Page 1


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

Editor-in-Chief Jan Cools (Leuven)

Deputy Editor Luca Malcovati (Pavia)

Managing Director Antonio Majocchi (Pavia)

Associate Editors Hélène Cavé (Paris), Ross Levine (New York), Claire Harrison (London), Pavan Reddy (Ann Arbor), Andreas Rosenwald (Wuerzburg), Juerg Schwaller (Basel), Monika Engelhardt (Freiburg), Wyndham Wilson (Bethesda), Paul Kyrle (Vienna), Paolo Ghia (Milan), Swee Lay Thein (Bethesda), Pieter Sonneveld (Rotterdam)

Assistant Editors Anne Freckleton (English Editor), Cristiana Pascutto (Statistical Consultant), Rachel Stenner (English Editor), Kate O’Donohoe (English Editor), Ziggy Kennell (English Editor)

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

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

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


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

Information for readers, authors and subscribers Haematologica (print edition, pISSN 0390-6078, eISSN 1592-8721) publishes peer-reviewed papers on all areas of experimental and clinical hematology. The journal is owned by a non-profit organization, the Ferrata Storti Foundation, and serves the scientific community following the recommendations of the World Association of Medical Editors (www.wame.org) and the International Committee of Medical Journal Editors (www.icmje.org). Haematologica publishes editorials, research articles, review articles, guideline articles and letters. Manuscripts should be prepared according to our guidelines (www.haematologica.org/information-for-authors), and the Uniform Requirements for Manuscripts Submitted to Biomedical Journals, prepared by the International Committee of Medical Journal Editors (www.icmje.org). Manuscripts should be submitted online at http://www.haematologica.org/. Conflict of interests. According to the International Committee of Medical Journal Editors (http://www.icmje.org/#conflicts), “Public trust in the peer review process and the credibility of published articles depend in part on how well conflict of interest is handled during writing, peer review, and editorial decision making”. The ad hoc journal’s policy is reported in detail online (www.haematologica.org/content/policies). Transfer of Copyright and Permission to Reproduce Parts of Published Papers. Authors will grant copyright of their articles to the Ferrata Storti Foundation. No formal permission will be required to reproduce parts (tables or illustrations) of published papers, provided the source is quoted appropriately and reproduction has no commercial intent. Reproductions with commercial intent will require written permission and payment of royalties. Detailed information about subscriptions is available online at www.haematologica.org. Haematologica is an open access journal. Access to the online journal is free. Use of the Haematologica App (available on the App Store and on Google Play) is free. For subscriptions to the printed issue of the journal, please contact: Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, E-mail: info@haematologica.org). Rates of the International edition for the year 2017 are as following: Print edition

Institutional Euro 500

Personal Euro 150

Advertisements. Contact the Advertising Manager, Haematologica Office, via Giuseppe Belli 4, 27100 Pavia, Italy (phone +39.0382.27129, fax +39.0382.394705, e-mail: marketing@haematologica.org). Disclaimer. Whilst every effort is made by the publishers and the editorial board to see that no inaccurate or misleading data, opinion or statement appears in this journal, they wish to make it clear that the data and opinions appearing in the articles or advertisements herein are the responsibility of the contributor or advisor concerned. Accordingly, the publisher, the editorial board and their respective employees, officers and agents accept no liability whatsoever for the consequences of any inaccurate or misleading data, opinion or statement. Whilst all due care is taken to ensure that drug doses and other quantities are presented accurately, readers are advised that new methods and techniques involving drug usage, and described within this journal, should only be followed in conjunction with the drug manufacturer’s own published literature. Direttore responsabile: Prof. Edoardo Ascari; Autorizzazione del Tribunale di Pavia n. 63 del 5 marzo 1955. Printing: Tipografia PI-ME, via Vigentina 136, Pavia, Italy. Printed in March 2017.


haematologica calendar of events

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

EHA Tutorial on Laboratory Hematology Chairs: MC Ar, T Celkan, S McCann, T Patıroğlu April 8-9, 2017 Çanakkale, Turkey

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

Hematology Society of Taiwan - Joint Symposium & EHA Special Lecture April 15-16, 2017 Taipei, Taiwan

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

Annual Meeting PETHEMA Group PETHEMA Foundation May 25-27, 2017 Caceres, Spain

Korean Society of Hematology - Joint Symposium May 26-27, 2017 Seoul, South Korea

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

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

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

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

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

Calendar of Events updated on March 1, 2017


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

Table of Contents Volume 102, Issue 4: April 2017 Cover Figure Late effects of blood and marrow transplantation (image generated by www.somersault1824.com)

Editorials 611

Acute lymphoblastic leukemia of the central nervous system: on the role of PBX1 Ameera Alsadeq and Denis M. Schewe

Review Article 614

Late effects of blood and marrow transplantation Yoshihiro Inamoto and Stephanie J. Lee

Guideline Article 626

Risk factors for mortality in adult patients with sickle cell disease: a meta-analysis of studies in North America and Europe Poulami Maitra et al.

Articles Hematopoiesis

637

Dual role of IL-21 in megakaryopoiesis and platelet homeostasis Salima Benbarche et al.

647

A population of hematopoietic stem cells derives from GATA4-expressing progenitors located in the placenta and lateral mesoderm of mice Ana Cañete et al.

Lysosomal Storage Disease

656

Efferocytosis is impaired in Gaucher macrophages Elma Aflaki et al.

Red Cell Biology & its Disorders

666

Associations between environmental factors and hospital admissions for sickle cell disease Frédéric B. Piel et al.

676

Erythrocyte survival is controlled by microRNA-142 Natalia Rivkin et al.

Hemostasis

686

Factor VIII/V C-domain swaps reveal discrete C-domain roles in factor VIII function and intracellular trafficking Eduard H.T.M. Ebberink et al.

Platelet Biology & its Disorders

695

The transcription factor GATA1 regulates NBEAL2 expression through a long-distance enhancer Anouck Wijgaerts et al.

Haematologica 2017; vol. 102 no. 4 - April 2017 http://www.haematologica.org/


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

707

CD244 maintains the proliferation ability of leukemia initiating cells through SHP-2/p27kip1 signaling Feifei Zhang et al.

719

Phase I study of the aurora A kinase inhibitor alisertib with induction chemotherapy in patients with acute myeloid leukemia Amir T. Fathi et al.

728

Lenalidomide combined with intensive chemotherapy in acute myeloid leukemia and higher-risk myelodysplastic syndrome with 5q deletion. Results of a phase II study by the Groupe Francophone Des MyĂŠlodysplasies Lionel Ades et al.

Acute Lymphoblastic Leukemia

736

CREBBP knockdown enhances RAS/RAF/MEK/ERK signaling in Ras pathway mutated acute lymphoblastic leukemia but does not modulate chemotherapeutic response Zach A. Dixon et al.

Chronic Lymphocytic Leukemia

746

miR-125b and miR-532-3p predict the efficiency of rituximab-mediated lymphodepletion in chronic lymphocytic leukemia patients. A French Innovative Leukemia Organization study Anne-Laure Gagez et al.

Non-Hodgkin Lymphoma

755

Inhibition of 4EBP phosphorylation mediates the cytotoxic effect of mechanistic target of rapamycin kinase inhibitors in aggressive B-cell lymphomas Chengfeng Bi et al.

765

Safety and efficacy of obinutuzumab with CHOP or bendamustine in previously untreated follicular lymphoma Andrew Grigg et al.

Plasma Cell Disorders

773

IL21R expressing CD14+CD16+ monocytes expand in multiple myeloma patients leading to increased osteoclasts Marina Bolzoni et al.

785

Serum B-cell maturation antigen: a novel biomarker to predict outcomes for multiple myeloma patients Michael Ghermezi et al.

Cell Therapy & Immunotherapy

796

Increased age-associated mortality risk in HLA-mismatched hematopoietic stem cell transplantation Daniel FĂźrst et al.

Letters to the Editor Letters are available online only at www.haematologica.org/content/102/4.toc

e120

Plerixafor and G-CSF combination mobilizes hematopoietic stem and progenitors cells with a distinct transcriptional profile and a reduced in vivo homing capacity compared to plerixafor alone Maria Rosa Lidonnici et al. http://www.haematologica.org/content/102/4/e120

e125

Acute myeloid leukemias with ring sideroblasts show a unique molecular signature straddling secondary acute myeloid leukemia and de novo acute myeloid leukemia Pedro Martin-Cabrera et al. http://www.haematologica.org/content/102/4/e125

Haematologica 2017; vol. 102 no. 4 - April 2017 http://www.haematologica.org/


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

e129

Impact of FLT3-ITD diversity on response to induction chemotherapy in patients with acute myeloid leukemia Mike Fischer et al. http://www.haematologica.org/content/102/4/e129

e132

Long-term relapse-free survival in a phase 2 study of blinatumomab for the treatment of patients with minimal residual disease in B-lineage acute lymphoblastic leukemia Nicola GĂśkbuget et al. http://www.haematologica.org/content/102/4/e132

e136

The central nervous system microenvironment influences the leukemia transcriptome and enhances leukemia chemo-resistance Jeffrey S. Gaynes et al. http://www.haematologica.org/content/102/4/e136

e140

Peripheral neuropathies in chronic lymphocytic leukemia: a single center experience on 816 patients Chiara Briani et al. http://www.haematologica.org/content/102/4/e140

e144

Integration of B-cell receptor-induced ERK1/2 phosphorylation and mutations of SF3B1 gene refines prognosis in treatment-naĂŻve chronic lymphocytic leukemia Chiara Cavallini et al http://www.haematologica.org/content/102/4/e144

e148

Integrative clinicopathological and molecular analyses of angioimmunoblastic T-cell lymphoma and other nodal lymphomas of follicular helper T-cell origin Maria Pamela Dobay et al. http://www.haematologica.org/content/102/4/e148

e152

Hematopoiesis in patients with mature B-cell malignancies is deregulated even in patients with undetectable bone marrow involvement Bokang Calvin Lenyeletse Maswabi et al. http://www.haematologica.org/content/102/4/e152

e156

Efficacy and safety of idelalisib in patients with relapsed, rituximab- and alkylating agent-refractory follicular lymphoma: a subgroup analysis of a phase 2 study Gilles Salles et al. http://www.haematologica.org/content/102/4/e156

Case Report Case Reports are available online only at www.haematologica.org/content/102/4.toc

e160

Dasatinib and allogeneic stem cell transplantation enable sustained response in an elderly patient with RCSD1-ABL1-positive acute lymphoblastic leukemia Miriam Frech et al. http://www.haematologica.org/content/102/4/e160

Haematologica 2017; vol. 102 no. 4 - April 2017 http://www.haematologica.org/


EDITORIALS Acute lymphoblastic leukemia of the central nervous system: on the role of PBX1 Ameera Alsadeq1 and Denis M. Schewe2 1

Institute of Immunology, University Medical Center Ulm and 2Department of Pediatrics I, ALL-BFM Study Group, Christian-Albrechts University Kiel and University Medical Center Schleswig-Holstein, Kiel, Germany E-mail: denis.schewe@uksh.de

C

doi:10.3324/haematol.2017.165142

entral nervous system (CNS) infiltration is rarely detected at initial diagnosis of pediatric acute lymphoblastic leukemia (ALL). This is mainly due to a lack of sensitivity of cerebrospinal fluid (CSF) diagnostics by cytology. Currently, all patients receive extensive CNS-directed chemotherapy regardless of the presence of leukemic cells in the CSF, which is associated with a number of neurological toxicities. Some high-risk patients also receive cranial irradiation, which carries an increased risk for secondary malignancies. On the other hand, omitting this type of CNS-directed therapy will lead to CNS relapses in a vast majority of the patients1 suggesting that CNS tropism is an omnipresent feature in ALL blasts. Even though the prognosis of pediatric ALL has been continuously improving on contemporary treatment protocols, the CNS compartment is affected in roughly one-third of ALL relapses.2 Importantly, CNS involvement at relapse occurs mainly in patients who were CNS negative at initial diagnosis.3 This suggests that ALL cells that are refractory to therapy or particularly receptive to microenvironment-derived protective signals are able to survive in the CNS niche for prolonged periods of time as extramedullary minimal residual disease. In order to solve the mechanistic problem of CNS disease in ALL, factors influencing the homing and the survival of leukemic cells in this protective sanctuary are increasingly being investigated. In this issue of Haematologica, Gaynes et al. report a set of CNS-related genes identified by a screening approach in NALM-6 B-cell precursor (BCP)-ALL cells recovered from the bone marrow and CNS compartments of NOD/SCIDgamma (NSG) mice.4 They further investigated the role of the homeobox gene PBX1, one of the genes up-regulated in the CNS by using BCP-ALL cell lines in a co-culture model of the blood brain barrier (BBB) and in xenografts. They found PBX1 protein levels to be up-regulated in different BCP-ALL cell lines in co-culture with murine choroid plexus cells and in the CNS of xenografts. In co-culture, PBX1 conferred chemotherapy resistance to cytarabine and methotrexate, both of which are important drugs for CNS-directed therapy in ALL. Targeting PBX1 by RNA interference decreased the protective effects induced in this co-culture model pointing at a role of PBX1 as a niche-specific survival factor. As expected, ectopic expression of PBX1 enhanced the colony-forming ability of ALL cell lines in vitro and, importantly, CNS infiltration in vivo. PBX1 is known to support long-term self-renewal abilities in hematopoietic stem cells and, as a translocation partner for E2A in t(1;19) positive BCP-ALL, it leads to an arrest of B cells at the pro-/pre-B-stage and is also able to enhance self-renewal capacity in pre-leukemic B-cell progenitors.5 It has been previously shown that enforced expression of PBX1 has no transforming potential in NIH3T3 cells.6 In contrast, the oncogenic potential of the E2A-PBX1 fusion has been demonstrated in a variety of experimental models.5,7 Interestingly, E2APBX1 rearranged BCP-ALL has a particular propensity to enter the CNS.8 The E2A-PBX1 fusion includes almost the haematologica | 2017; 102(4)

entire coding sequence of PBX1, is predominantly localized in the nucleus, and acts as a constitutive transcriptional activator.9 E2A-PBX1 binds a subset of targets also bound by PBX1, raising the possibility that both regulate similar genes, which may also be relevant for CNS disease in ALL. The TAMreceptor MER is over-expressed in E2A-PBX1 rearranged BCP-ALL and has been found to mediate a quiescent and chemo-resistant phenotype in leukemic cells in the CNS of pediatric t(1;19) positive patients.10 There is, however, no evidence for a direct regulation of MER by either E2A-PBX1 or PBX1. In contrast, E2A-PBX1 was found to induce pre-B-cell receptor (pre-BCR) signaling by activating the transcription of the pre-BCR kinases SYK, LCK and ZAP70 in a subset of murine leukemias in a transgenic E2A-PBX1 BCP-ALL mouse model5,11 and in human BCP-ALL.12 The E2A-PBX1 transgene caused the acquisition of a number of secondary activating mutations, most notably in the JAK/STAT and RAS pathways, suggesting that E2A-PBX1 induces genomic instability.5 Interestingly, activated JAK/STAT signaling and constitutive activation of the RAF/RAS/MEK/ERK pathway may be distinguishing features of certain leukemias invading the CNS: interleukin-15, shown to be predictive of CNS involvement and relapse in patients,13 induces both ERK1/2 and STAT5, thereby up-regulating leukocyte trafficking molecules.14 Activating mutations in the RAS pathway have been found to be highly prevalent in relapsing childhood ALL and led to CNS infiltration in xenografts which could be abrogated by MEK1/2 inhibition.15 Recently, ZAP70 kinase has been associated with CNS disease in ALL, in part via ERK1/2-mediated activation of the chemokine receptors CCR7 and CXCR4.16 A summary of the signaling pathways induced in E2A-PBX1 positive BCP-ALL and their potential role in CNS leukemia is presented in Figure 1. Whether CNS leukemogenesis by PBX1 without the E2A-PBX1 rearrangement is based on similar molecular mechanisms remains to be investigated. It has recently been suggested that CNS infiltration is a universal feature of leukemic blasts based on pre-clinical experimental data.17 While it is highly possible that the majority of ALL blasts are able to cross the BBB and home to the CNS, pre-clinical and clinical data suggest that ALL sub-clones expressing niche-specific survival factors may be more relevant from a mechanistic point of view as these cells may be quiescent and chemotherapy-resistant (Figure 2, model 1). On the other hand, it is also possible that only ALL cells expressing specific homing markers can enter the CNS compartment and that they activate pro-survival signaling in that niche in a second step (Figure 2, model 2). Whether PBX1-mediated CNS leukemia reflects the first or the latter mechanism remains to be elucidated using further in vivo modeling, primary patient samples, and clinical cohorts. The work of Gaynes et al., together with our own work and that of others, illustrates a number of challenges in the field of CNS leukemia research. First, there have been attempts to identify target genes in 611


Editorials

Figure 1. Surface receptors and signaling pathways involved in central nervous system (CNS) leukemia. E2APBX1 transcriptional targets are depicted in green. Many of the E2A-PBX1 target genes are important for pre-BCR signaling, which has also been connected to CNS leukemia. Molecules for which there is direct evidence for a role in CNS leukemia are illustrated with red stars. The overlap between genes regulated by E2A-PBX1 and by PBX1 alone remains unclear.

Figure 2. Models of central nervous system (CNS) leukemia. In model 1, every acute lymphoblastic leukemia (ALL) cell has the ability to migrate to the CNS, but only cells with up-regulated survival pathways cause CNS leukemia. In model 2, ALL cells with up-regulated homing markers (eg. chemokine receptors) can migrate to the CNS. In a second step, they activate the survival pathways necessary to maintain CNS disease. It is essential to clarify which model is true, as this has important implications for CNS-directed prophylaxis and therapy in ALL.

612

haematologica | 2017; 102(4)


Editorials

blasts directly isolated from patient CSF which could confirm, among others, JAK/STAT pathway18,19 as well as ERK and BCR activation.19 Due to limited availability of material and restricted cell integrity in primary patient CSF, model systems will increasingly be needed. Co-culture models of the BBB naturally do not reflect the complexity of its composition, and the leukemic BBB may be entirely different from a healthy one. Clearly, immunodeficiency of xenograft mice can affect the modeling of CNS leukemia20 and high degrees of immunodeficiency may favor heavy ALL infiltration in all target organs, including the CNS. Also, modeling of the leukemic CNS compartment in xenografts may have to take place during chemotherapy as it remains completely unclear at which time point before, during or after therapy initiation CNS infiltration occurs in patients. Second, the functions of novel biomarkers indicating CNS leukemia are sometimes not easy to distinguish. It has been shown that chemokine receptors are responsible for the homing of mainly T-ALL cells into the CNS,16,21 and this is also possible, although a lot less clear, in BCP-ALL. In fact, similarly to the situation for PBX1 in the current study, it is often obscure as to whether a parameter influences the homing and/or the survival of ALL cells in the CNS niche or both. Third, parameters identified in a pre-clinical model need to be confirmed in clinical patient series. Due to the rare detection of CNS infiltration at diagnosis, patient numbers are often too small to be able to come to any valid conclusions on a specific marker. Even patient cohorts of large study groups are biased due to selection of CNS positive patients. It is important that a CNS biomarker reliably diagnoses or excludes CNS infiltration. It must also be very clear if a novel biomarker indicates CNS infiltration at initial diagnosis or if it is predictive for CNS relapse. Both scenarios may be fundamentally different regarding their mechanism. A fourth and final point is that diagnostic strategies for the detection of CNS infiltration need to be refined independently of novel biomarkers. Microscopic examinations of CSF samples are the conventional gold standard; however, novel techniques such as flow cytometry in the CSF and genetic methods for the detection of minimal residual disease in the CNS compartment need to be thoroughly evaluated. Taken together this work provides important insights into the biology of CNS infiltration in ALL which is important in the context of the previous findings on CNS leukemia of the past few years. PBX1 is established as a novel mechanism of survival and chemotherapy resistance for ALL in the CNS compartment independent of the E2APBX1 translocation but with important implications for this entity. Work like this is fundamental in order to unravel the molecular mechanisms of CNS leukemia to ultimately identify two patient subgroups: 1) patients not needing such an intensive CNS-directed therapy, thereby avoiding significant toxicity; and 2) patients with ALL cells prone to survive in the CNS. These are patients who may benefit from intensive CNS-directed therapy and novel, more specific, approaches in the future.

haematologica | 2017; 102(4)

Acknowledgments DMS is supported by the Max-Eder group leader program by the Deutsche Krebshilfe e. V. and the Wilhelm-Sander Stiftung.

References 1. Evans AE, Gilbert ES, Zandstra R. The increasing incidence of central nervous system leukemia in children. (Children's Cancer Study Group A). Cancer. 1970;26(2):404-409. 2. Krishnan S, Wade R, Moorman AV, et al. Temporal changes in the incidence and pattern of central nervous system relapses in children with acute lymphoblastic leukaemia treated on four consecutive Medical Research Council trials, 1985-2001. Leukemia. 2010;24(2):450-459. 3. Burger B, Zimmermann M, Mann G, et al. Diagnostic cerebrospinal fluid examination in children with acute lymphoblastic leukemia: significance of low leukocyte counts with blasts or traumatic lumbar puncture. J Clin Oncol. 2003;21(2):184-188. 4. Gaynes JS, Jonart LM, Zamora EA, Naumann JA, Gossai NP, Gordon PM. The central nervous system microenvironment influences the leukemia transcriptome and enhances leukemia chemo-resistance. Haematologica. 2016 Dec 29. [Epub ahead of print] 5. Duque-Afonso J, Feng J, Scherer F, et al. Comparative genomics reveals multistep pathogenesis of E2A-PBX1 acute lymphoblastic leukemia. J Clin Invest. 2015;125(9):3667-3680. 6. Monica K, LeBrun DP, Dedera DA, Brown R, Cleary ML. Transformation properties of the E2a-Pbx1 chimeric oncoprotein: fusion with E2a is essential, but the Pbx1 homeodomain is dispensable. Mol Cell Biol. 1994;14(12):8304-8314. 7. LeBrun DP. E2A basic helix-loop-helix transcription factors in human leukemia. Front Biosci. 2003;8:s206-222. 8. Jeha S, Pei D, Raimondi SC, et al. Increased risk for CNS relapse in preB cell leukemia with the t(1;19)/TCF3-PBX1. Leukemia. 2009;23(8):14061409. 9. Aspland SE, Bendall HH, Murre C. The role of E2A-PBX1 in leukemogenesis. Oncogene. 2001;20(40):5708-5717. 10. Krause S, Pfeiffer C, Strube S, et al. Mer tyrosine kinase promotes the survival of t(1;19)-positive acute lymphoblastic leukemia (ALL) in the central nervous system (CNS). Blood. 2015;125(5):820-830. 11. Duque-Afonso J, Lin CH, Han K, et al. E2A-PBX1 Remodels Oncogenic Signaling Networks in B-cell Precursor Acute Lymphoid Leukemia. Cancer Res. 2016;76(23):6937-6949. 12. Geng H, Hurtz C, Lenz KB, et al. Self-enforcing feedback activation between BCL6 and pre-B cell receptor signaling defines a distinct subtype of acute lymphoblastic leukemia. Cancer Cell. 2015;27(3):409-425. 13. Cario G, Izraeli S, Teichert A, et al. High interleukin-15 expression characterizes childhood acute lymphoblastic leukemia with involvement of the CNS. J Clin Oncol. 2007;25(30):4813-4820. 14. Williams MT, Yousafzai Y, Cox C, et al. Interleukin-15 enhances cellular proliferation and upregulates CNS homing molecules in pre-B acute lymphoblastic leukemia. Blood. 2014;123(20):3116-3127. 15. Irving J, Matheson E, Minto L, et al. Ras pathway mutations are prevalent in relapsed childhood acute lymphoblastic leukemia and confer sensitivity to MEK inhibition. Blood. 2014;124(23):3420-3430. 16. Alsadeq A, Fedders H, Vokuhl C, et al. The role of ZAP70 kinase in acute lymphoblastic leukemia infiltration into the central nervous system. Haematologica. 2017;102(2):346-355. 17. Williams MT, Yousafzai YM, Elder A, et al. The ability to cross the blood-cerebrospinal fluid barrier is a generic property of acute lymphoblastic leukemia blasts. Blood. 2016;127(16):1998-2006. 18. van der Velden VH, de Launaij D, de Vries JF, et al. New cellular markers at diagnosis are associated with isolated central nervous system relapse in paediatric B-cell precursor acute lymphoblastic leukaemia. Brit J Haematol. 2016;172(5):769-781. 19. Hicks C, Sitthi-Amorn J, Douglas J, et al. Molecular Analysis of Central Nervous System Disease Spectrum in Childhood Acute Lymphoblastic Leukemia. Clin Med Insights Oncol. 2016;10:5-15. 20. Frishman-Levy L, Shemesh A, Bar-Sinai A, et al. Central nervous system acute lymphoblastic leukemia: role of natural killer cells. Blood. 2015;125(22):3420-3431. 21. Buonamici S, Trimarchi T, Ruocco MG, et al. CCR7 signalling as an essential regulator of CNS infiltration in T-cell leukaemia. Nature. 2009;459(7249):1000-1004.

613


REVIEW ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Late effects of blood and marrow transplantation Yoshihiro Inamoto1 and Stephanie J. Lee2

Department of Hematopoietic Stem Cell Transplantation, National Cancer Center Hospital, Tokyo, Japan and 2Clinical Research Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA

1

Haematologica 2017 Volume 102(4):614-625

Correspondence: yinamoto@fredhutch.org

Received: November 24, 2016. Accepted: January 20, 2017. Pre-published: February 23, 2017. doi:10.3324/haematol.2016.150250 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/614 Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

614

ABSTRACT

H

ematopoietic cell transplantation is a curative treatment for a variety of hematologic diseases. Advances in transplantation technology have reduced early transplant-related mortality and expanded application of transplantation to older patients and to a wider variety of diseases. Management of late effects after transplantation is increasingly important for a growing number of long-term survivors that is estimated to be half a million worldwide. Many studies have shown that transplant survivors suffer from significant late effects that adversely affect morbidity, mortality, working status and quality of life. Late effects include diseases of the cardiovascular, pulmonary, and endocrine systems, dysfunction of the thyroid gland, gonads, liver and kidneys, infertility, iron overload, bone diseases, infection, solid cancer, and neuropsychological effects. The leading causes of late mortality include recurrent malignancy, lung diseases, infection, secondary cancers and chronic graftversus-host disease. The aim of this review is to facilitate better care of adult transplant survivors by summarizing accumulated evidence, new insights, and practical information about individual late effects. Further research is needed to understand the biology of late effects allowing better prevention and treatment strategies to be developed.

Introduction Hematopoietic cell transplantation (HCT) is a curative treatment for a variety of hematologic diseases.1 The safety of HCT has improved over the decades,2 indications for HCT have expanded to older patients,3 and almost all patients are able to find suitable allogeneic donors by the growing use of cord blood4 and haploidentical transplantation.5 These current conditions have contributed to a growing number of HCT survivors, estimated to be half a million worldwide.6 Patients who are disease-free at two or five years after HCT have a greater than 80% subsequent 10-year survival rate,7-10 but many studies show that HCT survivors suffer from significant late effects that adversely affect morbidity, mortality, working status and quality of life.7-13 A prospective observational study of 1022 survivors who underwent HCT between 1974 and 1998 showed that 66% of the survivors had at least one chronic condition and 18% had severe or life-threatening conditions.14 A retrospective study of 1087 contemporary survivors also showed that the cumulative incidence of any non-malignant late effect at five years after HCT was 45% among autologous and 79% among allogeneic recipients, and 2.5% of autologous and 26% of allogeneic recipients had three or more late effects.15 Life expectancy among 5-year survivors remained 30% lower compared with the general population, regardless of their current ages and years since HCT.9 The leading causes of excess deaths in 5-year survivors included secondary malignancies (27%) and recurrent disease (14%), followed by infections (12%), chronic graft-versus-host disease (GvHD) (11%), cardiovascular diseases (11%), and respiratory diseases (7%).9 The aim of this review is to facilitate better care of adult HCT survivors by summarizing accumulated evidence, new insights, and practical information about individual late effects (Figure 1). Recurrent disease and chronic GvHD are not discussed and readers are referred to other reviews.16-20 haematologica | 2017; 102(4)


Late effects after HCT

Cardiovascular diseases

Pulmonary diseases

Cardiovascular diseases (CVD) after HCT include cardiomyopathy, congestive heart failure, valvular dysfunction, arrhythmia, pericarditis, and coronary artery disease.21 Their cumulative incidences were 5%-10% at ten years after HCT,22-24 accounting for 2%-11% of mortality among long-term survivors.8,9,25 The incidence of CVD and its associated mortality were 1.4-3.5-fold higher compared with the general population.8,9,24,25 HCT survivors are more likely to have conventional risk factors such as dyslipidemia and diabetes than the general population.26 Early diagnosis and treatment of modifiable risk factors is important. We usually treat hypertension more than 140/90 mmHg on 2 separate visits or more than 130/80 mmHg for patients with diabetes or renal disease.27 The first step is lifestyle modification including weight reduction, dietary sodium reduction and regular physical activity, followed by initiating antihypertensive drugs such as angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs). Anthracycline exposure and chest radiation are the major risk factors for CVD after HCT.21 Several studies showed that dexrazoxane, ACE inhibitors, ARBs and beta-blockers can prevent anthracycline-related cardiomyopathy in the non-HCT setting.28-32 Once cardiomyopathy is established, it is important to initiate appropriate treatment. ACE inhibitors and beta-blockers have been effective in improving left ventricular function.33

Non-infectious late complications of the lung include bronchiolitis obliterans syndrome (BOS), cryptogenic organizing pneumonia (COP) and pulmonary hypertension. BOS represents chronic GvHD of the lung, and is characterized by the new onset of fixed airflow obstruction after allogeneic HCT.34 According to the strict 2005 National Institutes of Health (NIH) diagnostic criteria for chronic GvHD, incidence of BOS was 5.5% and its prevalence was 15% among patients with chronic GvHD.35 Symptoms of BOS include dyspnea on exertion, cough and wheezing, but early BOS may be asymptomatic until significant lung function is lost.36 One study showed rapid decline in %FEV1 during the six months before BOS diagnosis, with a lower %FEV1 at diagnosis associated with worse survival.37 In our practice, we perform pulmonary function tests every three months including %FEV1 and FEV1/FVC among patients with active chronic GvHD. When testing shows significant new airflow obstruction, we repeat testing every month until stability is confirmed.38 Plasma matrix metalloproteinase 3 levels39 and parametric response mapping from CT scans40 might be useful diagnostic tests for BOS but these have not yet entered clinical practice. Standard treatment of BOS is prednisone at 1 mg/kg per day, followed by a taper to reach a lower, alternate-day regimen.38 A multicenter prospective study showed that addition of FAM (inhaled fluticasone propionate at 440 Îźg twice a day,

Figure 1. Late effects of blood and marrow transplantation.

haematologica | 2017; 102(4)

615


Y. Inamoto et al.

azithromycin at 250 mg taken 3 days per week, and montelukast at 10 mg nightly) to prednisone treatment stabilized pulmonary function in 70% of patients with newly diagnosed BOS and permitted systemic steroid exposure to be reduced.41 Cryptogenic organizing pneumonia is a disorder involving bronchioles, alveolar ducts, and alveoli, the lumen of which become filled with buds of granulation tissue consisting of fibroblasts.42 Clinical symptoms include dry cough, shortness of breath, and fever. Bronchoalveolar lavage is performed to exclude infection. Lung biopsy is required for definitive diagnosis, but an empiric diagnosis is often based on radiographic findings of diffuse, peripheral, fluffy infiltrates consistent with airspace consolidation. Pulmonary function testing shows restrictive changes and low diffusing capacity of the lungs for carbon monoxide. The incidence of COP is 2%-10%,43,44 and it is strongly associated with acute and chronic GvHD.45 COP usually responds within 5-7 days to prednisone at 1 mg/kg per day, which is continued for one month followed by a slow taper over five months because COP can often recur. Small case series suggest potential benefits of macrolides for treatment of COP.46 Pulmonary hypertension is an uncommon but potentially fatal complication after HCT, with a reported prevalence of 2.4%.47 The most common symptoms are hypoxia, tachypnea, dyspnea, and acute respiratory failure,48 and if untreated, pulmonary hypertension can result in a progressive increase in pulmonary vascular resistance, right ventricular failure and death. Since initial symptoms are non-specific, it is likely to be underdiagnosed after HCT. Although cardiac catheterization is the gold standard for diagnosis of pulmonary hypertension, high-resolution chest computed tomography and echocardiography are non-invasive and useful diagnostic modalities. The most common types are pulmonary arterial hypertension and pulmonary veno-occlusive disease, sometimes associated with transplant-associated microangiopathy and inherited or acquired hemolytic anemia.48 First-line therapies are supplemental oxygen and phosphodiesterase-5 inhibitors, followed by inhaled nitric oxide, diuretics, bipyridine inotropes and after-load reducing agents.48

Endocrine diseases Major late effects in the endocrine system include thyroid dysfunction, diabetes, dyslipidemia, and adrenal insufficiency. Hypothyroidism occurs in 30% of patients by 25 years after HCT.49 Risk factors include age under ten years, conditioning containing radiation, busulfan or cyclophosphamide, and hematologic malignancies.49,50 The international guidelines recommend checking serum thyroid-stimulating hormone and free thyroxine levels every year.21 For patients who received radiolabeled iodine antibody therapy, thyroid function should be checked earlier starting at three and six months after HCT, and other times as clinically indicated. Standard criteria are used to initiate replacement therapy for hypothyroidism. Some patients develop hyperthyroidism after HCT as a rare complication.51 Diabetes occurs in 8%-41% of patients after allogeneic HCT and in 3% of patients after autologous HCT.15,52,53 Its incidence after allogeneic HCT is 3.65 times higher compared with their siblings.54 Initial treatment is therapeutic lifestyle counseling, but many patients require hypoglycemic agents or insulin. 616

Dyslipidemia occurs in 9%-61% of HCT survivors.53,55 Despite no established consensus for management of dyslipidemia after HCT, our practice is to initiate therapeutic lifestyle counseling followed by statin therapy when LDL cholesterol exceeds 130-190 mg/dL according to the estimated risk of CVD, based on the National Cholesterol Education Program Adult Treatment Panel III guidelines56 and the recently suggested approach after allogeneic HCT.57 The 2013 ACC/AHA guidelines do not specify the targeted levels for LDL cholesterol, and addition of statin therapy is based on calculated risk for future cardiovascular events.58 Addition of omega-3-acid ethyl esters or fibrate is considered when fasting triglycerides exceed 200-499 mg/dL. Adrenal insufficiency occurs in 13% of patients after allogeneic HCT and 1% of patients after autologous HCT,15 and can be confirmed by a cortisol-stimulation test. Once adrenal insufficiency is diagnosed, physiological glucocorticoid replacement and a very slow terminal taper is needed. Patients should carry notification that they have adrenal insufficiency to alert emergency medical providers. For chronic GvHD therapy, the risk of adrenal insufficiency is lower with alternate-day administration of corticosteroids than with daily dosing,59 although patients with brittle diabetes need daily dosing to allow for optimal glucose control.

