%) 47 (7%) 394 (55%) 320 (45%) 309 (43%) 351 (49%) 54 (8%) 405 (57%) 124 (17%) 262 (37%) 243 (34%) 81 (11%) 4 (1%) 289 (41%)

<0.01 2000-2007 2008-2013

Patient Gender: Male Female Donor to Male Donor Type

Source

Pediatric patients (Age <21 years) University Hôpital P* of Minnesota Saint-Louis, Paris

MSD Other related donor Matched URD Mismatched URD UCB BM PBSC UCB Myeloablative RIC CsA/MMF CsA/MTX Other/unknown

ALL AML MDS/MPN/CML NHL/HL/CLL/Myeloma Other leukemia

11 (<1-20) 151 (62%) 92 (38%) 154 (63%) 93 (38%) 59 (24%) 14 (6%) 11 (5%) 1 (0%) 158 (65%) 70 (29%) 14 (6%) 159 (65%) 232 (95%) 11 (5%) 128 (53%) 79 (33%) 36 (15%) 139 (57%) 123 (51%) 81 (33%) 24 (10%) 15 (6%) 42 (17%)

15 (3-20) 156 (81%) 36 (19%) 121 (63%) 55 (29%) 77 (40%) 8 (4%) 46 (24%) 29 (15%) 32 (17%) 119 (62%) 41 (21%) 32 (17%) 186 (96%) 6 (4%) 9 (5%) 143 (75%) 40 (21%) 101 (53%) 89 (46%) 63 (33%) 32 (17%) 8 (4%) 32 (17%)

<0.01 0.94 0.04 <0.01

<0.01

0.34 0.17

0.86

<0.01 0.05 0.93 <0.01 <0.01

<0.01

<0.01 <0.01

0.90 <0.01

0.31

*P-value for between-treatment comparisons. Continuous variables were analyzed by general Wilcoxon test. Categorical variables were analyzed by χ2. ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; BM: bone marrow; CLL: chronic lymphocytic leukemia; CML: chronic myeloid leukemia; CMV: cytomegalovirus; CsA: cyclosporine; GvHD: graft versus host disease; HL: Hodgkin lymphoma; RIC: reduced intensity conditioning; MDS: myelodysplastic syndrome; MMF: mycophenolate mofetil; MPN: myeloproliferative neoplasm; MSD: matched sibling donor; MTX: methotrexate; NHL: non-Hodgkin lymphoma; PBSC: peripheral blood stem cells; UCB: umbilical cord blood; URD: unrelated donor.

seropositive, 61% had acute leukemia and 63% had standard-risk disease. As expected, there were significant differences in patient and treatment characteristics between the centers. More than half of the patients at the UMN received UCB as the graft source (53%), while PBSC was the most frequent graft source (56%) at Saint-Louis. Correspondingly, the most common donor type at SaintLouis was MSD (49%), which constituted 35% of donor types at the UMN. Myeloablative conditioning regimens were used in a majority of patients at both the institutions, although more commonly at Saint-Louis (64%) than at the UMN (55%). Correlating with the differences in donor/graft preferences, the choice of GVHD prophylaxis also differed between the centers. Two-thirds of patients at the UMN received mycophenolate mofetil with cyclosporine for GVHD prophylaxis, while methotrexate and cyclosporine (55%) was used most commonly at Saint-Louis. The median overall follow-up was 6.2 years (range, 0.3-14.4 years) and was similar in both sites.

Estimates of GRFS, DFS and OS in adults at one- and two-years post-transplantation The unadjusted Kaplan-Meier estimate of GRFS was 34% (95% confidence interval (C.I.) 31-36%) at 1-year and 27% (95% C.I. 25-29%) at 2-years. The estimates of 766

DFS were 54% (95% C.I. 51-57%) and 46% (95% C.I. 4349%) at 1-and 2-years, respectively, while those of OS were 62% (95% C.I. 60-65%) and 53% (95% C.I. 5156%) at 1-and 2-years, respectively (Figure 1A; Online Supplementary Table S1A and S2A). In univariate analysis, BM from MSD resulted in best GRFS at both 1-year (57%, 95% C.I. 48-65%) and at 2years (49%, 95% C.I. 40-57%) compared with any other donor/graft source, P<0.001 (Table 2A). After adjusting for age, conditioning regimens, diagnosis, disease risk and recipient CMV serostatus in multiple regression analysis stratified by center, donor/graft source was found to be an independent predictor of GRFS (Table 3A). Figure 2 shows adjusted survival and adjusted GRFS-defining events in adults with BM versus PBSC graft from MSD versus UCB grafts. The use of PBSC from either MSD, matched URD or mismatched URD, and the use of UCB grafts resulted in a 50-120% increased risk of GRFS events at both 1- and 2years compared with BM from MSD. However, BM from matched URD was associated with similar GRFS at 1- and 2-years as BM from MSD. The use of RIC was associated with a 20-30% lower incidence of GRFS events at 1- and 2-years compared with myeloablative regimens. The use of ATG with myeloablative regimens did not impact GRFS. Similarly, GRFS was not affected by age, after haematologica | 2016; 101(6)

BM graft is associated with best GRFS after allogeneic HCT

Table 2A. Univariate estimates of GRFS in adults (Ageâ&#x2030;Ľ21 years).

Factor

Strata

Overall Conditioning

Total N

MA w/o ATG MA w/ ATG RIC with ATG RIC without ATG Donor/Graft Type BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB Gender Male Female Diagnosis ALL AML MDS/MPN/CML NHL/HL/ CLL/MM Other Malignancy Disease Risk Standard-Risk High-Risk Recipient CMV SerostatusNegative Positive GvHD prophylaxis CsA/MMF CsA/MTX Tac +/- Sirolimus Other/unknown Year of 2000-2007 transplantation 2008-2013

1466 625 79 297 465 123 540 33 104 120 30 104 412 858 608 246 568 423 206 23 892 574 637 829 838 538 33 57 816 650

At 1-year At 2-years Survival Estimate Log Rank Survival Estimate Event N Estimate 95% CI P Event N Estimate 95% CI 973 413 45 184 331 53 372 24 63 78 18 65 300 584 389 174 352 286 141 20 572 401 406 567 585 334 22 32 560 413

34% 34% 43% 38% 29% 57% 31% 29% 40% 36% 43% 38% 27% 32% 36% 29% 38% 32% 32% 13% 36% 30% 36% 32% 30% 38% 33% 44% 31% 36%

31- 36% 34- 37% 32- 54% 33- 43% 25- 33% 48- 65% 27- 35% 14- 44% 31- 49% 26-44% 26- 60% 29-47% 23- 31% 29- 35% 32- 40% 24- 35% 34- 42% 28- 37% 25- 38% 3- 30% 33- 39% 26- 34% 32- 40% 28- 35% 27- 33% 34- 42% 18- 49% 31- 56% 28- 35% 33- 40%

0.02

<0.01

0.15 0.02

<0.01 0.07 <0.01

0.13

1065 449 53 207 356 63 411 25 70 83 20 69 324 644 421 194 382 311 156 22 626 439 447 618 637 365 22 41 618 447

27% 28% 33% 29% 23% 49% 23% 24% 32% 30% 33% 34% 21% 24% 31% 20% 32% 26% 24% 4% 29% 23% 29% 25% 24% 32% 33% 28% 24% 31%

25- 29% 25- 31% 23- 43% 24- 34% 20- 26% 40- 57% 20- 27% 11- 40% 23- 41% 22- 39% 18- 50% 25- 43% 17- 25% 22- 27% 27- 34% 15- 26% 29- 36% 22- 30% 18- 30% 0- 18% 26- 33% 20- 27% 26- 33% 22- 28% 21- 27% 28- 36% 18- 49% 17- 40% 21- 27% 27- 34%

Log Rank P 0.04

<0.01

0.04 <0.01

<0.01 0.08 <0.01

0.05

ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; ATG: anti-thymocyte globulin; BM: bone marrow; CI: confidence interval; CLL: chronic lymphocytic leukemia; CML: chronic myeloid leukemia; CMV: cytomegalovirus; CsA: cyclosporine; GvHD: graft versus host disease; HL: Hodgkin lymphoma; MA: myeloablative; RIC, reduced intensity conditioning; MDS: myelodysplastic syndrome; MM: multiple myeloma; MMF: mycophenolate mofetil; MPN: myeloproliferative neoplasm; MSD: matched sibling donor; MTX: methotrexate; NH: non-Hodgkin lymphoma; PBSC: peripheral blood stem cells; Tac: tacrolimus; UCB: umbilical cord blood; URD: unrelated donor.

adjusting for other variables. Patients with high-risk disease had a 30% higher risk of GRFS events at 1- and 2years compared to those with standard-risk disease. Recipient CMV serostatus did not impact GRFS.

Distribution of GRFS events in adults at one-year post-transplantation The distribution of GRFS events at 1-year varied significantly among different graft types and donor sources, P<0.01. In the setting of MSD transplantation, chronic GVHD was the most frequent GRFS event in recipients of either PBSC (35%) or BM (30%) grafts. In contrast, acute GVHD was the most common GRFS event in recipients of PBSC (49%) or BM (33%) from matched URD and PBSC (49%) from mismatched URD. Acute and chronic GVHD were common (28% each) in recipients of BM from mismatched URD. Death (29%) and relapse (27%) contributed to a majority of GRFS events after UCB transplantation while GVHD events were less frequent. Events occurring after the GRFS-defining events also varied by donor/graft source (Online Supplementary Figures S1 and S2). Table 4A depicts one-year cumulative incidences of grade III-IV acute GVHD, chronic GVHD, relapse and haematologica | 2016; 101(6)

death by donor/graft sources. Analyzing by conditioning regimens, deaths were more common in recipients of myeloablative compared with RIC regimens (23% versus 16%), while relapse (21% versus 34%) accounted for more GRFS events in those who received RIC regimens. Major causes of death were infections (30%) in recipients of BM from MSD, disease relapse (46%) in PBSC from MSD, acute GVHD in either PBSC (53%) or BM (31%) from matched URD, and disease relapse (33%) and infection (22%) in UCB recipients (Online Supplementary Figure S3A).

Estimates of GRFS, DFS and OS in pediatric patients at one- and two-years post-transplantation Pediatric patients had considerably superior estimated GRFS at 1-year (52%, 95% C.I. 48-57%) and at 2-years (46%, 95% C.I. 41-51%) compared with adults. Similarly, the estimates of DFS at 1-year (62%, 95% C.I. 57- 66%) and 2-years (56%, 95% C.I. 51- 61%) and OS at 1-year (69%, 95% C.I. 65- 74%) and at 2-years (63%, 95% C.I. 58- 68%) were higher than those in adults (Figure 1B; Online Supplementary Table S1B and S2B). 767

R.S. Mehta et al. Table 2B. Univariate estimates of GRFS in pediatric patients (Age <21 years).

Factor Overall Conditioning

Strata

Total N

MA RIC with ATG RIC without ATG Donor/graft type BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB Gender Male Female Diagnosis ALL AML MDS/MPN/CML NHL/HL/ CLL/MM Disease risk Standard-Risk High-Risk Recipient CMV serostatus Negative Positive GvHD prophylaxis CsA/MMF CsA/MTX Other/unknown Year of transplantation 2000-2007 2008-2013

435 418 5 12 124 12 22 34 23 18 12 190 275 160 212 144 56 23 361 74 195 240 137 222 76 307 128

At 1-year Survival Estimate Event N Estimate 95% CI 207 192 4 11 46 4 8 19 14 10 6 100 129 78 101 69 24 13 166 41 85 122 70 93 44 146 61

52% 54% 20% 8% 62% 72% 63% 45% 37% 45% 50% 49% 53% 51% 52% 52% 57% 43% 54% 45% 56% 49% 49% 58% 42% 52% 52%

48- 57% 49- 59% 1- 58% 1- 31% 54- 70% 48- 96% 43- 83% 29- 61% 17-57% 22- 68% 22-78% 42- 56% 47- 59% 43- 59% 45- 59% 44- 60% 43- 69% 23- 62% 49- 59% 33- 55% 49- 63% 43- 55% 40- 57% 51- 64% 31- 53% 47- 58% 43- 61%

P <0.01

<0.01

0.68 0.73

0.09 0.15 0.02

0.93

At 2-years Log Rank Survival Estimate Log Rank Event N Estimate 95% CI P 233 218 4 11 51 6 10 20 15 14 7 110 151 82 114 77 28 14 188 45 94 139 75 106 52 166 67

46% 48% 20% 8% 59% 50% 55% 41% 34% 22% 42% 42% 45% 49% 46% 46% 50% 39% 48% 39% 52% 42% 45% 52% 32% 46% 47%

41- 51% 43- 52% 1- 58% 1- 31% 49- 67% 21- 74% 32- 72% 25- 57% 16- 53% 7- 43% 15- 67% 35- 49% 39- 51% 41- 56% 39- 53% 38- 54% 36- 62% 20- 58% 42- 53% 28- 50% 44- 58% 35- 48% 37- 53% 45- 58% 22- 42% 40- 51% 38- 56%

<0.01

0.03

0.70 0.79

0.10 0.06 <0.01

0.93

ALL: acute lymphoblastic leukemia; AML: acute myeloid leukemia; ATG: anti-thymocyte globulin; BM: bone marrow; CI: confidence interval; CLL: chronic lymphocytic leukemia; CML: chronic myeloid leukemia; CMV: cytomegalovirus; CsA: cyclosporine; GvHD: graft versus host disease; HL: Hodgkin lymphoma; MA: myeloablative; RIC, reduced intensity conditioning; MDS: myelodysplastic syndrome; MM: multiple myeloma; MMF: mycophenolate mofetil; MPN: myeloproliferative neoplasm; MSD: matched sibling donor; MTX: methotrexate; MUD: matched unrelated donor; NHL: non-Hodgkin lymphoma; PBSC: peripheral blood stem cells; UCB: umbilical cord blood; URD: unrelated donor.