Male gonadal dysfunction and infertility Hypogonadism is common after HCT. Impaired spermatogenesis, erectile dysfunction, low testosterone, and low libido occur in male patients. Erectile dysfunction and low libido have been associated with both physical and psychosocial factors.60,61 Testosterone replacement may be considered for patients with low testosterone levels and has improved sexual function, libido and bone mass, although monitoring prostate-specific antigen and testosterone levels is necessary.62,63 Azoospermia occurred in 70% of male patients, and spermatogenesis recovered in 90% of patients conditioned with cyclophosphamide alone, in 50% of patients conditioned with cyclophosphamide plus busulfan or thiotepa, and in 17% of patients conditioned with total body irradiation (TBI).64 Semen banking or cryopreservation of testicular tissue should be discussed before HCT with patients desiring fertility.

Female gonadal dysfunction, infertility and pregnancy Ovarian insufficiency, vaginal changes and low libido occur in female patients. A historical study showed that ovarian failure occurred in more than 90% of female patients after HCT and recovered in 92% of patients conditioned with cyclophosphamide alone, but only in 24% of patients conditioned with cyclophosphamide and TBI.65 A pilot study showed that only 10% of patients had ovarian failure after reduced-intensity allogeneic HCT.66 The use of hormone replacement therapy for premature ovarian failure should be individualized based on the patient age, severity of menopausal symptoms, low bone density, risk of breast cancer, clotting predisposition and liver abnormalities.67 Since efficacy of gonadotropin-releasing hormone agonists in preserving fertility in cancer patients is controversial,68,69 cryopreservation of oocytes, ovarian tissue, or embryos should be discussed with patients desiring fertility.70 The largest study of pregnancy after HCT showed that haematologica | 2017; 102(4)


Late effects after HCT

0.87% of patients or their partners had pregnancies after allogeneic HCT, and 0.36% of those after autologous HCT.71 We generally recommend that women wait 2-5 years after HCT before attempting conception since rates of relapse are generally highest in the first two years after HCT. Another concern is the theoretical risk of recurrent malignancy because of disturbance of the graft-versusleukemia effect, and some cases of recurrent chronic myeloid leukemia after conception have been reported.71 Pregnancy outcomes are generally good with no increase in the risk of fetal malformations, although these pregnancies are considered high risk because of higher maternal risks of pregnancy complications.71

Iron overload Iron overload is rare after autologous HCT72 but common after allogeneic HCT.73,74 Previous prospective studies showed that 30%-60% of long-term survivors of allogeneic HCT had elevated serum ferritin levels and 25%-50% had elevated liver iron concentration on T2* magnetic resonance imaging (MRI).73,74 Since serum ferritin does not specifically reflect iron overload and can be elevated in hepatic and systemic inflammation, additional testing is required if the ferritin is elevated. We favor transferrin saturation, which is widely available and defined as the ratio of serum iron concentration divided by total iron-binding capacity.75 Normal transferrin saturation is less than 50% in males and less than 45% in females. Patients with iron overload usually have saturation more than 60%. HFE genotyping is considered in patients with a family history of hemochromatosis and in patients of Northern or Western European ethnicity. When saturation is not elevated, other etiologies for an elevated ferritin including inflammation, metabolic syndrome, and alcoholism should be ruled out. The most accurate test of tissue iron concentration is liver biopsy, but the procedure is invasive and may cause serious complications. Thus, T2* MRI and other modalities (FerriScan and superconducting quantum interference device) have been increasingly used.76 Importantly, liver tests are often normal among long-term survivors with iron overload, so hepatitis and GvHD should also be considered when results of liver tests are elevated.77 Iron overload may cause cardiomyopathy. Studies of thalassemia patients showed that cardiomyopathy typically took more than ten years to be clinically evident,78 and that many patients improved with intensive chelation therapy.79 Although a prospective study and a meta-analysis showed no statistical association of liver iron concentration with mortality after allogeneic HCT,80,81 our practice is to start phlebotomy of 5 mL/kg or 250-300 mL every 3-4 weeks as long as hematocrit is more than 35% until serum ferritin falls below 1000 ng/mL. Deferasirox, an oral chelating agent, is considered for patients with anemia precluding phlebotomy.

Liver diseases Late liver diseases include chronic hepatitis B, chronic hepatitis C, liver cirrhosis, nodular regenerative hyperplasia and focal nodular hyperplasia.77 Hepatitis B-infected patients have an increased risk of fulminant liver failure. One study reported a 35% risk of HBV reactivation after HCT even among patients with isolated anti-HBc antibodies, mostly during steroid treatment for GvHD.82 Patients treated with anti-CD20 antibodies have an increased risk of HBV reactivation. Antiviral prophylaxis using entecavir haematologica | 2017; 102(4)

or lamivudine will prevent almost all fulminant cases if initiated before the start of conditioning regimens in patients with positive blood HBV DNA levels.83 Patients with latent HBV (i.e. anti-HBc+/HBV DNA-) should be monitored monthly with HBV DNA levels after HCT and antiviral treatment should be initiated when viremia is detected.83 Hepatitis C virus infection in HCT survivors almost always results in chronic hepatitis.84,85 Typically, asymptomatic elevation of alanine aminotransferase occurs 2-4 months after HCT, coinciding with tapering of immunosuppressive medications. There may be little liver-related mortality in the first ten years after HCT,84 but liver cirrhosis occurs later with a cumulative incidence of 4%-24% at 20 years.85,86 A large retrospective study showed that hepatitis C-infected patients had an increased risk of 2-year non-relapse mortality due to hepatic problems and bacterial infection.87 Antiviral therapy for HCV has not been given early after HCT, but may improve both oncological and hepatic outcomes after HCT.88 Ribavirin and interferon-based therapy have been used for patients who have discontinued all immunosuppressive medications without active GvHD, but it can cause pancytopenia and GvHD. Recently, highly effective and well tolerated direct acting antiviral agents with more than 90% rates of sustained virological response have been developed, and interferonfree regimens are now the treatments of choice.89,90 Nodular regenerative hyperplasia is a rare liver condition characterized by a widespread benign transformation of the hepatic parenchyma into small regenerative nodules.77 This process is usually asymptomatic unless portal hypertension develops. Focal nodular hyperplasia occurs in 12% of HCT survivors, and possibly reflects sinusoidal injury caused by myeloablative conditioning regimens.91

Kidney diseases Chronic kidney disease (CKD) is defined as an elevated serum creatinine level, or a decreased glomerular filtration rate (GFR) less than 60 mL/min/1.73 m2 for three months or longer.92 CKD occurs in approximately 20% of HCT recipients.93-95 There are three major etiologies of CKD after HCT: thrombotic microangiopathy (TMA), nephrotic syndrome and idiopathic CKD. Other etiologies include persistent acute kidney injury and BK virus nephropathy.96 Whenever possible, renal biopsy should be considered to accurately diagnose the etiology of CKD and to provide appropriate management.97 Thrombotic microangiopathy occurs in 2%-21% of patients after HCT, and is characterized by renal dysfunction, thrombocytopenia, neurological dysfunction, hemolytic anemia with schistocytes, elevated lactate dehydrogenase and decreased haptoglobin.98,99 Risk factors of TMA include TBI, calcineurin inhibitors, and acute and chronic GvHD.100-102 TMA-related kidney injury often improves with tapering or stopping calcineurin inhibitors, but full renal function is rarely restored.103 In some cases TMA did not improve until GvHD was treated.104 Efficacy of plasma exchange is limited.105 Nephrotic syndrome occurs in 6%-8% of patients after allogeneic HCT.106,107 Membranous nephropathy comprised 61% of cases, and minimal change disease comprised 22% of cases, with a median onset of 14 months and eight months after HCT, respectively.108 Mechanisms of membranous nephropathy are thought to be formation of immune complexes through allo- or auto-antibodies 617


Y. Inamoto et al.

recognizing antigens expressed by the podocyte, while T cells are implicated with minimal change disease.109 Nephrotic syndrome after HCT is often associated with chronic GvHD and tapering of immunosuppressive medications. Initial treatment is prednisone 1 mg/kg/day in addition to calcineurin inhibitors. Complete response was observed in 90% of patients with minimal change in disease, but only in 27% of patients with membranous nephropathy.108 Refractory cases may be treated with rituximab or mycophenolate mofetil.110 Idiopathic CKD comprises most cases of CKD. Risk factors include acute GvHD, chronic GvHD, acute kidney injury, long-term use of calcineurin inhibitors and previous autologous HCT,94,111 suggesting that GvHD, accompanying treatment and inflammatory conditions may have pathogenic roles in this entity. Associations of TBI with risk of CKD have been controversial.94,112 ACE inhibitors and ARBs have been used to treat CKD and hypertension associated with CKD.113

Bone diseases Late complications of bone include osteopenia, osteoporosis and avascular necrosis (AVN).114 Osteoporosis has been reported in as many as 50% of HCT recipients.115,116 The diagnoses of osteopenia and osteoporosis are made by measuring T-scores with dual-energy X-ray absorptiometry. A T-score between -1.0 and -2.5 indicates osteopenia, and a T-score less than -2.5 or presence of a fragility fracture indicates osteoporosis.117 Multiple risk factors are implicated including chemotherapy, radiation, corticosteroids, calcineurin inhibitors, vitamin D deficiency, and gonadal failure.116,118 Bone loss occurs within 6-12 months after HCT, and recovery of bone mineral density (BMD) begins from the lumber spine, followed by a slower recovery in the femoral neck. The use of corticosteroids is the strongest risk factor for osteoporosis. General preventative recommendations include adequate intake of calcium of 1200 mg per day or over and vitamin D of 1000 IU (25 Îźg) per day or over, regular weight-bearing exercise, and avoidance of smoking and excessive alcohol. Bisphosphonates are the primary treatment for bone loss.119 Patients who are taking 5 mg or more daily prednisone-equivalent steroids for three months or more should have screening BMD tests for osteoporosis, and bisphosphonate treatment may be indicated until corticosteroid treatment is discontinued or for up to five years.120 Second-line treatment includes calcitonin, raloxifene, denusomab, romosozumab, and blosozumab, though their reported use in HCT recipients is limited and adverse effects may be more prominent than with the bisphosphonates. Avascular necrosis occurs in 4%-19% of HCT survivors with a cumulative incidence of 3%-10% at five years after HCT.121,122 AVN causes severe bone pain and bone destruction, causing significant impairment in quality of life. AVN typically affects the femoral heads, but sometimes affects other joints such as the knee and shoulders.21 Risk factors for AVN include corticosteroids, calcineurin inhibitors, older age and TBI conditioning.114 When AVN is suspected, diagnostic MRI should be performed. Early involvement of an orthopedic specialist is important for management of AVN, including conservative treatment, joint-preserving surgery and joint replacement surgery.21,114

Infectious diseases All HCT survivors have some degree of immunodeficiency, particularly during the first year after HCT.123 If patients 618

are able to stop immunosuppressive medications without GvHD or recurrent disease, many recover adequate immune function by one year after HCT. Patients with chronic GvHD, however, remain immunodeficient and have a high risk of infections. Common late infections are caused by Pneumocystis jirovecii, encapsulated bacteria, fungi, varicella-zoster virus (VZV), cytomegalovirus, and respiratory viruses. Patients may report more frequent episodes of upper respiratory infections and sinusitis. All patients should receive prophylaxis against Pneumocystis jirovecii for at least one year after HCT or until 3-6 months after all immunosuppressive medication is discontinued, whichever occurs later. The preferred drug is trimethoprim-sulfamethoxazole, but dapsone or atovaquone could be substituted for patients who are allergic to or intolerant of trimethoprim-sulfamethoxazole. In particular, patients with chronic GvHD are highly susceptible to encapsulated bacteria such as Streptococcus pneumoniae, Haemophilus influenzae and Neisseria meningitidis due to low levels of opsonizing antibodies, low CD4+ T-cell counts, poor reticuloendothelial function and suppressive effects of immunosuppressive medications on phagocytosis. Vaccination against these bacteria is recommended.124 Efficacy of vaccination in increasing antibody levels has been shown in several prospective studies.125,126 Chemoprophylaxis is always recommended due to the unpredictable protection provided by vaccination. The firstline drug is trimethoprim-sulfamethoxazole, but if it is not tolerated, penicillin or azithromycin is substituted until 3-6 months after discontinuation of all immunosuppressive medications. Invasive fungal infection occurs in 1% of patients after autologous HCT and in 6%-8% of patients after allogeneic HCT.127 GvHD and long-term use of corticosteroids have been a major risk factor associated with onset of invasive fungal infection.128 As recommended in the European guidelines, mold prophylaxis with posaconazole or voriconazole may be considered for patients with GvHD requiring highdose corticosteroid treatment.129 Varicella-zoster virus-seropositive patients should receive prophylaxis with acyclovir or valacyclovir during the first year after HCT or until six months after discontinuation of immunosuppressive medications. A standard dose of acyclovir is 800 mg twice daily,130 but some studies showed that 200 mg once daily was effective in preventing VZV reactivation.131 Acyclovir should be started empirically if the patient presents with an acute abdomen or hepatitis typical of fulminant visceral VZV infection.132 CMV monitoring in blood is continued beyond 100 days after HCT until one year for patients at risk of late CMV disease, including CMV-seropositive patients receiving high-dose corticosteroids, those who have already experienced CMV reactivation, and cord blood transplantation.133 Pre-emptive therapy is usually considered for CMV levels of 250 IU/mL or more (equivalent to ≼1000 copies/mL) or a positive antigenemia test. Community-acquired respiratory virus infections are an important cause of morbidity and mortality after HCT. The most frequent viruses include rhinovirus, respiratory syncytial virus (RSV), parainfluenza viruses (PIV), human metapneumovirus, and influenza viruses as these frequently cause lower respiratory tract disease associated with 12%-100% mortality.134 An immunodeficiency scoring index can predict severity of RSV infection.135 Aerosolized ribavirin showed efficacy in treating lower tract RSV after HCT.136 haematologica | 2017; 102(4)


Late effects after HCT

Combination therapy with immunomodulators such as intravenous immunoglobulin or palivizumab has been seen to have variable success.137 Treatment for PIV infection has not been established. Efficacy of ribavirin has been limited for patients with lower respiratory tract infection of PIV.138 Novel drugs such as a recombinant sialidase fusion protein and a hemagglutinin-neuraminidase inhibitor are under investigation.138

88%-100% for thyroid, testis and melanoma, approximately 50% for breast, mouth, soft tissue and female reproductive organs, and 20% or less for bone, lower gastrointestinal tract, and central nervous system.144 These rates were similar to those of de novo cancers, except that rates were lower for female reproductive organs, bone, colorectum, and central nervous system, although further studies are warranted to confirm this observation. There is emerging evidence that human papilloma virus (HPV) is involved in the pathogenesis of squamous cell cancer after HCT.145,146 The efficacy of HPV vaccination in preventing squamous cell cancer after HCT remains to be determined in prospective studies.147

Solid cancers There is an increased risk of solid cancers following both autologous and allogeneic HCT compared with the general population. The cumulative incidence is 1%-6% at ten years after HCT, and continues to rise over time without a plateau.139-142 The most common sites include oral cavity, skin, breast and thyroid, but rates are also elevated in esophagus, liver, nervous system, bone and connective tissues compared with the general population.143 Myeloablative TBI, young age at HCT, chronic GvHD and prolonged immunosuppressive medications beyond two years are well-documented risk factors for many types of cancers.143 All HCT recipients should be advised of the risk of second cancers and should be encouraged to undergo recommended screening tests based on their predisposition.143 The 5-year overall survival rates after diagnosis of solid cancers varied by cancer site, with

Neuropsychological effects Neuropsychological effects after HCT are being increasingly recognized and include, among others, depression, post-traumatic stress disorder, and neurocognitive deficits. Depression occurs in 12%-30% of HCT survivors and is more frequent in female patients, younger patients and those with poor social support, history of recurrent disease, chronic pain, and chronic GvHD.148 Post-traumatic stress disorder occurs in 28% of patients at six months after HCT and may persist for 5%-13% of cases, although its risk factors are not yet clear.148-150 Neurocognitive deficits, so called “chemo brain�, have

Table 1. Late effects after blood and marrow transplantation

Late effect Cardiovascular Pulmonary Bronchiolitis obliterans syndrome Cryptogenic organizing pneumonia Pulmonary hypertension Endocrine Thyroid dysfunction Diabetes Dyslipidemia Adrenal insufficiency Gonadal dysfunction/ infertility Iron overload Liver Hepatitis B Hepatitis C and cirrhosis Nodular regenerative hyperplasia Focal nodular hyperplasia Kidney Thrombotic microangiopathy Nephrotic syndrome Idiopathic chronic kidney disease Bone Osteoporosis/osteopenia Avascular necrosis Infection Solid cancer Neuropsychological Recurrent disease Chronic graft-versus-host disease

Incidence

Mortality

Morbidity

Treatable

Preventable

+

+

+

+

+

+ + +

++ + ++

++ + ++

+ ++ +

-

++ ++ ++ + +++ ++

+ -

-/+ + -/+ -/+ -

+++ +++ +++ +++ -/+ ++

-/+ -/+ -

+ + + +

-

+ + -

++ ++ -

+ -/+ -

+ + +

+ -

++ ++ ++

-/+ ++ +

-

++ + ++ + ++ ++ ++

+ ++ +++ +

++ + +++ ++ +++ ++

++ ++ +++ -/+ + -/+ +

+ + -

+ : <20%; ++ : 20%-50%; +++ : >50%.

haematologica | 2017; 102(4)

619


Y. Inamoto et al. Table 2. Tests, preventive approaches and treatment of late effects.

Late effect Cardiovascular Anthracycline-related cardiomyopathy Others Pulmonary Bronchiolitis obliterans syndrome Cryptogenic organizing pneumonia Pulmonary hypertension

Tests

Preventive approaches

Treatment

Physical exam, chest X-ray, electrocardiogram, echocardiogram, brain natriuretic peptide level Blood pressure, lipid panel, glucose level, HbA1c, glycoalbumin

Dexrazoxane, ACE inhibitors, ARBs, beta-blockers Lifestyle modification, ACE inhibitors, ARBs

ACE inhibitors, ARBs, beta-blockers

%FEV1, FEV1/FVC CT, lung biopsy High-resolution chest CT, echocardiography, cardiac catheterization

Prednisone, FAM Prednisone, macrolides Oxygen, phosphodiesterase-5 inhibitors, inhaled nitric oxide, diuretics, bipyridine inotropes, after-load reducing agents

Endocrine Hypothyroidism Diabetes

Thyroid-stimulating hormone, thyroxine levels Glucose level, HbA1c, glycoalbumin

Dyslipidemia

Lipid panel

Adrenal insufficiency

Cortisol-stimulation test

Gonadal dysfunction Male patients Female patients Infertility Iron overload

Liver Hepatitis B Hepatitis C and cirrhosis

Testosterone level Follicle-stimulating hormone, lutenizing hormone, estradiol levels Sperm test

Alternate-day regimen when corticosteroids are used

Replacement therapy Lifestyle modification, hyperglycemic agents, insulin Lifestyle modification, statins, fibrates, fish oil (omega-3 fatty acids), ezetimibe Hydrocortisone, low-dose prednisone

Reduced-intensity conditioning

Testosterone replacement Hormone replacement

Semen banking, cryopreservation of testicular or ovarian tissues

Assisted reproduction, surrogate pregnancy, adoption Phlebotomy, desferoxamine, deferasirox

Entecavir, lamivudine Direct acting antiviral agents

Entecavir, lamivudine Direct acting antiviral agents, interferon

Serum ferritin levels, transferrin saturation, Prussian blue-stained marrow biopsy, T2* MRI, FerriScan, SQUID ALT, HBV DNA levels ALT, HCV RNA levels

continued in the next page

adverse functional impacts on HCT survivors who return to work and daily activities that require short-term memory, information-processing speed, multitasking and co-ordination.151 Neuropsychological tests can help identify neurocognitive deficits. Most evidence is derived from studies of breast cancer survivors, with estimated rates of deficits ranging from 16% to 50% up to ten years after treatment.152,153 Potential mechanisms for chemotherapy-induced neurocognitive changes include cytokine and immune dysregulation, damage to DNA and telomere length through cytotoxic agents, oxidative stress and hormonal changes.154 In cases of HCT survivors, there may be additional deficits derived from neurological complications including nervous system infection (HHV-6, fungi, etc.), immune-mediated damage, and toxicities of calcineurin inhibitors such as TMA and posterior reversible encephalopathy syndrome. A prospective observational study showed that neurocognitive function declined substantially at 80 days after HCT, returned to pre-transplantation levels at one year, and continued to improve between one and five years after HCT, except for motor dexterity and verbal learning and retention.155 Mostly mild, neurocognitive dysfunction according to the Global Deficit Score persisted at five years in 42% of 620

long-term survivors.155 Rehabilitation programs have succeeded in improving neurocognitive functions,156 and methylphenidate and modafinil have demonstrated variable efficacies to improve neurocognitive function in nonHCT cancer patients.157,158 Efficacies of these interventions remaine to be determined among HCT survivors.

Influence of newer practices on late effects An understanding of the influence of newer practices such as cord blood transplantation, non-TBI or reducedintensity conditioning regimens and older patients on the incidence and severity of late effects awaits longer follow up. For example, TBI is associated with an increased risk of many late effects such as cardiovascular diseases, COP, hypothyroidism, diabetes, dyslipidemia, infertility, TMArelated kidney injury, bone density loss, avascular necrosis, and secondary solid cancer.49,54,100,102,114,118,143,159,160 The use of non-TBI conditioning regimens might reduce the burden of these late effects among HCT survivors. Some studies found that cumulative incidences of late effects did not differ much after reduced-intensity regimens compared with myeloablative regimens,15,161 and reduced-intensity conditioning was associated with a higher risk of recurrent malighaematologica | 2017; 102(4)


Late effects after HCT

continued from the previous page

Late effect

Tests

Kidney Thrombotic microangiopathy

Preventive approaches

CBC, schistocytes, serum creatinine, lactate dehydrogenase, haptoglobin, renal biopsy

Nephrotic syndrome

Urine protein, renal biopsy

Idiopathic chronic kidney disease

Renal biopsy

Bone Osteoporosis/osteopenia

Dual-energy X-ray absorptiometry

Avascular necrosis

MRI

Infection Pneumocystis jirovecii

Bronchoalveolar lavage, PCR, β-D-glucan

Encapsulated bacteria

Treatment Taper or stop calcineurin inhibitors, GvHD treatment required in some cases Prednisone, rituximab, mycophenolate mofetil, ACE inhibitors, ARBs

Calcium intake, vitamin D intake, bisphosphonate, estrogen, testosterone

Bisphosphonate, estrogen, testosterone, (calcitonin, raloxifene, denusomab, romosozumab, blosozumab) Conservative treatment, surgery

Trimethoprim-sulfamethoxazole, dapsone, atovaquone

Prednisone, trimethoprimsulfamethoxazole, atovaquone, pentamidine Antibiotics

Trimethoprim-sulfamethoxazole, penicillin, azithromycin, vaccination against Haemophilus influenzae type b, Neisseria meningitidis, Streptococcus pneumoniae Posaconazole, voriconazole Acyclovir, valacyclovir Ganciclovir, valganciclovir, foscarnet

Fungi Varicella-zoster virus Cytomegalovirus

Galactomannan assay, β-D-glucan, CT PCR PCR, antigenemia

Respiratory syncytial virus Influenza virus Solid cancer Neuropsychological

PCR Aerosolized ribavirin, palivizumab Immunoassay Vaccination Recommended screening tests (see reference 143) (Human papillomavirus vaccination) Neuropsychological test, MRI

Antifungal agents Acyclovir, valacyclovir Ganciclovir, valganciclovir, foscarnet Oseltamivir Site- and stage-specific treatment Rehabilitation, methylphenidate, modafinil

ACE: angiotension-converting enzyme; ARBs: angiotensin II receptor blockers; FAM: inhaled fluticasone propionate, azithromycin and montelukast; CT: computed tomography; MRI: magnetic resonance imaging; SQUID: superconducting quantum interference device; ALT: alanine aminotransferase; GvHD: graft-versus-host disease; PCR: polymerase chain reaction.

nancy among patients with myeloid malignancy.162 One study showed that the risk of AVN was elevated after cord blood transplantation, but graft source had a limited influence on other long-term health status and QOL.163

Consensus guidelines for late effects and prevention behaviors Incidence, mortality, morbidity and management of individual late effects are summarized in Tables 1 and 2. Recognizing the importance of managing late effects after HCT, the Center for International Blood and Marrow Transplant Research (CIBMTR), the European Group for Blood and Marrow Transplantation (EBMT), and the American Society for Bone Marrow Transplantation (ASBMT) developed recommendations in 2006 for screening and prevention practices for HCT survivors.164 Consensus recommendations were up-dated in 2011 including other international transplant communities.21 The NIH convened working groups to formulate late effects initiatives in 2015.148,165-169 Despite higher levels of engagement with health care providers, HCT survivors had similar health and prevention behaviors as matched untransplanted controls, suggesting the need for further education of both HCT survivors and health practitioners.170 Major modifiable predictors of lower haematologica | 2017; 102(4)

adherence to preventive care practices were concerns about medical costs and lack of knowledge.171

Conclusion While the number of HCT survivors is growing, there is no evidence that the burden of late effects is lessening. HCT survivors face myriad late effects that can limit their functioning, require prolonged or life-long medical treatment, reduce their quality of life and also shorten their survival. To the extent that the HCT procedure itself causes these late effects, the transplant community has a responsibility to appropriately monitor, treat and ultimately try to prevent late effects. Given the dispersion of survivors and the varied structure of health care, hematologists, oncologists, primary care physicians and medical subspecialists are all involved in providing this care. Further research is needed to understand the biology of late effects to help identify better prevention and treatment strategies.166 Funding This work was supported by grants from the Japan Society for the Promotion of Science (15K19563), from the Friends of Leukemia Research Fund, and from the National Institutes of Health (CA18953 and CA18029). 621


Y. Inamoto et al.

References 16. 1. Thomas E, Storb R, Clift RA, et al. Bonemarrow transplantation (first of two parts). N Engl J Med. 1975;292(16):832-843. 2. Gooley TA, Chien JW, Pergam SA, et al. Reduced mortality after allogeneic hematopoietic-cell transplantation. N Engl J Med. 2010;363(22):2091-2101. 3. Sorror ML, Sandmaier BM, Storer BE, et al. Long-term outcomes among older patients following nonmyeloablative conditioning and allogeneic hematopoietic cell transplantation for advanced hematologic malignancies. JAMA. 2011;306(17):1874-1883. 4. Gluckman E, Rocha V, Boyer-Chammard A, et al. Outcome of cord-blood transplantation from related and unrelated donors. Eurocord Transplant Group and the European Blood and Marrow Transplantation Group. N Engl J Med. 1997;337(6):373-381. 5. 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. 6. Gratwohl A, Pasquini MC, Aljurf M, et al. One million haemopoietic stem-cell transplants: a retrospective observational study. Lancet Haematol. 2015;2(3):e91-100. 7. Socie G, Stone JV, Wingard JR, et al. Longterm survival and late deaths after allogeneic bone marrow transplantation. Late Effects Working Committee of the International Bone Marrow Transplant Registry. N Engl J Med. 1999;341(1):14-21. 8. Bhatia S, Robison LL, Francisco L, et al. Late mortality in survivors of autologous hematopoietic-cell transplantation: report from the Bone Marrow Transplant Survivor Study. Blood. 2005;105(11):4215-4222. 9. Martin PJ, Counts GW Jr, Appelbaum FR, et al. Life expectancy in patients surviving more than 5 years after hematopoietic cell transplantation. J Clin Oncol. 2010;28(6) :1011-1016. 10. Wingard JR, Majhail NS, Brazauskas R, et al. Long-term survival and late deaths after allogeneic hematopoietic cell transplantation. J Clin Oncol. 2011;29(16):2230-2239. 11. Duell T, van Lint MT, Ljungman P, et al. Health and functional status of long-term survivors of bone marrow transplantation. EBMT Working Party on Late Effects and EULEP Study Group on Late Effects. European Group for Blood and Marrow Transplantation. Ann Intern Med. 1997;126(3):184-192. 12. Chiodi S, Spinelli S, Ravera G, et al. Quality of life in 244 recipients of allogeneic bone marrow transplantation. Br J Haematol. 2000;110(3):614-619. 13. Lee SJ, Logan B, Westervelt P, et al. Comparison of Patient-Reported Outcomes in 5-Year Survivors Who Received Bone Marrow vs Peripheral Blood Unrelated Donor Transplantation: Long-term Followup of a Randomized Clinical Trial. JAMA Oncol. 2016;2(12):1583-1589. 14. Sun CL, Francisco L, Kawashima T, et al. Prevalence and predictors of chronic health conditions after hematopoietic cell transplantation: a report from the Bone Marrow Transplant Survivor Study. Blood. 2010;116(17):3129-3139. 15. Khera N, Storer B, Flowers ME, et al. Nonmalignant late effects and compromised functional status in survivors of hematopoi-

622

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

etic cell transplantation. J Clin Oncol. 2012;30(1):71-77. Avigan D, Hari P, Battiwalla M, et al. Proceedings from the National Cancer Institute's Second International Workshop on the Biology, Prevention, and Treatment of Relapse after Hematopoietic Stem Cell Transplantation: part II. Autologous Transplantation-novel agents and immunomodulatory strategies. Biol Blood Marrow Transplant. 2013;19(12):1661-1669. Gress RE, Miller JS, Battiwalla M, et al. Proceedings from the National Cancer Institute's Second International Workshop on the Biology, Prevention, and Treatment of Relapse after Hematopoietic Stem Cell Transplantation: Part I. Biology of relapse after transplantation. Biol Blood Marrow Transplant. 2013;19(11):1537-1545. de Lima M, Porter DL, Battiwalla M, et al. Proceedings from the National Cancer Institute's Second International Workshop on the Biology, Prevention, and Treatment of Relapse After Hematopoietic Stem Cell Transplantation: part III. Prevention and treatment of relapse after allogeneic transplantation. Biol Blood Marrow Transplant. 2014;20(1):4-13. Jagasia MH, Greinix HT, Arora M, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: I. The 2014 Diagnosis and Staging Working Group Report. Biol Blood Marrow Transplant. 2015;21(3):389-401. Carpenter PA, Kitko CL, Elad S, et al. National Institutes of Health Consensus Development Project on Criteria for Clinical Trials in Chronic Graft-versus-Host Disease: V. The 2014 Ancillary Therapy and Supportive Care Working Group Report. Biol Blood Marrow Transplant. 2015;21(7): 1167-1187. Majhail NS, Rizzo JD, Lee SJ, et al. Recommended screening and preventive practices for long-term survivors after hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2012;18(3):348371. Tichelli A, Bucher C, Rovo A, et al. Premature cardiovascular disease after allogeneic hematopoietic stem-cell transplantation. Blood. 2007;110(9):3463-3471. Tichelli A, Passweg J, Wojcik D, et al. Late cardiovascular events after allogeneic hematopoietic stem cell transplantation: a retrospective multicenter study of the Late Effects Working Party of the European Group for Blood and Marrow Transplantation. Haematologica. 2008;93(8): 1203-1210. Chow EJ, Mueller BA, Baker KS, et al. Cardiovascular hospitalizations and mortality among recipients of hematopoietic stem cell transplantation. Ann Intern Med. 2011;155(1):21-32. Bhatia S, Francisco L, Carter A, et al. Late mortality after allogeneic hematopoietic cell transplantation and functional status of longterm survivors: report from the Bone Marrow Transplant Survivor Study. Blood. 2007;110(10):3784-3792. Chow EJ, Baker KS, Lee SJ, et al. Influence of conventional cardiovascular risk factors and lifestyle characteristics on cardiovascular disease after hematopoietic cell transplantation. J Clin Oncol. 2014;32(3):191-198. James PA, Oparil S, Carter BL, et al. 2014 evidence-based guideline for the management of high blood pressure in adults: report from the panel members appointed to the Eighth

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38. 39.

40.

41.

Joint National Committee (JNC 8). JAMA. 2014;311(5):507-520. Lipshultz SE, Scully RE, Lipsitz SR, et al. Assessment of dexrazoxane as a cardioprotectant in doxorubicin-treated children with high-risk acute lymphoblastic leukaemia: long-term follow-up of a prospective, randomised, multicentre trial. Lancet Oncol. 2010;11(10):950-961. Chow EJ, Asselin BL, Schwartz CL, et al. Late Mortality After Dexrazoxane Treatment: A Report From the Children's Oncology Group. J Clin Oncol. 2015;33(24):2639-2645. Asselin BL, Devidas M, Chen L, et al. Cardioprotection and Safety of Dexrazoxane in Patients Treated for Newly Diagnosed T-Cell Acute Lymphoblastic Leukemia or Advanced-Stage Lymphoblastic Non-Hodgkin Lymphoma: A Report of the Children's Oncology Group Randomized Trial Pediatric Oncology Group 9404. J Clin Oncol. 2016;34(8):854862. Cardinale D, Colombo A, Sandri MT, et al. Prevention of high-dose chemotherapyinduced cardiotoxicity in high-risk patients by angiotensin-converting enzyme inhibition. Circulation. 2006;114(23):2474-2481. Georgakopoulos P, Roussou P, Matsakas E, et al. Cardioprotective effect of metoprolol and enalapril in doxorubicin-treated lymphoma patients: a prospective, parallelgroup, randomized, controlled study with 36-month follow-up. Am J Hematol. 2010;85(11):894-896. Cardinale D, Colombo A, Lamantia G, et al. Anthracycline-induced cardiomyopathy: clinical relevance and response to pharmacologic therapy. J Am Coll Cardiol. 2010;55(3):213-220. Chien JW, Duncan S, Williams KM, Pavletic SZ. Bronchiolitis obliterans syndrome after allogeneic hematopoietic stem cell transplantation-an increasingly recognized manifestation of chronic graft-versus-host disease. Biol Blood Marrow Transplant. 2010;16(1 Suppl):S106-114. Au BK, Au MA, Chien JW. Bronchiolitis obliterans syndrome epidemiology after allogeneic hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2011;17(7):1072-1078. Dudek AZ, Mahaseth H, DeFor TE, Weisdorf DJ. Bronchiolitis obliterans in chronic graft-versus-host disease: analysis of risk factors and treatment outcomes. Biol Blood Marrow Transplant. 2003;9(10):657666. Cheng GS, Storer B, Chien JW, et al. Lung Function Trajectory in Bronchiolitis Obliterans Syndrome after Allogeneic Hematopoietic Cell Transplantation. Ann Am Thorac Soc. 2016;13(11):1932-1939. Flowers ME, Martin PJ. How we treat chronic graft-versus-host disease. Blood. 2015;125(4):606-615. Liu X, Yue Z, Yu J, et al. Proteomic Characterization Reveals That MMP-3 Correlates With Bronchiolitis Obliterans Syndrome Following Allogeneic Hematopoietic Cell and Lung Transplantation. Am J Transplant. 2016;16 (8):2342-2351. Galban CJ, Boes JL, Bule M, et al. Parametric response mapping as an indicator of bronchiolitis obliterans syndrome after hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2014;20(10): 1592-1598. Williams KM, Cheng GS, Pusic I, et al.

haematologica | 2017; 102(4)


Late effects after HCT

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Fluticasone, Azithromycin, and Montelukast Treatment for New-Onset Bronchiolitis Obliterans Syndrome after Hematopoietic Cell Transplantation. Biol Blood Marrow Transplant. 2016;22(4):710716. Epler GR, Colby TV, McLoud TC, Carrington CB, Gaensler EA. Bronchiolitis obliterans organizing pneumonia. N Engl J Med. 1985;312(3):152-158. Palmas A, Tefferi A, Myers JL, et al. Lateonset noninfectious pulmonary complications after allogeneic bone marrow transplantation. Br J Haematol. 1998;100(4):680687. Patriarca F, Skert C, Sperotto A, et al. Incidence, outcome, and risk factors of lateonset noninfectious pulmonary complications after unrelated donor stem cell transplantation. Bone Marrow Transplant. 2004;33(7):751-758. Freudenberger TD, Madtes DK, Curtis JR, Cummings P, Storer BE, Hackman RC. Association between acute and chronic graft-versus-host disease and bronchiolitis obliterans organizing pneumonia in recipients of hematopoietic stem cell transplants. Blood. 2003;102(10):3822-3828. Pathak V, Kuhn JM, Durham C, Funkhouser WK, Henke DC. Macrolide use leads to clinical and radiological improvement in patients with cryptogenic organizing pneumonia. Ann Am Thorac Soc. 2014;11(1):8791. Jodele S, Hirsch R, Laskin B, Davies S, Witte D, Chima R. Pulmonary arterial hypertension in pediatric patients with hematopoietic stem cell transplant-associated thrombotic microangiopathy. Biol Blood Marrow Transplant. 2013;19(2):202-207. Dandoy CE, Hirsch R, Chima R, Davies SM, Jodele S. Pulmonary hypertension after hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2013;19(11): 1546-1556. Sanders JE, Hoffmeister PA, Woolfrey AE, et al. Thyroid function following hematopoietic cell transplantation in children: 30 years' experience. Blood. 2009;113(2):306-308. 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. Aldouri MA, Ruggier R, Epstein O, Prentice HG. Adoptive transfer of hyperthyroidism and autoimmune thyroiditis following allogeneic bone marrow transplantation for chronic myeloid leukaemia. Br J Haematol. 1990;74(1):118-119. Majhail NS, Flowers ME, Ness KK, et al. High prevalence of metabolic syndrome after allogeneic hematopoietic cell transplantation. Bone Marrow Transplant. 2009;43(1):49-54. Abou-Mourad YR, Lau BC, Barnett MJ, et al. Long-term outcome after allo-SCT: close follow-up on a large cohort treated with myeloablative regimens. Bone Marrow Transplant. 2010;45(2):295-302. Baker KS, Ness KK, Steinberger J, et al. Diabetes, hypertension, and cardiovascular events in survivors of hematopoietic cell transplantation: a report from the bone marrow transplantation survivor study. Blood. 2007;109(4):1765-1772. Rovo A, Daikeler T, Halter J, et al. Late altered organ function in very long-term survivors after allogeneic hematopoietic stem cell transplantation: a paired comparison with their HLA-identical sibling donor.

haematologica | 2017; 102(4)

Haematologica. 2011;96(1):150-155. 56. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation and Treatment of High Blood Cholesterol in Adults. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002;106(25):3143-3421. 57. Griffith ML, Savani BN, Boord JB. Dyslipidemia after allogeneic hematopoietic stem cell transplantation: evaluation and management. Blood. 2010;116(8):11971204. 58. Stone NJ, Robinson JG, Lichtenstein AH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol. 2014;63(25 Pt B):2889-2934. 59. Ackerman GL, Nolsn CM. Adrenocortical responsiveness after alternate-day corticosteroid therapy. N Engl J Med. 1968;278(8):405-409. 60. Molassiotis A, van den Akker OB, Milligan DW, Boughton BJ. Gonadal function and psychosexual adjustment in male long-term survivors of bone marrow transplantation. Bone Marrow Transplant. 1995;16(2):253259. 61. Chatterjee R, Andrews HO, McGarrigle HH, et al. Cavernosal arterial insufficiency is a major component of erectile dysfunction in some recipients of high-dose chemotherapy/chemo-radiotherapy for haematological malignancies. Bone Marrow Transplant. 2000;25(11):1185-1189. 62. Behre HM, Kliesch S, Leifke E, Link TM, Nieschlag E. Long-term effect of testosterone therapy on bone mineral density in hypogonadal men. J Clin Endocrinol Metab. 1997;82(8):2386-2390. 63. Chatterjee R, Kottaridis PD, McGarrigle HH, Linch DC. Management of erectile dysfunction by combination therapy with testosterone and sildenafil in recipients of highdose therapy for haematological malignancies. Bone Marrow Transplant. 2002;29(7):607-610. 64. Anserini P, Chiodi S, Spinelli S, et al. Semen analysis following allogeneic bone marrow transplantation. Additional data for evidence-based counselling. Bone Marrow Transplant. 2002;30(7):447-451. 65. Sanders JE, Buckner CD, Amos D, et al. Ovarian function following marrow transplantation for aplastic anemia or leukemia. J Clin Oncol. 1988;6(5):813-818. 66. Phelan R, Mann E, Napurski C, et al. Ovarian function after hematopoietic cell transplantation: a descriptive study following the use of GnRH agonists for myeloablative conditioning and observation only for reduced-intensity conditioning. Bone Marrow Transplant. 2016;51(10):1369-1375. 67. Fish JD. Part 1: Hormone replacement for survivors of childhood cancer with ovarian failure--when is it worth the risk? J Pediatr Adolesc Gynecol. 2011;24(2):98-101. 68. Munster PN, Moore AP, Ismail-Khan R, et al. Randomized trial using gonadotropinreleasing hormone agonist triptorelin for the preservation of ovarian function during (neo)adjuvant chemotherapy for breast cancer. J Clin Oncol. 2012;30(5):533-538. 69. Demeestere I, Brice P, Peccatori FA, et al. No Evidence for the Benefit of Gonadotropin-

70.

71.

72.

73.

74.

75.

76.

77. 78. 79.

80.

81.

82.

83.

84.

Releasing Hormone Agonist in Preserving Ovarian Function and Fertility in Lymphoma Survivors Treated With Chemotherapy: Final Long-Term Report of a Prospective Randomized Trial. J Clin Oncol. 2016;34(22):2568-2574. Joshi S, Savani BN, Chow EJ, et al. Clinical guide to fertility preservation in hematopoietic cell transplant recipients. Bone Marrow Transplant. 2014;49(4):477-484. Salooja N, Szydlo RM, Socie G, et al. Pregnancy outcomes after peripheral blood or bone marrow transplantation: a retrospective survey. Lancet. 2001;358(9278):271276. Majhail NS, DeFor TE, Lazarus HM, Burns LJ. Iron-overload after autologous hematopoietic cell transplantation. Leuk Res. 2009;33(4):578-579. Rose C, Ernst O, Hecquet B, et al. Quantification by magnetic resonance imaging and liver consequences of post-transfusional iron overload alone in long term survivors after allogeneic hematopoietic stem cell transplantation (HSCT). Haematologica. 2007;92(6):850-853. Majhail NS, Lazarus HM, Burns LJ. A prospective study of iron overload management in allogeneic hematopoietic cell transplantation survivors. Biol Blood Marrow Transplant. 2010;16(6):832-837. Schrier SL, Bacon BR. Approach to the patient with suspected iron overload. In: UpToDate, Post TW (Ed.), UpToDate, Waltham, MA. Accessed on December 30, 2016. Au WY, Lam WM, Chu WC, et al. A magnetic resonance imaging study of iron overload in hemopoietic stem cell transplant recipients with increased ferritin levels. Transplant Proc. 2007;39(10):3369-3374. McDonald GB. Hepatobiliary complications of hematopoietic cell transplantation, 40 years on. Hepatology. 2010;51(4):1450-1460. Olivieri NF, Brittenham GM. Iron-chelating therapy and the treatment of thalassemia. Blood. 1997;89(3):739-761. Aldouri MA, Wonke B, Hoffbrand AV, et al. High incidence of cardiomyopathy in betathalassaemia patients receiving regular transfusion and iron chelation: reversal by intensified chelation. Acta Haematol. 1990;84(3):113-117. Trottier BJ, Burns LJ, DeFor TE, Cooley S, Majhail NS. Association of iron overload with allogeneic hematopoietic cell transplantation outcomes: a prospective cohort study using R2-MRI-measured liver iron content. Blood. 2013;122(9):1678-1684. Armand P, Kim HT, Virtanen JM, et al. Iron overload in allogeneic hematopoietic cell transplantation outcome: a meta-analysis. Biol Blood Marrow Transplant. 2014;20(8): 1248-1251. Picardi M, De Rosa G, Selleri C, Pane F, Rotoli B, Muretto P. Clinical relevance of intrahepatic hepatitis B virus DNA in HBsAg-negative HBcAb-positive patients undergoing hematopoietic stem cell transplantation for hematological malignancies. Transplantation. 2006;82(1):141-142. Liang R. How I treat and monitor viral hepatitis B infection in patients receiving intensive immunosuppressive therapies or undergoing hematopoietic stem cell transplantation. Blood. 2009;113(14):3147-3153. Strasser SI, Myerson D, Spurgeon CL, et al. Hepatitis C virus infection and bone marrow transplantation: a cohort study with 10year follow-up. Hepatology. 1999;29(6): 1893-1899.

623


Y. Inamoto et al. 85. Peffault de Latour R, Levy V, Asselah T, et al. Long-term outcome of hepatitis C infection after bone marrow transplantation. Blood. 2004;103(5):1618-1624. 86. Strasser SI, Sullivan KM, Myerson D, et al. Cirrhosis of the liver in long-term marrow transplant survivors. Blood. 1999;93(10): 3259-3266. 87. Nakasone H, Kurosawa S, Yakushijin K, et al. Impact of hepatitis C virus infection on clinical outcome in recipients after allogeneic hematopoietic cell transplantation. Am J Hematol. 2013;88(6):477-484. 88. Kyvernitakis A, Mahale P, Popat UR, et al. Hepatitis C Virus Infection in Patients Undergoing Hematopoietic Cell Transplantation in the Era of Direct-Acting Antiviral Agents. Biol Blood Marrow Transplant. 2016;22(4):717-722. 89. Panel AIHG. Hepatitis C guidance: AASLDIDSA recommendations for testing, managing, and treating adults infected with hepatitis C virus. Hepatology. 2015;62(3):932-954. 90. Torres HA, McDonald GB. How I treat hepatitis C virus infection in patients with hematologic malignancies. Blood. 2016;128 (11):1449-1457. 91. Sudour H, Mainard L, Baumann C, Clement L, Salmon A, Bordigoni P. Focal nodular hyperplasia of the liver following hematopoietic SCT. Bone Marrow Transplant. 2009;43(2):127-132. 92. National Kidney F. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002;39(2 Suppl 1):S1-266. 93. Choi M, Sun CL, Kurian S, et al. Incidence and predictors of delayed chronic kidney disease in long-term survivors of hematopoietic cell transplantation. Cancer. 2008;113(7):1580-1587. 94. Hingorani S, Guthrie KA, Schoch G, Weiss NS, McDonald GB. Chronic kidney disease in long-term survivors of hematopoietic cell transplant. Bone Marrow Transplant. 2007;39(4):223-229. 95. Hingorani S. Renal Complications of Hematopoietic-Cell Transplantation. N Engl J Med. 2016;374(23):2256-2267. 96. Verghese PS, Finn LS, Englund JA, Sanders JE, Hingorani SR. BK nephropathy in pediatric hematopoietic stem cell transplant recipients. Pediatr Transplant. 2009;13(7): 913-918. 97. Chang A, Hingorani S, Kowalewska J, et al. Spectrum of renal pathology in hematopoietic cell transplantation: a series of 20 patients and review of the literature. Clin J Am Soc Nephrol. 2007;2(5):1014-1023. 98. Ho VT, Cutler C, Carter S, et al. Blood and marrow transplant clinical trials network toxicity committee consensus summary: thrombotic microangiopathy after hematopoietic stem cell transplantation. Biol Blood Marrow Transplant. 2005;11(8):571575. 99. Ruutu T, Barosi G, Benjamin RJ, et al. Diagnostic criteria for hematopoietic stem cell transplant-associated microangiopathy: results of a consensus process by an International Working Group. Haematologica. 2007;92(1):95-100. 100. Cohen EP, Lawton CA, Moulder JE. Bone marrow transplant nephropathy: radiation nephritis revisited. Nephron. 1995;70(2): 217-222. 101. Pettitt AR, Clark RE. Thrombotic microangiopathy following bone marrow transplantation. Bone Marrow Transplant. 1994;14(4):495-504. 102. Fuge R, Bird JM, Fraser A, et al. The clinical features, risk factors and outcome of throm-

624

botic thrombocytopenic purpura occurring after bone marrow transplantation. Br J Haematol. 2001;113(1):58-64. 103. Abboud I, Peraldi MN, Hingorani S. Chronic kidney diseases in long-term survivors after allogeneic hematopoietic stem cell transplantation: monitoring and management guidelines. Semin Hematol. 2012;49(1):7382. 104. Oran B, Donato M, Aleman A, et al. Transplant-associated microangiopathy in patients receiving tacrolimus following allogeneic stem cell transplantation: risk factors and response to treatment. Biol Blood Marrow Transplant. 2007;13(4):469-477. 105. Christidou F, Athanasiadou A, Kalogiannidis P, et al. Therapeutic plasma exchange in patients with grade 2-3 hematopoietic stem cell transplantation-associated thrombotic thrombocytopenic purpura: a ten-year experience. Ther Apher Dial. 2003;7(2):259-262. 106. Srinivasan R, Balow JE, Sabnis S, et al. Nephrotic syndrome: an under-recognised immune-mediated complication of nonmyeloablative allogeneic haematopoietic cell transplantation. Br J Haematol. 2005;131(1):74-79. 107. Colombo AA, Rusconi C, Esposito C, et al. Nephrotic syndrome after allogeneic hematopoietic stem cell transplantation as a late complication of chronic graft-versushost disease. Transplantation. 2006;81(8): 1087-1092. 108. Brukamp K, Doyle AM, Bloom RD, Bunin N, Tomaszewski JE, Cizman B. Nephrotic syndrome after hematopoietic cell transplantation: do glomerular lesions represent renal graft-versus-host disease? Clin J Am Soc Nephrol. 2006;1(4):685-694. 109. Rao PS. Nephrotic syndrome in patients with peripheral blood stem cell transplant. Am J Kidney Dis. 2005;45(4):780-785. 110. Terrier B, Delmas Y, Hummel A, et al. Postallogeneic haematopoietic stem cell transplantation membranous nephropathy: clinical presentation, outcome and pathogenic aspects. Nephrol Dial Transplant. 2007;22(5):1369-1376. 111. Weiss AS, Sandmaier BM, Storer B, Storb R, McSweeney PA, Parikh CR. Chronic kidney disease following non-myeloablative hematopoietic cell transplantation. Am J Transplant. 2006;6(1):89-94. 112. Miralbell R, Bieri S, Mermillod B, et al. Renal toxicity after allogeneic bone marrow transplantation: the combined effects of totalbody irradiation and graft-versus-host disease. J Clin Oncol. 1996;14(2):579-585. 113. Cohen EP, Irving AA, Drobyski WR, et al. Captopril to mitigate chronic renal failure after hematopoietic stem cell transplantation: a randomized controlled trial. Int J Radiat Oncol Biol Phys. 2008;70(5):15461551. 114. McClune B, Majhail NS, Flowers ME. Bone loss and avascular necrosis of bone after hematopoietic cell transplantation. Semin Hematol. 2012;49(1):59-65. 115. Yao S, McCarthy PL, Dunford LM, et al. High prevalence of early-onset osteopenia/osteoporosis after allogeneic stem cell transplantation and improvement after bisphosphonate therapy. Bone Marrow Transplant. 2008;41(4):393-398. 116. Savani BN, Donohue T, Kozanas E, et al. Increased risk of bone loss without fracture risk in long-term survivors after allogeneic stem cell transplantation. Biol Blood Marrow Transplant. 2007;13(5):517-520. 117. Kanis JA. Assessment of fracture risk and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report.

WHO Study Group. Osteoporos Int. 1994;4(6):368-381. 118. Stern JM, Sullivan KM, Ott SM, et al. Bone density loss after allogeneic hematopoietic stem cell transplantation: a prospective study. Biol Blood Marrow Transplant. 2001;7(5):257-264. 119. Tauchmanova L, Selleri C, Esposito M, et al. Beneficial treatment with risedronate in long-term survivors after allogeneic stem cell transplantation for hematological malignancies. Osteoporos Int. 2003;14(12):10131019. 120. Recommendations for the prevention and treatment of glucocorticoid-induced osteoporosis: 2001 update. American College of Rheumatology Ad Hoc Committee on Glucocorticoid-Induced Osteoporosis. Arthritis Rheuma. 2001;44(7):1496-1503. 121. Enright H, Haake R, Weisdorf D. Avascular necrosis of bone: a common serious complication of allogeneic bone marrow transplantation. Ma J Med. 1990;89(6):733-738. 122. Socie G, Cahn JY, Carmelo J, et al. Avascular necrosis of bone after allogeneic bone marrow transplantation: analysis of risk factors for 4388 patients by the Societe Francaise de Greffe de Moelle (SFGM). Br J Haematol. 1997;97(4):865-870. 123. Tomblyn M, Chiller T, Einsele H, et al. Guidelines for preventing infectious complications among hematopoietic cell transplantation recipients: a global perspective. Biol Blood Marrow Transplant. 2009;15(10): 1143-1238. 124. Carpenter PA, Englund JA. How I vaccinate blood and marrow transplant recipients. Blood. 2016;127(23):2824-2832. 125. Cordonnier C, Ljungman P, Juergens C, et al. Immunogenicity, safety, and tolerability of 13-valent pneumococcal conjugate vaccine followed by 23-valent pneumococcal polysaccharide vaccine in recipients of allogeneic hematopoietic stem cell transplant aged >/=2 years: an open-label study. Clin Infect Dis. 2015;61(3):313-323. 126. Patel SR, Ortin M, Cohen BJ, et al. Revaccination with measles, tetanus, poliovirus, Haemophilus influenzae type B, meningococcus C, and pneumococcus vaccines in children after hematopoietic stem cell transplantation. Clin Infect Dis. 2007;44(5):625-634. 127. Kontoyiannis DP, Marr KA, Park BJ, et al. Prospective surveillance for invasive fungal infections in hematopoietic stem cell transplant recipients, 2001-2006: overview of the Transplant-Associated Infection Surveillance Network (TRANSNET) Database. Clin Infect Dis. 2010;50(8):1091-1100. 128. Marr KA, Carter RA, Boeckh M, Martin P, Corey L. Invasive aspergillosis in allogeneic stem cell transplant recipients: changes in epidemiology and risk factors. Blood. 2002;100(13):4358-4366. 129. Maertens J, Marchetti O, Herbrecht R, et al. European guidelines for antifungal management in leukemia and hematopoietic stem cell transplant recipients: summary of the ECIL 3--2009 update. Bone Marrow Transplant. 2011;46(5):709-718. 130. Boeckh M, Kim HW, Flowers ME, Meyers JD, Bowden RA. Long-term acyclovir for prevention of varicella zoster virus disease after allogeneic hematopoietic cell transplantation--a randomized double-blind placebocontrolled study. Blood. 2006;107(5):18001805. 131. Asano-Mori Y, Kanda Y, Oshima K, et al. Long-term ultra-low-dose acyclovir against varicella-zoster virus reactivation after allogeneic hematopoietic stem cell transplanta-

haematologica | 2017; 102(4)


Late effects after HCT tion. Am J Hematol. 2008;83(6):472-476. 132. David DS, Tegtmeier BR, O'Donnell MR, Paz IB, McCarty TM. Visceral varicellazoster after bone marrow transplantation: report of a case series and review of the literature. Am J Gastroenterol. 1998;93(5):810813. 133. Boeckh M, Leisenring W, Riddell SR, et al. Late cytomegalovirus disease and mortality in recipients of allogeneic hematopoietic stem cell transplants: importance of viral load and T-cell immunity. Blood. 2003;101 (2):407-414. 134. Forman SJ, Negrin RS, Antin JH, Appelbaum FR. Thomas' Hematopoietic Cell Transplantation, 5th Edition. Oxford, UK: Wiley-Blackwell, 2015. 135. Shah DP, Ghantoji SS, Ariza-Heredia EJ, et al. Immunodeficiency scoring index to predict poor outcomes in hematopoietic cell transplant recipients with RSV infections. Blood. 2014;123(21):3263-3268. 136. Boeckh M, Englund J, Li Y, et al. Randomized controlled multicenter trial of aerosolized ribavirin for respiratory syncytial virus upper respiratory tract infection in hematopoietic cell transplant recipients. Clin Infect Dis. 2007;44(2):245-249. 137. de Fontbrune FS, Robin M, Porcher R, et al. Palivizumab treatment of respiratory syncytial virus infection after allogeneic hematopoietic stem cell transplantation. Clin Infect Dis. 2007;45(8):1019-1024. 138. Shah DP, Shah PK, Azzi JM, Chemaly RF. Parainfluenza virus infections in hematopoietic cell transplant recipients and hematologic malignancy patients: A systematic review. Cancer Lett. 2016;370(2):358-364. 139. Curtis RE, Rowlings PA, Deeg HJ, et al. Solid cancers after bone marrow transplantation. N Engl J Med. 1997;336(13):897-904. 140. Bhatia S, Louie AD, Bhatia R, et al. Solid cancers after bone marrow transplantation. J Clin Oncol. 2001;19(2):464-471. 141. Rizzo JD, Curtis RE, Socie G, et al. Solid cancers after allogeneic hematopoietic cell transplantation. Blood. 2009;113(5):11751183. 142. Bilmon IA, Ashton LJ, Le Marsney RE, et al. Second cancer risk in adults receiving autologous haematopoietic SCT for cancer: a population-based cohort study. Bone Marrow Transplant. 2014;49(5):691-698. 143. Inamoto Y, Shah NN, Savani BN, et al. Secondary solid cancer screening following hematopoietic cell transplantation. Bone Marrow Transplant. 2015;50(8):1013-1023. 144. Ehrhardt MJ, Brazauskas R, He W, Rizzo JD, Shaw BE. Survival of patients who develop solid tumors following hematopoietic stem cell transplantation. Bone Marrow Transplant. 2016;51(1):83-88. 145. Herrero R, Castellsague X, Pawlita M, et al. Human papillomavirus and oral cancer: the International Agency for Research on Cancer multicenter study. J Natl Cancer Inst. 2003;95(23):1772-1783. 146. Savani BN, Stratton P, Shenoy A, Kozanas E, Goodman S, Barrett AJ. Increased risk of cervical dysplasia in long-term survivors of allogeneic stem cell transplantation--implications for screening and HPV vaccination. Biol Blood Marrow Transplant. 2008;14(9):1072-1075. 147. Tedeschi SK, Savani BN, Jagasia M, et al.

haematologica | 2017; 102(4)

Time to consider HPV vaccination after allogeneic stem cell transplantation. Biol Blood Marrow Transplant. 2010;16(8):1033-1036. 148. Bevans M, El-Jawahri A, Tierney DK, et al. National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: The Patient-Centered Outcomes Working Group Report. Biol Blood Marrow Transplant. 2016 Sep 19. [Epub ahead of print] 149. Rusiewicz A, DuHamel KN, Burkhalter J, et al. Psychological distress in long-term survivors of hematopoietic stem cell transplantation. Psychooncology. 2008;17(4):329-337. 150. El-Jawahri AR, Vandusen HB, Traeger LN, et al. Quality of life and mood predict posttraumatic stress disorder after hematopoietic stem cell transplantation. Cancer. 2016;122(5):806-812. 151. Andrykowski MA, Altmaier EM, Barnett RL, Burish TG, Gingrich R, HensleeDowney PJ. Cognitive dysfunction in adult survivors of allogeneic marrow transplantation: relationship to dose of total body irradiation. Bone Marrow Transplant. 1990;6(4):269-276. 152. Tannock IF, Ahles TA, Ganz PA, Van Dam FS. Cognitive impairment associated with chemotherapy for cancer: report of a workshop. J Clin Oncol. 2004;22(11):2233-2239. 153. Stewart A, Bielajew C, Collins B, Parkinson M, Tomiak E. A meta-analysis of the neuropsychological effects of adjuvant chemotherapy treatment in women treated for breast cancer. Clin Neuropsychol. 2006;20(1):76-89. 154. Ahles TA, Saykin AJ. Candidate mechanisms for chemotherapy-induced cognitive changes. Nat Rev Cancer. 2007;7(3):192-201. 155. Syrjala KL, Artherholt SB, Kurland BF, et al. Prospective neurocognitive function over 5 years after allogeneic hematopoietic cell transplantation for cancer survivors compared with matched controls at 5 years. J Clin Oncol. 2011;29(17):2397-2404. 156. Gehring K, Sitskoorn MM, Gundy CM, et al. Cognitive rehabilitation in patients with gliomas: a randomized, controlled trial. J Clin Oncol. 2009;27(22):3712-3722. 157. Meyers CA, Weitzner MA, Valentine AD, Levin VA. Methylphenidate therapy improves cognition, mood, and function of brain tumor patients. J Clin Oncol 1998;16(7):2522-2527. 158. Lundorff LE, Jonsson BH, Sjogren P. Modafinil for attentional and psychomotor dysfunction in advanced cancer: a doubleblind, randomised, cross-over trial. Palliat Med. 2009;23(8):731-738. 159. Meacham LR, Chow EJ, Ness KK, et al. Cardiovascular risk factors in adult survivors of pediatric cancer--a report from the childhood cancer survivor study. Cancer Epidemiol Biomarkers Prev. 2010;19(1):170181. 160. Nakasone H, Onizuka M, Suzuki N, et al. Pre-transplant risk factors for cryptogenic organizing pneumonia/bronchiolitis obliterans organizing pneumonia after hematopoietic cell transplantation. Bone Marrow Transplant. 2013;48(10):1317-1323. 161. Clavert A, Peric Z, Brissot E, et al. Late Complications and Quality of Life after Reduced-Intensity Conditioning Allogeneic Stem Cell Transplantation. Biol Blood

Marrow Transplant. 2017;23(1):140-146. 162. Scott BL, Pasquini M, Logan B, et al. Results of a Phase III Randomized, Multi-Center 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): (Abstract LBA-128). 163. Visentin S, Auquier P, Bertrand Y, et al. The Impact of Donor Type on Long-Term Health Status and Quality of Life after Allogeneic Hematopoietic Stem Cell Transplantation for Childhood Acute Leukemia: A Leucemie de l'Enfant et de L'Adolescent Study. Biol Blood Marrow Transplant. 2016;22(11): 2003-2010. 164. Rizzo JD, Wingard JR, Tichelli A, et al. Recommended screening and preventive practices for long-term survivors after hematopoietic cell transplantation: joint recommendations of the European Group for Blood and Marrow Transplantation, the Center for International Blood and Marrow Transplant Research, and the American Society of Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2006;12(2):138-151. 165. Armenian SH, Chemaitilly W, Chen M, et al. National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: The Cardiovascular Disease and Associated Risk Factors Working Group Report. Biol Blood Marrow Transplant. 2016;23(2):201210 166. Gea-Banacloche J, Komanduri K, Carpenter P, et al. National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: The Immune Dysregulation and Pathobiology Working Group Report. Biol Blood Marrow Transplant. 2016 Oct 14. [Epub ahead of print] 167. Hashmi SK, Bredeson C, Duarte RF, et al. National Institutes of Health Blood and Marrow Transplant Late Effects Initiative: The Healthcare Delivery Working Group Report. Biol Blood Marrow Transplant in press. 2016 Oct 3. [Epub ahead of print] 168. Morton LM, Saber W, Baker KS, et al. National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: Consensus Recommendations for Subsequent Neoplasms. Biol Blood Marrow Transplant. 2016 Sep 12. [Epub ahead of print] 169. Shaw BE, Hahn T, Martin PJ, et al. National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: The Research Methodology and Study Design Working Group Report. Biol Blood Marrow Transplant. 2017;23(1):10-23. 170. Bishop MM, Lee SJ, Beaumont JL, et al. The preventive health behaviors of long-term survivors of cancer and hematopoietic stem cell transplantation compared with matched controls. Biol Blood Marrow Transplant. 2010;16(2):207-214. 171. Khera N, Chow EJ, Leisenring WM, et al. Factors associated with adherence to preventive care practices among hematopoietic cell transplantation survivors. Biol Blood Marrow Transplant. 2011;17(7):995-1003.

625


GUIDELINE ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Risk factors for mortality in adult patients with sickle cell disease: a meta-analysis of studies in North America and Europe

Poulami Maitra,1 Melissa Caughey,2 Laura Robinson,3 Payal C. Desai,4 Susan Jones,5 Mehdi Nouraie,6 Mark T. Gladwin,7 Alan Hinderliter,3 Jianwen Cai1 and Kenneth I. Ataga5

Haematologica 2017 Volume 102(4):626-636

Department of Biostatistics, University of North Carolina, Chapel Hill; 2Division of Cardiology, University of North Carolina, Chapel Hill; 3Childrenâ&#x20AC;&#x2122;s Hospital of Philadelphia; 4 Division of Hematology, The Ohio State University, Columbus; 5Division of Hematology/Oncology, University of North Carolina, Chapel Hill; 6Department of Medicine, Howard University, Washington, DC and 7Department of Medicine, University of Pittsburgh, PA, USA 1

ABSTRACT

A

Correspondence: kataga@med.unc.edu

Received: August 10, 2016. Accepted: January 12, 2017. Pre-published: January 19, 2017.

lthough recent studies show an improved survival of children with sickle cell disease in the US and Europe, for adult patients mortality remains high. This study was conducted to evaluate the factors associated with mortality in adult patients following the approval of hydroxyurea. We first evaluated the association between selected variables and mortality at an academic center (University of North Carolina). Data sources were then searched for publications from 1998 to June 2016, with meta-analysis of eligible studies conducted in North America and Europe to evaluate the associations of selected variables with mortality in adult patients. Nine studies, combined with the UNC cohort (total n=3257 patients) met the eligibility criteria. Mortality was significantly associated with age (per 10-year increase in age) [7 studies, 2306 participants; hazard ratio (HR): 1.28; 95% confidence interval (CI): 1.10-1.50], tricuspid regurgitant jet velocity 2.5 m/s or more (5 studies, 1577 participants; HR: 3.03; 95%CI: 2.0-4.60), reticulocyte count (3 studies, 1050 participants; HR: 1.05; 95%CI: 1.01-1.10), log(N-terminal-probrain natriuretic peptide) (3 studies, 800 participants; HR: 1.68; 95%CI: 1.48-1.90), and fetal hemoglobin (7 studies, 2477 participants; HR: 0.97; 95%CI: 0.94-1.0). This study identifies variables associated with mortality in adult patients with sickle cell disease in the hydroxyurea era.

doi:10.3324/haematol.2016.153791 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/626 Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

626

Introduction Sickle cell disease (SCD) is characterized by multi-organ morbidity and an increased risk of early death. Several studies show that the survival of children with SCD has improved over the last decades.1 This improved survival is attributed to the implementation of newborn screening, use of prophylactic penicillin, and vaccinations against Haemophilus influenza type b and Streptococcus pneumonia,2-4 with possible contributions from advances in red blood cell transfusion medicine, iron chelation therapy, and transcranial Doppler screening to identify those at increased risk of stroke.5,6 The Cooperative Study of Sickle Cell Disease (CSSCD), a multicenter natural history study of SCD conducted between 1978 and 1988, reported median ages at death during the study period of 42 and 48 years, respectively, for male and female HbSS patients, and median ages at death of 60 and 68 years, respectively, for male and female HbSC patients.7 While many patients died following acute episodes of pain, acute chest syndrome or stroke, statistical modeling revealed that the acute chest syndrome, renal failure, seizures, high white blood cell count (WBC) and low level of fetal hemoglobin (HbF) were associated with an increased risk of early haematologica | 2017; 102(4)


Mortality in sickle cell disease

death in HbSS patients.7 However, this prospective study was conducted prior to the approval of hydroxyurea for the treatment of sickle cell anemia. More recent studies show associations of elevated echocardiography-derived tricuspid regurgitant jet velocity (TRV),8,9 pulmonary hypertension (PHT),10,11 elevated levels of N-terminal probrain natriuretic peptide (NT-pro-BNP),12 history of asthma and/or wheezing,13 end-stage renal disease requiring dialysis,14 severity of hemolysis,15 and prolongation of QTc interval16 with an increased risk of death in patients with SCD. Hydroxyurea, approved by the US Food and Drug Administration (FDA) in 1998, remains the only drug that has been shown to alter the natural history of SCD in peer-reviewed studies. In two randomized, placebo-controlled studies, hydroxyurea was shown to decrease the frequency of acute pain episodes, acute chest syndrome, red blood cell (RBC) transfusion and hospitalization rates in both adults and children with sickle cell anemia.17,18 Despite the studies showing an improved survival in adult patients with SCD following treatment with hydroxyurea,19-21 the mortality rate remains high for patients aged 18 years or older following transition to adult care,22 possibly reflecting a lack of access to high-quality care. The present study evaluated the association of selected clinical and laboratory variables with mortality in a cohort of adult patients with SCD followed at an academic medical center. With the relatively small number of patients evaluated in this single center study, we have conducted comprehensive meta-analyses of contemporary prospective and retrospective studies to evaluate the factors associated with mortality in adult patients with SCD following the approval of hydroxyurea. Our goal is to uncover the underlying causes of mortality in SCD.

Methods Patients and study design We evaluated a subset of patients followed at the Comprehensive Sickle Cell Clinic at the University of North Carolina (UNC), Chapel Hill between August 2004 and December 2014 (subsequently referred to as the UNC cohort). The data were collected as part of a prospective study of the natural history of pulmonary hypertension in SCD.9 The patients were evaluated while in a non-crisis, 'steady state'; they had not experienced an episode of acute chest syndrome in the four weeks preceding enrollment, and had no clinical evidence of congestive heart failure. This study was approved by the Institutional Review Board at UNC, Chapel Hill, and all subjects gave written informed consent to participate in accordance with the Declaration of Helsinki.

Sickle cell disease-related clinical complications and laboratory variables The presence or history of clinical complications, including the number of acute pain episodes (or vaso-occlusive crises) in the past year, history of acute chest syndrome, and use of hydroxyurea, was ascertained at the time of evaluation as well as through a review of the patients’ medical records. Acute pain episodes, acute chest syndrome, and other SCD-related complications were defined using accepted definitions.19,23 SCD genotypes were dichotomized into clinically severe (HbSS, HbSβ0 thalassemia and HbSD) and less severe (HbSC and HbSβ+ thalassemia) groups. The TRV was measured by continuous wave Doppler echocardiography. Multiple views were obtained to record optimal tricuspid haematologica | 2017; 102(4)

flow signals. Measurements were performed on at least three waveforms with well-defined velocity envelopes and an average value was used for data analysis. Patients with non-quantifiable TRVs were assumed to have normal estimated PASP. The echocardiograms were interpreted by a cardiologist blinded to all patient data. Laboratory measurements obtained in study subjects include complete blood counts with reticulocyte counts, serum creatinine, lactate dehydrogenase, total bilirubin, direct bilirubin, indirect bilirubin, D-dimer and NT-pro-BNP. Hemoglobin analysis was performed by high performance liquid chromatography to confirm the SCD diagnosis and ascertain HbF levels.