Outcomes after BM-MSD versus PBSC-MSD versus UCB HCT in adult patients

Figure 2. Adjusted GRFS, DFS, OS, grade III-IV acute GVHD, chronic GVHD and relapse among adults with BM graft from matched sibling donor versus PBSC from matched sibling donor versus UCB graft.

768

haematologica | 2016; 101(6)

BM graft is associated with best GRFS after allogeneic HCT

It is noteworthy that only 12 patients received PBSC from MSD contributing to limited GRFS events compared with 124 BM recipients from MSD. Within this limitation, GRFS differed significantly among various donor/graft sources in univariate analysis (Table 2B). However, in multiple regression analysis stratified by center, graft/donor source was not a statistically significant predictor of GRFS. All donor/graft sources had a similar risk of GRFS events at 1- or 2-years (overall P=0.17 and 0.15, respectively), excluding UCB, which was associated with 50-70% increased risk of GRFS events (Table 3B). Figure 3 shows the adjusted GRFS, DFS, OS and GRFS-defining events among recipients of BM graft from MSD compared with UCB graft â&#x20AC;&#x201C;the two most common graft sources used in pediatric patients. The use of RIC resulted in 2.5-2.8 fold higher risk of GRFS events at 1- and 2-years compared to myeloablative conditioning. Disease risk and recipient CMV seropositivity were not associated with GRFS.

Distribution of GRFS events in pediatric patients at one -year post-transplantation GRFS events differed across various graft/donor types, P<0.01. Relapse was the most frequent GRFS event in recipients of BM (59%) or PBSC (50%) from MSD. In the setting of matched URD transplantation, chronic GVHD (43%) in PBSC recipients and acute GVHD (47%) in BM recipients accounted for a majority of events. Recipients of BM from mismatched URD had either relapse (50%) or acute GVHD (50%) as the GRFS-defining event. Relapse (33% and 37%) and death (33% and 36%) added to more than two-thirds of GRFS events in recipients of PBSC from mismatched URD and UCB recipients, respectively. Online Supplementary Figures S1 and S2 depict the distribution of events occurring subsequent to GRFS-defining events by donor/graft source. In recipients of either myeloablative or RIC regimens, relapse (39% and 44%) was the most frequent event while chronic GVHD (14% and 3%) was the

Table 3A. Multiple regression analysis on risk of GRFS *(adult patients).

Factors

RR of events (95% CI)

MA w/o ATG MA w/ ATG RIC w/o ATG RIC with ATG

625 79 465 297

BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB Standard-risk High-risk

123 540 33 104 120 30 104 412 892 574

Conditioning

Age by decade Donor/graft type

Disease risk

At 1-year P

1.0 0.9 (0.6-1.2) 0.8 (0.7-1.0) 0.7 (0.5-0.8) 1.05 (1.0-1.1) 1.0 1.5 (1.1-2.1) 1.7 (1.0-2.8) 1.4 (0.9-2.0) 2.1 (1.5-3.0) 1.5 (0.9-2.6) 2.2 (1.5-3.2) 1.6 (1.2-2.2) 1.0 1.3 (1.1-1.4)

Reference 0.36 0.07 <0.01 0.13 Reference 0.01 0.04 0.10 <0.01 0.14 <0.01 0.01 Reference <0.01

Overall P

RR of events (95% CI)

<0.01

1.0 0.9 (0.7-1.3) 0.8 (0.7-1.0) 0.7 (0.6-0.8) 1.07 (1.0-1.1) 1.0 1.5 (1.2-2.0) 1.6 (1.0-2.6) 1.3 (0.9-1.9) 1.9 (1.4-2.7) 1.4 (0.9-2.4) 2.0 (1.4-2.8) 1.5 (1.1-2.1) 1.0 1.3 (1.1-1.4)

<0.01

At 2-years P

Reference 0.70 0.03 <0.01 0.07 Reference <0.01 0.05 0.12 <0.01 0.16 <0.01 0.01 Reference <0.01

Overall P <0.01

<0.01

*Model stratified by center since there appears to be a violation of the proportional hazards assumption. No other factors violated assumption. ATG: anti-thymocyte globulin; BM: bone marrow; CI: confidence interval; MA: myeloablative; RIC: reduced intensity conditioning; MSD: matched sibling donor; PBSC: peripheral blood stem cells; RR: relative risk; UCB: umbilical cord blood; URD: unrelated donor.

Table 3B. Multiple regression analysis on risk of GRFS * (Pediatric patients).

Factors

RR of events (95% CI)

MA RIC

418 17

BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB

124 12 22 34 23 18 12 190

1.0 2.8 (1.6-4.9) 1.4 (1.0-1.9) 1.0 0.7 (0.2-1.9) 1.0 (0.5-2.1) 1.6 (1.0-2.8) 1.6 (0.9-3.0) 1.7 (0.8-3.4) 1.2 (0.5-2.8) 1.5 (1.1-2.3)

Reference <0.01 0.05 Reference 0.42 0.97 0.07 0.14 0.14 0.74 0.03**

Conditioning Age by decade Donor/Graft Type

Overall P

0.17

RR of events (95% CI)

1.0 2.5 (1.4-4.4) 1.4 (1.1-1.9) 1.0 0.9 (0.4-2.3) 1.1 (0.6-2.2) 1.6 (0.9-2.7) 1.5 (0.8-2.8) 2.1 (1.2-3.9) 1.2 (0.5-2.8) 1.7 (1.1-2.4)

Reference <0.01 0.02 Reference 0.90 0.73 0.09 0.15 0.02** 0.61 0.01

Overall P

0.15

*Model stratified by center since there appears to be a violation of the proportional hazards assumption. No other factors violated assumption. BM: bone marrow; CI: confidence interval; MA: myeloablative; RIC: reduced intensity conditioning; MSD: matched sibling donor; PBSC: peripheral blood stem cells; RR: relative risk; UCB: umbilical cord blood; URD: unrelated donor.

haematologica | 2016; 101(6)

769

R.S. Mehta et al. Table 4A. Cumulative incidence/Kaplan-Meier estimate of each type of event (age â&#x2030;Ľ21 years).

Factor

Strata

Donor/graft type

BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB

By 1-year Grade III-IV Chronic Relapse Acute GVHD GVHD Total Event Estimate Event Estimate Event Estimate N N N N 123 540 33 104 120 30 104 412

11 86 4 22 38 5 32 81

9% 16% 12% 21% 32% 17% 31% 20%

20 168 10 15 27 6 14 80

16% 31% 30% 15% 23% 20% 13% 19%

13 141 7 20 19 5 20 98

11% 26% 21% 19% 16% 17% 19% 24%

Mortality Event N

Estimate

Total number of events

30 183 16 42 45 13 42 183

24% 34% 48% 41% 38% 43% 40% 44%

74 578 37 99 129 29 108 442

BM: bone marrow; MSD: matched sibling donor; PBSC: peripheral blood stem cells; UCB: umbilical cord blood; URD: unrelated donor.

Table 4B. Cumulative incidence/Kaplan-Meier estimate of each type of event (age â&#x2030;Ľ21 years).

Factor

Strata

Donor/graft type

BM-MSD PBSC-MSD Other related donor BM- Matched URD PBSC- Matched URD BM- Mismatched URD PBSC- Mismatched URD UCB

By 1-year Grade III-IV Chronic Relapse Acute GVHD GVHD Total Event Estimate Event Estimate Event Estimate N N N N 124 12 22 34 23 18 12 190

10 0 1 9 3 5 1 19

9% 0% 5% 26% 13% 28% 8% 10%

7 1 1 8 7 0 1 12

6% 8% 5% 24% 31% 0% 8% 6%

31 2 3 3 5 5 2 37

25% 17% 14% 9% 22% 28% 17% 19%

Mortality Event N

Estimate

Total number of events

26 2 6 13 5 7 4 70

21% 17% 27% 38% 22% 39% 33% 37%

74 5 11 33 20 17 8 138

BM: bone marrow; MSD: matched sibling donor; PBSC: peripheral blood stem cells; UCB: umbilical cord blood; URD: unrelated donor.

least frequent event contributing to GRFS; acute GVHD (23% and 25%) and death (24% and 19%) accounted for the rest. Disease relapse accounted for a vast majority of deaths in recipients of BM (69%) or PBSC (50%) from MSD (Online Supplementary Figure S3B). Cumulative incidences of grade III-IV acute GVHD, chronic GVHD, relapse and death at 1-year are shown in Table 4B.

Discussion In this study, we show that the use of BM from a MSD resulted in the best GRFS at 1-year (57%) and 2-years (49%) compared with other donor types and graft sources in adults. In multivariate analysis, BM from matched URD or even mismatched URD led to similar GRFS as BM from MSD, suggesting that BM may be the optimal graft source from adult HLA-matched related or unrelated donors in terms of GRFS. The use of UCB and PBSC from any donor were associated with 1.5-2 times higher risk of GRFS events than BM from MSD. These findings are of interest, especially in view of the reported results of the randomized controlled trial in adult patients receiving PBSC or BM from matched URD, conducted by the Blood and Marrow Transplant Clinical 770

Trials Network (BMT CTN 0201).7 That trial showed significantly higher incidence of chronic GVHD at 2-years in recipients of PBSC (53% versus 41%) and a higher incidence of graft failure in BM recipients (9% versus 3%), without any differences in acute GVHD, disease-free survival, or overall survival with BM versus PBSC.7 Despite the outcomes of BMT CTN 0201, PBSC continues to be the most frequently utilized graft source - representing about 65% of all grafts used for allogeneic HCT in recent years.8 In our study, both acute and chronic GVHD contributed to higher GRFS events in recipients of PBSC (49% and 22%) compared with BM (33% and 14%) from matched URD. Moreover, deaths due to GVHD were higher in PBSC (53%) than in BM (31%) recipients. Graft failures did not contribute to any deaths in recipients of either BM or PBSC from matched URD. Similar findings were noted in the MSD group where the proportion of acute or chronic GVHD events and deaths associated with them were higher in PBSC recipients compared with BM recipients. In addition to donor/graft source, other factors that had a favorable impact on GRFS included RIC regimen with ATG and standard risk disease in adults. Although these factors and the availability of a donor cannot be modified, GRFS could be favorably influenced by the preferential use of BM grafts rather than PBSC. haematologica | 2016; 101(6)

BM graft is associated with best GRFS after allogeneic HCT

Outcomes after BM-MSD versus UCB HCT in pediatric patients

Figure 3. Adjusted GRFS, DFS, OS, grade III-IV acute GVHD, chronic GVHD and relapse among pediatric patients with BM graft from matched sibling donor compared with UCB graft.

However, despite these data, transplant centers and donors may continue the preferential use of PBSC for many reasons.9-11 Novel GVHD prophylaxis regimens are therefore currently undergoing evaluation for their impact on GRFS after PSBC transplantation.12 On the other hand, in the pediatric population, existing data distinctly show that the use of PBSC from MSD is associated with a higher risk of chronic GVHD, treatment related mortality, treatment failure and mortality compared with BM grafts.13 Given these data, only 12 patients in our study received PBSC from MSD compared with 124 BM grafts from MSD. Consequently, within the limitations of the sample size, GRFS did not differ between different donor or graft sources in pediatric patients, although UCB was associated with inferior GRFS. The only factor that favorably affected GRFS was the use of myeloablative conditioning compared with the RIC regimen. However, patients who received RIC regimens had comorbidities or contraindications to receive myeloablative regimens, such as prior use of high dose radiation therapy, recent autologous HCT, suspected or proven fungal pneumonia, prior CNS malignancy or veno-occlusive disease. Two patients received sequential intensive chemotherapy and RIC HCT for high risk AML. It is important to recognize that GRFS is computed by a time-to-event analysis, which disregards all incidents happening after the onset of the first GRFS-defining event. Therefore, patients who develop GRFS-defining GVHD or relapse but completely recover with treatment may be morbidity-free at 1-year and are yet included among the GRFS events. To account for this shortcoming of GRFS analysis, real-time periodic assessments and analysis with specific survival models are needed. Another proposed composite outcome for GVHD treatment trials is failurefree survival,14-16 where the outcomes are measured from the time of onset of acute or chronic GVHD, which haematologica | 2016; 101(6)

include the absence of second-line systemic therapy for GVHD, non-relapse mortality and recurrent malignancy during initial treatment of GVHD. Yet, it does not consider events happening prior to the onset of GVHD. Also, as mentioned above, BM and UCB constituted the majority of the graft sources in the pediatric population, because of which caution is warranted when comparing other donor/graft sources. In this study, we were unable to include hematopoietic cell transplantation comorbidity index (HCT-CI)17 or the revised disease risk index18 for risk stratification of all patients. Our study also does not address the importance of GRFS in â&#x20AC;&#x153;alternative donorâ&#x20AC;? settings such as haploidentical HCT compared with UCB and mismatched HCT, which should be assessed in future studies. It is critical to consider GRFS when deciding about a donor/graft source, in addition to independent outcomes of GVHD, disease-free survival and overall survival, because GRFS potentially represents a surrogate endpoint for quality-of-life. In our study, although 64% of all patients were alive and 56% were disease-free at 1-year, only 38% were free of any GRFS defining event at that time. Analysis of GRFS outcomes in the BMT CTN 0201 trial or in larger datasets, such as those available through the Center for International Blood and Marrow Transplant Research, will shed even more light on this crucial question. These data reignite the question of the optimal donor/graft source in adults, challenge the current practice standards, and emphasize the need for continued studies to address the key events leading to morbidity and mortality in HCT recipients given that a minority of patients still survive 1 year without a GRFS event. Funding This work was supported in part by grants from the National Cancer Institute P01 CA65493 (C.G.B, B.R.B.). 771

R.S. Mehta et al.

References 1. Holtan SG, DeFor TE, Lazaryan A, et al. Composite end point of graft-versus-host disease-free, relapse-free survival after allogeneic hematopoietic cell transplantation. Blood. 2015;125(8):1333-1338. 2. ASBMT. American Society for Blood and Marrow Transplantation RFI 2006; 2006. 3. Kaplan EL, Meier P. Nonparametric Estimation from Incomplete Observations. J Am Stat Assoc. 1958;53(282):457-481. 4. Chang IM, Gelman R, Pagano M. Corrected group prognostic curves and summary statistics. J Chronic Dis. 1982;35(8):669-674. 5. Cox DR. Regression models and life tables. J R Stat Soc Series B Stat Methodol. 1972;34(2):187-220. 6. Collett D. Modelling Survival Data in Medical Research, Second Edition (Chapman & Hall/CRC Texts in Statistical Science) In: Chatfield C, ed; 2003. 7. Anasetti C, Logan BR, Confer DL, Blood, Marrow Transplant Clinical Trials N. Peripheral-blood versus bone marrow stem cells. N Engl J Med. 2013;368(3):288.

772

8. Pasquini MC, Zhu X. Current uses and outcomes of hematopoietic stem cell transplantation: 2014 CIBMTR Summary Slides. Available at: http://www.cibmtr.org; 2014. 9. Pulsipher MA, Chitphakdithai P, Logan BR, et al. Lower risk for serious adverse events and no increased risk for cancer after PBSC vs BM donation. Blood. 2014;123(23):36553663. 10. Pulsipher MA, Chitphakdithai P, Logan BR, et al. Acute toxicities of unrelated bone marrow versus peripheral blood stem cell donation: results of a prospective trial from the National Marrow Donor Program. Blood. 2013;121(1):197-206. 11. Young JH, Logan BR, Wu J, et al. Infections after Transplantation of Bone Marrow or Peripheral Blood Stem Cells from Unrelated Donors. Biol Blood Marrow Transplant. 2016;22(2):359-370. 12. BMT CTN Protocol 1203 (PROGRESS I), available online at https://web.emmes.com/study/bmt2/protocol/1203_protocol/1203_protocol.html; 2015. 13. Eapen M, Horowitz MM, Klein JP, et al. Higher mortality after allogeneic peripheral-blood transplantation compared with

14.

15.

16.

17.

18.

bone marrow in children and adolescents: the Histocompatibility and Alternate Stem Cell Source Working Committee of the International Bone Marrow Transplant Registry. J Clin Oncol. 2004;22(24):48724880. Inamoto Y, Flowers ME, Sandmaier BM, et al. Failure-free survival after initial systemic treatment of chronic graft-versus-host disease. Blood. 2014;124(8):1363-1371. Inamoto Y, Storer BE, Lee SJ, et al. Failurefree survival after second-line systemic treatment of chronic graft-versus-host disease. Blood. 2013;121(12):2340-2346. Palmer J, Chai X, Martin PJ, et al. Failurefree survival in a prospective cohort of patients with chronic graft-versus-host disease. Haematologica. 2015;100(5):690-695. Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. 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):36643671.

haematologica | 2016; 101(6)

ARTICLE

Stem Cell Transplantation

Expanding transplant options to patients over 50 years. Improved outcome after reduced intensity conditioning mismatched-unrelated donor transplantation for patients with acute myeloid leukemia: a report from the Acute Leukemia Working Party of the EBMT Bipin N. Savani,1,2 Myriam Labopin,2,3,4,5 Nicolaus Kröger,6 Jürgen Finke,7 Gerhard Ehninger,8 Dietger Niederwieser,9 Rainer Schwerdtfeger,10 Donald Bunjes,11 Bertram Glass,12 Gerard Socié,13 Per Ljungman,14 Charles Craddock,15 Frédéric Baron,16 Fabio Ciceri,17 Norbert Claude Gorin,18 Jordi Esteve,19 Christoph Schmid,20 Sebastian Giebel,21 Mohamad Mohty,2,3,4,5,* and Arnon Nagler2,22,*

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2016-780 Volume 101(6):773

1 Vanderbilt University Medical Center, Nashville, TN, USA; 2Acute Leukemia Working Party, EBMT Paris study office / CEREST-TC, France; 3Department of Haematology, Saint Antoine Hospital, Paris, France; 4INSERM UMR 938, Paris, France; 5Université Pierre et Marie Curie, Paris, France; 6 University Hospital Eppendorf, Department of Stem cell Transplantation, Hamburg, Germany; 7 University of Freiburg, Department of Medicine –Hematology, Oncology, Germany; 8 Universitaetsklinikum Dresden, MedizinischeKlinik und Poliklinik I, Germany; 9University Hospital Leipzig, Div.Hematology, Oncology and Hemostasiology, Germany; 10Helios-Klinikum Berlin-Buch, Dept. Hematology Berlin, Germany; 11Klinik fuer Innere Medzin III - Universitätsklinikum Ulm, Germany; 12Asklepios Klinik St. Georg - Department of Haematology, Hamburg, Germany; 13 Hopital St. Louis - Dept.of Hematology, Paris, France; 14Karolinska University Hospital, Department of Hematology, Stockholm, Sweden; 15Center for Clinical Hematology, Queen Elizabeth Hospital, Birmingham, UK; 16Department of Medicine, Division of Hematology, University of Liège, Belgium; 17Department of Hematology, Ospedale San Raffaele, Università degli Studi, Milan, Italy; 18Faculté de Médicine Saint-Antoine, Paris, France; 19Dept. of Hematology, Hospital Clinic, Barcelona, Spain; 20Klinikum Augsburg, Dept. of Hematology and Oncology, University of Munich, Augsburg, Germany; 21Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice Branch, Poland; and 22Hematology Division, Chaim Sheba Medical Center, Tel Hashomer, Israel

*MM and AN contributed equally to this work.

ABSTRACT

Correspondence: bipin.savani@vanderbilt.edu

he outcome of patients undergoing HLA-matched unrelated donor allogeneic hematopoietic cell transplantation following reduced-intensity conditioning or myeloablative regimens is reported to be equivalent; however, it is not known if the intensity of the conditioning impacts outcomes after mismatched unrelated donor transplantation for acute myeloid leukemia. Eight hundred and eighty three patients receiving reduced-intensity conditioning were compared with 1041 myeloablative conditioning regimen recipients in the setting of mismatched unrelated donor transplantation. The donor graft was HLAmatched at 9/10 in 872 (83.8%) and at 8/10 in 169 (16.2%) myeloablative conditioning recipients, while in the reduced-intensity conditioning cohort, 754 (85.4%) and 129 (14.6%) were matched at 9/10 and 8/10 loci, respectively. Myeloablative conditioning regimen recipients were younger, 70% being <50 years of age compared to only 30% in the reduced-intensity conditioning group (P=0.0001). Significantly, more patients had secondary acute myeloid leukemia (P=0.04) and Karnofsky Performance Status score <90% (P=0.02) in the reducedintensity conditioning group. Patients <50 and ≥50 years were analyzed separately. On multivariate analysis and after adjusting for differences between the two groups, reduced-intensity conditioning in patients age ≥50 years was associated with higher overall survival (HR 0.78; P=0.01), leukemia-free survival (HR 0.82; P=0.05), and decreased non-relapse mortality (HR 0.73; P=0.03). Relapse incidence (HR 0.91; P=0.51) and chronic graft-versus-host disease (HR 1.31; P=0.11) were, however, not significantly different. In patients <50 years old, there were no statistically significant differences in overall survival, leukemia-free survival, relapse incidence, non-relapse mortality, and chronic graft-versus-host-disease between the groups. Our study shows no significant outcome differences in patients younger than 50 years receiving reduced-intensity vs. myeloablative conditioning regimens after mismatched unrelated donor transplantation. Furthermore, the data support the superiority of reduced-intensity conditioning regimens in older adults receiving transplants from mismatched unrelated donors.

haematologica | 2016; 101(6)

Received: October 20, 2015. Accepted: March 2, 2016. Pre-published: March 8, 2016. doi:10.3324/haematol.2015.138180

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/773

773

B.N. Savani et al.

Introduction Allogeneic hematopoietic cell transplantation (HCT) can provide extended leukemia-free survival (LFS) for the majority of patients with acute myeloid leukemia (AML).1 However, only about 30-35% of patients have an HLAmatched sibling.2-5 When an HLA-matched donor (related or unrelated) is not available or not suitable to donate, alternative donors may be considered if the patient is likely to benefit from allogeneic transplantation. Alternative donor sources include HLA-mismatched adult unrelated donors, unrelated umbilical cord blood (UCB), and mismatched related (family) members (haploidentical donor). Increasing numbers of patients are receiving alternative donor HCT, preference is given to a centers experience, physician preferences and participation in clinical trials.6 Currently it is possible to find a stem cell source for virtually all patients who have an indication to receive HCT.7,8 Outcomes for matched or mismatched unrelated donor allografting have improved over time, likely because of, among other factors, better HLA-typing and matching and intensive supportive care.9,10 Outcomes for matched unrelated donor allografting now appear comparable with those seen with matched sibling donor HCT.3,4,11,12 Reduced-intensity conditioning (RIC) regimens have further extended the use of allogeneic HCT to older patients and those with significant pre-transplant comorbidities.9,13,14 Several previous studies comparing the clinical outcomes of RIC and MAC regimens after matched related or unrelated donor HCT in AML have shown similar outcomes;13-16 however, it is unknown in patients receiving mismatched unrelated donor (MM-URD) HCT. Therefore, we explored whether there were outcome differences for adults with AML after RIC or MAC regimens in the setting of MM-URD HCT using data reported to the European Society for Blood and Marrow Transplantation (EBMT).

Methods Study design and data collection This was a retrospective multicenter analysis. Data were provided and approved for this study by the Acute Leukemia Working Party (ALWP) of the EBMT group registry. The latter is a voluntary working group of more than 500 transplant centers that are required to report all consecutive stem cell transplantations and follow-ups once a year. Audits are routinely performed to determine the accuracy of the data. Since 1990, patients have provided informed consent authorizing the use of their personal information for research purposes. Eligibility criteria for this analysis included adult patients (age >18 years) with AML, transplanted between 2000 and 2012, from a HLA-mismatched unrelated donor with bone marrow (BM) or G-CSF-mobilized peripheral blood (PB) stem cells. All donors were HLA-mismatched at one or two loci (9/10 or 8/10) (-A, -B, -C, DRB1, -DQB1). Recipients of a previous allogeneic or cord blood transplant were excluded. Variables collected included recipient and donor characteristics (age, gender, CMV serostatus), disease status at transplant, transplant related-factors including conditioning regimen, immunosuppression (in vivo T-cell depletion vs. none), stem cells source (BM or PB), GVHD prophylaxis, and outcome variables (acute and chronic GVHD, relapse, non-relapse mortality [NRM], LFS, overall survival [OS] and causes of death). Regimens were classified as MAC or RIC based on published criteria.17 The list of institutions reporting data included in this study is provided in the Online Supplementary Appendix. 774

Statistical analysis The primary end points of the study were OS and LFS. Secondary endpoints included: disease relapse incidence (RI), NRM, engraftment and incidences and severity of acute and chronic GVHD. The starting point for time-to-event analysis was “date of transplantation”. OS was defined as the time to death from any cause. Surviving patients were censored at the time of their last follow-up. LFS was defined as survival without relapse or progression. Patients surviving in continuous CR were censored at the time of last follow-up. RI was defined as time to onset of leukemia recurrence. NRM was the competing risk, and patients surviving in continuous complete remission were censored at the last contact. NRM was defined as death without relapse/progression (relapse was the competing risk). The probabilities of OS and LFS were calculated by using the Kaplan-Meier estimator. The probabilities of chronic GVHD, NRM, and relapse were calculated by using the cumulative incidence estimator to accommodate competing risks. For chronic GVHD, death without the event was the competing risk. The 2 groups according to the conditioning regimen were compared by the Chi-square test for qualitative variables, whereas the Mann-Whitney test was applied for continuous parameters. Univariate comparisons were done using the log-rank test for OS, LFS, and the Gray’s test for RI, NRM and chronic GVHD cumulative incidences. Multivariate analyses were performed using logistic regression for acute GVHD and Cox proportional hazards model for all other endpoints. All factors known as potentially related to the outcome were included in the final model. All tests were two-sided. The type I error rate was fixed at 0.05 for the determination of factors associated with time to event outcomes. Statistical analyses were performed with SPSS 22.0 (IBM Corp., Armonk, NY, USA) and R 3.1.1 software packages (R Development Core Team, Vienna, Austria).