Search strategy and covariate abstraction A meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and MetaAnalyses guidelines and Meta-Analysis of Observational Studies in Epidemiology Group criteria.24 Published articles were identified using MEDLINE (PubMed), Web of Science and EMBASE search engines between January 1, 1998, and June 30, 2016. The search on PubMed was performed using the Medical Subject Headings (MeSH) “anemia, sickle cell [mesh] OR “sickle cell disease” OR "sickle cell anemia" AND “mortality OR survival.” The search on Web of Science was performed using the search terms “sickle cell” AND “mortality OR survival” and the search on EMBASE was performed using the search terms “sickle cell anemia OR sickle cell disease” AND “mortality OR survival”. In addition, we interrogated references from relevant original papers and review articles to identify further relevant studies. Reviews, editorials, duplicate citations, non-human studies, highly selected study populations, retrospective analyses of large administrative datasets and articles written in languages other than English were excluded. We limited our analysis to prospective and retrospective studies conducted in North America and Europe and published since 1998, the year hydroxyurea was approved by the FDA for treatment of sickle cell anemia. Multiple publications from the same patient population were narrowed down to the single publication with the most complete information. Eligible studies were combined with the UNC cohort for meta-analyses.

Statistical analysis Median, range and interquartile range for continuous variables and frequencies for categorical variables were provided in the UNC cohort. NT-pro-BNP was log-transformed based on previous literature.25 We evaluated the Spearman correlation coefficient of each pair of variables to check the collinearity. Median ages at death were calculated for those individuals who died during the study. Median survival age was calculated based on Kaplan-Meier estimates for the survival probability using age as the time scale taking into account the left truncation. Log-rank tests were conducted to compare survival probabilities among different groups. Cox proportional hazards regression was fit to the data and variables were selected by backward elimination at 5% level of significance to obtain the final model. Schoenfeld residuals were used to check the proportional hazards assumption. Hazard ratios (HRs) were calculated in the UNC cohort, controlling for age and sex. Mortality HRs were meta-analyzed using DerSimonian and Laird26 random effects models26,27 due to the diversity of individual studies and the small number of studies included. The Knapp and Hartung27 method was used to adjust the standard errors of the estimated coefficients. Unadjusted HRs in the studies were used for our meta-analyses whenever they were reported. The degree of heterogeneity across the studies was examined using I2 values,28 classifying more than 50% as moderate heterogeneity, and more than 75% as high heterogeneity. Meta-analysis was performed if 627


P. Maitra et al.

HRs were available from a minimum of 3 studies. We considered the Voskaridou et al. study21 to contribute two separate HRs depending on whether patients received hydroxyurea. Similarly, from the Elmariah et al. study,25 two estimates were used based on the use of hydroxyurea. All the analyses were performed using SAS statistical software v.9.4 (Cary, NC, USA). The graphs for the meta-analyses were made using the R Project for Statistical Computing.

Results Study population One hundred and sixty-one subjects in the UNC cohort [female: 97 (60.3%)] with SCD (SS: 119; SC 19; Sβ0: 12; Sβ+: 10; SD: 1), and a median age of 36 years (range 18-71 years) (Table 1) were followed for a median duration of 7.2 years (range 0.06-10.3 years) and a total of 954.0 person-years. The majority of the subjects [159 (99%)] were self-reported African Americans. There were 29 (18%) deaths during the period of follow up (SS: 20, SC: 4, Sβ0: 3,

Table 1. Baseline demographic and laboratory characteristics of UNC cohort.

Characteristic Age (years) Sex Male Female Race Black Hispanic Genotype SS Sβ0 Sβ+ SC Other (SD) Use of hydroxyurea Yes No White blood cell count (x109/L) Hemoglobin (g/dL) Platelet count (x109/L) Reticulocyte count (%) Absolute reticulocyte count Fetal hemoglobin (%) Creatinine (mg/dL) Lactate dehydrogenase (U/L) Total bilirubin (mg/dL) Indirect bilirubin (mg/dL) Direct bilirubin (mg/dL) D-dimer (ng/mL) Tricuspid regurgitant jet velocity (m/s) Number of pain episodes in previous year History of acute chest syndrome Yes No 628

Number

Number (%)/median (interquartile range)

161 161

36 (27-44) 64 (40%) 97 (60%)

161

Sβ+: 2). Amongst those who died during the study, the median age of death was 48.0 years (range 29.8-70.2 years), while the median age in the subjects alive at the time of last contact was 38.7 years (range 22.6-79.4 years). When the subjects were grouped based on presumed disease severity (SS/Sβ0/SD vs. SC/Sβ+), the median age at the time of death in the SS/Sβ0/SD group was 44.7 years (range 24.1-79.4 years), while the median age at the time of death in the SC/Sβ+ group was 59.4 years (range 29.869.2 years). The median survival age for all subjects in the UNC cohort was 50.2 years [95% confidence interval (CI): 45.2-62.3 years) (Figure 1), with a median survival age in the SS/Sβ0/SD group of 49.0 years (95%CI: 44.9-68.6 years). With the small total number of subjects and deceased subjects in the SC/Sβ+ group, the median survival age in this group cannot be reliably estimated.

Factors associated with mortality in UNC cohort Deaths in the UNC cohort were ascertained from hospital records as well as vital statistics data from the North Carolina Center for Health Statistics. The variables in the initial model are shown in Table 2. When covariates were grouped based on quartiles, age (P=0.036), TRV (P=0.012), hemoglobin (P=0.018), creatinine (P=0.047) and log(NTpro-BNP) (P=0.0022) were associated with an increased risk of death (Figure 2A-E) in univariate analysis of a subset in this cohort (SS/Sβ0/SD group). There were no associations between WBC, platelet count, reticulocyte count, HbF, lactate dehydrogenase, the number of acute pain episodes in the previous year, history of acute chest syndrome, or hemoglobin genotype with mortality in univariate analysis of patients in the SS/Sβ0/SD group.

159 (99%) 2 (1%)

Multivariable analysis

119 (73.9%) 12 (7.5%) 10 (6.2%) 19 (11.8%) 1 (0.6%)

We found that the correlation coefficient between each pair of variables was less than 0.6 and, hence, the variables are not collinear (Online Supplementary Table S1). Carrying out the usual backward variable selection, TRV [hazard ratio (HR): 22.7, 95%CI: 5.88-87.5; P<0.0001], frequency of acute pain episodes in the previous year (HR: 1.2, 95%CI: 1.06-1.35; P=0.0038) and D-dimer (HR: 1.24 for 1000 ng/mL increment in D-dimer, 95%CI: 1.07-1.44; P=0.0039) were

161

161

161 161 159 159 158 158 161 158 161 158 158 121 110

93 (58%) 68 (42%) 9.3 (7.2-11.2) 8.9 (7.9-10.3) 408 (303-498) 6.3 (4.3-9.2) 176.35 (125.4-241.6) 5.6 (2.6-10.7) 0.7 (0.6-0.9) 847.5 (644-1136) 1.9(1.1-3.0) 1.8(1.1-2.7) 0.1(0.09-.1) 1234 (740-2023.3) 2.4(2.16-2.8)

161

3 (1-5)

161 135(84%) 26(16%)

Figure 1. Kaplan-Meier survival curves for all subjects and subjects in the SS/Sβ0/SD group in the UNC Cohort. The median age of survival for all subjects was 50.2 years (95%CI: 45.2–62.3 years). The median age of survival of subjects in the SS/Sβ0/SD group was 49.0 years (95%CI: 44.9–68.6 years).

haematologica | 2017; 102(4)


Mortality in sickle cell disease

found to be associated with an increased risk of death. The HR estimates indicate that after controlling for the other variables, the hazard of dying was approximately 23 times higher for every 1 m/s increase in the TRV; for each additional pain episode requiring an emergency department visit or hospitalization, the hazard of dying was increased by 20%, and the hazard of dying was approximately 1.24 times higher for every 1000 ng/mL increase in the D-dimer. When a sensitivity analysis was performed with an assigned TRV value of 1.69 m/s (lower than the lowest measurable TRV value in the cohort) for all missing values, TRV was no longer significantly associated with an increased risk of death. When D-dimer was excluded from the model due to missing values in some subjects, age (HR: 1.04, 95%CI: 1.009-1.073; P=0.012), reticulocyte count (HR: 1.09, 95%CI: 1.015-1.17; P=0.018) and log(NT-pro-BT) (HR: 1.62, 95%CI: 1.21-2.17; P=0.001) were associated with an increased risk of death.

Meta-analysis Our initial search retrieved 1490 articles from PubMed, 1668 articles from Web of Science and 2444 articles from

EMBASE (Figure 3). Nine studies11,21,24,29-34 met our inclusion criteria and were analyzed together with the UNC cohort for a total of 3257 participants. The quality of individual studies was assessed using the NewcastleOttawa Scale (Table 3). We were unable to evaluate publication bias due to the small number of studies included in the meta-analysis.35 Because the Walk-PHaSST cohort31 included patients as young as 12 years old, we restricted our analyses to only those patients who were at least 18 years old. The number of patients in the meta-analyses of the selected variables are shown in Table 3. The number of studies included in the meta-analyses varied from 3 to 9 depending on what had been reported for each variable in the publications. The median age of death of deceased subjects in these studies (available in only 5 of the 9 published studies) ranged between 39.7 and 53 years. The median age of survival for the subjects with severe SCD genotypes (SS/Sβ0/SD), available in only 3 studies (2 of the published 9 studies plus the UNC cohort), ranged between 49.0 and 60.8 years. We found statistically significant associations of age (per 10-year increase in age) (HR: 1.28; 95%CI: 1.10-1.50), TRV 2.5

Table 2. Univariate and multivariate Cox proportional hazards regression analysis of mortality for demographic, laboratory and clinical variables in the UNC cohort.

Covariate

Total number Number of of patients deceased patients

Age (per 10 years) + Sex Male vs. female + Genotype SS/Sβ0/SD SC/Sβ+ + Use of hydroxyurea Yes No + Acute chest syndrome Yes No + Tricuspid regurgitant jet velocity (dichotomized) ≥2.5 m/s <2.5 m/s) Tricuspid regurgitant jet velocity (continuous) Number of pain episodes in previous year White blood cell count Hemoglobin Platelet count Absolute reticulocyte count Fetal hemoglobin Creatinine Reticulocyte count (%) Lactate dehydrogenase Total bilirubin D-dimer (per 1000 ng/mL) Log(NT-pro-BNP) Ferritin

Hazard Ratio (95% CI)a

Hazard Ratio (95% CI)b

161

29

1.59 (1.21-2.09)

1.59 (1.21-2.09)

161 161

29

0.83 (0.39-1.79) 1.12 (0.44-2.81)

0.85 (0.40-1.83) 0.83 (0.34-2.03)

0.592 (0.28-1.25)

0.51 (0.25-1.08)

1.50 (0.51-4.42)

1.15 (0.40-3.31)

2.79 (0.88-8.84)

3.82 (1.24-11.75)

161 17 161

23 6 (29) 12

Hazard Ratio (95% CI)c

110

(29) 25 4 (17)

110

13 4 17

4.98 (1.65-15.02)

6.26 (2.50-15.63)

22.7 (5.88-87.5)

161 161 161 159 158 158 161 159 158 161 121 158 84

29 29 29 29 28 29 29 29 29 29 26 28 8

1.096 (1.01-1.2) 1.11 (0.97-1.27) 0.72 (0.57-0.9) 1.002 (0.99-1.01) 1.002 (1.00-1.01) 0.952 (0.9-1.01) 1.67 (1.04- 2.67) 1.1 (1.02-1.18) 1.00 (0.99-1.0) 1.06 (0.88-1.27) 1.11 (1.001-1.22) 1.59 (1.22-2.08) 1.001 (1.0-1.0)

1.05 (0.97-1.15) 1.05 (0.92-1.19) 0.72 (0.57-0.90) 1.000 (0.998-1.003) 1.001 (0.999-1.004) 0.95 (0.89-1.01) 1.744 (1.14-2.67) 1.05 (0.98-1.12) 1.00 (1.000-1.001) 0.97 (0.81-1.16) 1.083 (0.99-1.19) 1.77 (1.39-2.25) 1.00 (1.000-1.001)

1.20 (1.06-1.35)

1.24 (1.07-1.44)

Hazard ratio adjusted for age and sex; bhazard ratio based on the univariable model; chazard ratio based on the multivariable model. +Dichotomized variables.

a

haematologica | 2017; 102(4)

629


P. Maitra et al.

m/s or more (HR: 3.03; 95%CI: 2.0-4.60), reticulocyte count (HR: 1.05; 95%CI: 1.01-1.10), HbF (HR: 0.97; 95%CI: 0.94-1.00), and log(NT-pro-BNP) (HR: 1.68; 95%CI: 1.48-1.90) with mortality (Figure 4A-E). This means that the weighted hazard of dying was approximately 30% higher for every 10-year increase in age, approximately 200% higher for TRV 2.5 m/s or more compared to lower values, approximately 5% higher for every 1% increase in reticulocyte count; approximately

3% lower for every 1% increase in HbF, and approximately 100% higher for every 1 unit increase in log(NTpro-BNP). No statistically significant associations were found between WBC (HR: 1.05; 95%CI: 0.97-1.14), hemoglobin (HR: 0.86; 95%CI: 0.68-1.09), platelet count (HR: 1.00; 95%CI: 1.00-1.00), ferritin (HR: 1.00; 95%CI: 0.99-1.01), creatinine (HR: 1.20; 95%CI: 0.95-1.52), lactate dehydrogenase (HR: 1.00; 95%CI: 1.00-1.00), total bilirubin (HR:

A

B

C

D

E

Figure 2. Kaplan-Meier survival curves for variables (grouped as quartiles) associated with increased mortality on univariate analysis in the UNC Cohort. (A) Increased age is significantly associated with mortality (log-rank test statistic: 8.53; P=0.036). (B) Increased tricuspid regurgitant jet velocity (TRV) is associated with mortality (log-rank test statistic: 10.93; P=0.012). (C) Decreased hemoglobin is significantly associated with mortality (log-rank test statistic: 10.12; P=0.018). (D) Increased creatinine is significantly associated with mortality (log-rank test statistic: 7.95; P=0.047). (E) Increased log(NT-pro-BNP) is associated with mortality (log-rank test statistic: 14.6; P=0.0022).

630

haematologica | 2017; 102(4)


Mortality in sickle cell disease

1.08; 95%CI: 0.93-1.25), use of hydroxyurea (HR: 0.70; 95%CI: 0.41-1.22), history of acute chest syndrome (HR: 1.14; 95%CI: 0.50-2.59), and male sex (HR: 1.47; 95%CI: 0.41-5.31) with mortality (Online Supplementary Figure S1A-J). All of the studies, except that of Karacaoglu et al.,34 used the time from entry to event as the response variable. We performed a sensitivity analysis by removing this study from the meta-analysis. The results are similar [with and without: WBC 1.05 (0.97, 1.14) vs. 1.03 (0.93, 1.14); hemoglobin 0.87 (0.61, 1.25) vs. 0.92 (0.62, 1.36); and HbF 0.97 (0.94, 1.00) vs. 0.97 (0.94, 1.00)]. The study was, therefore, maintained in the meta-analysis.

Discussion As patients with SCD age, they begin to manifest evidence of end-organ damage which contributes to increased morbidity and early mortality. In random effects meta-analyses, we find statistically significant associations of increased reticulocyte count, TRV and log(NT-pro-BNP) with increased mortality, whereas decreased HbF was associated with increased mortality. While no significant associations were seen between either hemoglobin or lactate dehydrogenase and the risk of death, the association of increased reticulocyte count with mortality suggests that hemolysis may contribute to the increased mortality

observed in SCD. The intensity of hemolytic anemia, evaluated using a composite variable derived from several individual markers of hemolysis (reticulocyte count, serum lactate dehydrogenase, aspartate aminotransferase and total bilirubin concentrations), was associated with an increased risk of death in a previous study of patients with SCD.15 Hemolysis with the subsequent release of cell-free hemoglobin results in the generation of reactive oxygen species which is a potent scavenger of nitric oxide.36 This appears to predispose patients to a vasculopathy, characterized by systemic and pulmonary hypertension, endothelial dysfunction, and proliferative changes in the intima and smooth muscle of blood vessels.37 In addition to their relationship with hemolysis, reticulocytes are more adhesive than mature erythrocytes due to increased expression of the integrin complex, ι4β1 (which binds to both fibronectin and vascular cell adhesion molecule-1, an adhesion receptor expressed on the surface of endothelial cells) as well as expression of CD36 by a subpopulation of sickle reticulocytes.38 The degree of adhesiveness is reportedly correlated with the severity of disease in patients with SCD.39 We further confirm the association of both elevated echocardiography-derived TRV and NT-pro-BNP levels with increased mortality in adult SCD patients. As increased TRV is not sufficiently specific to confidently establish the presence of PHT,10 a right heart catheterization (RHC) is required in patients with elevated TRV val-

Figure 3. Search strategy and results. Our analysis was limited to studies conducted in North America and Europe and published between January 1, 1998 and June 30, 2016.

haematologica | 2017; 102(4)

631


P. Maitra et al. Table 3. Summary of publications in meta-analysis.

Publication

Sickle cell disease genotypes

Variable with Hazard Ratio

*Voskaridou et al., 201021 131

HbSS, HbSβ0, HbSβ+

**Voskaridou et al., 201021 199

HbSS, HbSβ0, HbSβ+

Saraf et al., 201129

306

HbSS, HbSβ0

Mehari et al., 201311

84

Cabrita et al., 201332

164

HbSS, HbSC, HbSβ0, HbSβ+, other genotypes not defined HbSS, HbSC, HbSβ0, HbSβ+, HbSD, HbS/HPFH

Elmariah et al., 201424

542

HbSS, HbSC, HbSβ0, HbSβ+

Gladwin et al., 201431

605

HbSS, HbSC

Age (Median: 33.0, SD: 11.2, HR: 0.99; N/A 0.94-1.05) LDH (Mean: 753.88, SD: 413.62, HR: 0.99; 0.99-1.0) Reticulocyte count (%) (Mean: 10.03, SD: 7.32, HR: 1.01; 0.94-1.09) Hemoglobin (Mean: 9.10, SD: 1.39, HR: 0.88; 0.60-1.3) HbF (Mean: 7.09, SD: 5.26, HR: 0.95; 0.82-1.09) Bilirubin (Mean: 1.92, SD: 3.27, HR : 1.09; 0.99-1.2) Age (Median: 35.0, SD: 12.8, N/A HR: 0.98; 0.93-1.04) LDH (Mean: 701.16, SD: 330.16, HR: 1.0; 1.0-1.0) Reticulocyte count (%) (Mean: 6.88, SD: 6.38, HR: 1.06; 1.02-1.09) Hemoglobin (Mean: 9.43, SD: 1.67, HR: 0.74; 0.62-0.89) HbF (Mean: 5.58, SD: 6.22, HR: 0.92; 0.86-0.99) Bilirubin (Mean: 1.80, SD: 1.61, HR: 1.31; 1.16-1.49) Age (HR: 1.03; 1.0-1.05) N/A Creatinine (HR: 1.14; 0.99-1.3) HbF (HR: 0.99; 0.95-1.04) WBC (HR: 0.94; 0.87-1.01) Age (HR: 1.0; 0.97-1.03) N/A Creatinine (HR: 1.09; 0.71-1.67) Ferritin (HR: 1.26; 1.04-1.53) Age (Median: 42.3, IQR: 33, 50, 49 (range: 25-82) HR: 1.04; 1.0-1.08) Creatinine (μmol/L) (Median: 70, IQR: 59, 85, HR: 5.24; 0.41-67.9) Platelets (HR: 0.99; 0.99-1.0) LDH (Median: 606, IQR: 385, 879, HR: 0.99; 0.99-1.0) Hemoglobin (g/L) (Median: 94, IQR: 80, 110, HR: 0.62; 0.45-0.86) HbF (HR: 1.04; 0.9-1.2) WBC (Median: 9.1, IQR: 7.4, 10.9, HR: 0.93; 0.77-1.13) TRV (≥2.5 m/s: 48/164 {29.27%}, HR: 2.71; 0.98-7.5) Creatinine (HR: 1.39; 1.25-1.53) 45 (range: 20-86) **Platelets (HR: 1.0; 0.95-1.05) Use of hydroxyureaa (Yes: 184/441 {41.7%}, HR: 0.82; 0.55-1.22) History of ACS (Yes: 354/475 {74.5%}, HR: 1.27; 0.81-1.99) **Hemoglobin (HR: 1.21; 1.06-1.38) *Hemoglobin (HR: 1.36; 1.07-1.72) **WBC (HR: 1.06; 1.0-1.11) Bilirubin (HR: 1.01; 0.88-1.16) Log(NT-pro-BNP) (HR: 1.62; 0.66-3.97) Age (Mean: 38.0, SD: 12.57, 40.8 (range: 21.3-65.3) HR: 1.03; 0.99-1.06) Creatinine (Mean: 17.07, SD: 35.45, HR: 0.99; 0.97-1.01) Platelets (Mean: 350.63, SD: 134.85, HR: 0.99; 0.99-1.0)

632

# of subjects

Median/mean age Median age Duration of Quality at death of deceased of survival follow up appraisal subjects (years) of SS/Sβ0/SD (Newcastlesubjects Ottawa Scale (years) [NOS]) N/A

5 years (range: 0.1-18)

6

N/A

8 years (range: 0.1-17)

6

N/A

1/1/1997 6/30/2009

8

N/A

4.7 years 6 (maximum follow up of 11 years) N/A 68.1 months 7 (IQR: 48-78)

58

9.3 years (range: 2.7-10.5)

7

60.8

2.4 years (range: 0.04-3.4)

8

continued on the next page

haematologica | 2017; 102(4)


Mortality in sickle cell disease

continued from the previous page

Schimmel et al., 201533

85

Damy et al., 201630

656

HbSS, HbSC, HbSβ0, HbSβ+ HbSS, HbSβ0

Karacaoglu et al., 201634

324

HbSS, HbSβ0, HbSβ+

UNC cohort, 2015

161

HbSS, HbSC, HbSβ0, HbSβ+, HbSD

Use of hydroxyurea (Yes: 230/605 {38.02%}, HR: 1.62; 0.70-3.73) History of ACS (Yes: 384/605 {63.47%}, HR: 0.67; 0.29-1.55) Ferritin (Mean: 607.88, SD: 1297.61, HR: 1.0; 1.0-1.0) LDH (Mean: 451.82, SD: 298.79, HR: 1.0; 1.0-1.0) Sex (Male: 278/605 {45.95%} HR: 2.52; 1.03-6.19) Reticulocyte count (Mean: 8.91, SD: 5.99, HR: 1.03; 0.96-1.1) Hemoglobin (Mean: 9.31, SD: 1.92, HR: 0.9; 0.71-1.13) HbF (Mean: 7.30, SD: 7.64, HR: 0.96; 0.9-1.04) WBC (Mean: 9.68, SD: 3.66, HR: 1.1; 1.0-1.23) Bilirubin (Mean: 14.66, SD: 31.73, HR: 0.97; 0.93-1.01) Log(NT-pro-BNP) (Mean: 4.24, SD: 1.56, HR: 1.73; 1.41-2.12) TRV (≥2.5 m/s: 347/562 {61.74%}, HR: 3.5; 1.03-11.9) TRV (≥2.5 m/s: 25/81 {30.86%}, 53 (IQR: 37-60) HR: 1.1; 0.3-3.7) Age (HR: 1.03; 0.99-1.06) N/A Sex (HR: 1.60; 0.91-2.75) Hemoglobin(g/L) (HR: 1.44; 0.52-3.97) HbF (HR: 0.97; 0.86-1.10) TRV (HR: 3.92; 1.97-7.78) HbF (HR: 0.94; 0.74-1.2) 36.6 ± 13 Hemoglobin (HR: 0.62; 0.45-0.83) WBC (HR: 1.10; 1.05-1.16) Age (Mean: 36.83, SD: 12.37, 48.0 HR: 1.05; 1.02-1.08) (range: Creatinine (Mean: 0.85, SD: 0.52, 29.8-70.2) HR: 1.67; 1.04-2.67) Platelets (Mean: 411.45, SD: 156.80, HR: 1.0; 0.99-1.01) Use of hydroxyurea (Yes: 93/161 {57.76%}, HR: 0.59; 0.28-1.25) History of ACS (Yes: 135/161 {83.85%}, HR: 1.49; 0.51-4.42) Ferritin (Mean: 631.76, SD: 813.42, HR: 1.0; 1.0-1.0) LDH (Mean: 984.03, SD: 496.54, HR: 1.0; 0.99-1.0) Gender (Male: 64/161 {39.75%}, HR: 0.83; 0.39-1.79) Reticulocyte count (Mean: 7.38, SD: 4.68, HR: 1.1; 1.02-1.18) Hemoglobin (Mean: 9.15, SD: 1.75, HR: 0.72; 0.57-0.9) HbF (Mean: 8.24, SD: 7.52, HR: 0.95; 0.9-1.0) WBC (Mean: 9.40, SD: 2.81, HR: 1.1; 0.97-1.27) Bilirubin (Mean: 2.49, SD: 2.34, HR: 1.06; 0.88-1.27) Log(NT-proBNP) (Mean: 4.74, SD: 1.09, HR: 1.59; 1.22-2.08) TRV (≥2.5 m/s): 47/110 {42.73%}, HR: 2.79; 0.88-8.83)

N/A

82 months (IQR: 75-85) 48 months (range: 32-59)

N/A

N/A

7 7

66 ± 44 months 7 (range: 3 – 148 months)

49.0 (95% CI: 44.9-68.6)

6.4 years (range: 0.06-10.3)

7

*With Hydroxyurea; ** Without Hydroxyurea; LDH: lactate dehydrogenase; HbF: fetal hemoglobin; HR: hazard ratio; Pain Crisis: frequency of Pain Crisis; TRV: tricuspid regurgitant jet velocity; IQR: interquartile range; CI: confidence interval. *The paper by Voskariduo et al.21 reports OR for Cox PH model: we have treated them as HRs. Here, reported median is presumed to be the same as the mean; data have been pooled to obtain overall mean. aOnly HbSS/HbSβ0 thalassemia patients evaluated.

ues to confirm the diagnosis. With the strength of data showing that increased TRV and PHT are associated with mortality in SCD, we recommend that patients undergo periodic echocardiographic screening for assessment of mortality risk, consistent with recently published clinical practice guidelines.40 NT-pro-BNP levels assess ventricular strain and levels greater than 160 pg/mL are reported to be an independent risk factor for mortality in SCD.41,42 The utility of NT-pro-BNP alone as a screening tool for PHT has not been studied, although in combination with TRV haematologica | 2017; 102(4)

and 6-minute walk distance, it appears to increase the predictive capacity for PHT.10 In the UNC cohort, the effect of all the variables that were selected in the final model were stronger in the multivariable model than in the univariable model, without any change in the direction. D-dimer and the numbers of pain episodes were not significantly associated with mortality when they were considered individually, but the effects became significant after controlling for each other, TRV and genotype. The effect for TRV was statistically 633


P. Maitra et al. A

B

C

Figure 4. Forest plots of hazard ratios for variables significantly associated with mortality in random effects meta-analyses (continued on next page). (A) Age (per 10-year increase in age) was significantly associated with mortality [hazard ratio (HR): 1.28; 95% confidence interval (CI): 1.10-1.50]. (B) TRV ≥ 2.5 m/s (HR: 3.03; 95%CI: 2.0-4.60). (C) Reticulocyte count was significantly associated with mortality (HR: 1.05; 95%CI: 1.01-1.10);

significant in univariable analysis and it became much stronger in the multivariable model. This meta-analysis provides additional evidence that decreased HbF level is associated with increased mortality in SCD. There is strong epidemiological and clinical evidence that elevated HbF level is associated with relatively mild clinical manifestations of SCD.7 The reduction of disease severity by elevated levels of HbF is the basis for the development of HbF-inducing drugs. The beneficial effect of hydroxyurea in SCD is thought to be largely due to the induction of HbF following ‘stress erythropoiesis’, although the production of nitric oxide and the soluble guanylyl cyclase and cGMP-dependent protein kinase 634

pathway may play a role in induced expression of the γglobin gene.43 Although no significant association was observed between the use of hydroxyurea and mortality in the meta-analysis, there are not sufficient data on adherence to hydroxyurea or the doses of hydroxyurea in the individual studies. Sickle cell disease has been described as an inflammatory disease.44 While not statistically significant, there was a trend towards an association between WBC and mortality in the meta-analysis. Although leukocytes contribute to disease pathophysiology,45 it remains uncertain whether a decrease in the leukocyte count is associated with clinical benefit.17,19 It has been suggested that the association haematologica | 2017; 102(4)


Mortality in sickle cell disease

D

E

Figure 4. (continued). Forest plots of hazard ratios for variables significantly associated with mortality in random effects meta-analyses. (D) Fetal hemoglobin was significantly associated with mortality (HR: 0.97; 95% CI: 0.94-1.0). (E) Log(NT-pro-BNP) was significantly associated with mortality (HR: 1.68; 95%CI: 1.48-1.90).

between neutropenia and the frequency of pain episodes in the Multicenter Study of Hydroxyurea (MSH)17 was 'forced' due to the study design requiring a titration of hydroxyurea to maximum tolerated dose. This appears to be supported by the MSH follow-up study, during which patients received doses of hydroxyurea that were less than the maximum tolerated doses, with no notable effects on leukocyte counts.19 In addition to its inflammatory nature, SCD is characterized by hypercoagulability.46,47 There is increasing evidence that coagulation activation plays a role in disease pathophysiology46 and may be associated with complications of SCD.47,48 Consistent with the hypercoagulable state in SCD, venous thromboembolism is common and is associated with increased mortality in this patient population.49 Further studies are required to define the contribution of coagulation activation to SCD pathogenesis as well as its effects on adverse outcomes. The present study appears to confirm previous reports of early death in patients with SCD. It is conceivable that the current practice of commencing hydroxyurea in young children, possibly combined with subsequent decreases in both morbidity and end-organ damage, may result in greater longevity of SCD patients. A prospective, multicenter study will be required to confirm whether the survival of SCD patients has improved in the hydroxyurea era. This study has several limitations. The studies in the meta-analysis were conducted at referral centers which usually see only the more severe cases; the studies may, therefore, not be representative of the general SCD popuhaematologica | 2017; 102(4)

lation. The number of eligible contemporary studies evaluating mortality in adult patients with SCD was relatively small and meta-analyses were not performed for several clinically important complications such as acute pain episodes and stroke. Although considerable effort was made to study unique patient populations, we recognize that some individual patients may be represented in more than one cohort or publication. Patients with varied SCD genotypes were enrolled in the individual studies, with inadequate data on concomitant alpha thalassemia, and dosing and adherence to hydroxyurea. We did not include studies from Africa and other developing countries due to our concerns regarding the limited healthcare infrastructure in many of these countries and the limited use of hydroxyurea compared with the United States and Europe. Despite these limitations, this meta-analysis of 9 published studies as well as the UNC cohort represents the largest number of SCD patients (n=3257) in whom risk factors for mortality have been evaluated in the hydroxyurea era. With the increased availability and utility of hydroxyurea, a prospective study to evaluate the factors associated with mortality in adult patients is warranted. Acknowledgments We acknowledge Lara Handler for assistance with the search strategy, and the Clinical and Translational Research Center at the University of North Carolina at Chapel Hill, which is funded by NIH grant UL1RR025747. 635


P. Maitra et al.

References 1. Quinn CT, Rogers ZR, Buchanan GR. Survival of children with sickle cell disease. Blood. 2004;103(11):4023-4027. 2. Gaston MH, Verter JI, Woods G, et al. Prophylaxis with oral penicillin in children with sickle cell anemia. A randomized trial. N Engl J Med. 1986;314(25):1593-1599. 3. Vichinsky E, Hurst D, Earles A, Kleman K, Lubin B. Newborn screening for sickle cell disease: effect on mortality. Pediatrics. 1988;81(6):749-755. 4. Vichinsky EP. Comprehensive care in sickle cell disease: its impact on morbidity and mortality. Semin Hematol. 1991;28(3):220-226. 5. Vichinsky EP, Earles A, Johnson RA, Hoag MS, Williams A, Lubin B. Alloimmunization in sickle cell anemia and transfusion of racially unmatched blood. N Engl J Med. 1990;322(23):1617-1621. 6. Adams RJ, McKie VC, Hsu L, et al. Prevention of a first stroke by transfusions in children with sickle cell anemia and abnormal results on transcranial Doppler ultrasonography. N Engl J Med. 1998;339(1):5-11. 7. Platt OS, Brambilla DJ, Rosse WF, et al. Mortality in sickle cell disease. Life expectancy and risk factors for early death. N Engl J Med. 1994;330(23):1639-1644. 8. Gladwin MT, Sachdev V, Jison ML, et al. Pulmonary hypertension as a risk factor for death in patients with sickle cell disease. N Engl J Med. 2004;350(9):886-895 9. Ataga KI, Moore CG, Jones S, et al. Pulmonary hypertension in patients with sickle cell disease: a longitudinal study. Br J Haematol. 2006;134(1):109-115. 10. Parent F, Bachir D, Inamo J, et al. A hemodynamic study of pulmonary hypertension in sickle cell disease. N Engl J Med. 2011;365(1):44-53. 11. Mehari A, Alam S, Tian X, et al. Hemodynamic predictors of mortality in adults with sickle cell disease. Am J Respir Crit Care Med. 2013;187(8):840-847. 12. Machado RF, Anthi A, Steinberg MH, et al. N-terminal-pro-brain natriuretic peptide levels and risk of death in sickle cell disease. JAMA. 2006;296(3):310-318. 13. Boyd JH, Macklin EA, Strunk RC, DeBaun MR. Asthma is associated with increased mortality in individuals with sickle cell anemia. Haematologica. 2007;92(8):1115-1118. 14. McClellan AC, Luthi JC, Lynch JR, et al. High one year mortality in adults with sickle cell disease and end-stage renal disease. Br J Haematol. 2012;159(3):360-367. 15. Nouraie M, Lee JS, Zhang Y, et al. The relationship between the severity of hemolysis, clinical manifestations and risk of death in 415 patients with sickle cell anemia in the US and Europe. Haematologica. 2013;98(3): 464-472. 16. Upadhya B, Ntim W, Brandon Stacey R, et al. Prolongation of QTc intervals and risk of death among patients with sickle cell disease. Eur J Haematol. 2013;91(2):170-178. 17. Charache S, Terrin ML, Moore RD, et al. Effect of hydroxyurea on the frequency of painful crises in sickle cell anemia. Investigators of the Multicenter Study of Hydroxyurea in Sickle Cell Anemia. N Engl

636

J Med. 1995;332(20):1317-1322. 18. Wang W, Ware RE, Miller ST, et al. Hydroxycarbamide in very young children with sickle-cell anaemia: a multicenter, randomized, controlled trial (BABY HUG). Lancet. 2011;377(9778):1663-1672. 19. Steinberg MH, Barton F, Castro O, et al. Effect of hydroxyurea on mortality and morbidity in adult sickle cell anemia: risks and benefits up to 9 years of treatment. JAMA. 2003;289(13):1645-1651. 20. Steinberg MH, McCarthy WF, Castro O, et al. The risks and benefits of long-term use of hydroxyurea in sickle cell anemia: A 17.5 year follow-up. Am J Hematol. 2010;85(6): 403-408. 21. Voskaridou E, Christoulas D, Bilalis A, et al. The effect of prolonged administration of hydroxyurea on morbidity and mortality in adult patients with sickle cell syndromes: results of a 17-year, single-center trial (LaSHS). Blood. 2010;115(12):2354-2363. 22. Lanzkron S, Carroll CP, Haywood C Jr. Mortality rates and age at death from sickle cell disease: U.S., 1979-2005. Public Health Rep. 2013;128(2):110-116. 23. Ballas SK, Lieff S, Benjamin LJ, et al. Definitions of the phenotypic manifestations of sickle cell disease. Am J Hematol. 2010;85(1):6-13. 24. Elmariah H, Garrett ME, De Castro LM, et al. Factors associated with survival in a contemporary adult sickle cell disease cohort. Am J Hematol. 2014;89(5):530-535. 25. Stroup DF, Berlin JA, Morton SC, et al. Metaanalysis of observational studies in epidemiology: a proposal for reporting. Metaanalysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008-2012. 26. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled clinical trials. 1986;7(3):177-188. 27. Knapp G, Hartung J. Improved tests for a random effects meta-regression with a single covariate. Stat Med. 2003;15;22(17): 2693-2710. 28. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in metaanalyses. Br Med J. 2003;327(7414):557-560. 29. Saraf S, Farooqui M, Infusino G, et al. Standard clinical practice underestimates the role and significance of erythropoietin deficiency in sickle cell disease. Br J Haematol. 2011;153(3):386-392. 30. Damy T, Bodez D, Habibi A, et al. Haematological determinants of cardiac involvement in adults with sickle cell disease. Eur Heart J. 2016;37(14):1158-1167. 31. Gladwin MT, Barst RJ, Gibbs JS, et al. Risk factors for death in 632 patients with sickle cell disease in the United States and United Kingdom. PLoS One. 2014;9:e99489. 32. Cabrita IZ, Mohammed A, Layton M, et al. The association between tricuspid regurgitation velocity and 5-year survival in a North West London population of patients with sickle cell disease in the United Kingdom. Br J Haematol. 2013;162(3):400-408. 33. Schimmel M, van Beers EJ, van Tuijn CF, et al. N-terminal pro-B-type natriuretic peptide, tricuspid jet flow velocity, and death in adults with sickle cell disease. Am J Hematol. 2015;90(4):E75-76.