Results Patient, disease and transplant characteristics Details of patients, disease and transplant characteristics are summarized in Table 1. One thousand nine hundred and twenty four patients with AML were included in the study. One thousand and forty one patients received MAC and 883 RIC regimens before MM-URD HCT between 2000 and 2012. RIC recipients were older with a median age of 57 years (range, 18-75) in comparison to 43 years (range, 18-72) for the MAC group (P<0.0001). Only thirty percent of the patients were ≤50 years of age in the RIC group versus 70% in the MAC group (P<0.0001). The median time from diagnosis to transplant was similar in the MAC and RIC groups (240 vs. 235 days, respectively; P=0.89). The median follow-up of surviving patients in the MAC group was 27 (IQR, 12-48) months, while that of the RIC group was 23 (IQR, 5-44) months (P=0.004). Significantly higher numbers of patients had secondary AML (13 vs. 10%, P=0.04) and KPS<90% (30 vs. 25%, P=0.02) in the RIC group. There were no significant differences in the distribution of advanced disease and poor risk cytogenetic among regimens. The proportions of CMV positive recipients and donors were similar in the MAC and RIC groups (67% vs. 66%; P=0.84 and 44% vs. 45%; P=0.52, respectively). Commonly used MAC regimens were TBI based (n=369), BuCy (n=354) and Bu-Flu (n=143); in the RIC group the most commonly used regimens were low dose TBI based (n=275), Bu-Flu (n=312) and Flu-Mel (n=178). haematologica | 2016; 101(6)

Mismatched unrelated donor transplantation for AML

Among the MAC recipients, 872 (83.8%) received 9/10 and 169 (16.2%) 8/10 HLA-matched donors, and in the RIC cohort, 754 (85.4%) received 9/10 and 129 (14.6%) 8/10 matched donors (P=0.33). The percentage of patients receiving in vivo T-cell depletion was not significantly different between the two groups (P=0.18, Table 1). In the MAC group bone marrow was used more frequently as the stem cell source (20 vs. 9%; P<0.0001). Apart from in vivo T-cell depletion, GvHD prophylaxis consisted of the combination of one calcineurin inhibitor with mycophenolate mofetil (MMF) alone, or in association with methotrexate (MTX). A calcineurin inhibitor + MMF and/or MTX were used in 96% and 95% of the

patients in the MAC and RIC groups, respectively. The choice of conditioning and GvHD prophylaxis was dependent on transplant center protocols and strategies for transplantation.

Engraftment and GvHD Conditioning regimen specific engraftment and GvHD data are summarized in Table 2. Ninety five percent of patients in the MAC group engrafted versus 96% in the RIC group (P=0.45). The median day to absolute neutrophil count (ANC) >500/mL was 16 in both groups. The percentage of grade II-IV (33% vs. 32%; P=0.55) and III-IV (12% vs. 14%; P=0.38) acute GvHD were not significantly

Table 1. Patients, disease and transplant characteristics.

Patient characteristics Recipient age at HCT (median, IQR), years <50 â&#x2030;Ľ50 Recipient gender, n (%) Male Female Unknown Interval from diagnosis to HCT (median, IQR), days Donor age (years, range) Donor gender, n (%) Male Female Unknown Female donor to male recipient, n (%) Disease status at HCT, n (%) CR1 â&#x2030;ĽCR2 Active disease Secondary AML Karnofsky at HCT, <90%, n (%) Patient positive CMV serology Donor positive CMV serology Human leukocyte antigen matching 9/10 match 8/10 match Stem cells source BM PB Conditioning regimen, n Bu-Cy Bu-Flu Flu-Mel TBI-MAC TBI-RIC Others regimens In vivo T-cell depletion, n (%)

MAC (n=1041)

RIC (n=883)

43 (32-52) 731 (70.2%) 310 (29.8%)

57 (47-63) 267 (30.2%) 616 (69.8%)

<10-4

523 (50.2%) 518 (49.8%) 0 240 (158-478) 36 (19-70)

455 (51.7%) 425 (48.3%) 3 235 (156-505) 35 (20-61)

651 (64.3%) 361 (35.7%) 29 157 (15.5%)

567 (65.2%) 303 (34.8%) 13 132 (15.2%)

523 (50.2%) 261 (25.1%) 257 (24.7%) 104 (10%) 230/917 (25.1%) 670/1008 (66.5%) 444/1014 (43.8%)

408 (46.2%) 226 (25.6%) 249 (28.2%) 115 (13%) 236/787 (30%) 573/868 (66%) 393/868 (45.3%)

872 (83.8%) 169 (16.2%)

754 (85.4%) 129 (14.6%)

207 (19.5%) 834 (80.1%)

84 (9.5%) 799 (90.5%)

354 143 36 369

312 178

0.52

0.89 0.66 0.7

0.86 0.14

0.04 0.02 0.84 0.52 0.33

<0.0001

139 841/1029 (81.7%)

275 118 734/878 (83.6%)

0.18

AML: acute myeloid leukemia; CMV: cytomegalovirus; CR: complete remission; RIC: reduced intensity conditioning; MAC: myeloablative conditioning; HCT: allogeneic stem cell transplantation; BM: bone marrow; PB: peripheral blood; Bu- busulfan; Cy: cyclophosphamide; Flu: fludarabine; Mel: melphalan; TBI: total body irradiation; some percentages do not add up to 100% because of rounding.

haematologica | 2016; 101(6)

775

B.N. Savani et al.

different between the groups. As shown in Table 3, in multivariate analysis, the factors associated with increased risk of grade II-IV acute GvHD were active disease (OR 1.48; 95% CI, 1.05-2.11; P=0.03) in patients <50 years and female donor to male recipient (OR 1.56; 95% CI, 1.042.35; P=0.03) in patients ≥50 years. In vivo T-cell depletion was associated with a decreased risk of acute GVHD in patients in both the <50 years (OR 0.67; 95% CI, 0.470.97; P=0.03) and ≥50 years groups (OR 0.70; 95% CI, 0.48-1.03; P=0.07). The two-year cumulative incidence of chronic GVHD was higher after the RIC regimen (34% [95% CI, 31-38] vs. 29% [95% CI, 26-32] in the MAC, P=0.04) (Table 4). In multivariate analysis, chronic GVHD was not significantly different between MAC and RIC in patients <50 years (HR 0.91; 95% CI, 0.66-1.26; P=0.58) and in ≥50 years groups (HR 1.31; 95% CI, 0.95-1.81; P=0.11). The factor associated with chronic GVHD were female donor to male recipient (HR 1.49; 95% CI, 1.08-2.07; P=0.02) and in vivo T-cell depletion (HR 0.51; 95% CI, 0.37-0.68; P=0.00001) in patients <50 years and (HR 0.63; 95% CI, 0.45-0.89; P=0.01) in patients ≥50 years (Table 3).

NRM There was no difference in NRM at 2 years between the MAC and RIC groups in univariate analysis- (28%; 95% CI, 25-30 after MAC vs. 27%; 95% CI, 24-30 after RIC; P=0.76) (Table 4). When analyzing patients <50 and ≥50 years separately, only the older cohort in the MAC group showed a higher risk of NRM (36%; 95% CI, 30-41 for MAC vs. 30%; 95% CI, 24-35; P=0.05) (Table 4). In multivariate analysis, NRM was not significantly different between the MAC and RIC in patients <50 years (HR 0.91; 95% CI, 0.65-1.26; P=0.56), however in the ≥50 years group, the RIC regimen was independently associated with decreased NRM (HR 0.73; 95% CI, 0.56-0.97; P=0.03). The other factors associated with NRM were active disease and secondary AML in those <50 ((HR 1.86; 95% CI, 1.32-2.62; P=0.0004; HR 1.65; 95% CI, 1.06-2.58; P=0.03, respectively) and ≥50 years (HR 1.45; 95% CI, 1.06-1.98; P=0.02; HR 1.41; 95% CI, 1.01-1.96; P=0.04, respectively) and CMV seropositivity (HR 1.37; 95% CI, 1.01-1.86; P=0.04) in patients <50 years and age at HCT (by +10 years) (HR 1.15; 95% CI, 1.04-1.28; P=0.01) in patients ≥50 years (Table 3).

Leukemia-free survival There was no difference in LFS at 2 years between the MAC and RIC groups in univariate analysis (43%; 95% CI, 39-46; after MAC vs. 40%; 95% CI, 36-44 after RIC; P=0.34) (Table 4). When analyzing patients <50 and ≥50 years separately, an advantage was seen for LFS in the MAC group only for younger patients (<50 years) (46%; 95% CI, 42-50 for MAC vs. 39%; 95% CI, 32-45; P=0.05) and in the RIC group for the older patients (34%; 95% CI, 29-40 for MAC vs. 40%; 95% CI, 36-45; P=0.03) (Table 4). In multivariate analysis, LFS was not significantly different between MAC and RIC in patients <50 years (HR 1.02; 95% CI, 0.83-1.26; P=0.83). Among older cohorts (age ≥50), there was an advantage of RIC regimen with higher LFS (HR 0.82; 95% CI, 0.68-1.00; P=0.05). The other factors associated with LFS were active disease in <50 (HR 2.48; 95% CI, 2.00-3.06; P<10-4) and ≥50 years groups (HR 1.93; 95% CI, 1.56-2.38; P<10-4) and secondary AML (HR 1.36; 95% CI, 1.00-1.85; P=0.05) in patients <50 years and age at HCT (by +10 years) (HR 1.09; 95% CI, 1.02-1.17; P=0.02) in patients ≥50 years (Table 3).

Overall survival There was no difference in OS at 2 years between the MAC and RIC groups in univariate analysis (45%; 95% CI, 42-49; after MAC vs. 45%; 95% CI, 41-48 after RIC; P=0.81) (Table 4). When analyzing patients <50 and ≥50 years separately, RIC had an advantage in older cohorts (37%; 95% CI, 31-43 for MAC vs. 45%; 95% CI, 40-49; P=0.01) (Table 4). In multivariate analysis, OS was not significantly different between MAC and RIC in patients <50 years old (HR 0.96; 95% CI, 0.77-1.19; P=0.71). Among older cohorts (age ≥50), there was an advantage of RIC regimen with higher OS (HR 0.78; 95% CI, 0.66-0.95; P=0.01). The other factors associated with OS were active disease in patients <50 (HR 2.37; 95% CI, 1.90-2.95; P<0.0001) and ≥50 years old (HR 1.82; 95% CI, 1.46-2.26; P<0.0001), secondary AML (HR 1.41; 95% CI, 1.03-1.94; P=0.03), higher HLA-mismatching (8/10 vs. 9/10) (HR 1.27; 95% CI, 1.011.61; P=0.05) in patients <50 years old and age at HCT (by +10 years) (HR 1.14; 95% CI, 1.05-1.22; P=0.0008) in patients ≥50 years (Table 3).

Relapse There was no difference in RI at 2 years between the MAC and RIC groups in univariate analysis (30%; 95% CI, 27-33; after MAC vs. 33%; 95% CI, 30-36 after RIC; P=0.11) (Table 4). When analyzing patients <50 and ≥50 years separately, an advantage of receiving a MAC regimen in reducing relapse risk was only observed in the younger cohort (<50 years) (30%; 95% CI, 26-34 for MAC vs. 40%; 95% CI, 33-46; P=0.008) (Table 4); In multivariate analysis, RI was not different between MAC and RIC in both <50 years and ≥50 years groups (Table 3). The factors associated with RI were active disease in both <50 (HR 2.96; 95% CI, 2.25-3.88; P<0.0001) and ≥50 years group (HR 2.46; 95% CI, 1.85-3.27; P<0.0001), CMV positive recipients (HR 0.78; 95% CI, 0.60-1.00; P=0.05) and higher HLA- mismatching (8/10 vs. 9/10) (HR 0.69; 95% CI, 0.48-1.00; P=0.05) in patients ≥50 years (Table 3). 776

Table 2. Conditioning regimen specific engraftment and GvHD data.

Engraftment Yes No Unknown, n Acute GvHD Grade II-IV, n (%) Grade III-IV, n (%) Chronic GvHD* Limited, n Extensive, n Unknown, n

MAC (n=1041)

RIC (n=883)

976 (95.2%) 49 (4.8%) 16

826 (95.9%) 35 (4.1%) 22

0.45

330/997 (33.1%) 122/997 (12.2%) 29.2% [26-32.3] 119 123 7

271/852 (31.8%) 116/852 (13.6%) 34.2% [30.6-37.8] 125 88 23

0.55 0.38 0.04

GvHD: graft-versus-host disease; *2-year cumulative incidence; RIC: reduced intensity conditioning; MAC: myeloablative conditioning; some percentages do not add up to 100% because of rounding.

haematologica | 2016; 101(6)

Mismatched unrelated donor transplantation for AML

Interaction between the conditioning regimen and HLA-mismatch

Discussion

There is no significant interaction between the conditioning regimen and the degree of HLA-mismatch (P=0.32) or status at transplant (P=0.29). Interactions remained insignificant when data were analyzed separately for patients <50 (P=0.30 and P=0.82 with HLA-mismatches and with status at transplant, respectively) and ≥50 years (P=0.38 and P=0.29 for HLA mismatch and status at transplant, respectively). Results were similar among HLA-mismatched pairs, and recipients with advanced disease, although this study was not designed to detect potential differences within these subsets.

This large, multicenter, registry study did not show significant outcome difference between transplant recipients who received RIC and those who received MAC regimen followed by HLA MM-URD HCT for AML in patients younger than 50 years. Moreover, data support the superiority of RIC regimen in patients ≥50 years receiving transplant from a MM-URD. This finding is novel and clinically very important since many older adults are not transplanted due to the lack of a 10/10 HLA-matched donor. We investigated patient, disease, and transplantation factors affecting survival, LFS, relapse, and NRM in a well-

Table 3. Multivariate analysis- comparison between RIC vs. MAC regimen and significant factors associated with outcome.

P Age <50 years

HR (95% CI)

P Age ≥50 years

HR (95% CI)

0.50 <10-4

1.10 (0.84-1.43) 2.96 (2.25-3.88)

RIC vs. MAC Active disease (ref CR1) CMV+ donor HLA-match 8/10 vs. 9/10

0.51 <10-4 0.05 0.05

0.91 (0.70-1.20) 2.46( 1.85-3.27) 0.78 (0.60-1.00) 0.69 (0.48-1.00)

0.56 0.0004 0.03 0.04

0.91 (0.65-1.26) 1.86 (1.32-2.62) 1.65 (1.06-2.58) 1.37 (1.01-1.86)

RIC vs. MAC Active disease (ref CR1) Secondary AML Age at SCT (+10 years)

0.03 0.02 0.04 0.01

0.73 (0.56-0.97) 1.45 (1.06-1.98) 1.41 (1.01-1.96) 1.15 (1.04-1.28)

0.27 0.03 0.03

0.83 (0.59-1.16) 1.48 (1.05-2.11) 0.67 (0.47-0.97)

RIC vs.MAC Female►male In vivo T-cell depletion

0.29 0.03 0.07

1.20 (0.86-1.67) 1.56 (1.04-2.35) 0.70 (0.48-1.03)

0.58 0.02 0.00001

0.91 (0.66-1.26) 1.49 (1.08-2.07) 0.51 (0.37-0.68)

RIC vs. MAC In vivo T-cell depletion

0.11 0.01

1.31 (0.95-1.81) 0.63 (0.45-0.89)

0.83 <10-4 0.05

1.02 (0.83-1.26) 2.48 (2.00-3.06) 1.36 (1.00-1.85)

RIC vs. MAC Active disease (ref CR1) Age at SCT (+10 years)

0.05 <10-4 0.02

0.82 (0.68-1.00) 1.93 (1.56-2.38) 1.09 (1.02-1.17)

0.71 <10-4 0.03 0.05

0.96 (0.77-1.19) 2.37 (1.90-2.95) 1.41 (1.03-1.94) 1.27 (1.01-1.61)

RIC vs. MAC Active disease (ref CR1) Age at SCT (+10 years)

0.01 <10-4 0.0008

0.78 (0.66-0.95) 1.82 (1.46-2.26) 1.14 (1.05-1.22)

Relapse RIC vs. MAC Active disease (ref CR1)

NRM RIC vs. MAC Active disease (ref CR1) Secondary AML CMV+ patient Acute GvHD* RIC vs. MAC Active disease (ref CR1) In vivo T-cell depletion Chronic GVHD RIC vs. MAC Female►male In vivo T-cell depletion LFS RIC vs. MAC Active disease (ref CR1) Secondary AML OS RIC vs. MAC Active disease (ref CR1) Secondary AML HLA-match 8/10 vs. 9/10

HR-hazard ratio; CR: complete remission; GvHD: graft-versus-host-disease; LFS: leukemia-free survival; NRM: non-relapse mortality; RIC: reduced intensity conditioning; MAC: myeloablative conditioning; OS: overall survival; Female►male: female donor for male recipient.

Table 4. Outcomes at 2 years, by age at transplantation.

Disease status All patients* Age <50 years*~ Age ≥50 years*

Patients group

NRM

LFS

cGvHD

MAC RIC P MAC RIC P MAC RIC P

29.9% [27.2-32.9] 33% [29.6-36.4] 0.11 29.9% [26.4-33.5] 39.7% [33.3-46.1] 0.008 29.8% [24.5-35.4] 30% [26.2-34] 0.95

27.5% [24.7-30.4] 27.1% [24.3-30] 0.76 24.1% [20.9-27.4] 21.4% [18.3-24.6] 0.50 35.8% [30.1-41.4] 29.6% [24.2-35.3] 0.05

42.5% [39.2-45.7] 39.9% [36.3-43.5] 0.34 45.9% [42-49.7] 38.9% [32.3-45.4] 0.05 34.3% [28.5-40.1] 40.3% [36-44.6] 0.03

45.4% [42.1-48.7] 44.6% [40.9-48.3] 0.81 48.8% [44.9-52.8] 44.6% [37.8-51.3] 0.24 37.1% [31.3-43] 44.6% [40.2-49] 0.01

29.2% [26-32.3] 34.2% [30.6-37.8] 0.04 31.2% [27.4-35] 26.2% [20.4-32.5] 0.18 24.3% [19-29.9] 37.6% [33.2-42.1] 0.0009

Data are % (95% CI), unless otherwise specified; *2-year outcome; cGvHD: chronic graft-versus-host-disease; LFS: leukemia-free survival; NRM: non-relapse mortality; MAC: myeloablative conditioning; OS: overall survival; RI: relapse incidence; RIC: reduced intensity conditioning; ~higher advanced disease patients in RIC group (38% vs. 21% in MAC group, P<10-4).

haematologica | 2016; 101(6)

777

B.N. Savani et al.

studies note similar outcomes for those in complete remission.22-24 An unanswered question is whether HLA match requirements should differ based on the conditioning regimen. With increasing numbers of reduced-intensity conditioning transplantations being performed for AML, our study is timely and explores the impact of HLA mismatching and the intensity of conditioning intensity in the current era. It is possible that the historically higher relapse rate after RIC compared to MAC can be abrogated by the potent graft-versus-leukemia (GVL) effect induced by greater HLA disparity after RIC MM-URD transplant.25,26 In multivariate analysis, our data showed that the relapse rate was not different between MAC and RIC in both <50 years and â&#x2030;Ľ50 years groups. However, in older patients (â&#x2030;Ľ50 years), the superiority of the RIC regimen was due to the additional benefits of decreased NRM compared to patients receiving MAC. The protective effect of in vivo T-cell depletion on the incidence of GVHD without compromising transplant outcome was reaffirmed in our study.27 Given the multiple

Cumulative incidence of NRM Overall survival

Leukemia-free survival

Cumulative incidence of relapse

characterized population of nearly 2000 adult AML patients receiving MAC or RIC MM-URD HCT. Overall, nearly 45% of patients (Figure 1,2) transplanted with MMURD survived beyond 2 years and the intensity of the conditioning regimen did not significantly influence LFS or OS. Two-year survival after RIC regimen was favorable (45%) compared with MAC regimens (37%, P=0.01). Similarly, the risk of relapse was not different between the two groups in multivariate analysis. These promising results compare favorably with outcomes after HLAmatched donor transplant, yet the heterogeneity in the subjects likely contribute to the differences.2,4,18 Another important finding is the worsening outcome with age in the older cohort which is consistent with many previous reports.9 Clear data on the value of the conditioning regimen intensity for AML are still lacking, though the use of RIC has extended the availability of allogeneic HCT to older patients.13 Many retrospective studies have highlighted that the lower risks of NRM are offset by the increased rates of relapse in RIC with similar OS.13,19-21 More recent

Figure 1. Probability of overall survival (OS), leukemia-free survival (LFS), non-relapse mortality (NRM) and relapse incidence (RI) after myeloablative or reduced intensity conditioning regimen for acute myeloid leukemia (age <50 years) after mismatched unrelated donor transplantation. 778

haematologica | 2016; 101(6)

Mismatched unrelated donor transplantation for AML

Cumulative incidence of NRM

answer questions in patients receiving MM-URD HCT. Despite the inherent limitations of our retrospective registry study and in the absence of the prospect for prospective data in the near future, it is reasonable to consider RIC regimen for patients receiving MM-URD HCT for AML in transplant-indicated patients. Published data support any one of three alternative donor HCT options for the patients without matched donors considered optimal.2,4,32-34 Only through the conduct of well-designed clinical trials can we understand and appreciate the complexities of donor choice and their impact on outcome after HCT for AML. Unfortunately, there are no ongoing trials that compare outcomes after MM-URD with that of related mismatched or UCB transplantation. Therefore, in the absence of any prospect of such a comparative study, our data support the use of RIC MM-URD HCT for patients with AML when a suitably matched donor is unavailable.

Overall survival

Leukemia-free survival

Cumulative incidence of relapse

adverse long-term implications of chronic GVHD on survival and quality of life (QOL) this is an important observation.28,29 Chronic GVHD can impair QOL and is associated with significant morbidity and mortality among HCT recipients. However, the costs, economic burden and resource utilizations to manage long term complications associated with cGVHD have not been well described.30 More research is needed to better understand the costs of GVHD to patients, centers and the health care system and to determine if the lower incidence and severity of GVHD with in vivo T-cell depletion leads to long-term resource savings. Recently presented results of a randomized trial within the Blood and Marrow Transplant Clinical Trials Network 0901 showed that RIC regimens result in higher relapse rates and lower TRM compared to MAC, with a statistically significant advantage in relapse-free survival for patients receiving MAC regimens.31 The study is closed to accrual and reports of the study data are unlikely to

Figure 2. Probability of overall survival (OS), leukemia-free survival (LFS), non-relapse mortality (NRM) and relapse incidence (RI) after myeloablative or reduced intensity conditioning regimen for acute myeloid leukemia (age â&#x2030;Ľ50 years) after mismatched unrelated donor transplantation. haematologica | 2016; 101(6)

779

B.N. Savani et al.

Acknowledgments We thank all European Group for Blood and Marrow Transplantation (EBMT) centers and national registries for contributing patients to the study and data managers for their super work. Supplementary information is available at the EBMT Web

References 1. Zuckerman T, Rowe JM. Transplantation in acute myeloid leukemia. Hematol Oncol Clin North Am. 2014;28(6):983-994. 2. Kekre N, Antin JH. Hematopoietic stem cell transplantation donor sources in the 21st century: choosing the ideal donor when a perfect match does not exist. Blood. 2014;124(3):334-343. 3. Kanakry CG, Luznik L. Are alternative donors really still "alternative?". Biol Blood Marrow Transplant. 2014;20(10):1463-1464. 4. Raiola AM, Dominietto A, di Grazia C, et al. Unmanipulated Haploidentical Transplants Compared with Other Alternative Donors and Matched Sibling Grafts. Biol Blood Marrow Transplant. 2014. 5. Pidala J, Kim J, Schell M, et al. Race/ethnicity affects the probability of finding an HLA-A, -B, -C and -DRB1 allele-matched unrelated donor and likelihood of subsequent transplant utilization. Bone Marrow Transplant. 2013;48(3):346-350. 6. Passweg JR, Baldomero H, Bader P, et al. Hematopoietic SCT in Europe 2013: recent trends in the use of alternative donors showing more haploidentical donors but fewer cord blood transplants. Bone Marrow Transplant. 2015;50(4):476-482. 7. Gragert L, Eapen M, Williams E, et al. HLA match likelihoods for hematopoietic stemcell grafts in the U.S. registry. N Engl J Med. 2014;371(4):339-348. 8. Ballen KK, King RJ, Chitphakdithai P, et al. The national marrow donor program 20 years of unrelated donor hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2008;14(9 Suppl):2-7. 9. Sengsayadeth S, Savani BN, Blaise D, Malard F, Nagler A, Mohty M. Reduced intensity conditioning allogeneic hematopoietic cell transplantation for adult acute myeloid leukemia in complete remission - a review from the Acute Leukemia Working Party of the EBMT. Haematologica. 2015;100(7):859-869. 10. Nagler A, Savani BN, Labopin M, et al. Outcomes after use of two standard ablative regimens in patients with refractory acute myeloid leukaemia: a retrospective, multicentre, registry analysis. The Lancet Haematology;2(9):e384-e392. 11. Lee SJ, Klein J, Haagenson M, et al. High-resolution donor-recipient HLA matching contributes to the success of unrelated donor marrow transplantation. Blood. 2007;110(13):4576-4583. 12. Raiola A, Dominietto A, Varaldo R, et al. Unmanipulated haploidentical BMT following non-myeloablative conditioning and post-transplantation CY for advanced Hodgkin's lymphoma. Bone Marrow Transplant. 2014;49(2):190-194. 13. Reshef R, Porter DL. Reduced-intensity conditioned allogeneic SCT in adults with AML. Bone Marrow Transplant. 2015;50(6):759769. 14. Abdul Wahid SF, Ismail NA, Mohd-Idris

780

15.

16.

17.

18.

19.

20.

21.

22.

23.

site. The list of institutions reporting data included in this study are summarized in the Online Supplementary Appendix. BNS thanks Dr. Katy Rezvani, MD, PhD (Houston, TX, USA) and Prof John Greer (Nashville, TN, USA) for critical reading of the manuscript.