34. Karacaoglu PK, Asma S, Korur A, et al. East Mediterranean region sickle cell disease mortality trial: retrospective multicenter cohort analysis of 735 patients. Ann Hematol. 2016;95(6):993-1000. 35. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ. 2007;176(8):1091-1096 36. Rother RP, Bell L, Hillmen P, Gladwin MT. The clinical sequelae of intravascular hemolysis and extracellular plasma hemoglobin: a novel mechanism of human disease. JAMA. 2005;293(13):1653-1662. 37. Kato GJ, Gladwin MT, Steinberg MH. Deconstructing sickle cell disease: reappraisal of the role of hemolysis in the development of clinical subphenotypes. Blood Rev. 2007;21(1):37-47. 38. Bunn HF. Pathogenesis and treatment of sickle cell disease. N Engl J Med. 1997;337(11):762-769. 39. Hebbel RP, Boogaerts MA, Eaton JW, Steinberg MH. Erythrocyte adherence to endothelium in sickle cell anemia: a possible determinant of disease severity. N Engl J Med. 1980;302(18):992-995. 40. Klings ES, Machado RF, Barst RJ, et al. An official American Thoracic Society clinical practice guideline: diagnosis, risk stratification, and management of pulmonary hypertension of sickle cell disease. Am J Respir Crit Care Med. 2014;189(6):727-740. 41. Machado RF, Anthi A, Steinberg MH, et al. N-terminal pro-brain natriuretic peptide levels and risk of death in sickle cell disease. JAMA 2006;296(3):310-318. 42. Machado RF, Hildesheim M, Mendelsohn L, Remaley AT, Kato GJ, Gladwin MT. NT-pro brain natriuretic peptide levels and the risk of death in the cooperative study of sickle cell disease. Br J Haematol. 2011;154(4): 512520. 43. McGann PT, Ware RE. Hydroxyurea for sickle cell anemia: what have we learned and what questions still remain? Curr Opin Hematol. 2011;18(3):158-165. 44. Hebbel RP, Osarogiagbon R, Kaul D. The endothelial biology of sickle cell disease: inflammation and a chronic vasculopathy. Microcirculation. 2004;11(2):129-151. 45. Manwani D, Frenette PS. Vaso-occlusion in sickle cell disease: pathophysiology and novel targeted therapies. Blood. 2013;122 (24):3892-3898. 46. Chantrathammachart P, Mackman N, Sparkenbaugh E, et al. Tissue factor promotes activation of coagulation and inflammation in a mouse model of sickle cell disease. Blood. 2012;120(3):636-646. 47. Ataga KI, Brittain JE, Desai P, et al. Association of coagulation activation with clinical complications in sickle cell disease. PLoS One. 2012;7:e29786. 48. Arumugam PI, Mullins ES, Shanmukhappa SK, et al. Genetic diminution of circulating prothrombin ameliorates multiorgan pathologies in sickle cell disease mice. Blood. 2015;126(15):1844-1855. 49. Naik RP, Streiff MB, Haywood C Jr, Segal JB, Lanzkron S. Venous thromboembolism incidence in the Cooperative Study of Sickle Cell Disease. J Thromb Haemost. 2014; 12(12):2010-2016.

haematologica | 2017; 102(4)


ARTICLE

Hematopoiesis

Dual role of IL-21 in megakaryopoiesis and platelet homeostasis

Salima Benbarche, Catherine Strassel, Catherine Angénieux, Léa Mallo, Monique Freund, Christian Gachet, François Lanza, Henri de la Salle

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Université de Strasbourg, INSERM, EFS Grand-Est, BPPS UMR-S 949, FMTS, F-67000, France

ABSTRACT

G

ene profiling studies have indicated that in vitro differentiated human megakaryocytes express the receptor for IL-21 (IL-21R), an immunostimulatory cytokine associated with inflammatory disorders and currently under evaluation in cancer therapy. The aim of this study was to investigate whether IL-21 modulates megakaryopoiesis. We first checked the expression of IL-21 receptor on human bone marrow and in vitro differentiated megakaryocytes. We then investigated the effect of IL-21 on the in vitro differentiation of human blood CD34+ progenitors into megakaryocytes. Finally, we analyzed the consequences of hydrodynamic transfection-mediated transient expression of IL-21, on megakaryopoiesis and thrombopoiesis in mice. The IL-21Rα chain was expressed in human bone marrow megakaryocytes and was progressively induced during in vitro differentiation of human peripheral CD34+ progenitors, while the signal transducing γ chain was down-regulated. Consistently, the STAT3 phosphorylation induced by IL-21 diminished during the later stages of megakaryocytic differentiation. In vitro, IL-21 increased the number of colony-forming unit megakaryocytes generated from CD34+ cells and the number of megakaryocytes differentiated from CD34+ progenitors in a JAK3- and STAT3-dependent manner. Forced expression of IL-21 in mice increased the density of bi-potent megakaryocyte progenitors and bone marrow megakaryocytes, and the platelet generation, but increased platelet clearance with a consequent reduction in blood cell counts. Our work suggests that IL-21 regulates megakaryocyte development and platelet homeostasis. Thus, IL-21 may link immune responses to physiological or pathological platelet-dependent processes.

Haematologica 2017 Volume 102(4):637-646

Correspondence: henri.delasalle@efs.sante.fr

Received: February 15, 2016. Accepted: January 4, 2017. Pre-published: January 5, 2017. doi:10.3324/haematol.2016.143958

Introduction Megakaryopoiesis is mainly controlled by thrombopoietin (TPO). In vitro, TPO is essential to differentiate hematopoietic progenitors into megakaryocytes, a differentiation enhanced by cytokines such as IL-6, IL-1β, IL-3, IL-9 and IL-11.1 In vivo, megakaryopoiesis occurs in the bone marrow (BM), a complex environment in which innate and adaptive immune cells produce cytokines regulating this process, some positively (such as IL-6, TNF-α and IL-1β2,3), others negatively (IL-10, IL-4 and TGF-β4-6). This influence is exemplified by reactive thrombocytosis, attributed to infections or inflammatory diseases7 and largely mediated by IL-6.8 More recently, IL-1α was shown to enhance thrombopoiesis.9 Whether other cytokines are involved in the regulation of megakaryopoiesis requires further investigation in order to better understand inflammatory thrombocytosis. To address this issue, we compared published gene profiles of CD34+ progenitors and in vitro differentiated megakaryoblasts.10 This analysis indicated the presence of IL-21 receptor (IL-21R) on megakaryocytes. IL-21R is a heterodimer composed of a specific alpha chain (IL-21Rα) and the haematologica | 2017; 102(4)

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/637 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

637


S. Benbarche et al.

common gamma chain (IL-2Rγ) required for signal transduction.11 IL-21 is produced by subsets of natural killer (NK) T cells and helper CD4+ T cells, in particular follicular Th cells and Th17 cells. In healthy individuals, the BM contains T cells producing IL-21,12 probably follicular Th cells whose frequency may increase in pathological states.13 IL-21 controls a variety of responses of different immune cells such as B, NK and T lymphocytes, macrophages and dendritic cells, and also vascular endothelial cells.11 IL-21 is associated with the development of autoimmune diseases and inflammatory disorders, and hence, like IL-6, could play a role in reactive thrombocytosis. Notably, IL-6 induces the production of IL-21 by CD4+ T cells. Recently, IL-21 was found to be expressed by mouse hematopoietic stem cells and progenitors when stimulated by TLR activators released by apoptotic cells.14 The aim of this study was to investigate the role of IL-21 on megakaryopoiesis, using in vitro assays with human cells and in vivo experiments in mice.

Flow cytometry For flow cytometry (FC), cells were labeled as described in the Online Supplementary Appendix and analyzed on a Gallios or a BD LSRFortessa cytometer; data were analyzed with Kaluza (Beckman Coulter) or FACSDiva (BD Biosciences) software, respectively.

Hydrodynamic transfections Murine IL-21 cDNA was cloned into the pLIVE expression vector (Mirus Bio LLC). Empty or recombinant plasmids were intravenously injected into mice.15 Plasma samples were stored at -80°C before quantification of IL-21 concentration.

Mouse platelets The percentage of reticulated platelets was checked by FC after staining with Thiazole Orange (TO) and anti-CD42c mAb. To measure platelet survival, washed EGFP+ platelets were retroorbitally injected into mice five days after hydrodynamic transfection. The ratio of EGFP+ transfused to EGFP-endogenous platelets was determined by FC.

Statistical analysis Methods Techniques are described in detail in the Online Supplementary Appendix.

All values are reported as the mean±Standard Error of Mean (SEM). Statistical analyses were performed with GraphPad software (Prism v.5.0) using Student t-test, or one-way or two-way ANOVA and a Bonferroni post-hoc test.

Antibodies Mouse anti-human CD41/CD61, CD42b and CD42d and rat anti-mouse and human CD42c antibodies were prepared in our laboratory; other antibodies were available commercially.

In vitro differentiation of human MKs Peripheral CD34+ progenitors from healthy blood donors were isolated and cultured in serum free media and appropriate cytokine combinations.

Colony-forming unit megakaryocyte assays Colony-forming unit megakaryocyte (CFU-MK) assays were performed using MegaCult™-C Kits with cytokines (StemCell Technologies) according to the manufacturer’s instructions. Three independent experiments were performed in quadruplicate.

Quantification of proplatelet-bearing human MKs On day 13 of culture, images of ten different wide fields in 24well plates were acquired using an inverted microscope coupled to a camera (Zeiss). Round and proplatelet-bearing megakaryocytes were counted.

RT-PCR analyses Semi-quantitative RT-PCRs were performed using total RNA of CD34+ progenitors and cultured CD41/CD61+ cells. The identity of the RT-PCR products was confirmed by DNA sequencing.

Immunofluorescence microscopy Human BM specimens from the iliac crest were obtained from individuals having a normal megakaryocytic lineage. Mouse femora, spleens and livers were harvested, fixed, then decalcified (femora). Samples were embedded in OCT compound and cryosectioned at 8 µm. On day 13 of culture, megakaryocytes were fixed and cytospun onto poly L-lysine-coated slides. IL-21Rα was revealed using a tyramide amplification technique (TSA PLUS Fluorescence Kit, Perkin Elmer). MKs and macrophages were counterstained with antiCD42c or F4/80 antibody, respectively, before analysis by confocal microscopy (TCS SP5, Leica Microsystems). 638

Results Human megakaryocytes express the IL-21 receptor Peripheral blood CD34+ progenitors were differentiated into megakaryocytes using a 2-phase protocol optimized to generate large numbers of megakaryocytes: firstly, seven days of culture in the presence of TPO, IL-6, IL-9 and SCF, to allow the expansion of megakaryocyte progenitors, and secondly, six days of culture in the presence of TPO alone, to generate mature megakaryocytes. RNA was extracted from freshly isolated CD34+ cells and from purified CD41/CD61+ cells, isolated on days 4, 7 and 10 of culture (Figure 1A, Scheme 1, without IL-21). Semi-quantitative RTPCR analyses showed that IL21R was not expressed in freshly isolated CD34+ cells but was progressively induced during their megakaryocytic differentiation. In contrast, IL2RG transcripts were detected in the progenitor cells, their numbers relatively increased during the first days of culture, and then decreased (Figure 1B). FC analysis of in vitro differentiated megakaryocytes confirmed the progression of IL-21Rα chain expression on CD41/CD61+ cells (Figure 1C). In human BM samples, IL-21Rα was expressed on a subpopulation of mature megakaryocytes identifiable owing to their large size (> 20 μm) and CD42c and DAPI staining (Figure 1D). Among 211 megakaryocytes from 3 individuals, 37%±7% expressed IL-21R. The final steps of megakaryopoiesis are characterized by proplatelet formation followed by platelet release. At day 13 of culture, fixed and cytospun megakaryocytes were analyzed by immunofluorescence microscopy after labeling with anti-CD42c and anti-IL-21R mAbs (Online Supplementary Figure S1A). IL-21Rα was detected on the cellular body of about 65% of the cells but not on proplatelets. Accordingly, IL-21Rα could not be revealed on the surface of human blood platelets, in the resting state as previously described16 nor after thrombin-stimulation, nor intracellularly (Online Supplementary Figure S1B). haematologica | 2017; 102(4)


Modulation of platelet homeostasis by IL-21

A

B

C

D

Figure 1. Human in vitro differentiated and bone marrow megakaryocytes (BM MKs) express the IL-21 receptor. (A) Human CD34+ cells were isolated from adult peripheral blood and cultured in a serum-free medium, in the presence of Megakaryocyte Expansion Supplement (CC220 cocktail, containing TPO, SCF, IL-6 and IL9) for the first 7-day phase and of thrombopoietin (TPO) alone for the second 6-day phase. (B) Gel electrophoresis analysis of the RT-PCR products of IL2RG and IL21R mRNA and 18S rRNA (internal control) extracted from freshly isolated CD34+ cells (day 0) and CD41/CD61+ sorted cells on days 4, 7 and 10 of culture. A representative gel from three independent experiments is shown. RT0 is the negative control without cDNA. (C) Cells were labeled with anti-CD41/CD61-Alexa 488, -IL21Rα-APC antibodies, 7AAD and analyzed by flow cytometry (FC). Representative FC dot plot analyses of the CD41/CD61 and IL-21Rα distribution among live cells (7AAD-) from day 0, 4, 7 and 10 cultures are shown. The percentages of live cells are depicted in each quadrant as the mean±Standard Error of Mean (SEM) in three independent experiments. (D) Human BM samples were fixed and labeled with an anti-IL-21Rα antibody, stained with FITC using a tyramide-based amplification method (green) and counterstained with an anti-CD42c-Alexa 555 antibody (red). Nuclei were stained with DAPI (blue). Scale bar: 20 μm. The Figure shows two representative images of samples from 3 individuals. Arrows denote IL-21Rα+ MKs. Inserts represent merged views (0.5-fold smaller scale).

Altogether, these results confirm the expression of IL-21 receptor on human megakaryocytes and suggest a role of IL-21 during megakaryopoiesis but not in platelet functions.

IL-21R activity changes during the in vitro differentiation of megakaryocytes In order to check the functionality of IL-21R in CD34+ cell-derived megakaryocytes, we focused on STAT3 and STAT5 phosphorylation, which occurs after stimulation of T cells by IL-2111 and of megakaryocytes by TPO.17 At days 7 and 12 of megakaryocyte differentiation, corresponding to the ends of the two culture phases, the cells were starved of cytokines for 5 hours, then incubated for 15 minutes with IL-21 and/or TPO, before fixation and immunostaining with anti-CD41/CD61, -pSTAT3 and -pSTAT5 antibodies. FC analysis revealed that IL-21 induced the phosphorylation of only STAT3 in CD41/CD61+ cells, the ratio of responding cells being higher on day 7 (46%±4.9%, n=3) than on day 12 (15%±2.9%, n=3) (Figure 2). TPO induced mainly STAT5 phosphorylation on day 7 of culture, and both STAT3 and STAT5 on day 12. On day 7, IL-21 and TPO signaling were haematologica | 2017; 102(4)

additive. To confirm that STAT3 phosphorylation was mediated by IL-21R, we used the neutralizing anti-IL-21R recombinant monoclonal antibody ATR-107.18 As expected, this antibody completely inhibited the IL-21-induced phosphorylation of STAT3 in CD41+ cells at day 7 of culture when under our experimental conditions signal transduction was maximal (Online Supplementary Figure S2). Thus, IL-21R is functional during in vitro differentiation of megakaryocytes but, probably due to downregulation of the γ chain, tends to lose its functionality during their maturation.

IL-21 promotes the in vitro proliferation of megakaryocyte progenitors through the JAK3/STAT3 pathway We first measured the effect of IL-21 on the megakaryocytic potential of CD34+ progenitor cells in collagenbased medium containing TPO, IL-3 and IL-6. IL-21 increased the number of medium-sized (21-50 cells per colony) and large (>50 cells per colony) colonies, on average by approximately 30% and 90%, respectively (n=3) (Figure 3A). We then investigated the effects of IL-21 on the in vitro differentiation of CD34+ progenitors into 639


S. Benbarche et al.

megakaryocytes. When blood CD34+ cells were cultured for seven days in the presence of TPO alone (Figure 1A, scheme 2), addition of IL-21 increased the total number of viable cells by 50% on average and doubled the number of CD41+ cells (Figure 3B-D). Specificity of the IL-21-induced cellular responses was confirmed by blocking IL-21R with anti-IL-21R antibody ATR-107: addition of this antibody on days 1 and 3 completely blocked IL-21-induced increase in the yield of the total number of cells and the percentage of CD41+ cells obtained at day 7 (Figure 3E). IL-6 in combination with soluble IL-6R has been shown to increase the yield of the in vitro differentiation of CD34+ progenitors into megakaryocytes in the presence of TPO alone.19 IL-21 increased this effect without impacting the ratio of CD41+, CD42+ or GPV+ cells as assessed in 7-day culture assays (Figure 3F and data not shown). However, when CD34+ cells were cultured in the presence of the Megakaryocyte Expansion Supplement cocktail (Figure 1A, scheme 3, CC220), addition of IL-21 in the first step did not improve the number of viable cells and of CD41+ cells obtained at day 7 (data not shown). In order to obtain larger numbers of megakaryocyte progenitors, CD34+ cells were differentiated in the presence of CC220 cocktail according to cell culture scheme 1. Cells obtained after seven days of culture were cultured in the presence of TPO for six additional days in the presence of increasing concentrations of IL-21. Under these experimental conditions, IL-21 increased the number of megakaryocytes in a dose-dependent manner, reaching 1.9±0.1-fold increase at 100 ng/mL (n=6) (Figure 3G). The cell viability was not modified by the presence of IL-21 (data not shown). Among the Janus kinases, JAK3 is the only molecule linked to IL-2Rγ, mediating signal transduction via STAT3 phosphorylation. Since IL-21 induced STAT3 phosphorylation in differentiating megakaryocytes, we checked whether its effect on megakaryocyte generation depended on the JAK3/STAT3 pathway. CD34+ progenitors were cultured according to scheme 1; on day 7, several concentrations of Tofacitinib (a JAK3 inhibitor) and Stattic (a STAT3 inhibitor), or their vehicle (DMSO, 0.25%), were added to the cultures, supplemented or not with IL-21 (100 ng/mL). Analysis of cells five days later revealed that 300 nM Stattic or 1500 nM Tofacitinib had no impact on the number of megakaryocytes obtained in the presence of TPO and the vehicle (DMSO) (Figure 3H). The combination of TPO+IL-21 afforded a 1.5±0.03-fold (n=5) increase in the number of megakaryocytes, as compared to TPO with the vehicle. Addition of Stattic or Tofacitinib abolished this enhancement (1.16±0.03- and 1.06±0.05fold, respectively; n=5). FC analysis revealed that the inhibitors, at these concentrations, did not affect the differentiation of megakaryocytes since in all conditions 90% of the cells were CD41+ CD42+ (data not shown). Altogether, these results indicate that IL-21 promotes the in vitro proliferation of megakaryocyte progenitors through the JAK3/STAT3 pathway.

Megakaryocyte phenotype and platelet production are not modified by IL-21 At the end of the second phase of culture (scheme 1), we evaluated the influence of IL-21 on the megakaryocyte phenotype (Online Supplementary Figure S3). IL-21 did not significantly affect the surface expression of CD41, CD42c or CD42d on viable cells, their ploidy, the proportion of 640

proplatelet-bearing megakaryocytes, or the number of platelets released per megakaryocyte. Thus, IL-21 had no obvious effect on the phenotype of mature megakaryocytes or on platelet production.

A

B

Figure 2. IL-21R activity changes during in vitro differentiation of megakaryocytes (MKs). CD34+ cells were cultured as described in Figure 1, scheme 1. On day 7 or 12 of culture, the cells were pre-incubated for five hours in serum-free medium without cytokines, before stimulation with thrombopoietin (TPO) and/or IL-21. The cells were then fixed, permeabilized and labeled with antiCD41/CD61-ECD, -pSTAT3-Alexa 647 and -pSTAT5-Alexa 488 antibodies. Representative flow cytometry (FC) dot plots showing STAT3 and STAT5 phosphorylation in CD41/CD61+ cells after stimulation as indicated on days 7 and 12 of culture. Values indicate the mean±Standard Error of Mean (SEM) of the percentage of events in each quadrant in three independent experiments.

haematologica | 2017; 102(4)


Modulation of platelet homeostasis by IL-21

IL-21 stimulates megakaryopoiesis in vivo To appraise the in vivo relevance of these observations, additional experiments were performed in mice. Firstly, we confirmed that IL-21 increased the number of megakaryocytes differentiated in vitro from murine BM Lin– cells in the presence of TPO (1.4±0.15-fold increase; n=5), without affecting the phenotype of the cells (expres-

A

C

F

sion of CD41/CD61 heterodimer and CD42c, and polyploidy) (Online Supplementary Figure S4). This positive effect could be inhibited by the ATR-107 mAb, which also antagonizes mouse IL-21R. IL-21 was then transiently expressed in mice using hydrodynamic transfection of an IL-21 expression vector (pLIVE-IL-21). In negative control mice receiving the empty vector (pLIVE), the plasma con-

B

E

D

G

H

Figure 3. IL-21 promotes in vitro proliferation of megakaryocyte (MK) progenitors through the JAK3/STAT3 pathway. CD34+ cells were grown in the presence (filled bars) or absence (empty bars) of IL-21 (100 ng/mL). (A) 2500 CD34+ cells were plated in quadruplicate on collagen-based medium containing thrombopoietin (TPO), IL-3 and IL-6 for ten days. (Top) Representative CD41+ MK colony types. Scale bar: 200 μm. (Bottom) Numbers of MK colony types per plate; n=3 in quadruplicates. (B-D) CD34+ cells were cultured for seven days in the presence of TPO, with or without IL-21 (Figure 1A, scheme 2). (B) Representative FACS plots of CD41 and CD34 distribution among 7AAD– cells. (C) Total live cells were counted at day 7 and reported to the number of seeded CD34+ cells (n=11). (D) Numbers of CD41+ live cells per seeded CD34+ cells were deduced after flow cytometry (FC) analysis (n=11). (E) CD34+ cells were cultured in the presence of TPO (100 ng/mL), 100 μg/mL polyclonal human immunoglobulins, in the absence or the presence of IL-21 (100 ng/mL) and ATR-107 mAb (100 ng/mL) (n=3, in duplicates). Viable cells were counted on day 7, and results were normalized relatively to the mean number of cells obtained at day 7 in the presence of TPO alone (relative proliferation index, RPI). Then, the cells were labeled with anti-CD41 mAb and 7AAD to exclude dead cells. The percentage of CD41+ cells among viable cells was determined by FC. (F) CD34+ cells were cultured in duplicates for seven days in the presence of different combinations of TPO (40 ng/mL), IL-6 (100 ng/mL) and soluble IL-6R (200 ng/mL) and IL-21 (100 ng/mL). On day 7, the RPI and the ratio of CD41+ cells among viable cells were calculated by FC (n=3, in duplicates). (G and H) CD34+ cells were cultured for 12 days as described in Figure 1A, scheme 1. (G) IL-21 was added at the indicated concentrations and analyzed on day 12 as in (C) (n=6). (H) On day 7 of culture, the cells were washed, pre-incubated for 1 hour in cytokine-free medium containing Stattic (300 nM), Tofacitinib (1500 nM) or vehicle (DMSO, 0.25%) and then cultured in the presence of TPO, with or without IL-21 (n=5). Values are reported as the mean±Standard Error of Mean (SEM). *P<0.05, **P<0.01, ***P<0.001; (A, CE) t-tests, (F) one-way and (G) two-way ANOVA followed by a Bonferonni post-hoc test.

haematologica | 2017; 102(4)

641


S. Benbarche et al.

centration of IL-21 remained below the detection limit (<64 pg/mL). In contrast, one day after transfection of pLIVE-IL-21, the plasma concentration of IL-21 reached 3227±952 pg/mL, then progressively declined to 494±97 pg/mL at day 7, reaching 124±17 pg/mL by day 16 post transfection (n=7) (Online Supplementary Figure S5). Animals were analyzed during the first seven days following the hydrodynamic transfer. We first analyzed the numbers of megakaryocyte progenitors in the BM of the transfected mice on day 6. In the mouse, these cells belong to the Lin–Sca–IL7Rα– cKit+CD150+CD16/32lo population. Within this subset, CD9hiCD105lo and CD9loCD105hi cells represent megakaryocyte (PreMK) and erythroid (PreE) precursors, respectively, while CD9loCD105lo cells are bi-potent ery-

throid-megakaryocyte progenitors (BEMP).20 The percentages of PreEs and BEMPs were significantly increased in pLIVE-21 transfected animals, while PreMKs were not significantly expanded (Figure 4A). In the spleen, an important secondary hematopoietic organ, the Lin–Sca–IL7Rα– cKit+CD150+ CD16/32lo progenitors appeared to be rarer; nevertheless their number was increased by IL-21 expression (Figure 4B). To complete the FC analysis, we compared BM and spleen cells in CFU-MK assays using a collagen-based medium containing TPO, IL-3 and IL-6. These assays showed that the forced expression of IL-21 resulted in increased numbers of CFUs of MK-containing mixed colonies and non-megakaryocyte colonies in the BM and the spleen, while numbers of CFUs of pure MK colonies

A

B

C

D

Figure 4. Hydrodynamic gene transfer of IL-21 promotes megakaryopoiesis in vivo. The IL-21 expression vector (pLIVE-IL21) or control empty vector (pLIVE) were hydrodynamically transferred into mice. Six days later, bone marrow (BM) and spleens were recovered and hematopoietic precursor cells were analyzed by flow cytometry (FC). Representative dot plot analyses of bi-potent erythroid-megakaryocyte progenitors (BEMP), megakaryocytic (PreMK) and erythroid (PreE) BM precursors (A) and corresponding spleen cells (B). Values are the mean±Standard Error of Mean (SEM) of the percentages of gated cells among total viable cells: (A) n=9, data were analyzed using a t-test; (B) n=3. (C) BM cells and spleen cells of hydrodynamically transfected mice were plated in quadruplicate in a collagen-based medium in the presence of TPO, IL-6 and IL-3. Megakaryocytes (MK) were identified by cytochemical staining of acetylcholinesterase activity. (Top) Representative images of BM-derived pure MK colonies, mixed MK-containing colonies and non-MK colonies. (Bottom) Number of colonies of each type per 50,000 plated BM cells (left), and 106 plated spleen cells (right) (n=6). Values are the mean±Standard Error of Mean (SEM). Data were analyzed using a t-test; *P<0.05, **P<0.01, ***P<0.001. (D) Femurs were harvested six days after hydrodynamic transfection. (Left) Representative DAPI (blue) and CD42c staining (green) of transversal sections of femoral BM. Bar: 100 μm. (Right) Quantification of the number of MKs per mm2 of BM cross-section. Three 120 µm spaced whole transversal sections per femur from 6 mice per group were analyzed. Data were analyzed using a t-test.

642

haematologica | 2017; 102(4)


Modulation of platelet homeostasis by IL-21

A

C

B

D

Figure 5. IL-21 promotes platelet clearance in vivo. The IL-21 expression vector (pLIVE-IL21) or control empty vector (pLIVE) were hydrodynamically transferred into mice. (A) On the indicated days after transfer, the total platelet count (left), the percentage of TObright platelets (middle) were measured and the absolute count of reticulated TObright platelets was deduced (right). (B) Five days after hydrodynamic transfection, EGFP+ platelets were transfused and their clearance was monitored by flow cytometry (FC). (C) Six days after hydrodynamic transfection of pLIVE-IL21 or pLIVE, the spleens (left) and livers (right) were harvested, weighed, fixed, immunostained and analyzed by immunofluorescence microscopy and CD42c+ platelet and F4/80+ macrophage areas were measured. Three sections per organ and a minimum of three micrographs per section were analyzed from 3 animals per group. Data were analyzed using a t-test; *P<0.05, **P<0.01, ***P<0.001. (D) Mice were splenectomized and seven days after surgery were hydrodynamically transfected with pLIVE (open symbols, n=6) or pLIVE-IL-21 (gray symbols, n=7). Total (tot) and reticulated (ret) platelet counts were determined at indicated days. Differences in counts of reticulated platelet were analyzed using a t-test; *P<0.05, ***P<0.0001.

were increased only in the spleen (Figure 4C). The ratios of CFU-MKs (pure and mixed) to the number of seeded cells from control and IL-21 expressing mice were approximately 47- and 28-fold higher for BM than for spleen cells, respectively, indicating the rarity of megakaryocyte progenitors in the spleen and suggesting a minor contribution of spleen to megakaryopoiesis under these experimental conditions. Finally, immunofluorescence analysis of BM sections stained with an anti-CD42c mAb revealed an increased MK density in mice expressing IL-21 as compared to control animals (160±8 vs. 129±4 MKs per mm2, respectively; n=6) (Figure 4D). Altogether, these data showed that IL-21 expression increases the number of MK progenitors in the BM and the spleen and enhances megakaryopoiesis, confirming the in vivo relevance of our in vitro observations.

IL-21 expression increases platelet clearance One day after hydrodynamic transfection, total platelet counts decreased by 50% (Figure 5A) and a similar decrease was observed after injection of the transfer solution or saline alone (data not shown). Thereafter, in pLIVEreceiving control animals, platelet counts progressively increased to recover their baseline value six days later (n=6) (Figure 5A, left). Conversely, in pLIVE-IL-21 transfected mice, platelet counts unexpectedly remained haematologica | 2017; 102(4)

reduced. However, in agreement with the observed higher megakaryocyte density in the BM, the number of TObright young platelets was increased, compared to control mice (Figure 5A, middle and right). On day 6, the phenotype of the platelets from pLIVE and pLIVE-IL-21 transfected mice and their responses to thromboxane and PAR4 agonists were similar; the responses of youngest (TObright) and older (TOdim) platelets to agonists were also identical (Online Supplementary Figure S6). This prolonged thrombocytopenia in mice transfected with pLIVE-IL21 could result from a peripheral mechanism, probably an increased clearance. To test this hypothesis, washed EGFP+ platelets from untreated mice were transfused into untransfected, control transfected or pLIVE-IL-21 transfected EGFP– mice and their counts were followed for five days. Two days after transfusion, the percentage of circulating EGFP+ platelets was lower in mice expressing IL-21 (n=6) as compared to control transfected (n=7) or untransfected animals (n=6) [17%±2% vs. 45±2% (P<0.001), and vs. 52%±1% (P<0.001), respectively] (Figure 5B), confirming that IL-21 expression increased platelet clearance. Since the spleen can retain a pool of platelets which increases with splenomegaly,21 we weighed the spleens and found that six days after transfection, the spleens of mice expressing IL-21 were larger than those of control animals (163±8 mg vs. 88±6 mg, respectively; n=9) (Figure 643


S. Benbarche et al.

5C, Spleens panel, left). Additional experiments showed that splenomegaly was already maximal by day 3 and stable for at least nine additional days (data not shown). The distribution of macrophages and platelets in the spleen were analyzed by immunofluorescence microscopy (Figure 5C, Spleens panel, right, and Online Supplementary Figure S7A). The areas of F4/80+ macrophages in spleen cryosections from control mice and animals expressing IL21 were similar. The area of platelets (CD42c+) was slightly but significantly increased in mice expressing IL-21 (Figure 5C, Spleens panel, middle), indicating an increased accumulation of platelets. To confirm the participation of the spleen in IL-21induced persistent thrombocytopenia, complementary experiments were performed with splenectomized mice. In these animals, IL-21 expression also increased the number of circulating TObright platelets (Figure 5D), confirming that IL-21 enhanced BM megakaryopoiesis. Nevertheless, total platelet counts increased similarly in mice expressing IL-21 and control animals, with 1802±63 x 109 (n=7) versus 1755±78 x 109 (n=6) platelets/L respectively, five days after transfection. This latter observation, together with the findings in non-splenectomized animals, indicated that the spleen participates in IL-21-induced platelet clearance. In hydrodynamically transfected splenectomized mice, control mice and those expressing IL-21 displayed similar total platelet counts, despite the increased TObright platelets, suggesting increased platelet clearance by liver macrophages. We thus also analyzed the livers in transfected animals. IL-21 expression did not impact the weight of the livers, but significantly increased the numbers of macrophages and platelets present in this tissue (Figure 5C, Livers panel, and Online Supplementary Figure S7B). Altogether, these observations suggest that macrophages in the spleen and the liver could be involved in the increase platelet clearance mediated by IL-21.

Discussion We here showed that IL-21R is expressed when human CD34+ progenitors are cultured in the presence of TPO. The receptor was expressed on megakaryocytic CD41+ cells, but not on CD41– cells (Figure 1C). Similar observations have likewise been described.10,16 TPO alone also induced the expression of IL-21R on 40%-50% of CD41+ within seven days of culture (data not shown). The progressive IL-21Rα expression during megakaryocyte differentiation and its presence on human BM megakaryocytes strongly suggest that this expression results from commitment of the progenitor cells to the megakaryocytic lineage. On the other hand, IL-21 increased the numbers of medium and large megakaryocyte colonies obtained in CFU-MK assays with blood CD34+ progenitors, which suggests that this cytokine could also promote the expansion of committed cells. IL-21Rα was not detected on human and mouse platelets. Another example of discordance between megakaryocytes and platelets in the expression of a membrane-associated protein is the tyrosine phosphatase receptor CD45, which is likewise present on human BM megakaryocytes but not on platelets.22 IL-21 induced the phosphorylation of STAT3 in CD41+ cells, the percentage of responding cells being higher on day 7 than on day 12, in agreement with the downregulation of the common γ 644

chain. TPO induced the phosphorylation of only STAT5 on day 7 of in vitro human megakaryocyte differentiation, but both STAT3 and STAT5 phosphorylation was observed during the later stages of differentiation. This finding is consistent with studies documenting the TPOinduced phosphorylation of STAT3 and STAT5 in megakaryocytic cell lines23,24 and blood platelets.17,25 Thus, at late stages of megakaryocyte differentiation, the downregulation of IL-21R-mediated signaling could be compensated by TPO activity. In vitro, in the presence of TPO alone, IL-21 increased in a dose-dependent manner the number of megakaryocytic cells generated. This effect was inhibited using the blocking anti-IL21R ATR-107 mAb and was dependent on IL-21R-mediated activation of JAK3/STAT3 pathway. These observations are in line with the documented role of STAT3 in megakaryopoiesis.26,27 In the presence of TPO, IL-21 increased the capacity of IL-6 and soluble IL-6R combination to improve the in vitro differentiation of CD34+ cells into megakaryocytes, indicating complementary physiological functions. IL-21 and IL-21R deficiencies cause immunodeficiency syndromes in man,28,29 disturb B, T and NK lymphocyte functions in mice11 but do not impact platelet counts, indicating that IL-21 is not essential for megakaryopoiesis and thrombopoiesis under healthy conditions. Our observations suggest that IL-21 could compensate for increased platelet clearance in pathologies characterized by an increased production of IL-21. Thus, a hydrodynamic transfection method was used to over-express IL-21 in vivo, in hepatocytes.15 Plasmatic IL-21 concentrations was enhanced for two weeks, while IL-1α, IL-1β and IL-6 remained undetectable between day 1 and 7 (i.e. <16 pg/mL), in agreement with previous studies,30,31 and no significant TPO variations were noticed (pLIVE vs. pLIVE-IL21 transfected animals, n=3 per day) (data not shown). In IL-21-expressing mice, among the BM Lin–Sca–IL7Rα– cKit+CD150+CD16/32lo progenitors, the ratio of BEMPs was expanded. Consistently, clonogenic assays revealed that IL-21 enhanced the number of megakaryocytic mixed colonies obtained from BM cells. MK are thought to derive from PreMKs, which differentiate from BEMPs.32 Because the density of BM megakaryocytes was increased, the observed numbers of BEMPs and PreMKs may reflect differences in the dynamics of the maturation of these two types of progenitors. In the spleen, the numbers of Lin–Sca–IL7Rα– cKit+CD150+ CD16/32lo progenitors and of pure and mixed CFU-MK were increased by IL-21 expression. The number of CFU-MK per seeded cell was 30-50-fold lower in the spleen than in the BM, suggesting that in our experimental conditions IL-21 has a minor effect on spleen megakaryopoiesis. Our results are in agreement with and complement previous work, where splenomegaly and increased numbers of Lin–cKit+Sca1+ hematopoietic stem cells and CFU-GEMMs (granulocyte-erythrocyte-megakaryocytemacrophages) were observed in the BM and spleen of mice expressing IL-21 after hydrodynamic transfection.33 Accordingly, thrombopoiesis was enhanced in transfected mice expressing IL-21, as indicated by their increased counts of young reticulated platelets. However, total platelet counts did not follow this trend. The day after hydrodynamic transfection thrombocytopenia was noticed, probably due to vascular damage and consequently, platelet consumption.15 Normal platelet count was restored between days 4 and 7 in mice transfected haematologica | 2017; 102(4)


Modulation of platelet homeostasis by IL-21

with the empty plasmid, but not in pLIVE-IL-21 transfected animals. An increased clearance of platelets could explain this absence of recovery (Figure 5A and B). On day 6, greater numbers of macrophages and platelets were found in the liver of mice expressing IL-21. On cryosections of the enlarged spleen of these animals, macrophages and platelets occupied similar and higher surface areas, respectively (Figure 5C). Thus, increased clearance of platelets could occur in the spleen, or in the liver, another major site for the elimination of old or activated platelets.34 IL-21 expression in splenectomized animals was accompanied with increased thrombopoiesis, but no thrombocytopenia, indicating that: i) IL-21-mediated enhanced megakaryopoiesis mainly occurs in the BM; and ii) that a major part of the thrombocytopenia provoked by IL-21 expression is mediated by the spleen. The absence of increased platelet counts in IL-21 splenectomized animals, although thrombopoiesis is increased, could be explained by increased clearance by liver macrophages, since IL-21 expression increased the density of platelets and macrophages in this organ. Thus, additional investigations are necessary to elucidate the mechanisms responsible for the increased platelet clearance induced by IL-21. This study in the mouse is reminiscent of a pre-clinical study in macaques, which periodically received non-glycosylated recombinant IL-21 and experienced cycles of moderate anemia and thrombocytopenia followed by

References 8. 1. Cortin V, Garnier A, Pineault N, Lemieux R, Boyer L, Proulx C. Efficient in vitro megakaryocyte maturation using cytokine cocktails optimized by statistical experimental design. Exp Hematol. 2005; 33(10):11821191. 2. Dan K, Gomi S, Inokuchi K, et al. Effects of interleukin-1 and tumor necrosis factor on megakaryocytopoiesis: mechanism of reactive thrombocytosis. Acta Haematol. 1995; 93(2-4):67-72. 3. Lotem J, Shabo Y, Sachs L. Regulation of megakaryocyte development by interleukin6. Blood. 1989;74(5):1545-1551. 4. Catani L, Amabile M, Luatti S, et al. Interleukin-4 downregulates nuclear factorerythroid 2 (NF-E2) expression in primary megakaryocytes and in megakaryoblastic cell lines. Stem Cells. 2001;19(4):339-347. 5. Mitjavila MT, Vinci G, Villeval JL, et al. Human platelet alpha granules contain a nonspecific inhibitor of megakaryocyte colony formation: its relationship to type beta transforming growth factor (TGF-beta). J Cell Physiol. 1988;134(1):93-100. 6. Sosman JA, Verma A, Moss S, et al. Interleukin 10-induced thrombocytopenia in normal healthy adult volunteers: evidence for decreased platelet production. Br J Haematol. 2000;111(1):104-111. 7. Griesshammer M, Bangerter M, Sauer T, Wennauer R, Bergmann L, Heimpel H. Aetiology and clinical significance of thrombocytosis: analysis of 732 patients with an

haematologica | 2017; 102(4)

9.

10. 11.

12.

13.

14.

15.

polycythemia and thrombocythemia.31 Because the clearance of the recombinant IL-21 was rapid, it is not possible to compare these observations with ours, which are based on the in vivo expression of IL-21. Of note, cancer patients receiving IL-21 could also develop reversible thrombocytopenia.35 It is tempting to speculate that in humans receiving recombinant IL-21, megakaryopoiesis and platelet production are increased, but platelet clearance is also enhanced and the net result is thrombocytopenia. This study demonstrates that IL-21R is progressively induced during human megakaryopoiesis. IL-21 appears to enhance the proliferation of megakaryocyte progenitors through the JAK3/STAT3 pathway, resulting in increased in vivo megakaryopoiesis. Since IL-21 is mainly produced by T cells, our work provides new insights into the regulation of megakaryopoiesis through adaptive immunity. Our observations also suggest that the IL-21-mediated increased megakaryopoiesis could be a compensatory mechanism to counteract enhanced platelet clearance during immune responses characterized by increased IL-21 expression. Acknowledgments The authors would like to thank Catherine Ziessel for expert technical assistance, Arnaud Dupuis (HĂ´pital Universitaire de Strasbourg) for providing human BM samples, Nathalie Brouard for helpful discussions and technical advice, and Juliette Mulvihill for reviewing the English of the manuscript.

elevated platelet count. J Intern Med. 1999;245(3):295-300. Kaser A, Brandacher G, Steurer W, et al. Interleukin-6 stimulates thrombopoiesis through thrombopoietin: role in inflammatory thrombocytosis. Blood. 2001; 98(9): 2720-2725. Nishimura S, Nagasaki M, Kunishima S, et al. IL-1alpha induces thrombopoiesis through megakaryocyte rupture in response to acute platelet needs. J Cell Biol. 2015;209(3):453-466. Ferrari F, Bortoluzzi S, Coppe A, et al. Genomic expression during human myelopoiesis. BMC Genomics. 2007;8:264. Spolski R, Leonard WJ. Interleukin-21: a double-edged sword with therapeutic potential. Nat Rev Drug Discov. 2014;13(5):379-395. Hodge LS, Ziesmer SC, Yang ZZ, et al. IL-21 in the BM microenvironment contributes to IgM secretion and proliferation of malignant cells in Waldenstrom macroglobulinemia. Blood. 2012; 120(18):3774-3782. Yu H, Zhang J, Fu R, et al. Increased frequency of BM T follicular helper cells in patients with immune-related pancytopenia. Clin Dev Immunol. 2013;2013:730450. Chen CI, Zhang L, Datta SK. Hematopoietic stem and multipotent progenitor cells produce IL-17, IL-21 and other cytokines in response to TLR signals associated with late apoptotic products and augment memory Th17 and Tc17 cells in the BM of normal and lupus mice. Clin Immunol. 2016;162:926. Suda T, Liu D. Hydrodynamic gene delivery:

16.

17.

18.

19.

20.

21. 22.

its principles and applications. Mol Ther. 2007;15(12):2063-2069. Sun S, Wang W, Latchman Y, Gao D, Aronow B, Reems JA. Expression of plasma membrane receptor genes during megakaryocyte development. Physiol Genomics. 2013;45(6):217-227. Miyakawa Y, Oda A, Druker BJ, et al. Thrombopoietin induces tyrosine phosphorylation of Stat3 and Stat5 in human blood platelets. Blood. 1996;87(2):439-446. Zhu M, Pleasic-Williams S, Lin TH, Wunderlich DA, Cheng JB, Masferrer JL. pSTAT3: a target biomarker to study the pharmacology of the anti-IL-21R antibody ATR-107 in human whole blood. J Transl Med. 2013;11:65. Sui X, Tsuji K, Ebihara Y, et al. Soluble interleukin-6 (IL-6) receptor with IL-6 stimulates megakaryopoiesis from human CD34(+) cells through glycoprotein (gp)130 signaling. Blood. 1999;93(8):2525-2532. Ng AP, Kauppi M, Metcalf D, Di Rago L, Hyland CD, Alexander WS. Characterization of thrombopoietin (TPO)-responsive progenitor cells in adult mouse BM with in vivo megakaryocyte and erythroid potential. Proc Natl Acad Sci USA. 2012;109(7):2364-2369. Penny R, Rozenberg MC, Firkin BG. The splenic platelet pool. Blood. 1966;27(1):1-16. Tomer A. Human marrow megakaryocyte differentiation: multiparameter correlative analysis identifies von Willebrand factor as a sensitive and distinctive marker for early (2N and 4N) megakaryocytes. Blood. 2004; 104(9):2722-2727.

645


S. Benbarche et al.

23. Bacon CM, Tortolani PJ, Shimosaka A, Rees RC, Longo DL, O'Shea JJ. Thrombopoietin (TPO) induces tyrosine phosphorylation and activation of STAT5 and STAT3. FEBS Lett. 1995;370(1-2):63-68. 24. Drachman JG, Kaushansky K. Dissecting the thrombopoietin receptor: functional elements of the Mpl cytoplasmic domain. Proc Natl Acad Sci USA. 1997;94(6):2350-2355. 25. Majka M, Ratajczak J, Villaire G, et al. Thrombopoietin, but not cytokines binding to gp130 protein-coupled receptors, activates MAPKp42/44, AKT, and STAT proteins in normal human CD34+ cells, megakaryocytes, and platelets. Exp Hematol. 2002;30(7):751-760. 26. Kirito K, Osawa M, Morita H, et al. A functional role of Stat3 in in vivo megakaryopoiesis. Blood. 2002;99(9):3220-3227. 27. Jenkins BJ, Roberts AW, Greenhill CJ, et al. Pathologic consequences of STAT3 hyperac-

646

28.

29.

30.

31.

tivation by IL-6 and IL-11 during hematopoiesis and lymphopoiesis. Blood. 2007;109(6):2380-2388. Kotlarz D, Zietara N, Uzel G, et al. Loss-offunction mutations in the IL-21 receptor gene cause a primary immunodeficiency syndrome. J Exp Med. 2013;210(3):433443. Salzer E, Kansu A, Sic H, et al. Early-onset inflammatory bowel disease and common variable immunodeficiency-like disease caused by IL-21 deficiency. J Allergy Clin Immunol. 2014;133(6):1651-1659 e1612. Wang G, Tschoi M, Spolski R, et al. In vivo antitumor activity of interleukin 21 mediated by natural killer cells. Cancer Res. 2003; 63(24):9016-9022. Waggie KS, Holdren MS, Byrnes-Blake K, et al. Preclinical safety, pharmacokinetics, and pharmacodynamics of recombinant human interleukin-21 in cynomolgus macaques

32.

33.

34. 35.

(Macaca fascicularis). Int J Toxicol. 2012; 31(4):303-316. Tsang AP, Fujiwara Y, Hom DB, Orkin SH. Failure of megakaryopoiesis and arrested erythropoiesis in mice lacking the GATA-1 transcriptional cofactor FOG. Genes Dev. 1998;12(8):1176-1188. Ozaki K, Hishiya A, Hatanaka K, et al. Overexpression of interleukin 21 induces expansion of hematopoietic progenitor cells. Int J Hematol. 2006;84(3):224-230. Grozovsky R, Hoffmeister KM, Falet H. Novel clearance mechanisms of platelets. Curr Opin Hematol. 2010;17(6):585-589. Hashmi MH, Van Veldhuizen PJ. Interleukin-21: updated review of Phase I and II clinical trials in metastatic renal cell carcinoma, metastatic melanoma and relapsed/refractory indolent non-Hodgkin's lymphoma. Expert Opin Biol Ther. 2010; 10(5):807-817.

haematologica | 2017; 102(4)


ARTICLE

Hematopoiesis

A population of hematopoietic stem cells derives from GATA4-expressing progenitors located in the placenta and lateral mesoderm of mice

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Ana Cañete,1-2 Rita Carmona,1-2 Laura Ariza,1-2 María José Sánchez,3 Anabel Rojas4 and Ramón Muñoz-Chápuli1-2

1 Department of Animal Biology, University of Málaga; 2Andalusian Center for Nanomedicine and Biotechnology (BIONAND), Málaga; 3Centro Andaluz de Biología del Desarrollo (CABD), Consejo Superior de Investigaciones Científicas (CSIC), Universidad Pablo de Olavide (UPO), Seville; 4Andalusian Center of Molecular Biology and Regenerative Medicine (CABIMER) and Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas-CIBERDEM, Seville, Spain

Haematologica 2017 Volume 102(4):647-655

ABSTRACT

G

ATA transcription factors are expressed in the mesoderm and endoderm during development. GATA1-3, but not GATA4, are critically involved in hematopoiesis. An enhancer (G2) of the mouse Gata4 gene directs its expression throughout the lateral mesoderm and the allantois, beginning at embryonic day 7.5, becoming restricted to the septum transversum by embryonic day 10.5, and disappearing by midgestation. We have studied the developmental fate of the G2-Gata4 cell lineage using a G2-Gata4Cre;R26REYFP mouse line. We found a substantial number of YFP+ hematopoietic cells of lymphoid, myeloid and erythroid lineages in embryos. Fetal CD41+/cKit+/CD34+ and Lin–/cKit+/CD31+ YFP+ hematopoietic progenitors were much more abundant in the placenta than in the aorta-gonad-mesonephros area. They were clonogenic in the MethoCult assay and fully reconstituted hematopoiesis in myeloablated mice. YFP+ cells represented about 20% of the hematopoietic system of adult mice. Adult YFP+ hematopoietic stem cells constituted a long-term repopulating, transplantable population. Thus, a lineage of adult hematopoietic stem cells is characterized by the expression of GATA4 in their embryonic progenitors and probably by its extraembryonic (placental) origin, although GATA4 appeared not to be required for hematopoietic stem cell differentiation. Both lineages basically showed similar physiological behavior in normal mice, but clinically relevant properties of this particular hematopoietic stem cell population should be checked in physiopathological conditions.

Correspondence: chapuli@uma.es

Received: September 2, 2016. Accepted: December 28, 2016. Pre-published: January 5, 2017. doi:10.3324/haematol.2016.155812 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/647

Introduction The six transcription factors belonging to the GATA family in mammals play important roles in mesoderm and endoderm development. GATA1-3, but not GATA4-6, play critical roles in hematopoiesis.1 Mice deficient for GATA4 show defects in the heart and intestine and die around embryonic day (E) 13.5.2-4 A mesodermal-specific enhancer of Gata4, called G2, drives Gata4 expression in the lateral mesoderm and allantois starting at E7.5. Later, this activity is restricted to the septum transversum and ceases by E12.5.5 Our previous work using G2Cre;R26REYFP mice has shown that the cell lineage where Gata4 is activated by G2 contributes to hepatic stellate cells. Inactivation of Gata4 using this G2Cre driver is lethal by midgestation. The anemia observed in the G2Cre;Gata4flox/flox embryos was attributed to a failure in the expansion of the hematopoietic progenitors in the fetal liver. Interestingly, a small population of hepatic YFP+ cells from G2Cre;R26REYFP embryos was positive for leukocyte and megakaryocyte markers.6 We have performed a comprehensive analyhaematologica | 2017; 102(4)

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

647


A. Cañete et al.

sis of hematopoietic G2-Gata4 lineage cells in mice. A significant fraction of the fetal and adult blood cells derives from this lineage, despite the short embryonic period in which the G2 enhancer is active. Thus, the adult hematopoietic stem cells (HSCs) would belong to two distinct mesodermal lineages, depending on whether they derive from progenitors expressing GATA4 under control of the G2 enhancer or not. This observation raises multiple questions about the embryonic origin of this lineage, the hypothetical role played by GATA4 in hematopoiesis or the existence of different properties of the adult HSCs depending on their embryonic lineage. Answers to the these questions have been the aims of the study herein, which has confirmed the existence of a long-term repopulating adult HSC population derived from the embryonic G2-Gata4 lineage. The experimental evidence collected suggests that this hematopoietic lineage has a placental origin, but GATA4 appears dispensable for its differentiation.

Methods Transgenic mouse lines The animals used in our research program were handled in compliance with the institutional and European Union guidelines for animal care and welfare and housed in the animal facility of the University of Málaga under controlled standard conditions. The procedure was approved by the Committee on the Ethics of Animal Experiments of the University of Malaga (procedure code 2015-0028). Additional animals were maintained in the CABD animal care facility with the approval of the ethical committee of CSIC and the University of Pablo de Olavide. All embryos were staged from the time point of vaginal plug observation, which was designated as E0.5. G2Cre and Gata4 floxed mice were generated as previously described.5,7 G2Cre mice were crossed with the reporter line Rosa26REYFP (B6.129X1-Gt(ROSA)26Sortm1(EYFP)Cos/J). The resulting G2Cre/+;R26REYFP mice constitutively express YFP in all the lineage of the cells where the enhancer G2 has been activated. The tamoxifen-inducible Scl-CreERT mouse line was generated by Dr. Joachim Göthert.8 Blood and tissue samples from embryos and adult mice were obtained as described in the Online Supplementary Methods.

RT-PCR Bone marrow (BM) cells obtained from the femur were washed in phosphate buffered saline (PBS) and homogenized in 1mL of Tri Reagent (Sigma). Aorta-gonad-mesonephros (AGM) and fetal liver were used as positive controls. Details of reverse transcriptionpolymerase chain reaction (RT-PCR) and primer sequences are shown in the Online Supplementary Methods.

Flow cytometry Cell suspensions were incubated with fluorochrome-conjugated antibodies for 30 min on ice. After washing, the labelled cells were analyzed in a FACSverse flow cytometer. For fluorescence activated cell sorting, the labelled cells were resuspended in 1 mL of working solution and sorted in a MoFlo cell sorter. Hoechst 33342 staining of bone marrow cells to identify the side population was performed according to standard protocols.9 A list of the antibodies used is shown in the Online Supplementary Table S1.

Colony assay in MethoCult Cells obtained from tissue homogenization were cultured in the semisolid medium MethoCult (GF M3434, Stem Cell Technologies 648

Inc. Vancouver, Canada). Ectoplacental cones together with the allantois were obtained by fine dissection of E8.0 embryos and preincubated for 5 days in 24-well multidishes (Nunc). After 7-14 days the colonies were photographed. When the fluorescence of the cells had to be detected, the colonies were picked and adhered to a slide via cytospin. More details are provided in the Online Supplementary Methods.

Adult bone marrow cells transplant into irradiated adult mice

2,5 million bone marrow cells were suspended in 300 μL of Dulbecco’s phosphate buffered saline (DPBS) +1% fetal calf serum (FCS) +1% penicillin/streptomycin (P/S) and injected in the tail vein of C57Bl/6 x CBA irradiated mice as described in the Online Supplementary Methods.

Adult bone marrow and embryonic cells transplant in newborn mice treated with Busulfan Busulfan (1, 4-Butanediol dimethanesulfonate, Sigma-Aldrich) is a myeloablative agent used to improve the efficiency of a hematopoietic graft in newborn mice.10,11 Pregnant females are injected by intraperitoneal (i.p.) injection on days 17 and 18 of gestation with a 15 mg busulfan/Kg dose. Newborns at stage P1 are injected through the facial vein with cells suspended in 50 μL of DPBS + 1% FCS + 1% P/S.12

Statistical analysis Quantitative data were always expressed as mean±SEM. Statistical comparison of values was performed using the Student’s t-test.

Results A fraction of adult HSC derive from the G2-GATA4 embryonic lineage We analyzed the peripheral blood (PB) and bone marrow of adult G2Cre;R26REYFP mice and investigated the presence of YFP+ cells. After erythrocyte lysis, we found that about 20% of leukocytes (CD45+), circulating monocytes (CD11b+), T lymphocytes (CD3+) and B lymphocytes (B220+) in blood samples were YFP+ (Figure 1A; Table 1). In bone marrow samples, 20% of all the cells and 24% of all the CD45+ cells were YFP+. Similar percentages were found for B220+, CD11b+ and CD41+ cells. Erythroblasts (Ter119+) and lymphocytes (CD3+, CD4/CD8+) or lymphoid progenitors (CD127+, Online Supplementary Figure S1) showed a lower and higher proportion of YFP+ cells, respectively, but the differences were not significant. Similar proportions of YFP+ cells were found in both the spleen and thymus (Table 1). We then identified the bone marrow KSL population, i.e., cells expressing c-Kit and Sca1 but not lineage markers (Lin–, defined as negative for CD3, Ly-6G/Ly-6C, CD11b, CD45R and Ter-119). The KSL population constitutes about 0.075±0.02% of all the bone marrow cells and includes hematopoietic stem and progenitor cells. 26,6±2,8% of the bone marrow KSL cells were YFP+. The 1:4 proportion between YFP+ and YFP– cells was also observed between the CD135/Flk2– and CD135/Flk2+ KSL cells, representing longterm and short-term HSCs, respectively (Table 1 and Figure 1B). Additionally we performed an experiment to estimate the percentage of YFP+ cells within the bone marrow side population identified by Hoechst 33342 staining.9 As shown in Figure 1C, 18.5% of this side population was YFP+. The presence of YFP+ cells in the bone marrow was conhaematologica | 2017; 102(4)


Adult HSC from GATA4 expressing embryonic lineage

firmed by confocal microscopy. We observed colocalization of YFP with c-Kit and also with CD44, a marker present in most hematopoietic cells (Figure 2A,B). About 30% of the SCA1+ cells from the bone marrow were YFP+ (Table 1). Thus, we checked the colocalization of YFP with other mesenchymal stem cell markers. We found a relatively high percentage of YFP+ cells within the CD73+, CD90+ and CD105+ Lin negative bone marrow populations, reaching 30-40%. Furthermore, half of the Lin–/Sca1+/PDGFRα+ population, enriched in mesenchymal stem cells,13 was YFP+ (Online Supplementary Figure S2). Thus, we cannot discard a contribution of the G2-Gata4 cell lineage to the bone marrow mesenchymal stem cells. Finally, we checked the expression of GATA4 in the adult bone marrow in order to disregard postnatal reactivation of the G2 enhancer. Adult bone marrow cells do not express GATA4 (Figure 1D).

G2Cre;R26REYFP bone marrow cells contain transplantable, long-term repopulating hematopoietic stem cells 2.5x106 bone marrow cells from G2Cre;R26REYFP mice were injected in irradiated adult recipient mice. About 20% of the injected cells were derived from the G2-Gata4 lineage. Multilineage contribution from YFP+ progenitors was determined at long-term (4 months posttransplantation) (Table

2). YFP+ cells were identified in peripheral blood, bone marrow and the spleen. In lysed peripheral blood, 17.7±6.2% of all the cells were YFP+. This percentage was higher in the cases of the T and B lymphocytes, reaching 25% for B220+ and 28% for the CD3+ population, respectively (Table 2). In bone marrow, the proportion of YFP+ cells was well correlated with that found in peripheral blood, reaching 14.4±8.0% of the total cells, again with a higher proportion of CD3 lymphocytes (25%) and a lower fraction of erythroid cells (5.5%). The different contribution of YFP+ cells to the CD3+ and Ter119+ populations was statistically significant (Student's t-test, P value=0.04). Hematopoietic progenitors, including HSCs, have been shown to contribute to vascular endothelial cells in transplantation assays.14 Therefore, we analyzed the distribution of YFP+ cells in non-hematopoietic organs, such as the heart, kidneys, lungs and liver. Most of the YFP+ cells found in these organs expressed CD45 and were found close to the walls of some vessels, probably indicating foci of extramedullary hematopoiesis. A number of endothelial (CD31+/CD45–) YFP+ cells were localized in the heart and liver vessels, including the endocardium (Figure 2N-O). In samples of lysed peripheral blood 28.4±9.1% (n=2) of the CD45–/CD31+ cells were YFP+ (Online Supplementary Figure S3). These cells could be putative circulating endothelial progenitors. We also tested the long-term multilineage repopulation

+

A

B

D

C

haematologica | 2017; 102(4)

Figure 1. Analytical flow cytometry analysis of peripheral blood and bone marrow from G2Cre;R26REYFP mice. A: Gating strategy for the identification of different subpopulations of YFP+ cells in lysed peripheral blood B: Gating strategy for the identification of c(KSL) Kit+/SCA1+/Lin– hematopoietic cells in bone marrow (BM). A fraction of the KSL progenitors from bone marrow was YFP+, and 80% of them were Flk2–, representing long-term hematopoietic stem cells. Detailed data of these experiments are shown in Table 1. C: The side population of the bone marrow identified by Hoechst 33342 staining includes a fraction of YFP+ cells. D: GATA4 is not expressed in bone marrow (BM) cells by RT-PCR. As a positive control, GATA4 expression was checked in fetal liver (FL) and aorta-gonadmesonephros tissue (AGM). SSC-A: side scatter; FSC-A: forward scatter; FLK2: fetal liver kinase-2; SP: side population.

649


A. Cañete et al. Table 1. Frequency of YFP+ cells within different cell populations in G2Cre;R26REYFP mice.

Tissue Peripheral blood Nº experiments % total cells % YFP+ Bone Marrow Nº experiments % total % YFP+ Spleen Nº experiments % total % YFP+ Thymus Nº experiments % total % YFP+

All cells

CD3

B220

CD45

CD11b

10

5 19.3±3.5 21.2±5.8 6 2.7±0.4 26.4±4.9 4 26.3±0.6 19.4±2.8 4 14.2±0.7 19.2±1.9

9 52.5±1.2 20.1±2.0 6 18.9±2.8 25.1±3.5 4 52.5±8.6 21.3±3.3

10 83.3±5.5 18.8±1.5 16 74.1±3.3 24.1±1.6

5 21.5±2.8 20.3±1.8 2 51,6±14,7 24,6±0,54

17.7±2.1 17 19.9±1.1 4 20.6±2.9 4 18.9±2.9

Ter119

CD127

CD41

14 29.3±2.8 19.1±2.4 4 24.3±9.9 27.3±4.6

2 8.5±1.3 23.1±0.9 2 15.2±1.2 26.8±2.1 2 5.9±2.5 19.8±0.8

7 3.8±0.3 24.9±2.4

Sca1

KSL-

4 6 9.19±0.35 0.075±0.02 30.4±0.66 26.6±2.8

The first row of each tissue (% total cells) indicates the frequency of the population identified with each marker with respect to the total number of cells analyzed. The second row (% YFP+) shows the proportion of YFP+ cells within each population (mean±SEM). KSL: c-Kit+/Sca1+/Lin–; Sca1: stem cell antigen 1.

Table 2. Frequency of YFP+ cells after transplantation of bone marrow from G2Cre;R26REYFP mice into irradiated adult mice and into busulfan-treated newborns.

Peripheral blood Total BM into irradiated 4 months 17.7±6,2

CD45

BM into busulfan 1 month 17.7±1.0

17.3±1.1

4-9 months 16.8±3.3

17.2±3.4

B220 25.1±7.2

Bone marrow

Thymus Spleen

T cells Total CD45 CD11b T cells Ter119 KSL Total Total a a 28.1±7.9 14.4±8.0 23.8±7.8 16.1±14.4 25.3±6.2 5.5±3,4 23.1±6.5 (CD3) (CD3) 7.6±1,6b (CD4+CD8) 14.4±2,8b 14.6±2.7 15.3±2.9 17.1±3.6 14.2±2.8 13.4±2.6 18.6±4.5 16.0±5.6 16.2±2.2 (CD4+CD8) (CD4+CD8)

The table shows the frequency (%) of YFP+ cells in peripheral blood, bone marrow, the thymus and spleen after transplantation of bone marrow from G2Cre;R26REYFP mice into irradiated adults (N=4) and into busulfan-treated newborns (N=8). Data represent mean±SEM. aThe difference found between the % of YFP+ cells from the CD3 and the Ter119 populations was statistically significant (Student's t-test, P value=0.04). bThe difference found between the % of YFP+ cells in the CD4+CD8 populations in the short- and long-term was statistically significant (Student's t-test, P value=0.03). BM: bone marrow; KSL: c-Kit+/Sca1+/Lin–.

potential of bone marrow cells from G2Cre;R26REYFP mice by performing transplantation into myeloablated newborn mice, a model previously shown to allow hematopoietic reconstitution by early embryonic as well as adult HSC types.15 Busulfan-treated newborn mice received 5x106 bone marrow cells (i.e., about 106 YFP+ cells). Peripheral blood was analyzed one month (short-term) and 4-9 months (long-term) after the transplant (Table 2). Out of 14 mice transplanted, 8 showed YFP+cells in their peripheral blood, with a mean percentage of YFP+ nucleated cells of 17.7±1.0% a month after the transplantation. This proportion was similar in the long-term, and also in the bone marrow, thymus and spleen. No differences were found in the proportion of YFP+/Ter119+ cells as compared with other lineages. CD45+ cells showed similar percentages in the longand short-term, but YFP+ T lymphocytes were less represented in the blood samples obtained in the short-term (7.6±1.6%) compared with the long-term (14.4±2.8%). This difference was statistically significant (Student's t-test, P value=0.03). In the long-term in busulfan-treated chimaeras injected with bone marrow, CD31+/CD45–/YFP+ endothelial cells were also detected in the lungs and liver (data not shown).