MR, et al. Comparison of reduced-intensity and myeloablative conditioning regimens for allogeneic hematopoietic stem cell transplantation in patients with acute myeloid leukemia and acute lymphoblastic leukemia: a meta-analysis. Stem Cells Dev. 2014;23(21):2535-2552. Ringden O, Labopin M, Ehninger G, et al. Reduced intensity conditioning compared with myeloablative conditioning using unrelated donor transplants in patients with acute myeloid leukemia. J Clin Oncol. 2009;27(27):4570-4577. Zeng W, Huang L, Meng F, Liu Z, Zhou J, Sun H. Reduced-intensity and myeloablative conditioning allogeneic hematopoietic stem cell transplantation in patients with acute myeloid leukemia and myelodysplastic syndrome: a meta-analysis and systematic review. Int J Clin Exp Med. 2014;7(11):43574368. Bacigalupo A, Ballen K, Rizzo D, et al. Defining the intensity of conditioning regimens: working definitions. Biol Blood Marrow Transplant. 2009;15(12):1628-1633. Bashey A, Zhang X, Sizemore CA, et al. Tcell-replete HLA-haploidentical hematopoietic transplantation for hematologic malignancies using post-transplantation cyclophosphamide results in outcomes equivalent to those of contemporaneous HLA-matched related and unrelated donor transplantation. J Clin Oncol. 2013;31(10):1310-1316. Aoudjhane M, Labopin M, Gorin NC, et al. Comparative outcome of reduced intensity and myeloablative conditioning regimen in HLA identical sibling allogeneic haematopoietic stem cell transplantation for patients older than 50 years of age with acute myeloblastic leukaemia: a retrospective survey from the Acute Leukemia Working Party (ALWP) of the European group for Blood and Marrow Transplantation (EBMT). Leukemia. 2005;19(12):2304-2312. Shimoni A, Hardan I, Shem-Tov N, Yerushalmi R, Nagler A. Allogeneic hematopoietic stem-cell transplantation in AML and MDS using myeloablative versus reduced-intensity conditioning: long-term follow-up. Leukemia. 2010;24(5):1050-1052. Shimoni A, Hardan I, Shem-Tov N, et al. Allogeneic hematopoietic stem-cell transplantation in AML and MDS using myeloablative versus reduced-intensity conditioning: the role of dose intensity. Leukemia. 2006;20(2):322-328. Luger SM, Ringden O, Zhang MJ, et al. Similar outcomes using myeloablative vs reduced-intensity allogeneic transplant preparative regimens for AML or MDS. Bone Marrow Transplant. 2012;47(2):203211. Martino R, Valcarcel D, Brunet S, Sureda A, Sierra J. Comparable non-relapse mortality and survival after HLA-identical sibling blood stem cell transplantation with reduced or conventional-intensity preparative regimens for high-risk myelodysplasia or acute

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

myeloid leukemia in first remission. Bone Marrow Transplant. 2008;41(1):33-38. Bornhauser M, Kienast J, Trenschel R, et al. Reduced-intensity conditioning versus standard conditioning before allogeneic haemopoietic cell transplantation in patients with acute myeloid leukaemia in first complete remission: a prospective, open-label randomised phase 3 trial. Lancet Oncol. 2012;13(10):1035-1044. Warren EH, Deeg HJ. Dissecting graft-versus-leukemia from graft-versus-host-disease using novel strategies. Tissue Antigens. 2013;81(4):183-193. Stevanovic S, van Schie ML, Griffioen M, Falkenburg JH. HLA-class II disparity is necessary for effective T cell mediated Graftversus-Leukemia effects in NOD/scid mice engrafted with human acute lymphoblastic leukemia. Leukemia. 2013;27(4):985-987. Finke J, Bethge WA, Schmoor C, et al. Standard graft-versus-host disease prophylaxis with or without anti-T-cell globulin in haematopoietic cell transplantation from matched unrelated donors: a randomised, open-label, multicentre phase 3 trial. Lancet Oncol. 2009;10(9):855-864. El-Jawahri A, Pidala J, Inamoto Y, et al. Impact of age on quality of life, functional status, and survival in patients with chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2014;20(9):1341-1348. Savani BN. How can we improve life expectancy and quality of life in long-term survivors after allogeneic stem cell transplantation? Semin Hematol. 2012;49(1):1-3. Majhail NS, Mau LW, Chitphakdithai P, et al. National Survey of Hematopoietic Cell Transplantation Center Personnel, Infrastructure, and Models of Care Delivery. Biol Blood Marrow Transplant. 2015;21(7):1308-1314. Scott BL, Pasquini MC, Logan B, et al. Results of a Phase III Randomized, MultiCenter Study of Allogeneic Stem Cell Transplantation after High Versus Reduced Intensity Conditioning in Patients with Myelodysplastic Syndrome (MDS) or Acute Myeloid Leukemia (AML): Blood and Marrow Transplant Clinical Trials Network (BMT CTN) 0901. Blood. 2015;126(23):LBA8-LBA-8. Weisdorf D, Eapen M, Ruggeri A, et al. Alternative donor transplantation for older patients with acute myeloid leukemia in first complete remission: a center for international blood and marrow transplant researcheurocord analysis. Biol Blood Marrow Transplant. 2014;20(6):816-822. Eapen M, O'Donnell P, Brunstein CG, et al. Mismatched related and unrelated donors for allogeneic hematopoietic cell transplantation for adults with hematologic malignancies. Biol Blood Marrow Transplant. 2014;20(10):1485-1492. Bachanova V, Burns LJ, Wang T, et al. Alternative donors extend transplantation for patients with lymphoma who lack an HLA matched donor. Bone Marrow Transplant. 2015;50(2):197-203.

haematologica | 2016; 101(6)

ARTICLE

Quality of Life

Prospective international validation of the Quality of Life in Myelodysplasia Scale (QUALMS)

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Gregory A. Abel,1,2 Fabio Efficace,3 Rena J. Buckstein,4 Sara Tinsley,5 Joseph G. Jurcic,6 Yolanda Martins,1 David. P Steensma,2 Corey D. Watts,1 Azra Raza,6 Stephanie J. Lee,7 Alan F. List,5 and Robert J. Klaassen8

Division of Population Sciences, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA; 2Center for Leukemia, Department of Medical Oncology, DanaFarber Cancer Institute, Boston, MA, USA; 3Italian Group for Adult Hematologic Diseases (GIMEMA), Data Center and Health Outcomes Research Unit, Rome, Italy; 4Department of Medical Oncology/Hematology, Odette Cancer and Sunnybrook Health Sciences Center, Toronto, Canada; 5Department of Malignant Hematology, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, USA; 6Division of Hematology/Oncology, Department of Medicine, Columbia University Medical Center, New York, NY, USA; 7Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA; and 8Department of Pediatrics, Children's Hospital of Eastern Ontario, Ottawa, ON, Canada 1

Haematologica 2016 Volume 101(6):781-788

ABSTRACT

isease-specific measures of quality of life can improve assessment of disease-related symptoms and psychosocial sequelae. We report on the development and validation of the Quality of Life in Myelodysplasia Scale (QUALMS), a 38-item assessment tool for patients with myelodysplastic syndromes (MDS). In 2014-2015, a multicenter cohort of patients with myelodysplasia completed the QUALMS, as well as the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (QLQ-C30) and the Functional Assessment of Cancer Therapy Anemia Scale (FACT-An); a second administration was undertaken three to six months later. A total of 255 patients from the United States, Canada and Italy participated. Median age was 72 years, 56.1% were men, and the International Prognostic Scoring System distribution was 40.4% low, 42.0% intermediate-1, 13.3% intermediate-2 and 2.3% high. QUALMS scores ranged from 24 to 99 (higher scores are better), with a mean of 67.2 [standard deviation (SD)=15.2]. The measure was internally consistent (α=0.92), and moderately correlated with the multi-item QLQ-C30 scales and the FACT-An (r=-0.65 to 0.79; all P<0.001). Patients with hemoglobin of 8 g/dL or under scored lower than those with hemoglobin over 10 g/dL (61.8 vs. 71.1; P<0.001), and transfusion-dependent patients scored lower than transfusion-independent patients (62.4 vs. 69.7; P<0.01). Principal components analysis revealed “physical burden”, “benefit-finding”, and “emotional burden” subscales. There was good overall test-retest reliability among those with stable hemoglobin (r=0.81), and significant changes for patients hospitalized or with infections between administrations (both P<0.01). These data suggest the QUALMS is a valuable tool for assessing MDS-specific quality of life in the modern treatment era. Introduction The myelodysplastic syndromes (MDS) are a heterogeneous group of acquired hematopoietic stem cell disorders characterized by bone marrow failure and a tendency to transform to acute myeloid leukemia (AML). While supportive care with transfusions, hematopoietic growth factors and antimicrobial agents had long been the standard treatment approach,1 three disease-modifying therapies are now approved by the US Food and Drug Administration (FDA) for use in MDS, and haematologica | 2016; 101(6)

Correspondence: gregory_abel@dfci.harvard.edu

Received: December 1, 2015. Accepted: February 29, 2016. Pre-published: March 4, 2016. doi:10.3324/haematol.2015.140335

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/101/6/781

781

G.A. Abel et al. many other agents are under investigation.2 Unfortunately, although hematopoietic cell transplantation (HCT) is potentially curative, it is not available to many patients due to their advanced age, comorbidities, or lack of an appropriate donor.3 The syndromes also have a variable course, with many patients living relatively symptom-free for many years, and only gradually developing problems such as fatigue, infections, and bleeding. This chronic nature of the syndromes conspires with the lack of curative options to make patients’ quality of life (QOL) a major focus of treatment decisions.4 Given these considerations, and the importance of QOL to patients with MDS,5 its rigorous measurement has been recognized as an MDS research imperative.6-8 Indeed, patients with MDS suffer from a wide variety of symptoms including fatigue, anxiety, insomnia, and dyspnea.9 It is essential to understand such patient-reported outcomes for their own value, and not just for their potential contribution to disease risk. Although an Internet-based survey of MDS patients’ QOL has been published,10 no MDS-specific QOL measure has been widely adopted for clinical or research use. Researchers aiming to assess the impact of MDS and its treatments9,11-18 on QOL have most often used generic measures such as the Short-Form Health Survey (SF-36),19 or cancer-specific scales such as the European Organization for Research and Treatment of Cancer’s QLQ-C3020 and the Functional Assessment of Cancer Therapy-Anemia (FACT-An).21 While these measures are useful, they are not specific enough to contain all of the elements important to MDS-related QOL, and may contain several less relevant items. In contrast to generic QOL questionnaires, disease-specific measures have the potential to more accurately reflect the full breadth of functional limitations and symptoms experienced by specific cancer populations.22 For example, recognition of the need for a disease-specific measure for another bone marrow stem cell disorder, myelofibrosis, led to the creation of the Myeloproliferative Neoplasm Symptom Assessment Form Total Symptom Score (MPN-SAF TSS),23 which has been critical to clinical research efforts for that disease and to regulatory approval of its first specific therapy, ruxolitinib.24,25 In contrast, after many studies with generic instruments, the impact of many routine treatments on the QOL of MDS patients, such as erythropoiesis-stimulating agents26 and red cell transfusions,15 has yet to be clearly demonstrated. We hypothesized that an MDS-specific measure of QOL would allow for a more relevant and complete assessment of the impact of interventions in both clinical and research settings. With that goal in mind, we sought the structured input of patients, caregivers, and health care providers to create and pilot a new MDS-specific measure of QOL: the Quality of Life in Myelodysplasia Scale (QUALMS). This instrument has now undergone prospective validation in an international cohort of patients with MDS, and we report on its validity, reliability and responsiveness.

Methods Instrument development In 2011, using the guidelines outlined by Guyatt,27 and working with the FDA recommendations for developing patient-reported 782

outcome measures,28 we developed the QUALMS29 with structured input from MDS patients, caregivers, and clinicians (n=32) at Dana-Farber Cancer Institute (DFCI) in Boston, MA, USA. We

Table 1. Description of the QUALMS validation cohort. All Age at diagnosis (years) <65 65-69 70-74 75-79 80-84 85+ Male gender White race Married Center Columbia (New York, NY, USA) Dana-Farber (Boston, MA, USA) GIMEMA (Rome and Calgiari, Italy) Moffitt (Tampa, FL, USA) Odette (Toronto, ON, Canada) Current MDS sub-type RA RT RARS RMCD RAEB-1 RAEB-2 MDS-5qMDS-U IPSS LOW INT-1 INT-2 HI IPSS-R Very low Low Intermediate High Very high Treatment (current or ever) Erythropoiesis-stimulating agent (ESA) Hypomethylating agent Lenalidomide RBC transfusion dependent comorbidities 0 1 2 3 4 or more At follow up, patients experiencingb progression to transfusion dependence Hospitalization Infections Bleeding AML Transplant Death

255

100

46 41 58 53 40 17 143 239 168

18 16.1 22.7 20.8 15.7 6.7 56.1 93.7 65.9

30 51 91 32 51

11.8 20 35.7 12.5 20

31 2 32 94 27 29 11 36

12.2 0.8 12.5 36.9 10.6 11.4 4.3 14.1

103 107 34 6

40.4 42 13.3 2.3

22 136 59 16 17

8.8 54.4 23.6 6.4 6.8

140 78 36 75

54.9 30.5 14.2 29.4

95 78 40 18 24

37.3 30.6 15.7 17.1 9.4

16 49 36 20 7 10 9

7.8 23.5 17.3 9.6 3.4 4.8 4.3

Numbers may not add up to 255 (100%) for all variables due to missing data. Since enrollment (n=208 at follow up, median interval of 4.2 months).

haematologica | 2016; 101(6)

Validation of the QUALMS

piloted the instrument with additional DFCI MDS patients (n=20), making several changes, and producing a 38-question measure containing 33 core questions and 5 opt-out questions that took an average of 7.5 minutes to complete. Further development methods are detailed in the Online Supplementary Appendix.

structure to identify potential subscales, and assess its concurrent validity, known groups validity, test-retest reliability (stability) and responsiveness. Each institution obtained study approval from its respective Institutional Review Board before enrolling patients, and all enrolled patients signed informed consent.