Hematopoietic progenitors from the G2-Gata4 lineage are mainly localized in the placenta We performed a confocal microscopy study of the embryonic hematopoietic tissues (AGM, placenta, fetal liver) to study the distribution of the YFP+ cells (Figure 2C650

M). Most of the YFP+ cells located close to the aorta by E10.5 are found in its dorsal part, around the notochord, and they mainly represent sclerotome cells. A smaller population of YFP+ cells can also be seen in the mesentery and ventral part of the aorta, sometimes in the aortic endothelium (Figure 2C-E). Colocalization of YFP with CD31 demonstrated the endothelial nature of these cells (Figure 2C). YFP colocalizes with GATA4 protein in the mesothelium of the periaortic area, suggesting that the enhancer G2 is activating Gata4 expression in these cells (Figure 2D). Colocalization with RUNX1 also suggests that these YFP+ cells are hemogenic endothelium (Figure 2E). In the placenta, YFP+ cells are very abundant in the chorionic plate (Figure 2F-I), mainly around large vessels. Colocalization of YFP with CD41 and CD31 was frequent. Some YFP+ cells were observed apparently detaching from the vascular endothelium (Figure 2G). As described for the aorta, YFP+ cells also showed expression of GATA4 and RUNX1, mainly in the vascular walls (Figure 2H,I). However, YFP+ cells were GATA2-negative in the placenta (Figure 2L). Finally, colocalization of YFP and GATA4 was very frequent in the hepatic mesothelium and submesothelial areas. As previously described,6 these cells represent mesothelialderived cells invading the liver to contribute to the hepatic stroma (Figure 2J). Some of the YFP+ cells in the fetal liver showed expression of endothelial and hematopoietic markers, including RUNX1 (Figure 2K) and GATA2 (Figure 2M). We performed several experiments to establish the haematologica | 2017; 102(4)


Adult HSC from GATA4 expressing embryonic lineage

Figure 2. Localization of G2-GATA4 lineage cells (YFP+) in adult and embryonic tissues by confocal microscopy. A,B: YFP+ cells are localized in the bone marrow. Some of these YFP+ cells express c-Kit (arrows in A, separate channels are shown in insert) and CD44 (arrows in B). C-E: YFP+ cells are also localized close to the embryonic aorta (AO) by stage E10. YFP+ cells are abundant around the notochord (NC) and they are also observed in the dorsal mesentery (DM) and in the adjacent coelomic epithelium (CE) (arrowhead in C). Some YFP+ cells can be seen in the aorta, expressing the endothelial marker CD31 (arrow in C). Expression of GATA4 is prominent in the coelomic epithelium, suggesting activity of the enhancer G2 in specific areas of this tissue (D). Some YFP+ cells of the aortic endothelium are GATA4 immunoreactive (arrow and insert in D). The hematopoietic marker RUNX1 is expressed in some YFP+ cells of the aortic endothelium (arrows in E). F-I: YFP+ cells are also abundant in the placenta at stage E12.5. A general view of the placenta (F) shows the distribution of the YFP+ cells, more abundant in the chorionic plate (CP) and around the large vessels (V). The insert shows colocalization with the hematopoietic marker CD41 in the labyrinthine zone (LZ). YFP+ cells are very abundant in the vascular walls, sometimes forming part of the CD31+ endothelium of the large vessels and apparently detaching from it (insert in G). Colocalization with GATA4 and RUNX1 was observed in the vascular walls (arrows in H,I). Note the population of cells expressing GATA4 without activation of the G2 enhancer (arrowhead in H). J,K: YFP+ cells were also found in the fetal liver at stages E11.5 (J) and 13.5 (K). GATA4 is expressed by coelomic epithelium (CE) and mesenchymal cells of the liver, and most of them are YFP+ (arrows in J). Some YFP+ cells also express the hematopoietic marker RUNX1 (arrows in K). L,M: Immunolocalization of GATA2 in the placenta (L) and liver (M) of E11.5 and E12.5 embryos, respectively. No colocalization of GATA2 and YFP is evident in the placenta, but some YFP+ cells are GATA2+ cells in the liver (arrows) and others are not expressing GATA2 (arrowhead). N,O: YFP+ cells in the heart (N) and liver (O) of irradiated mice after four months of injections of bone marrow cells from G2Cre;R26REYFP mice. A number of these cells express the endothelial marker CD31 but they are negative for CD45. Some of these cells are integrated in the endocardium (EN). Clusters of hematopoietic cells appear in the wall of some hepatic sinusoids (arrow in O). P,Q: Busulfan-treated mice injected with cells obtained from G2Cre;R26REYFP embryos. After 4-7 months, putative endothelial CD31+/CD45–/YFP+ cells can be seen in the liver after injection of aorta-gonad-mesonephros (P) and placental (Q) cells (arrows). Bars represent 33 μm except for F (200 μm), L (50 μm) and M (25 μm). The images were acquired as 3-channel images by a Leica SP5 II confocal microscope (Leica, Heidelberg, Germany) using LAS AF software and 40X and 63X oil immersion objectives (numerical apertures 1.35 and 1.40, respectively). Levels were adjusted for the entire images using Photoshop 8.0.1.

haematologica | 2017; 102(4)

651


A. Cañete et al. Table 3. Absolute and relative number of HSCs/YFP+ cells found in hematopoietic embryonic tissues.

Tissue E11-5

Total number of cells/tissue (x10000) (n=5)

Fetal liver

27.2±14.7

AGM Placenta YS

21±5,6 379.2±83.2 100

% of CD41+ CD34+cKit+ Estimated number cells within of CD41+ CD34+ cells the P1 gate (n=4) by organ (n=4) YFP+ 0.29±0.03 (n=3) 0.02±0.01 0.90±0.21 0.32±0.12

YFP– 2.29±0.88 (n=3) 0.07±0.02 2.55±0.64 0.27±0.07

YFP+ YFP– 514±193 3813±996 (n=3) (n=3) 29±13 108±31 2508±363 7151±1531 1297±470 1099±197

% of LIN–CD31+cKit+ cells within the P1 gate (n=3)

Estimated number of LIN–CD31+cKit+ cells by organ (n=3)

YFP+ 0,45±0,17

YFP– 12.56±1.96

YFP+ 1016±607

YFP– 26689±4235

0.003±0.003 0.14±0.05 0.13±0.07

0.12±0.07 1.19±0.45 0.83±0.47

11±8 359±163 479±208

181±104 3322±2367 3096±1316

The table shows the estimated absolute and relative number of YFP+ and YFP– hematopoietic progenitors (defined by two criteria: cKit+/CD41+/CD34+ and cKit+/CD31+/Lin–) in different embryonic tissues (mean±SEM). The number of cells in the yolk sac is an estimation based on our previous work. HSCs: hematopoietic stem cells; AGM: aorta-gonadmesonephros: YS: yolk sac.

embryonic origin of the HSCs of the G2-Gata4 lineage. The cells used for these experiments were obtained from the fetal liver, AGM, placenta and yolk sac of E11.5 G2Cre;R26REYFP embryos, when definitive HSCs have already emerged.16 We considered two criteria to identify the fetal hematopoietic progenitors: CD41+/cKit+/CD34+ and Lin–/cKit+/CD31+.17,18 The four embryonic organs examined contained a population of YFP+ cells in all the stages studied, but the highest absolute and relative numbers of CD41+/cKit+/CD34+/YFP+ cells were found in the placenta (Table 3 and Online Supplementary Figure S4). The Lin–/cKit+/CD31+/YFP+ cells were also relatively abundant in the placenta, while they were virtually absent from the AGM region, where Lin–/cKit+/CD31+/YFP– cells were more frequent (Table 3 and Online Supplementary Figure S4). We have assayed the potential of the G2Cre;R26REYFP cells isolated from hematopoietic tissues of E10-E11 embryos to form hematopoietic colonies in the MethoCult assay. In three independent experiments, and seeding five plates by organ (10000 cells in each experiment), the cells obtained from fetal liver exhibited the highest clonogenic ability (Table 4). 29.2% of them were YFP+. Interestingly, the placenta showed a higher potential than the AGM to form colonies in vitro, and most (81.3%) of these placenta-derived colonies were YFP+ versus only 20% of the AGM-derived colonies. Thus, YFP+ colony forming cells are far more abundant in the placenta than in the AGM region. Furthermore, the culture of ten E8 ectoplacental cones in MethoCult also gave rise in one case to four YFP+ colonies, suggesting that G2-Gata4 hematopoietic progenitors are present in the early placenta, even before the onset of circulation. In order to directly compare the hematopoietic potential of the G2-Gata4 lineage cells obtained from different organs, we isolated the YFP+ fraction from each organ through fluorescence-activated cell sorting (FACS, E11.5 and E12.5) and seeded the purified cells in MethoCult. The result of three experiments is shown in Table 4. Fetal liver contained the highest number of YFP+ colony-forming cells, followed by the placenta. However, only a few colonies were obtained when the same numbers of YFP+ cells isolated from the AGM region were cultured. We next checked the potential of the embryonic G2-Gata4 lineage cells to repopulate the adult hematopoietic system, using the busulfan-treated newborn mouse model. We transplanted AGM, fetal liver and placental cells from stages E11.5-E12.5, as well as purified YFP+ cells sorted from the same tissues (Table 5). After four weeks we observed that the level of reconstitution of the hematopoi652

etic system, as assessed by the percentage of YFP+ cells in peripheral blood, was similar when AGM, fetal liver and placental cells were injected in the newborn mice, between 6.7% (placenta) and 11.0% (fetal liver). This percentage reached 19.2% in the AGM-injected chimaeras after 4-9 months, but the increase was small or absent in the chimaeras injected with placental or liver cells. The percentages of YFP+ cells in the bone marrow, thymus and spleen were similar in all the experiments. A significant difference between the short-term and longterm reconstitution of the T-lymphoid compartment was observed (Table 5), as described above for the busulfantreated mice injected with adult bone marrow. The percentage of YFP+ T lymphocytes increased between the shortand long-term independently of the origin of the cells. Considering all the experiments, the frequency of the YFP+ lymphocytes increased three times, from 4% to 12% (Student's t-test, P value=0.015), while the increase of the CD45+/YFP+ cells was not significant. When busulfan-treated newborns were injected with YFP+ cells purified by FACS from the three embryonic tissues, the fetal liver showed the highest ability to reconstitute the hematopoietic system, followed again by the placental cells (Table 6). YFP+ cells isolated from the AGM region showed a very weak ability to reconstitute the hematopoietic system in these mice, never reaching >1% of the nucleated blood cells. Thus, YFP+ cells from the placenta showed a much higher short- and long-term reconstitution potential than YFP+ cells isolated from the AGM region. In these mice, we again observed a statistically significant difference between short- and long-term percentages of lymphoid YFP+ cells (Table 6). We also analyzed the contribution of YFP+ cells to nonhematopoietic organs in the long-term busulfan-treated chimaeras injected with embryonic tissues. Endothelial cells of the G2-Gata4 lineage can be seen in many organs, and they were more abundant in the liver sinusoids (Figure 2P,Q).

GATA4 is not directly involved in the differentiation of hematopoietic progenitors To ascertain if GATA4 is required for embryonic hematopoiesis we have studied the conditional deletion of GATA4 in Scl-expressing cells (SCLCreERT;Gata4flox/flox). When GATA4 is deleted in the cells of the G2-Gata4 lineage, mice develop anemia.6 We inactivated GATA4 in cells of the SCL lineage through tamoxifen induction at different stages (E9.5-E11.5), but no differences were observed in the fetal hematopoiesis at different ages (E12.5-E14.5) (Online haematologica | 2017; 102(4)


Adult HSC from GATA4 expressing embryonic lineage

Table 4. Number of colonies generated in MethoCult after seeding of embryonic cells from different origins.

Type of cells

Tissue

Stage

Number of colonies

Analyzed by cytospin

YFP+ colonies

Total Total Total YFP+ purified YFP+ purified YFP+ purified YFP+ purified YFP+ purified YFP+ purified

Fetal liver AGM Placenta Fetal liver Fetal liver AGM AGM Placenta Placenta

E10-11 E10-11 E10-11 E11,5 E12,5 E11,5 E12,5 E11,5 E12,5

88.4±4.8 1.7±0.89 3.4±1.0 55.9±7.3 44.3±6.7 0.33±0.12 0.1±0.058 5.7±0.64 1.0±0.38

65 10 16

19 (29,2%) 2 (20%) 13 (81,3%)

The number of colonies generated in MethoCult after seeding of cells obtained from three hematopoietic tissues is presented. 10000 total cells were obtained from the tissues (E10-11, N=5) or 2000-14000 YFP+ cells were purified for each tissue (N=3, results are normalized to 1000 cells). In both cases the cells were seeded and the number of colonies counted. In the case of total cells seeded, the frequency of the YFP+ colonies was analyzed by cytospin (mean±SEM). AGM: aorta-gonad-mesonephros.

Table 5. Frequency of YFP+ cells after transplant of embryonic cells in busulfan-treated newborns.

Analyzed tissue

Donor tissue

Term

All cells

CD45

CD4+CD8a

Ter119

KSL

Peripheral blood

Fetal liver AGM Placenta Fetal liver AGM Placenta Fetal liver AGM Placenta Fetal liver AGM Placenta Fetal liver AGM Placenta

ST ST ST LT LT LT LT LT LT LT LT LT LT LT LT

11.0±2.1 (2) 8.4±0.57 (6) 6.7±0.45 (4) 10.7±4.5 (2) 19.2±4.5 (6) 8.2±3.6 (4) 10.7±7.4 (2) 16.7±6.4 (6) 18.5±17.1 (4) 26.6 (1) 29.4±16.4 (4) 20.1±18.9 (2) 17.0 (1) 16.3±7.4 (4) 17.5±8.3 (2)

10.7±2.3 (2) 10.1±1.4 (5) 6.5±0.6 (4) 10.8±4.5 (2) 18.4±6.2 (6) 8.5±3.8 (4) 11.0±10.5 (2) 22.8±24.6 (6) 18.6±33.9 (4)

5.1±0.78 (2) 1.8±0.61 (6) 5.3±2.9 (4) 11.6±4.7 (2) 14.1±4.0 (6) 7.9±5.5 (4) 27.6 (1)b 17.2±18.9 (4)b 24.3±30.4 (2)b

10.5±11.3 (2) 17.1±10.7 (6) 18.5±32.0 (4)

11.7±9.7 (2) 16.8±6.1 (6) 25.4±21.1 (4)

Bone marrow

Thymus

Spleen

The table shows the frequency of YFP+ cells (mean±SEM) in peripheral blood in the short- and long-term (1 and 4-9 months, respectively), bone marrow, the thymus and spleen (only long-term) after transplantation of cells from G2Cre;R26REYFP embryos into busulfan-treated newborns. The number of experiments is between parentheses. aThe difference found between the % of YFP+ cells in the CD4+CD8 populations in the short- and long-term was statistically significant when the experiments were gathered (ST: 4.1±1.2%, LT: 11.8±2.6%, Student's t-test, P value=0.015). bCD3 was used instead of CD4+CD8 in these samples. AGM: aorta-gonad-mesonephros; ST: short-term; LT: long-term; KSL: c-Kit+/Sca1+/Lin–.

Supplementary Table S2).

Discussion Early embryonic hematopoiesis occurs in three main territories; the yolk sac, AGM region and placenta. It has been well established that the hematopoietic progenitors originated in the yolk sac give rise to a transient wave of blood cells, mainly erythroid. Moreover, definitive HSCs are considered to derive from the AGM region, since the role played by the placenta in definitive hematopoiesis is more controversial, as discussed below. The lack of markers allowing one to distinguish between placental and intraembryonic HSCs has hampered the solving of this controversy. We have identified two distinct lineages in the adult blood cells of mice, characterized by the activity, during embryonic life, of an enhancer of the Gata4 gene (G2-Gata4) that drives its expression in the allantois/placenta and lateral mesoderm, starting at E7.5 and ceasing by midgestation. About 20% of the murine adult blood cells belong to this lineage, and this percentage shows a remarkhaematologica | 2017; 102(4)

able consistency between individuals. The G2-Gata4 lineage of blood cells derive from bone marrow HSCs which can be transplanted and reconstitute, over the long-term, the hematopoietic system of both lethally irradiated and busulfan-treated mice. The HSCs belonging to the G2Gata4 lineage give rise to all kinds of blood cells and even to putative circulating endothelial progenitors and longterm engrafted vascular endothelial cells, and they show basically the same properties as the other HSCs. An exception was the lower percentages of YFP+ T lymphocytes found in the short-term reconstitution of the hematopoietic system after myeloablation. This difference might be explained by the persistence of the host lymphocytes after the transplantation. T lymphocytes half-life in humans span for weeks,19 thus, host lymphocytes of myeloablated mice are probably diluting the YFP+ lymphocytes in the shortterm and reducing their relative abundances. On the other hand, the smaller frequency of YFP+ cells found among the Ter119+ bone marrow population can be explained by downregulation of the expression of YFP as the erythroblasts differentiate. The existence of heterogeneity among populations of 653


A. Cañete et al. Table 6. Frequency of YFP+ cells after transplant of YFP+ cells purified from embryonic tissues in busulfan-treated newborns.

Donor tissue and number of YFP+ cells transplanted

Fetal liver (20000 cells)

Number Survivors of transplants

Mice with >1% YFP+ ST

Mice with >1% YFP+ LT

4

4

2

2

AGM (10000-25000 cells) 6 Placenta (13000-25000 cells) 9

4b 7

0 4

0 3

Bone marrow Thymus Spleen All cells ST LT ST 26.0± 55,6± 19.6± 4.1 12.7 6.6 19.9± 43.5± 7.7± 10.3 26.1 4.7

Lysated PB CD4+CD8a CD45 LT 70.9± 4.1 40.9± 24.6

ST 31.7± 6.0 19.8± 10.6

LT 55.6± 5.0 42.9± 25.6

B220 CD11b

All cells

All cells

LT LT 68.1± 47.8± 4.7 9.1 46.8± 54.3± 28.8 33.2

LT 60.3± 10.6 43.6± 23.5

LT LT 90.8± 76.4± 0.78 2.8 62.6± 47.7± 31.3 20.9

All cells

The table shows the frequency of YFP+ cells (mean±SEM) in peripheral blood in the short- and long-term (ST and LT, 1 and 4-9 months, respectively), bone marrow, the thymus and spleen after transplantation into busulfan-treated newborns of YFP+ cells purified from tissues from G2Cre;R26REYFP embryos. aThe difference found between the % of YFP+ cells in the CD4+CD8 populations in the short- and long-term was statistically significant when the experiments were gathered (ST: 12.4±4.1%, LT: 52.9±13.3%, Student's t-test, P value=0.036). bThe percentages of YFP+ cells found in these mice were between 0.01% and 0.47%. AGM: aorta-gonad-mesonephros; ST: short-term; LT: long-term; PB: peripheral blood.

adult HSCs has been well established, but their causes and their potential relationships with different pathways of developmental hematopoiesis are unknown.20-23 As acknowledged by Kaimakis et al., the hematopoietic progenitors generated in the embryo result as being more diverse than previously appreciated.24 These authors described a population of embryonic hematopoietic progenitors cells (HPCs) that did not express GATA2, and they suggested that GATA4 is expressed and might play some role in this specific population. This proposal might be related with our observation of a lineage of HSCs derived from GATA4 expressing progenitors. In fact, we found no expression of GATA2 in the YFP+ cells from the placenta, although part of the fetal liver YFP+ cells were GATA2+ (Figure 2L,M). Thus, we do not disregard that the G2-Gata4 hematopoietic lineage can overlap totally or partially with the already described GATA2-independent population of HPCs.24 An important point revealed in our study is the embryonic origin of the G2-Gata4 HSCs. The G2 enhancer activates GATA4 expression in hematopoietic cells from both the placenta and AGM, although YFP+ cells were far more abundant in the placenta. The yolk sac also showed a significant population of presumptive YFP+ hematopoietic progenitors, but they probably do not originate in the yolk sac, since the enhancer G2 is not driving GATA4 expression in that tissue.5 In fact, conditional deletion of GATA4 in the G2 domain does not cause abnormalities in the yolk sac.6 Thus, we compared the number of YFP+ hematopoietic progenitors identified under two different criteria, (CD41+/cKit+/CD34+ and Lin–/CD31+/cKit+),17,18 in the placenta and in the AGM, and they were much more abundant in the former. In fact, Lin-/CD31+/cKit+/YFP+ cells were virtually lacking in the AGM region, whereas a significant number of Lin–/CD31+/cKit+/YFP– cells were identified. YFP+ colonies were four times more frequent in MethoCult when we seeded placental cells than when AGM cells were seeded. When unfractionated AGM and placental cells were transplanted into busulfan-treated newborn mice, the contribution to the YFP+ blood cells was basically similar. Interestingly, we recorded a two-fold increase of the AGM contribution to the reconstituted YFP+ population in peripheral blood in the long-term as compared with the shortterm, although this increase was not significant and it was not evident in bone marrow. Furthermore, YFP+ cells from 654

the placenta, but not those isolated from AGM, reconstituted the hematopoietic system of busulfan-treated newborns. We cannot exclude that YFP+ cells from the AGM require accessory cells for engraftment and/or a process of maturation and expansion to acquire a potential of reconstitution similar to that exhibited by the placental YFP+ cells.25 However, we think that our experimental evidence clearly supports a placental origin for most YFP+ hematopoietic progenitors, and presumably also for most of the adult HSCs derived from the G2-Gata4 lineage. Placental hematopoiesis has been well described,16,26-27 although the primary origin of placental HSCs is controversial.28 The allantois from embryos E8 pre-cultured in toto prior to seeding in a semisolid medium has demonstrated containing hematopoietic potential.29 Using this experimental system, we performed the MethoCult assay with ectoplacental cones and allantois by E8, before the onset of the circulation, and have obtained YFP+ colonies in one case. This supports the ability of the placenta to generate definitive HSCs, as suggested by other reports.30 To date, there is no report of GATA4 being involved in hematopoiesis. We previously suggested that the anemia phenotype found in mouse embryos with conditional deletion of GATA4 in the lateral mesoderm domain was due to the failure of the expansion of hematopoietic progenitors in the liver, rather than from a direct impact of a lack of GATA4 in hematopoietic cells.6 Our new results confirm this idea, as no hematopoietic defects were found when we inactivated Gata4 either in the G2-GATA4 lineage, or in the hematopoietic progenitors using an inducible SclCreERT driver. However, redundancy among members of the GATA family have been previously described,31-34 and we have discussed above how Kaimakis et al. do not disregard some redundancy between GATA4 and GATA2 in a subpopulation of embryonic hematopoietic progenitors cells.24 The embryonic origin of the different populations of bone marrow mesenchymal stem cells is still poorly known.35,36 This was not an aim of our study, but we observed a contribution of YFP+ cells to the lineage-negative bone marrow cells within the CD90+, CD73+ and CD105+ subpopulations, classical markers of mesenchymal stem cells. This contribution was higher (30-40%) than that obtained for hematopoietic stem cells. Furthermore, 50% of the lineage-negative SCA1+/PDGFRα+ population, that is haematologica | 2017; 102(4)


Adult HSC from GATA4 expressing embryonic lineage

enriched in mesenchymal stem cells,13 was also YFP+. Further study would be necessary to establish the relevance of these observations, possibly related to the origin of a mesenchymal stem cell subpopulation from the lateral mesoderm. In summary, we have identified a distinct lineage of adult HSCs characterized by its derivation of progenitors where Gata4 expression is activated by a specific enhancer. Most adult HSCs belonging to this lineage probably originate in the placenta. Despite the relatively normal behavior of this specific lineage, further experiments would be required to examine whether they exhibit differences in response to different physiopathological conditions.

References 12. 1. Gao J, Chen YH, Peterson LC. GATA family transcriptional factors: emerging suspects in hematologic disorders. Exp Hematol Oncol. 2015;4:28. 2. Kuo CT, Morrisey EE, Anandappa R , et al. GATA4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev. 1997;11(8):1048-1060. 3. Molkentin JD, Lin Q, Duncan SA, Olson EN. Requirement of the transcription factor GATA4 for heart tube formation and ventral morphogenesis. Genes Dev. 1997; 11(8):1061-1072. 4. Narita N, Bielinska M, Wilson DB. Wildtype endoderm abrogates the ventral developmental defects associated with GATA-4 deficiency in the mouse. Dev Biol. 1997;189(2):270-274. 5. Rojas A, De Val S, Heidt AB, Xu SM, Bristow J, Black BL. Gata4 expression in lateral mesoderm is downstream of BMP4 and is activated directly by Forkhead and GATA transcription factors through a distal enhancer element. Development. 2005;132(15):3405-3417. 6. Delgado I, Carrasco M, Cano E, et al. GATA4 loss in the septum transversum mesenchyme promotes liver fibrosis in mice. Hepatology. 2014;59(6):2358-2370. 7. Pu WT, Ishiwata T, Juraszek AL, Ma Q, Izumo S. GATA4 is a dosage sensitive regulator of cardiac morphogenesis. Dev Biol. 2004;275(1):235-244. 8. Göthert JR, Gustin SE, Hall MA, et al. In vivo fate-tracing studies using the Scl stem cell enhancer: embryonic hematopoietic stem cells significantly contribute to adult hematopoiesis. Blood. 2005;105(7):27242732. 9. Rossi L, Challen GA, Sirin O, et al. Hematopoietic stem cell characterization and isolation. Methods Mol Biol. 2011; 750:47–59. 10. Garcia-Ortega AM, Cañete A, Quinter C, et al. Enhanced hematovascular contribution of SCL 3' enhancer expressing fetal liver cells uncovers their potential to integrate in extramedullary adult niches. Stem Cells. 2010;28(1):100-112. 11. Cañete A, Comaills V, Prados I, et al. Characterization of a fetal liver cell population endowed with long-term multiorgan

haematologica | 2017; 102(4)

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

Funding This study was supported (in part) by research funding from grants BFU2014-52299-P (The Spanish Ministry of Economy), RD12/0019-0022 (Instituto de Salud Carlos III-TERCEL network), and P11-CTS-07564 (Junta de Andalucía) to RM-C and PI14-00804 (Instituto de Salud Carlos III cofounded by FEDER funding) to AR. Acknowledgments The authors would like to thank Dr. Brian Black for providing the G2-Gata4Cre mice, Dr. Agustín Zapata for his valuable comments and suggestions and David Navas and John Pearson for technical support.

endothelial reconstitution potential. Stem Cells. 2016 Sep 12. [Epub ahead of print] Sands MS, Barker JE. Percutaneous intravenous injection in neonatal mice. Lab Anim Sci. 1999;49(3):328-330. Houlihan DD, Mabuchi Y, Morikawa S, et al. Isolation of mouse mesenchymal stem cells on the basis of expression of Sca-1 and PDGFR-α. Nat Protoc. 2012;7(12):21032111. Bailey AS, Willenbring H, Jiang S, et al. Myeloid lineage progenitors give rise to vascular endothelium. Proc Natl Acad Sci USA. 2006;103(35):13156–13161. Yoder, MC, Hiatt K, Mukherjee P. In vivo repopulating hematopoietic stem cells are present in the murine yolk sac at day 9.0 postcoitus. Proc Natl Acad Sci USA. 1997;94(13): 6776-6780. Golub R, Cumano A. Embryonic hematopoiesis. Blood Cells Mol Dis. 2013;51(4):226-231. McKinney-Freeman SL, Naveiras O, Yates F, et al. Surface antigen phenotypes of hematopoietic stem cells from embryos and murine embryonic stem cells. Blood. 2009;114(2):268-278. Baumann CI, Bailey AS, Li W, et al. PECAM1 is expressed on hematopoietic stem cells throughout ontogeny and identifies a population of erythroid progenitors. Blood. 2004;104(4):1010-1016. Hellerstein M, Hanley MB, Cesar D, et al. Directly measured kinetics of circulating T lymphocytes in normal and HIV-1-infected humans. Nat Med. 1999;5(1):83-89. Dykstra B, Kent D, Bowie M, et al. Longterm propagation of distinct hematopoietic differentiation programs in vivo. Cell Stem Cell. 2007;1(2):218-229. Copley MR, Beer PA, Eaves CJ. Hematopoietic stem cell heterogeneity takes center stage. Cell Stem Cell. 2012;10(6):690697. Benz C, Copley MR, Kent DG, et al. Hematopoietic stem cell subtypes expand differentially during development and display distinct lymphopoietic programs. Cell Stem Cell. 2012;10(3):273-283. Crisan M, Kartalaei PS, Vink CS, et al. BMP signalling differentially regulates distinct haematopoietic stem cell types. Nat Commun. 2015;6:8040. Kaimakis P, de Pater E, Eich C, et al. Functional and molecular characterization of

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

mouse Gata2-independent hematopoietic progenitors. Blood. 2016;127(11):1426-1437. Taoudi S, Gonneau C, Moore K, et al. Extensive hematopoietic stem cell generation in the AGM region via maturation of VE-cadherin+CD45+ pre-definitive HSCs. Cell Stem Cell. 2008;3(1):99-108. Robin C, Bollerot K, Mendes S, et al. Human placenta is a potent hematopoietic niche containing hematopoietic stem and progenitor cells throughout development. Cell Stem Cell. 2009;5(4):385-395. Ottersbach K, Dzierzak E. The placenta as a haematopoietic organ. Int J Dev Biol. 2012;54(6-7):1099-1106. Mikkola HK, Gekas C, Orkin SH, DieterlenLievre F. Placenta as a site for hematopoietic stem cell development. Exp Hematol. 2005;33(9):1048-1054. Corbel C, Salaun J, Belo-Diabangouaya P, Dieterlen-Lievre F. Hematopoietic potential of the pre-fusion allantois. Dev Biol. 2007;301(2):478–488. Rhodes KE, Gekas C, Wang Y, et al. The emergence of hematopoietic stem cells is initiated in the placental vasculature in the absence of circulation. Cell Stem Cell. 2008;2(3):252–263. Xin M, Davis CA, Molkentin JD, et al. A threshold of GATA4 and GATA6 expression is required for cardiovascular development. Proc Natl Acad Sci USA. 2006; 103(30):11189-11194. Zhao R, Watt AJ, Battle MA, Li J, Bondow BJ, Duncan SA. Loss of both GATA4 and GATA6 blocks cardiac myocyte differentiation and results in acardia in mice. Dev Biol. 2008;317(2):614-619. Xuan S, Borok MJ, Decker KJ, et al. Pancreas-specific deletion of mouse Gata4 and Gata6 causes pancreatic agenesis. J Clin Invest. 2012;122(10):3516-3528. Carrasco M, Delgado I, Soria B, Martín F, Rojas A. GATA4 and GATA6 control mouse pancreas organogenesis. J Clin Invest. 2012; 122(10):3504-3515. Crisan M, Yap S, Casteilla L, et al. A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell. 2008;3(3):301-313. Isern J, García-García A, Martín AM, et al. The neural crest is a source of mesenchymal stem cells with specialized hematopoietic stem cell niche function. Elife. 2014; 3:e03696.

655


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Lysosomal Storage Disease

Ferrata Storti Foundation

Haematologica 2017 Volume 102(4):656-665

Efferocytosis is impaired in Gaucher macrophages

Elma Aflaki,1 Daniel K. Borger,1 Richard J. Grey,1 Martha Kirby,2 Stacie Anderson,2 Grisel Lopez1 and Ellen Sidransky1*

1 Section of Molecular Neurogenetics, Medical Genetics Branch and 2Flow Cytometry Core, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD, USA

ABSTRACT

G

aucher disease, the inherited deficiency of lysosomal glucocerebrosidase, is characterized by the presence of glucosylceramideladen macrophages resulting from impaired digestion of aged erythrocytes or apoptotic leukocytes. Studies of macrophages from patients with type 1 Gaucher disease with genotypes N370S/N370S, N370S/L444P or N370S/c.84dupG revealed that Gaucher macrophages have impaired efferocytosis resulting from reduced levels of p67phox and Rab7. The decreased Rab7 expression leads to impaired fusion of phagosomes with lysosomes. Moreover, there is defective translocation of p67phox to phagosomes, resulting in reduced intracellular production of reactive oxygen species. These factors contribute to defective deposition and clearance of apoptotic cells in phagolysosomes, which may have an impact on the inflammatory response and contribute to the organomegaly and inflammation seen in patients with Gaucher disease.

Correspondence: sidranse@mail.nih.gov

Received: August 25, 2016. Accepted: December 16, 2016. Pre-published: December 23, 2016. doi:10.3324/haematol.2016.155093 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/656

©2017 NIH (National Institutes of Health)

656

Introduction Gaucher disease (GD) is a lysosomal storage disease caused by inherited deficiency of the lysosomal hydrolase glucocerebrosidase, which leads to accumulation of glucosylceramide and glucosylsphingosine in lysosomes. Macrophages are the primary cell type affected in GD, and inefficient breakdown of glycolipid-rich membranes of phagocytosed cells leads to accumulation of lipid-laden GD macrophages in organs of the reticuloendothelial system.1 The accumulation of these GD macrophages is thought to contribute to chronic inflammation in patients with GD, but the mechanism underlying this link has remained largely undetermined. Recently, using primary macrophages derived from patients with GD, we demonstrated that lipid storage and subsequent lysosomal dysfunction leads to impaired macroautophagy in GD macrophages, contributing to dysregulation of pro-inflammatory cytokine processing and release by these cells.2 These findings led us to explore whether phagocytosis, a process vital to macrophages and mechanistically similar to autophagy, may also be affected in GD macrophages. While phagocytosis is important for the response of phagocytes to pathogens, phagocytosis of apoptotic cells – a process known as efferocytosis – is vital for tissue homeostasis.3 In this process, apoptotic cells release signaling molecules that recruit tissue-resident macrophages to the site of cell death4,5 and prompt phagocytosis of the apoptotic cell.6,7 Following phagocytosis, the nascent phagosome undergoes a process of maturation, facilitating digestion of the apoptotic cell.8 Clearance of apoptotic cells via efferocytosis is required to maintain immunological homeostasis, and the breakdown of this process is associated with inflammatory disease and autoimmunity.9 Hence, impairment of apoptotic cell engulfment by GD macrophages could contribute to chronic inflammation and organomegaly in GD. While studies using primary macrophages from patients with GD have indicated that GD macrophages have defects in the digestion of phagocytosed microbes, no studies have examined efferocytosis by GD macrophages.10 Here, we used primary macrophages from patients with GD to assess efferocytosis in GD. We found that while recognition and uptake of apoptotic cells by macrophages is not impaired in GD, digestion of engulfed cells is severely affected haematologica | 2017; 102(4)


Impaired efferocytosis in Gaucher disease

in the disease. This is caused by aberrant recruitment of phagosome-associated proteins, leading to substantially impaired phagosome maturation and phagosome-lysosome fusion. Failure of efferocytosis may, therefore, contribute to storage in GD macrophages and exacerbate the inflammatory and hematologic aspects of this disease.

Methods Peripheral blood mononuclear cell collection and isolation and differentiation of macrophages GD macrophages were derived from peripheral blood monocytes isolated from 25 patients with type 1 GD seen at the National Institutes of Health (Bethesda, MD, USA). GBA1 genotyping, performed as described elsewhere,11 revealed that two patients had genotype N370S/L444P, one had N370S/c.84dupG and 22 had N370S/N370S. Informed consent was obtained in accordance with a National Human Genome Research Institute Internal Review Board-approved clinical protocol. Thirty-two control macrophage samples were derived from monocytes isolated from blood from healthy donors provided by the Blood Bank at the National Institutes of Health Clinical Center. Peripheral blood mononuclear cells were isolated as described previously.12 Briefly, peripheral blood mononuclear cells were isolated using a Ficoll gradient, and monocytes were enriched via CD16 depletion. Macrophages were differentiated from purified monocytes using macrophage colony-stimulating factor (10 ng/mL) (R&D Systems, Minneapolis, MN, USA) in RPMI 1640 medium (ThermoFisher Scientific, Waltham, MA, USA), supplemented with 10% fetal calf serum (FCS) (Invitrogen, Carlsbad, CA, USA) and 1% Pen Strep (ThermoFisher Scientific). On days 3 and 6, the medium was replaced, and experiments were performed on day 7. Jurkat cells were a gift from Dr. Martin Playford (National Heart, Lung, and Blood Institute, National Institutes of Health). Erythrocyte ghosts, added to the cultured cells as a source of lipid, were prepared from blood collected from a patient with GD as previously described.12

Quantification of efferocytosis by flow cytometry Monocyte-derived macrophages were cultured in 12-well plates. Jurkat cells were incubated with CytoTracker Green (Molecular Probes, Eugene, OR, USA) at 37°C with 5% CO2 for 30 min. To induce apoptosis in Jurkat cells, cells were washed twice in RPMI without FCS. They were irradiated with ultraviolet light at 30 mJ/cm2 at a wavelength of 254 nm using a StrataLinker UV cross-linker (Stratagene, La Jolla, CA, USA) and then incubated at 37°C with 5% CO2 for 4 h. Macrophages were transferred to RPMI without FCS. Apoptotic Jurkat cells were added at a ratio of five apoptotic Jurkat cells: one macrophage and incubated together at 37°C with 5% CO2 for 1 h. Cells were then quickly washed with cold phosphate-buffered saline (PBS) to remove non-efferocytosed apoptotic Jurkat cells and scraped in PBS with 10% FCS. For antibody staining, allophycocyanin-conjugated CD11b antibodies (BD Biosciences, San Jose, CA, USA) were added to cells at a 1:200 dilution for 20 min. Cells were subsequently spun down and fresh PBS with 10% FCS was added. Flow cytometry was performed using a FACSCalibur (BD Biosciences).

(Alabaster, AL, USA). 100 mol% PC and an 80 mol% PC:20 mol% PS mixture were dried under N2, suspended in Dulbecco PBS without calcium to a concentration of 200 nM, and sonicated for 2 min. 6.4 μm glass beads or 2 μm green microsphere beads (Bangs Laboratory, Inc., Fishers, IN, USA) were centrifuged and washed three times, suspended in Dulbecco PBS and sonicated for 2 min to form a single-bead suspension. Beads were then added to the lipid suspensions (1×105 beads/nmol lipid), vortexed vigorously for 2 min, and incubated at room temperature for 10 min. The resulting lipid-coated beads were centrifuged, washed three times and stored at 4°C in Dulbecco PBS. The coated beads were stained with annexin V (Roche Life Science, Indianapolis, IN, USA) and analyzed using a FACSCalibur (BD Biosciences).