Instrument validation: overview

Instrument validation: psychometric analysis

In 2014-2015, we assessed the psychometric properties of the QUALMS in a cohort of patients with biopsy-proven MDS who were not involved in its development. Subjects came from five centers in the United States, Canada, and Italy (Table 1). All filled out the QUALMS, the QLQ-C30, and the FACT-An twice (at baseline and 3-6 months later) accompanied by reviews of their medical record. Additional study details are included in the Online Supplementary Appendix. We used the resulting data to assess the QUALMS’ internal consistency reliability, explore its underlying

Descriptive statistics were performed using baseline data, followed by exploratory principal components analysis (PCA) with oblique rotation to identify the factor structure underlying the baseline QUALMS data and create subscales. We made an a priori decision to retain all factors that accounted for at least 5% of variance, if confirmed by the scree plot. We would then retain all questions with moderately high loading on each factor (r≥0.50) for subscales, and for the overall scale, retain all with modestly high loadings (r≥0.30) on at least one factor.30 We created subscale

Table 2. 3-factor principal components analysis rotated structure matrix loadings and component correlation matrix used to derive the QUALMS subscales.*

QUALMS Items Q24 Q9 Q23 Q26 Q20 Q25 Q11 Q33 Q8 Q7 Q6 Q10 Q18 Q13 (R) Q22 Q29 (R) Q30 (R) Q17 (R) Q31 Q28 Q3 Q4 Q32 Q1 Q14 Q5 Q19 Q27 Q15 Q12 Q2 Q16 Q21

Too tired for prior responsibilities Low energy change schedule Weak Unable participate in activities Take into account might be fatigued Worry about becoming burden Felt hopelessness Change in bowels Shortness of breath Change long-term plans due to health Trouble concentrating Life organized around medical Nauseated Energy for routine tasks Family relationships strained Grateful for tomorrow Get quality information Gratitude when prior took for granted Bruising Avoid crowds Could not do anything about disease Disease unpredictable Lack of concrete answers No clear information Afraid of dying Difficulty explaining MDS to others Worry progressing/leukemia Anxious about tests or lab results Angry about diagnosis Worried infection Limited emotional support available Worried bleeding Concerned financial burden

1: “QUALMS-P”

Component 2: “QUALMS-BF”

3: “QUALMS-E”

0.88 0.83 0.78 0.78 0.75 0.73 0.65 0.63 0.62 0.57 0.57 0.56 0.53 0.52 0.48 0.12 0.09 -0.01 0.32 0.26 0.48 0.40 0.24 0.33 0.32 0.26 0.33 0.46 0.43 0.33 0.37 0.21 0.40

-0.02 0.03 0.09 -0.17 0.02 -0.03 -0.02 -0.16 -0.04 -0.27 0.09 -0.28 -0.11 0.09 -0.08 0.66 0.65 0.57 -0.47 -0.38 0.03 -0.06 -0.09 0.05 -0.20 0.04 -0.19 -0.16 -0.10 -0.42 -0.06 -0.45 -0.17

0.50 0.47 0.34 0.35 0.47 0.51 0.60 0.37 0.38 0.50 0.56 0.42 0.20 0.17 0.47 0.05 0.22 -0.09 0.37 0.37 0.67 0.66 0.65 0.63 0.62 0.61 0.60 0.58 0.58 0.58 0.53 0.48 0.48

*Question numbering reflects the placement of the question in the QUALMS instrument. In bold and italics: items that were used in the calculation of the subscale scores. R: reversescored items.

haematologica | 2016; 101(6)

783

G.A. Abel et al.

scores based on each of the factors. In the rare case that an item loaded moderately highly on more than one factor, we re-examined the inter-item correlations of the subscales, calculated internal consistency reliabilities, and discussed the theoretical implications of including the item in each of the factors before making a final decision about the subscale where the item would reside. Finally, we assessed if the subscales’ internal consistency improved with items removed; if it did, we planned to remove items accordingly. We used correlation analyses to assess the concurrent validity of the QUALMS with other QOL measures that are theoretically related. Specifically, we correlated mean QUALMS scores with scores on the QLQ-C30 and on FACT-An, and used Fisher’s r to z test to examine differences among correlations. We identified groups known to differ on clinical markers (e.g. hemoglobin) and compared scores on the QUALMS to assess known-groups validity. This was completed using t-tests or analyses of variance (ANOVA) f tests. Next, utilizing baseline and follow-up data, we assessed the stability of the QUALMS by correlating the two scores. To assess responsiveness, we compared mean difference in QUALMS scores for patients with significant clinical events since baseline (bleeding, infection or hospitalization) to mean difference in scores for those without. Finally, we conducted exploratory validity analyses for the QUALMS subscales.

Results Subjects Two-hundred and fifty-five MDS patients (56% male) participated, from across five centers (Columbia 12%; Dana-Farber 20%; GIMEMA 36%; Moffitt 12%; Odette 20%). Patients were primarily white (95%), non-Hispanic (95%), and ranged in age from 28 to 92 years (mean=2, SD=0.8). The mean time elapsed between MDS diagnosis and enrolment was 3.6 years. Ninety-two percent of subjects were either fully active, or ambulatory but restricted in strenuous physical activity [Eastern Cooperative Oncology Group Performance Status (ECOG) scores31 of 0 or 1]. Twenty patients had psychiatric comorbidities (defined as depression or anxiety requiring psychiatric counseling or treatment),32 and 29 had a history of a solid malignancy at some point. Twenty-four patients had secondary MDS. Additional baseline characteristics are included in Table 1. Of note, 208 subjects (81.5%) completed a second QUALMS administration after a median interval of 4.3 months.

Descriptive analyses Examination of individual QUALMS items indicated that none had floor or ceiling effects. A missing values analysis showed no identifiable pattern in missing values and indicated that across all 33 core items, fewer than 5% of responses were missing, and for 29 of 33 (88%), there were 2% or fewer with missing data. An analysis of the 5 potential opt-out questions revealed that the range of missing data or opt out was higher, 27% (“too tired to drive”) to 75% (“afraid of losing your job”). We thus retained only the 33 core QUALMS items for analyses after this step. To score the QUALMS, answers for each question (all have 5-point Likert-type answers) were assigned a value with a potential range of 0 (worst) to 100 (best) as follows: Never=100; Rarely=75; Sometimes=50; Often=25 and Always=0. Four items were scored in the opposite direction such that Always=100 and Never=0. The QUALMS 784

total score was calculated by averaging the scores on items 1-33, so the potential range of scores was 0 (worst) to 100 (best). Higher scores mean better QOL. Internal consistency reliability analysis of the QUALMS using the 33 items revealed a Cronbach’s alpha of 0.92. Moreover, we found no further improvement to internal consistency with any items removed, so we retained all items. Overall QUALMS scores ranged from 24 to 99, with a mean score of 67.2 (SD=15.2). No significant differences were found in mean QUALMS scores of patients from the different MDS centers (P=0.09): Columbia 64.5 (±15.7); DFCI 66.5 (±14.5); GIMEMA 66.5 (±16.3); Moffitt 67.0 (±17.0); Odette 72.5 (±11.2).

Exploratory principal components analysis Analysis of sampling adequacy indicated that the QUALMS questions were appropriate for factor analysis (Kaiser-Meyer-Olkin measure of sampling adequacy=0.88; Bartlett’s test of sphericity X2(528)=2889; P<0.01). On the basis of the pre-specified criteria (see Figure 1 for scree plot) the exploratory principal components analysis was constrained to a 3-factor solution that explained 43% of the variance, including “Physical Burden” (30% variance; QUALMS-P), “Benefit Finding” (“silver linings” associated with disease; 7% variance; QUALMS-BF) and “Emotional Burden” (5.5% variance; QUALMS-E). All 33 core items had acceptable factor loadings (r≥0.30)30 on at least one factor and were thus retained for inclusion in the calculation of the overall scale score (Table 2). We also created subscale scores based on each of the factors. Cronbach’s alphas for the final subscales were as follows: QUALMS-P a = 0.91; QUALMS-BF a = 0.62; QUALMS-E a = 0.84. The subscales’ internal consistency did not improve when items were removed, thus all items were retained. Correlation analyses revealed that the overall mean QUALMS had strong positive correlations with both the QUALMS-P (r=0.92, P<0.001) and the QUALMSE (r=0.87, P<0.001), and a small but consistent positive correlation with the QUALMS-BF (r=0.17, P<0.05). The QUALMS-P and QUALMS-E were moderately correlated with each other (r=0.67, P<0.001), but not with the QUALMS-BF [r’s = 0.06 and 0.03, respectively, not significant (ns)].

Table 3. Correlations between overall QUALMS scores, EORTC QLQ30 and FACT-An.

Correlation with QUALMS (r)* EORTC QLQ-C30 Global health Physical function Role function Emotional function Cognitive function Social function Fatigue Nausea Pain FACT-An Fact-An total score Anemia Subscale (AnS) Trial Outcome Index (TOI)

0.59 0.58 0.61 0.68 0.60 0.61 -0.65 -0.37 -0.43 0.79 0.74 0.78

*All correlations were significant at P<0.001 (two-tailed test).

haematologica | 2016; 101(6)

Validation of the QUALMS

Concurrent validity The overall QUALMS score was moderately correlated with the global QLQ-C30 and its eight additional multiitem subscales (r’s=-0.65 to 0.68; P<0.01 for all) (Table 3), and had slightly stronger correlations with the FACT scores (e.g. r’s=0.74 to 0.79; P<0.01 for all).

Known groups validity Patients who were transfusion-dependent had significantly lower overall QUALMS scores (worse QOL) than those who were not transfusion-dependent (Table 4). A similar pattern was seen comparing those who had ever had a transfusion with those who had not, and those who ever had treatment with those who had not. While variability in performance status was low, ECOG scores 2 or greater were highly associated with worse mean QUALMS scores compared to ECOG scores of 0 or 1 (52.2 vs. 68.0; P<0.001). Patients with Hb values greater than 10.0 g/dL had significantly higher scores compared to those who had values between 8.1 g/dL and 10.0 g/dL or 8.0 g/dL or lower. Higher scores were also found in patients who had lower marrow blasts and lower IPSS-R scores, but not for those with platelets greater than 50x109/L or ANC of 1x109/L or greater [an effect was seen at lower platelet (20x109/L) and ANC (0.5x109/L) thresholds; P<0.01 for both)]. Analyses using FACT-An scores as a dependent variable obtained mixed results. Although FACT-An total scores differed between patients with regard to transfusion use, dependence and hemoglobin (Hb) levels, they were not significantly discernible as a function of patients’ blast percentage, whether they ever had treatment, or their IPSS-R categories.

Test-retest reliability The overall QUALMS scores were relatively stable: r=0.76, P<0.001. Stability was strongest among those patients who showed no change in Hb from baseline (within ± 1pt) (r=0.81, P<0.001). Scores among patients whose Hb declined by 1 g/dL or more between baseline and follow up were less stable (r=0.60, P<0.05).

Responsiveness t-test analyses supported responsiveness in overall QUALMS score for patients who experienced infection or hospitalization between baseline and follow up (P<0.01

for both), but not for those who experienced bleeding (Table 5).

Preliminary validation of subscales Concurrent validation analyses showed that QUALMSP had moderately strong correlations with the QLQ-C30’s “Global Health” (r=0.67, P<0.001), “Physical Function” (r=-0.70, P<0.001) and “Fatigue” (r=-0.75, P<0.001) scores, along with the FACT-An (r=0.85, P<0.001). The QUALMS-E had moderate correlations with the QLQC30’s “Emotional Function” subscale (r=-0.58, P<0.001) and the FACT-An (r=0.57, P<0.001). The QUALMS-BF showed small, but consistent correlations with the EORTC QLQ-C30’s “Global Health” score (r=0.19, P<0.01) and the FACT-An (r=0.18, P<0.05). Finally, QUALMS-P scores significantly varied among patients who had different levels of transfusion exposure, dependence, treatment history, Hb, blast cells, IPSS scores, and baseline comorbidities (Table 4).

Discussion We have developed a comprehensive instrument for capturing the critical QOL issues faced by patients with MDS. As a patient-reported outcome measure, the QUALMS demonstrates robust internal consistency, strong concurrent validity, and excellent differentiation between many known groups. It is reliable, and shows good responsiveness for patients who have undergone major clinical events between administrations. It is practical in that it takes less than ten minutes to complete, and contains clinically useful subscales. We were surprised to find a factor emerge in our PCA that contained questions relating to so-called “benefit finding” (the QUALMS-BF). Benefit finding has been studied in other cancers,33-36 where high levels may be associated with coping better and decreased levels of psychosocial stress; however, studies are not conclusive.37 Interestingly, unlike the overall QUALMS, or the QUALMS-P, the QUALMS-BF did not show strong correlations with the other QOL scales studied, arguing that this is a new dimension of MDS-related QOL that has not been previously captured. On the other hand, the QUALMS-BF was also less well-correlated with the overall QUALMS scores, which perhaps reflects a difficulty in assessing this domain

Figure 1. Scree plot of exploratory principal components analysis: eigenvalues as a function of components. A scree plot is used in principal components analysis (PCA) to visually determine which underlying components explain most of the variability in the data. Generally, components are retained on the steep part of the curve, and once the line starts to straighten (after the “elbow”), it is felt that subsequent components do not explain much of the variability. For the QUALMS data, that point was seen after the 3rd component (arrow). haematologica | 2016; 101(6)

785

G.A. Abel et al.

given that there are only 3 contributing questions. Moreover, we do not know how benefit finding may change over the course of disease, whether it may peak soon after diagnosis, or if it may increase temporarily with each new treatment or disease-specific event. While we acknowledge that more work is needed to determine precise thresholds for the overall QUALMS and its subscales, our data suggest that a clinically meaningful difference on the overall QUALMS may be between 5 and 10 points. We were able to distinguish between patients at baseline with differences in this range, and for those patients who had changes on their QUALMS scores at follow up, differences were also in this range. Moreover, a distribution-based method would argue that a meaningful difference would be a half standard deviation,38 which in the case of the QUALMS would be 7.6. In the setting of a clinical trial, it is often desirable to minimize participant response burden when measuring QOL, especially if measured at many time points. We

found that the QUALMS-P, a 14-item subscale focused on physical factors, had excellent internal consistency and performed well, distinguishing between many clinical known groups. In contrast, the 5 final QUALMS questions with opt-out options have not been studied beyond the development phase. Four of these (questions about sexual function, ability to work and drive, and take care of others) are arguably more relevant to younger patients with MDS (patients aged <65 years made up only 18% of our sample). Since little is known about the MDS-related QOL of this group, we suggest that these questions be retained and scored individually; they are clearly an area for further study. We recognize several other areas that merit further investigation. First, while our study included five centers in three countries, we did not undertake a formal crosscultural validation, and the instrument has only been translated into one other language besides English (Italian). Clearly more work is needed to characterize potential cul-

Table 4. Known groups analysis for the overall QUALMS and the QUALMS-P.*

Variable Transfusion dependence No Yes Ever had pRBC transfusion No Yes Blast cell % < 5% 5% – 10% >10% Hemoglobin ≤ 8 g/dL 8.1 – 10 g/dL > 10 g/dL Platelets ≤ 50x109/L > 50 ANC <1x109/L ≥1x109/L Albumin ≤ 4g/dL > 4g/dL IPSS Low Int-1 Int-2 High IPSS-R Very low Low Intermediate High Very high Any MDS treatment None Past or current Comorbidities 0 1 2 or more

Overall QUALMS Mean (SD)

QUALMS-P Mean (SD)

69.7 (14.6) 62.4 (15.8)

<0.01

68.1 (18.5) 55.7 (19.6)

<0.01

70.6 (14.1) 65.2 (15.9)

<0.01

69.7 (17.6) 60.3 (20.2)

<0.01

68.6 (15.3) 65.9 (15.2) 60.1 (13.6)

0.02

65.3 (19.5) 63.5 (20.2) 56.7 (17.3)

<0.01

61.8 (14.8) 64.8 (16.7) 71.1 (15.2)

<0.001

55.1 (19.6) 60.5 (20.6) 70.3 (16.5)

<0.001

64.7 (15.5) 67.7 (15.2)

0.27

61.9 (20.3) 64.6 (19.5)

0.44

66.1 (16.5) 67.6 (14.9)

0.48

64.4 (20.5) 64.1 (19.3)

0.93

62.1 (16.3) 67.5 (15.1)

0.02

57.2 (21.3) 64.6 (19.4)

<0.05

68.8 (15.3) 68.1 (14.3) 59.0 (16.2) 60.3 (7.1)

<0.01

64.6 (19.9) 66.3 (19.0) 54.7 (18.4) 57.7 (14.2)

<0.05

72.3 (13.0) 67.8 (15.3) 65.8 (15.6) 67.2 (14.7) 57.3 (12.5)

0.03

70.6 (14.1) 64.6 (20.1) 63.1 (20.8) 64.6 (16.4) 51.3 (14.8)

<0.05

69.5 (15.7) 65.8 (14.4)

0.06

67.6 (18.8) 62.4 (19.1)

0.04

70.3 (14.2) 66.9 (13.5) 64.1 (17.2)

0.02

69.4 (18.5) 63.3 (16.5) 58.9 (21.9)

<0.01

*All tests were t-tests or ANOVA f-tests.

786

haematologica | 2016; 101(6)

Validation of the QUALMS

tural and race-ethnic differences, and further international validation is ongoing. Second, while we found evidence of responsiveness, a 4-month observational interval is too short to fully assess changes in MDS-related QOL,39 and pre- and post-intervention or longer-term observational studies using the QUALMS will need to be performed. Third, our data were captured with in-person and paper administrations of the QUALMS. Although patients with MDS are in general elderly and may be less comfortable with contemporary technology, electronic and online versions of the QUALMS should be evaluated in the future. Finally, as only 15.6% of patients in our cohort were in higher-risk IPSS categories, we note that further validation work will be necessary to investigate performance of our questionnaire in this higher risk population. This skew toward lower risk participants likely occurred because Intermediate-2 (INT-2) and High (HI) risk patients do not live as long, and are thus less likely to be captured in a study of our duration (enrolling for one year, with 6month follow-up window). Indeed, in the cohort used to create the IPSS-R (n=7008, enrolling over many years, with median follow-up of approx. 4 years),40 there were only 23% INT-2 and HI risk patients, which supports the idea that most people living with MDS are in the risk categories that were more heavily represented in our analysis. Still, while we are confident that our study demonstrates that there will be differences in QUALMS scores as patients with MDS move between lower and higher risk disease states (and potentially back again), we acknowledge that more research is needed to characterize how the measure differs among those with higher-risk disease. In conclusion, testing in a relatively large international cohort of MDS patients appears to show the QUALMS to be a valid measure of MDS-specific QOL. The measure and its subscales have the potential to be used to assess QOL throughout the MDS disease course. We envision

References 7. 1. Greenberg PL, Baer MR, Bennett JM, et al. Myelodysplastic syndromes clinical practice guidelines in oncology. J Natl Compr Canc Netw. 2006;4(1):58-77. 2. Garcia-Manero G. Myelodysplastic syndromes: 2014 update on diagnosis, riskstratification, and management. Am J Hematol. 2014;89(1):97-108. 3. Abel GA, Koreth J. Optimal positioning of hematopoietic stem cell transplantation for older patients with myelodysplastic syndromes. Curr Opin Hematol. 2013;20(2):150-156. 4. Heptinstall K. Quality of life in myelodysplastic syndromes. A special report from the Myelodysplastic Syndromes Foundation, Inc. Oncology (Williston Park). 2008;22(2 Suppl Nurse Ed):13-18; discussion 19. 5. Sekeres MA, Stone RM, Zahrieh D, et al. Decision-making and quality of life in older adults with acute myeloid leukemia or advanced myelodysplastic syndrome. Leukemia. 2004;18(4):809-816. 6. Caocci G, La Nasa G, Efficace F. Healthrelated quality of life and symptom assessment in patients with myelodysplastic syn-

haematologica | 2016; 101(6)

10.

11.

Table 5. Responsiveness analysis for overall QUALMS scores.*

Variable Bleeding No Yes Infection No Yes Hospitalization No Yes

QUALMS Mean difference (SD)

-0.21 (10.3) -0.09 (13.3)

0.96

0.6 (10.6) -5.0 (9.1)

<0.01

0.8 (10.6) -4.5 (9.5)

<0.01

*All tests were t-tests.

the QUALMS as a valuable tool for use in clinical and research settings when evaluating MDS symptoms, making treatment decisions, and informing efficacy and effectiveness results from clinical trials and health services research. Acknowledgments The authors would like to thank Dr. Giovanni Caocci and Dr. Pasquale Niscola, respectively from the Hematology Unit, Department of Internal Medicine, University of Cagliari (Italy) and Department of Hematology, â&#x20AC;&#x153;S. Eugenio Hospitalâ&#x20AC;?, Rome, (Italy), and Dr. Michael Constantine, from the Department of Medical Oncology, Milford Hospital (USA), for their important contribution in enrolling patients for this study. Funding We are grateful for funding from an Aplastic Anemia and MDS International Foundation grant, a Leukemia and Lymphoma Society Clinical Scholar Award, and a quality of life Grant of the Canadian Cancer Society in memory of James Tyrell (grant # 701837).

dromes. Expert Rev Hematol. 2009;2(1):6980. Estey E. Acute myeloid leukemia and myelodysplastic syndromes in older patients. J Clin Oncol. 2007;25(14):19081915. Malcovati L, Hellstrom-Lindberg E, Bowen D, et al. Diagnosis and treatment of primary myelodysplastic syndromes in adults: recommendations from the European LeukemiaNet. Blood. 2013;122(17):29432964. Efficace F, Gaidano G, Breccia M, et al. Prevalence, severity and correlates of fatigue in newly diagnosed patients with myelodysplastic syndromes. Br J Haematol. 2015;168(3):361-370. Steensma DP, Heptinstall KV, Johnson VM, et al. Common troublesome symptoms and their impact on quality of life in patients with myelodysplastic syndromes (MDS): results of a large internet-based survey. Leuk Res. 2008;32(5):691-698. Casadevall N, Durieux P, Dubois S, et al. Health, economic, and quality-of-life effects of erythropoietin and granulocyte colony-stimulating factor for the treatment of myelodysplastic syndromes: a randomized, controlled trial. Blood. 2004;104(2):321-327.

12. Efficace F, Santini V, La Nasa G, et al. Health-related quality of life in transfusiondependent patients with myelodysplastic syndromes: a prospective study to assess the impact of iron chelation therapy. BMJ Support Palliat Care. 2016;6(1):80-88. 13. Kornblith AB, Herndon JE 2nd, Silverman LR, et al. Impact of azacytidine on the quality of life of patients with myelodysplastic syndrome treated in a randomized phase III trial: a Cancer and Leukemia Group B study. J Clin Oncol. 2002;20(10):2441-2452. 14. Lubbert M, Suciu S, Baila L, et al. Low-dose decitabine versus best supportive care in elderly patients with intermediate- or highrisk myelodysplastic syndrome (MDS) ineligible for intensive chemotherapy: final results of the randomized phase III study of the European Organisation for Research and Treatment of Cancer Leukemia Group and the German MDS Study Group. J Clin Oncol. 2011;29(15):1987-1996. 15. Pinchon DJ, Stanworth SJ, Doree C, Brunskill S, Norfolk DR. Quality of life and use of red cell transfusion in patients with myelodysplastic syndromes. A systematic review. Am J Hematol. 2009;84(10):671677. 16. Revicki DA, Brandenburg NA, Muus P, Yu R, Knight R, Fenaux P. Health-related quali-

787

G.A. Abel et al.

17.

18.

19.

20.

21.

22. 23.

788

ty of life outcomes of lenalidomide in transfusion-dependent patients with Low- or Intermediate-1-risk myelodysplastic syndromes with a chromosome 5q deletion: results from a randomized clinical trial. Leuk Res. 2013;37(3):259-265. Ryblom H, Hast R, Hellstrom-Lindberg E, Winterling J, Johansson E. Self-perception of symptoms of anemia and fatigue before and after blood transfusions in patients with myelodysplastic syndromes. Eur J Oncol Nurs. 2015;19(2):99-106. Efficace F, Gaidano G, Breccia M, et al. Prognostic value of self-reported fatigue on overall survival in patients with myelodysplastic syndromes: a multicentre, prospective, observational, cohort study. Lancet Oncol. 2015;16(15):1506-1514. McHorney CA, Ware JE Jr, Lu JF, Sherbourne CD. The MOS 36-item ShortForm Health Survey (SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups. Med Care. 1994;32(1):40-66. Aaronson NK, Ahmedzai S, Bergman B, et al. The European Organization for Research and Treatment of Cancer QLQC30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst. 1993;85(5):365-376. Cella D. The Functional Assessment of Cancer Therapy-Anemia (FACT-An) Scale: a new tool for the assessment of outcomes in cancer anemia and fatigue. Semin Hematol. 1997;34(3 Suppl 2):13-19. Velikova G, Stark D, Selby P. Quality of life instruments in oncology. Eur J Cancer. 1999;35(11):1571-1580. Emanuel RM, Dueck AC, Geyer HL, et al. Myeloproliferative neoplasm (MPN) symptom assessment form total symptom score:

24.

25.

26.

27. 28.

29.

30. 31.

32.

prospective international assessment of an abbreviated symptom burden scoring system among patients with MPNs. J Clin Oncol. 2012;30(33):4098-4103. Mesa RA, Gotlib J, Gupta V, et al. Effect of ruxolitinib therapy on myelofibrosis-related symptoms and other patient-reported outcomes in COMFORT-I: a randomized, double-blind, placebo-controlled trial. J Clin Oncol. 2013;31(10):1285-1292. Verstovsek S, Mesa RA, Gotlib J, et al. A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N Engl J Med. 2012;366(9):799-807. Rizzo JD, Somerfield MR, Hagerty KL, et al. Use of epoetin and darbepoetin in patients with cancer: 2007 American Society of Hematology/American Society of Clinical Oncology clinical practice guideline update. Blood. 2008;111(1):25-41. Guyatt GH, Bombardier C, Tugwell PX. Measuring disease-specific quality of life in clinical trials. CMAJ. 1986;134(8):889-895. Guidance for Industry. Patient-Reported Outcome Measures: Use in Medical Product Development to Support Labeling Claims. Rockville, MD: U.S. Department of Health and Human Services, Food and Drug Administration, 2009. Abel GA, Klaassen R, Lee SJ, et al. Patientreported outcomes for the myelodysplastic syndromes: a new MDS-specific measure of quality of life. Blood. 2014;123(3):451-452. Nunnally J, Bertnstein I. Psychometric Theory New York: McGraw-Hill, 1994. Oken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5(6):649-655. Sorror ML, Maris MB, Storb R, et al. Hematopoietic cell transplantation (HCT)-

33.

34.

35.

36.

37.

38.

39.

40.

specific comorbidity index: a new tool for risk assessment before allogeneic HCT. Blood. 2005;106(8):2912-2919. Cavell S, Broadbent E, Donkin L, Gear K, Morton RP. Observations of benefit finding in head and neck cancer patients. Eur Arch Otorhinolaryngol. 2016;273(2)479-485. Jansen L, Hoffmeister M, Chang-Claude J, Brenner H, Arndt V. Benefit finding and post-traumatic growth in long-term colorectal cancer survivors: prevalence, determinants, and associations with quality of life. Br J Cancer. 2011;105(8):1158-1165. Mols F, Vingerhoets AJ, Coebergh JW, van de Poll-Franse LV. Well-being, posttraumatic growth and benefit finding in long-term breast cancer survivors. Psychol Health. 2009;24(5):583-595. Rinaldis M, Pakenham KI, Lynch BM. Relationships between quality of life and finding benefits in a diagnosis of colorectal cancer. Br J Psychol. 2010;101(Pt 2):259275. Pascoe L, Edvardsson D. Benefit finding in cancer: a review of influencing factors and health outcomes. Eur J Oncol Nurs. 2013;17(6):760-766. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003;41(5):582-592. Buckstein R, Alibhai S, Khalaf D, et al. The Independent Effects of Frailty and Comorbidity On the Quality of Life in MDS Patients. Blood. 2012; 120 (21):699.(Abstract). Greenberg PL, Tuechler H, Schanz J, et al. Revised international prognostic scoring system for myelodysplastic syndromes. Blood. 2012;120(12):2454-2465

haematologica | 2016; 101(6)