Immunocytochemistry For microsphere bead assays, cells were plated on glass coverslips in 12-well plates. Lipid-coated beads were added to the macrophages. The plates were spun at 4°C at 200 rpm for 2 min and incubated for 10 min at 4°C to allow the beads to bind to macrophages. Cells were washed with cold PBS and incubated at 37°C for the indicated times. Cells were fixed for 1 min with icecold methanol and then blocked in PBS containing 0.1% saponin, 100 μM glycine, and 2% donkey serum. For the assays of apoptotic Jurkat cells, macrophages were plated on glass chamber slides. Cells were incubated with apoptotic Jurkat cells labeled with CytoTracker green (Molecular Probes) for 60 min, fixed with 4% paraformaldehyde for 30 min, permeabilized with 0.1% Triton X-100 for 10 min and blocked. Cells were then incubated with the primary antibody diluted in PBS containing 0.1% saponin and 0.1% bovine serum albumin. The following primary antibodies were used for immunocytochemistry: goat α-p67phox antibody (1:200, Santa Cruz Biotechnology, Dallas, TX, USA), rabbit α-Rab5 (1:100, Abcam, Cambridge, UK), mouse α-Rab7 (1:100; Santa Cruz Biotechnology), rabbit α-Rab7 (1:100; Abcam), mouse α-Lamp1 (1:200, hybridoma), mouse α-Lamp2 (1:200, hybridoma) and goat α-Cathepsin D (1:200, R&D Systems). Cells were washed and incubated with donkey α-mouse, α-rabbit, or α-goat secondary antibodies conjugated to Alexa Fluor® 488, Alexa Fluor® 555, Alexa Fluor® 647 (Invitrogen) or Alexa Fluor® 647 phalloidin (Molecular Probes). Cells were mounted with ProLong® Gold antifade reagent with DAPI (Molecular Probes), and Z-stack images were acquired with a Zeiss 510 META laser scanning microscope (Carl Zeiss MicroImaging, Inc., Jena, Germany using a 488-nm argon, a 543nm HeNe and an ultraviolet laser. Images were acquired using a Plan Neofluar 63×/1.42 oil DIC objective (Carl Zeiss MicroImaging, Inc.). The intensity of each protein in the phagosomes was analyzed using IMARIS software (Bitplane, Belfast, Ireland).

Measurement of intracellular reactive oxygen species Macrophages were cultured in 96-well, black, clear-bottom plates in the presence and absence of erythrocyte ghosts.12 Apoptotic Jurkat cells were then added for 60 min. In addition, cells were incubated with apoptotic Jurkat cells supplemented with ATP (0.5 mM). Cells were stained with CellROXTM Deep Red Reagent (Molecular Probes) at a final concentration of 5 μM, and incubated for 30 min at 37°C. Cells were washed, and fluorescence was evaluated at 640/665 nm using a FlexStation 3 microplate reader (Molecular Devices).

Lipid coating of beads This procedure was adapted from a published protocol.13 1,2dioleoyl-sn-glycero-3-[phospho-L-serine] (PtdSer) (18:1 PS) and 1,2-dioleoyl-sn-glycero-3-phosphocholine (PtdCho) (18:1 PC) dissolved in chloroform were obtained from Avanti Polar Lipids, Inc. haematologica | 2017; 102(4)

Protein isolation and immunoblotting Cells were harvested and sonicated at 4°C in RIPA lysate buffer [50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5% Na deoxycholate, 0.1% SDS and protease inhibitor]. Protein quantifi657


E. Aflaki et al.

cation was performed using a BCA protein assay (Thermo Scientific, Waltham, MA, USA). Lysates were run on 4% - 20% polyacrylamide gels (BioRad, Hercules, CA, USA) and transferred to polyvinylidene difluoride membranes using the Trans-Blot Turbo system (BioRad). Blots were dried, reactivated in methanol and blocked using a mixture of one part of PBS with one part of Odyssey Blocking Buffer (LI-COR Biosciences, Lincoln, NE, USA) for 1 h at room temperature. The primary antibodies used were: αp67phox (Abcam), α-Rab5 (Abcam), α-Rab7 (Abcam and Santa Cruz Biotechnology), α-β-actin (Abcam) and α-GAPDH (GeneTex, Irvine, CA, USA). Primary antibodies were diluted 1:1000, except for β-actin, which was diluted 1:4000. Dilution was done in blocking buffer containing 0.1% Tween-20. Blots were incubated with the primary antibody dilution overnight at 4°C, followed by 5 min washes in PBST (0.1% Tween-20). IRDye® 680RD goat anti-rabbit and IRDye® 800CW donkey anti-mouse secondary antibodies (LiCOR Bioscience) were diluted 1:10,000 in blocking buffer containing 0.1% Tween-20 and 0.01% SDS. Blots were incubated with secondary antibody for 1 h at room termperature, followed by 5 min washes in PBST. Blots were imaged using the Li-COR Odyssey imaging system (Li-COR Bioscience) and quantified using Image Studio (Li-COR Bioscience).

Statistical analysis Statistical analyses were performed using GraphPadPrism 6.0 software. Significance was determined by a Student t-test. Data from two groups or more than two independent variables were analyzed by ANOVA, followed by the Bonferroni post hoc test. Data are presented as mean values ± standard deviation. Significance levels for differences between controls and patients’ macrophages were set when P<0.05(*), P<0.01(**), and P<0.001(***) for different conditions.

Results Gaucher macrophages engulf apoptotic cells Phagocytic clearance of apoptotic cells is initiated by the migration of macrophages to the site where apoptotic cells are located in the response to specific attraction signals released by the cells as they undergo apoptosis. The second signal directing efferocytosis is the juxtacrine “eat me” signal, which is displayed on the surface of apoptotic cells in order to initiate the process of their internalization and digestion. The primary “eat me” signal is exposed phosphatidylserine (PtdSer), a phospholipid typically found on the cytosolic surface of the plasma membrane, but which diffuses to the outer surface during apoptosis. To evaluate the response of macrophages to this juxtacrine signaling, GD and control macrophages were incubated with apoptotic Jurkat cells at a high confluence (Figure 1A). Macrophages were detached, stained for CD11b, and the engulfment of apoptotic cells at different time points was evaluated by flow cytometry (Figure 1B,C). The engulfment of both apoptotic Jurkat cells and 2 µm beads coated with PtdSer (Figure 1E) did not differ significantly between control and GD macrophages, and pre-incubation with erythrocyte ghosts (to mimic the in vivo condition)12 had no impact on the macrophages’ efferocytic ability (Figure 1D).

Impaired efferocytosis is due to defective phagosome maturation in Gaucher macrophages Both impaired recognition and clearance of apoptotic cells may contribute to inflammation in GD.2 Following engulfment, deposition of ingested cells within phago658

somes is achieved by gradual acidification of the phagosome and fusion with lysosomes. Rab5 and Rab7 play coordinated roles in the process of phagosome maturation. Rab5 accumulates at the cytosolic surface of the nascent phagosome during engulfment of apoptotic cells, and mediates phagosome maturation.14,15 We studied phagosome maturation at different time points after adding glass beads coated with a mixture of PtdSer and PtdCho. Glass beads were used to rule out the participation of ligands other than PtdSer that might be present on the surface of apoptotic cells. Another advantage of these glass beads is that they do not introduce additional proteins and they are not broken down or degraded within the phagosome. The presence of PtdSer-coated beads was confirmed by annexin V (Figure 1F). Control and GD1 macrophages were incubated with PtdSer-coated beads for different time intervals, and immunofluorescence staining was performed to identify proteins associated with specific stages of phagosome maturation. Rab5 and Rab7 are indicators of early and late endosomes, respectively, while cathepsin D (CathD) and Lamp2 were used as lysosomal markers. Rab5 is responsible for the recruitment of Rab7 to the surface of phagosomes, and Rab7, in turn, mediates the fusion of phagosomes with lysosomes.16,17 Figure 2A shows that Rab5 accumulated on phagosomes 5 min after phagocytosis, and this accumulation decreased gradually in both control and GD macrophages. The recruitment of Rab5 to the phagosome occurred similarly in control and GD macrophages (Figure 2A). However, in control macrophages, Rab7 appeared 10 min after phagocytosis, remaining for 120 min, while the amount of Rab7 on phagosomes was significantly reduced in GD macrophages compared to controls (Figure 2A-D). In addition, levels of CathD and Lamp2, two different lysosomal markers, increased in the phagosomes of control macrophages 40 min after engulfing PtdSer-coated beads, while they remained low for 120 min in GD macrophages (Figure 2B,E,F). Macrophages were also stained for Rab5, Lamp2 and CathD 120 min after efferocytosis, demonstrating the absence of Rab5 on the phagosome 120 min after engulfing the beads (Figure 2B, bottom panel). We next studied the expression of Rab7 in macrophages fed with apoptotic Jurkat cells. Rab7 levels were lower in GD macrophages than in control macrophages both 30 min (Figure 3A) and 60 min (Figure 3B) after efferocytosis of apoptotic Jurkat cells. When macrophages were fed with erythrocyte ghosts prior to efferocytosis, Rab7 levels were significantly lower in GD macrophages (48%±0.09) than control macrophages (Figure 3B,C).

The respiratory burst is diminished in Gaucher macrophages We previously demonstrated that the digestion of erythrocyte ghosts was defective in GD macrophages due to reduced production of reactive oxygen species (ROS).12 During phagocytosis, phagosomes actively recruit elements of the NADPH-oxidase complex (NOX2), which generates ROS within the phagosomal lumen and mediates the degradation of phagosomal contents. Among the five components of NOX2, gp91phox and p22phox are present in the membrane, and, upon phagocytosis, they recruit the cytosolic components p47phox and p67phox.18 We found that at baseline, GD macrophages had reduced p67phox levels compared to control macrophages. Twenty minutes after efferocytosis, p67phox levels increased marginally in control haematologica | 2017; 102(4)


Impaired efferocytosis in Gaucher disease

macrophages, but decreased further in GD macrophages (Figure 3A). Feeding GD macrophages with ghosts prior to efferocytosis resulted in an even more profound reduction in p67phox (Figure 3B,C). Following efferocytosis, in control macrophages, p67phox translocated to phagosomes containing apoptotic Jurkat cells or beads coated with PtdSer (Figure 3D), while in GD macrophages no significant translocation was observed. Next we studied the recruitment of p67phox to phagosomes

containing PtdSer-coated glassbeads. Macrophages were fixed at different time points after engulfing PtdSer-coated beads and were co-stained for p67phox and Rab5 or Rab7. Immunostaining showed that 10 min after engulfing beads coated with PtdSer, p67phox was present on the early phagosome (stained with Rab5) in control, but not GD macrophages (Figure 4A-C). In addition, the level of p67phox increased significantly in control macrophages during phagosome maturation when co-stained with Rab7 (120

A

B

D

C

E

F

Figure 1. Efferocytosis in Gaucher macrophages. (A) Jurkat cells were UV-irradiated, incubated at 37°C for 4 h, stained for annexin V/7AAD and analyzed by flow cytometry. Early apoptotic cells represent 66.1% of the population. (B) Control and Gaucher macrophages were co-cultured with GFP-labeled apoptotic Jurkat cells for different time intervals. Non-efferocytosed apoptotic Jurkat cells were removed by washing with PBS, and macrophages were detached and analyzed by flow cytometry. Three independent experiments were performed on samples from three different patients with Gaucher disease (GD). (C-D) GFP-labeled apoptotic Jurkat cells (AJ) were added to macrophages both with and without the addition of erythrocyte ghosts for 60 min. They were detached after extensive washing, stained for CD11b, and analyzed by flow cytometry. The graph represents data from six independent experiments. (E) 2 μm FluoSpheres sulfate beads, coated with PtdSer, were added to macrophages for 60 min and analyzed by flow cytometry. (F) 5 μm glass beads were coated with PtdSer and PtdCho, stained with annexin V and analyzed by flow cytometery. Dashed red and black lines represent unstained coated beads, while solid red and black lines indicate stained beads coated with PtdCho and PtdCho-PtdSer.

haematologica | 2017; 102(4)

659


E. Aflaki et al.

A

B

C

E

660

D

F

Figure 2. Phagosome maturation is impaired in Gaucher macrophages. (A,B) Control and Gaucher macrophages were co-cultured with PtdSercoated glass beads (shown by *) for different time intervals. Cells were fixed and were stained with (A) Rab5 (green), Rab7 (white) and CathD (red) or (B) Rab7 (green), Lamp2 (white) and CathD (red). (C-F) Using IMARIS software, the intensity of Rab5 (C), Rab7 (D), Lamp2 (E) and CathD (F) was measured on the phagosomes. Graphs represent data from 20 images at each time point. This experiment was performed using cells from four different individuals with genotype N370S/N370S, two with genotype N370S/L444P and six controls.

haematologica | 2017; 102(4)


Impaired efferocytosis in Gaucher disease

min), while no significant translocation of p67phox to the phagosome was observed in GD macrophages (Figure 4AC). Moreover, immunostaining showed that after engulfing apoptotic Jurkat cells, p67phox was present and co-localized with Rab7 on late phagosomes in control, but not GD macrophages (Figure 4D). The nucleotide ATP is a strong “find-me” signal, and small amounts of ATP are released during early apoptosis. Extracellular ATP stimulates ROS production by activating NADPH oxidase.19,20 We, therefore, studied the translocation of p67phox to phagosomes in the presence of ATP by costaining for p67phox and Rab5 or Rab7 at different time points. A significant increase in translocation of p67phox to the early phagosomes (Figure 5A) was observed in the presence of ATP. Recruitment of p67phox to late phagosomes containing coated beads (co-stained with Rab7) increased significantly in control, but not GD macrophages (Figure 5B). Figure 5C also shows that in the presence of ATP, cytosolic p67phox protein levels increased notably in control macrophages, while only slightly increased levels were observed in GD macrophages. It has been reported that efferocytosis leads to increased ROS production in macrophages.21 Intercellular ROS was measured using a fluorescent redox sensitive dye. Hydrogen peroxide levels increased marginally in control macrophages after engulfing apoptotic Jurkat cells, and increased significantly after the addition of ATP. This elevation was more profound when macrophages were fed with

C

A

B

D

haematologica | 2017; 102(4)

ghosts prior to efferocytosis or in the presence of ATP. However, ROS production did not change in GD macrophages (Figure 6). Thus, in GD macrophages the failure of p67phox to translocate to the phagosome after phagocytosing PtdSer coated-beads, and in response to ATP, led to less ROS production. These data suggest that digestion of apoptotic cells is impaired in GD macrophages.

Discussion The most common feature of GD is splenomegaly, which at times can be massive. It has been shown that the amount of glucosylceramide accumulation in this organ does not account for the extensive degree of enlargement, and thus other factors must be involved. In this study we show that macrophages from patients with GD manifest impaired efferocytosis (clearance of apoptotic cells), as reflected by delayed phagosome-lysosome fusion. Impaired clearance of apoptotic cells can result in splenomegaly due to an imbalance between apoptosis and phagocytosis.22 Apoptosis is vital for the removal of problematic cells. Despite the constant turnover of cells through apoptosis, apoptotic cells are rarely seen under physiological conditions. This suggests that at steady state, the rate of removal of apoptotic cells is high, and the dying cells are adequately removed by tissue-resident professional phagocytes such as macrophages.23 Apoptotic cells recruit

Figure 3. Rab7 and p67phox are reduced in Gaucher macrophages. (A, B) Control and Gaucher macrophages were co-cultured with apoptotic Jurkat cells (AJ). Macrophages were lysed after engulfing the apoptotic Jurkat cells for (A) 20, 30 min or (B) 60 min in the presence and absence of erythrocyte ghosts. Lysates were probed for Rab7 and p67phox. In (B) the upper blot shows cells from a patient with genotype N370S/N370S and the lower blot cells from a patient with genotype N370S/c.84dupG. (C) Quantification of band intensity 60 min after efferocytosis. Data represent the average ± standard deviation from three individuals with genotype N370S/N370S and two with N370S/L444P. (D) GFPlabeled apoptotic Jurkat cells (AJ) or green microbeadsPtdSer were added to control and Gaucher macrophages for 60 min, washed, fixed, and stained for p67phox (red), F-actin (purple), apoptotic Jurkat cells (AJ) or PtdSer-coated microsphere beads (green) and DAPI (blue). Z-stack images were acquired using a Zeiss 510 confocal microscope. Scale bars, 5 μm. Images represent five different independent experiments using N370S/ N370S macrophages.

661


E. Aflaki et al.

macrophages by releasing chemotactic factors, and the nucleotide ATP serves as a key mediator for this recruitment. This process requires caspase-mediated activation of pannexin 1 channels to release ATP from apoptotic cells.24 Subsequently, nucleotides are detected by purogenic receptors (such as P2Y2) on monocytes and macrophages.25 The process of efferocytosis depends on recognition of apoptotic cells by phagocytes. Phagocytes target PtdSer, a ubiquitous hallmark of apoptosis independent of the cell type and the form of cell death. PtdSer can mediate tethering of apoptotic cells to phagocytes, delivering a â&#x20AC;&#x153;tickleâ&#x20AC;? signal to phagocytes to stimulate the internalization of apoptotic cells by engaging different receptors.26 However, no significant differences were observed in the ability of control and GD macrophages to engulf apoptotic Jurkat cells or microspheres coated with PtdSer (Figure 2). While recognition and engulfment of apoptotic cells have been extensively studied, little is known about how

apoptotic cells are degraded. Since the efficiency of engulfment appeared intact, we studied phagosome maturation in Gaucher macrophages, both in the presence and absence of erythrocyte ghosts. After engulfing apoptotic cells, nascent phagosomes fuse with early endosomes, late endosomes and ultimately lysosomes, the terminal degradative compartment.7 These fusion events change the composition of the phagosome membrane and lumen, load digestive enzymes and trigger phagosome acidification.27 Rab GTPases target and tether specific organelles to phagosomes.28 Rab5 is essential for tethering the surface of early endosomes to nascent phagosomes29 and Rab7, located on late endosomes, mediates the tethering and fusion between late endosomes and lysosomes.7,30 It has been shown that when Rab7 is absent, apoptotic Jurkat cells are engulfed normally, but are not degraded.15 Our study shows that recruitment of Rab7 to late phagosomes containing glass microsphere beads coated with PtdSer was

A

B

C

662

D

Figure 4. p67phox recruitment to phagosomes is diminished in Gaucher macrophages. (A) Control and Gaucher macrophages were co-cultured with PtdSer-coated beads and fixed at different time points. Cells were co-stained with p67phox (red) and Rab5 (green) or Rab7 (green). DAPI (blue) staining demonstrates the position of the nucleus. Images are representative of 25 pictures taken at each time point in four different individuals with genotype N370S/N370S, two with N370S/L444P mutations and six controls. (B, C) The intensity of each protein in the phagosome was measured using IMARIS software. (D) GFP-labeled apoptotic Jurkat cells were added to control and Gaucher macrophages for 60 min, washed, fixed and stained for p67phox (white), Rab7 (red) and apoptotic Jurkat cells (green). Images represent 40 pictures taken in four independent experiments (63X magnification, scale bars: 5 Îźm). Insets show higher magnifications of the areas outlined in the images.

haematologica | 2017; 102(4)


Impaired efferocytosis in Gaucher disease

impaired significantly in GD macrophages (Figure 2A-D). Moreover, while the translocation of cathepsin D to the phagosome increased gradually in control macrophages 40 min after efferocytosis, it was significantly delayed in GD macrophages (Figure 2B,F). When GD macrophages were fed with erythrocyte ghosts prior to efferocytosis, Rab7 levels remained low (Figure 3). These findings indicate impaired lysosome-phagosome fusion in GD macrophages, resulting in delayed degradation of phago-

cytes. A clinical observation substantiating the inefficient efferocytosis observed31 is the finding that many patients with GD exhibit autoantibodies reactive to autoantigens. Testing for autoantigens in sera from 43 patients with type 1 GD revealed autoreactivity in all but ten patients. Eleven of the subjects reacted to eight to 12 antigens, although the antibodies were not generally associated with immune manifestations.32 Other studies have found antiphospho-

A

B

Figure 5. Impaired p67phox translocation to the phagosome in the presence of ATP in Gaucher macrophages. (A, B) In the presence of ATP (5 mM for 30 min), PtdSercoated glass beads were added to control and Gaucher macrophages for different time intervals. Cells were washed, fixed and stained with p67phox (red), Rab5 or Rab7 (green) and DAPI (nuclear stain, blue), and imaged by confocal microscopy. Scale bars, 5 Îźm. Images represent 15 pictures at each time point from three independent experiments. (C) Cytosolic fractions from Gaucher macrophages (N370S/N370S) and control macrophages stained for p67phox in the presence and absence of ATP (5 mM for 30 min). The graph presents the results of three independent experiments from patients with genotype N370S/N370S.

C

A

B

Figure 6. The respiratory burst is impaired in Gaucher macrophages. (A, B). The production of reactive oxygen species (ROS) production was measured in Gaucher macrophages [(A) genotype N370S/N370S or (B) N370S/L444P] after engulfing apoptotic Jurkat cells (AJ) in the presence or absence of erythrocyte ghosts. In addition, 5 mM ATP was added to serum-free media (after centrifugation to remove media from the apoptotic Jurkat cells) in the presence or absence of added apoptotic Jurkat cells.

haematologica | 2017; 102(4)

663


E. Aflaki et al.

lipid, lupus anticoagulant and anticardiolipid in patients with GD. Impaired efferecytosis by GD macrophages could result in undigested material spilling into the blood, where they induce autoantibody formation.33,34 We previously demonstrated that GD macrophages manifest the M1 (activated macrophage) phenotype, as indicated by the secretion of inflammatory cytokines including interleukin-1β and interleukin-6 in the presence of the TLR4 ligand lipopolysaccharide.2 In fully differentiated M1 macrophages treated with lipopolysaccharide, the kinetics of phagosome maturation are delayed,35 while in M2 macrophages (alternatively activated macrophages), early phagosomes undergo a rapid transition to phagolysosomes with enhanced proteolytic activity.36 There is evidence that the differential rates of maturation are influenced by the luminal pH of the phagosome. Phagosome acidification is mainly mediated by VATPases, which translocate protons from the cytosol into the lumen of phagosomes and lysosomes. In general, M1 macrophages display a slower rate of acidification, and have a less acidic phagosomal pH,37 while in M2 macrophages, phagosomes are acidified immediately in order to clear apoptotic bodies rapidly and effectively.36 Once the recycling proteins have been removed, macrophages recruit the NADPH-oxidase complex to ensure the acidification of phagosomes and the killing of the phagocyte products through the production of oxygen

References 1. Lee RE. The fine structure of the cerebroside occurring in Gaucher's disease. Proc Natl Acad Sci USA. 1968;61(2):484-489. 2. Aflaki E, Moaven N, Borger DK, et al. Lysosomal storage and impaired autophagy lead to inflammasome activation in Gaucher macrophages. Aging Cell. 2016;15 (1):77-88. 3. Gardai SJ, McPhillips KA, Frasch SC, et al. Cell-surface calreticulin initiates clearance of viable or apoptotic cells through trans-activation of LRP on the phagocyte. Cell. 2005;123(2):321-334. 4. Franz S, Gaipl US, Munoz LE, et al. Apoptosis and autoimmunity: when apoptotic cells break their silence. Curr Rheumatol Rep. 2006;8(4):245-247. 5. Greenberg ME, Sun M, Zhang R, Febbraio M, Silverstein R, Hazen SL. Oxidized phosphatidylserine-CD36 interactions play an essential role in macrophage-dependent phagocytosis of apoptotic cells. J Exp Med. 2006;203(12):2613-2625. 6. Fadok VA, Henson PM. Apoptosis: giving phosphatidylserine recognition an assist-with a twist. Curr Biol. 2003;13(16):R655657. 7. Desjardins M, Huber LA, Parton RG, Griffiths G. Biogenesis of phagolysosomes proceeds through a sequential series of interactions with the endocytic apparatus. J Cell Biol. 1994;124(5):677-688. 8. Vieira OV, Botelho RJ, Grinstein S. Phagosome maturation: aging gracefully. Biochem J. 2002;366(Pt 3):689-704. 9. Serhan CN, Savill J. Resolution of inflammation: the beginning programs the end. Nat Immunol. 2005;6(12):1191-1197.

664

radicals.38 The production of superoxide by NOX2 in the lumen of phagosomes affects phagosomal pH, as protons are used when superoxide dismutates into hydrogen peroxide. NOX2 assembly and expression increase significantly in M1 macrophages treated with lipopolysaccharide, which allows rapid and robust production of ROS in response to inflammatory stimuli,39 while M2 macrophages show reduced NOX2 expression.36 Although ROS production by NOX2 limits acidification of phagosomes in dendritic cells, it does not affect phagosome acidification in macrophages, due to extensive recruitment of V-ATPases and limited ROS production.40 We found that p67phox is significantly reduced in GD macrophages, even with the addition of erythrocytes ghosts prior to efferocytosis (Figure 4). Moreover, in GD macrophages this subunit fails to translocate to the phagosome, which leads to impaired NOX2 assembly (Figures 4 and 5) and reduced ROS production (Figure 6). Together these findings confirm impaired efferocytosis and degradation of apoptotic cells in GD macrophages. These aspects of the disease pathogenesis may contribute to inflammation and splenomegaly in patients with GD. Acknowledgments This work was supported by the Intramural Research Programs of the National Human Genome Research Institute and the National Institutes of Health.

10. Marodi L, Kaposzta R, Toth J, Laszlo A. Impaired microbicidal capacity of mononuclear phagocytes from patients with type I Gaucher disease: partial correction by enzyme replacement therapy. Blood. 1995;86(12):4645-4649. 11. Stone DL, Tayebi N, Orvisky E, Stubblefield B, Madike V, Sidransky E. Glucocerebrosidase gene mutations in patients with type 2 Gaucher disease. Hum Mutat. 2000;15(2):181-188. 12. Aflaki E, Stubblefield BK, Maniwang E, et al. Macrophage models of Gaucher disease for evaluating disease pathogenesis and candidate drugs. Sci Transl Med. 2014;6(240): 240ra73. 13. Buranda T. Biomimetic molecular assemblies on glass and mesoporous silica microbeads for biotechnology. Langmuir. 2003;19 (5):1654â&#x20AC;&#x201C;1663. 14. Kitano M, Nakaya M, Nakamura T, Nagata S, Matsuda M. Imaging of Rab5 activity identifies essential regulators for phagosome maturation. Nature. 2008;453(7192):241245. 15. Kinchen JM, Doukoumetzidis K, Almendinger J, et al. A pathway for phagosome maturation during engulfment of apoptotic cells. Nat Cell Biol. 2008;10(5): 556-566. 16. Harrison RE, Bucci C, Vieira OV, Schroer TA, Grinstein S. Phagosomes fuse with late endosomes and/or lysosomes by extension of membrane protrusions along microtubules: role of Rab7 and RILP. Mol Cell Biol. 2003;23(18):6494-6506. 17. Yu X, Lu N, Zhou Z. Phagocytic receptor CED-1 initiates a signaling pathway for degrading engulfed apoptotic cells. PLoS Biol. 2008;6(3):e61. 18. Vergnaud S, Paclet MH, El Benna J, Pocidalo

19.

20.

21.

22.

23.

24.

25.

26.

MA, Morel F. Complementation of NADPH oxidase in p67-phox-deficient CGD patients p67-phox/p40-phox interaction. Eur J Biochem. 2000;267(4):1059-1067. Cruz CM, Rinna A, Forman HJ, Ventura AL, Persechini PM, Ojcius DM. ATP activates a reactive oxygen species-dependent oxidative stress response and secretion of proinflammatory cytokines in macrophages. J Biol Chem. 2007;282(5):2871-2879. Moore SF, MacKenzie AB. NADPH oxidase NOX2 mediates rapid cellular oxidation following ATP stimulation of endotoxinprimed macrophages. J Immunol. 2009; 183(5):3302-3308. Yvan-Charvet L, Pagler TA, Seimon TA, et al. ABCA1 and ABCG1 protect against oxidative stress-induced macrophage apoptosis during efferocytosis. Circ Res. 2010;106(12):1861-1869. Yang A, Dai J, Xie Z, et al. High molecular weight kininogen binds phosphatidylserine and opsonizes urokinase plasminogen activator receptor-mediated efferocytosis. J Immunol. 2014;192(9):4398-4408. Ravichandran KS. Find-me and eat-me signals in apoptotic cell clearance: progress and conundrums. J Exp Med. 2010;207(9):18071817. Chekeni FB, Elliott MR, Sandilos JK, et al. Pannexin 1 channels mediate 'find-me' signal release and membrane permeability during apoptosis. Nature. 2010;467(7317):863867. Elliott MR, Chekeni FB, Trampont PC, et al. Nucleotides released by apoptotic cells act as a find-me signal to promote phagocytic clearance. Nature. 2009;461(7261):282-286. Somersan S, Bhardwaj N. Tethering and tickling: a new role for the phosphatidylserine receptor. J Cell Biol. 2001;155(4):501-504.

haematologica | 2017; 102(4)


Impaired efferocytosis in Gaucher disease

27. Zhou Z, Yu X. Phagosome maturation during the removal of apoptotic cells: receptors lead the way. Trends Cell Biol. 2008;18(10): 474-485. 28. Zerial M, McBride H. Rab proteins as membrane organizers. Nat Rev Mol Cell Biol. 2001;2(2):107-117. 29. Duclos S, Diez R, Garin J, et al. Rab5 regulates the kiss and run fusion between phagosomes and endosomes and the acquisition of phagosome leishmanicidal properties in RAW 264.7 macrophages. J Cell Sci. 2000;113(Pt 19):3531-3541. 30. Vieira OV, Bucci C, Harrison RE, et al. Modulation of Rab5 and Rab7 recruitment to phagosomes by phosphatidylinositol 3kinase. Mol Cell Biol. 2003;23(7):2501-2514. 31. Machaczka M, Klimkowska M, Regenthal S, Hagglund H. Gaucher disease with foamy transformed macrophages and erythrophagocytic activity. J Inherit Metab Dis. 2011;34(1):233-235. 32. Shoenfeld Y, Beresovski A, Zharhary D, et al. Natural autoantibodies in sera of patients

haematologica | 2017; 102(4)

33.

34.

35.

36.

37.

with Gaucher's disease. J Clin Immunol. 1995;15(6):363-372. Barone R, Giuffrida G, Musso R, Carpinteri G, Fiumara A. Haemostatic abnormalities and lupus anticoagulant activity in patients with Gaucher disease type I. J Inherit Metab Dis. 2000;23(4):387-390. Granel B, Serratrice J, Swiader L, et al. [Antiphospholipid antibodies and Gaucher's disease. A case report]. Rev Med Interne. 2002;23(12):1037-1039. Yates RM, Hermetter A, Taylor GA, Russell DG. Macrophage activation downregulates the degradative capacity of the phagosome. Traffic. 2007;8(3):241-250. Balce DR, Li B, Allan ER, Rybicka JM, Krohn RM, Yates RM. Alternative activation of macrophages by IL-4 enhances the proteolytic capacity of their phagosomes through synergistic mechanisms. Blood. 2011;118 (15):4199-4208. Ghigo E, Capo C, Tung CH, Raoult D, Gorvel JP, Mege JL. Coxiella burnetii survival in THP-1 monocytes involves the

impairment of phagosome maturation: IFN-gamma mediates its restoration and bacterial killing. J Immunol. 2002;169(8): 4488-4495. 38. Zhan Y, Virbasius JV, Song X, Pomerleau DP, Zhou GW. The p40phox and p47phox PX domains of NADPH oxidase target cell membranes via direct and indirect recruitment by phosphoinositides. J Biol Chem. 2002;277(6):4512-4518. 39. Amezaga MA, Bazzoni F, Sorio C, Rossi F, Cassatella MA. Evidence for the involvement of distinct signal transduction pathways in the regulation of constitutive and interferon gamma-dependent gene expression of NADPH oxidase components (gp91phox, p47-phox, and p22-phox) and highaffinity receptor for IgG (Fc gamma R-I) in human polymorphonuclear leukocytes. Blood. 1992;79(3):735-744. 40. Savina A, Jancic C, Hugues S, et al. NOX2 controls phagosomal pH to regulate antigen processing during crosspresentation by dendritic cells. Cell. 2006;126(1):205-218.

665


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Red Cell Biology & its Disorders

Ferrata Storti Foundation

Associations between environmental factors and hospital admissions for sickle cell disease Frédéric B. Piel,1,2 Sanjay Tewari,3 Valentine Brousse,4 Antonis Analitis,5 Anna Font,6 Stephan Menzel,3 Subarna Chakravorty,3 Swee Lay Thein,3,7 Baba Inusa,8 Paul Telfer,9 Mariane de Montalembert,4 Gary W. Fuller,6 Klea Katsouyanni5,6 and David C. Rees3

Department of Epidemiology and Biostatistics, MRC-PHE Centre for Environment and Health, School of Public Health, Imperial College London, UK; 2Department of Zoology, University of Oxford, UK; 3Department of Molecular Haematology, King’s College London School of Medicine, King’s College Hospital, UK; 4Reference Centre for Sickle-Cell Disease, Pediatric Department, Hôpital Universitaire Necker-Enfants Malades, APHP, Paris, Université Paris Descartes, France; 5Department of Hygiene, Epidemiology and Medical Statistics, Medical School, National and Kapodistrian University of Athens, Greece; 6Environmental Research Group, MRC-PHE Centre for Environment and Health, King's College London, UK; 7National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD, USA; 8Department of Paediatric Haematology, Evelina Children's Hospital, King's College London, UK and 9Department of Paediatric Haematology and Oncology, Barts Health NHS Trust, Royal London Hospital, UK 1

Haematologica 2017 Volume 102(4):666-675

ABSTRACT

S

Correspondence: f.piel@imperial.ac.uk

Received: August 8, 2016. Accepted: November 25, 2016. Pre-published: December 1, 2016. doi:10.3324/haematol.2016.154245 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/666 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

666

ickle cell disease is an increasing global health burden. This inherited disease is characterized by a remarkable phenotypic heterogeneity, which can only partly be explained by genetic factors. Environmental factors are likely to play an important role but studies of their impact on disease severity are limited and their results are often inconsistent. This study investigated associations between a range of environmental factors and hospital admissions of young patients with sickle cell disease in London and in Paris between 2008 and 2012. Specific analyses were conducted for subgroups of patients with different genotypes and for the main reasons for admissions. Generalized additive models and distributed lag non-linear models were used to assess the magnitude of the associations and to calculate relative risks. Some environmental factors significantly influence the numbers of hospital admissions of children with sickle cell disease, although the associations identified are complicated. Our study suggests that meteorological factors are more likely to be associated with hospital admissions for sickle cell disease than air pollutants. It confirms previous reports of risks associated with wind speed (risk ratio: 1.06/standard deviation; 95% confidence interval: 1.00-1.12) and also with rainfall (1.06/standard deviation; 95% confidence interval: 1.01-1.12). Maximum atmospheric pressure was found to be a protective factor (0.93/standard deviation; 95% confidence interval: 0.88-0.99). Weak or no associations were found with temperature. Divergent associations were identified for different genotypes or reasons for admissions, which could partly explain the lack of consistency in earlier studies. Advice to patients with sickle cell disease usually includes avoiding a range of environmental conditions that are believed to trigger acute complications, including extreme temperatures and high altitudes. Scientific evidence to support such advice is limited and sometimes confusing. This study shows that environmental factors do explain some of the variations in rates of admission to hospital with acute symptoms in sickle cell disease, but the associations are complex, and likely to be specific to different environments and the individual’s exposure to them. Furthermore, this study highlights the need for prospective studies with large numbers of patients and standardized protocols across Europe.

haematologica | 2017; 102(4)


Enviromental factors and admissions for SCD

Introduction The clinical severity of sickle cell disease (SCD) is extremely variable.1 Genetic and genome-wide association studies have so far only explained a small fraction of this phenotypic variability.2-4 Investigations of the impact of environmental factors, including meteorological factors and air quality, on the severity of the disease conducted across a range of countries have provided inconsistent results partly because of: (i) the use of potentially inaccurate coded data (e.g. International Classification of Disease 10) rather than specific hospital records; (ii) the intricate relationships between weather and air quality exposure variables; and (iii) the use of different modeling approaches to assess such interactions.5-9 Furthermore, the impact of environmental factors on different types of SCD (HbSS versus HbSC) and on the specific clinical complications leading to hospital admissions has not been previously reported (i.e. all genotypes and clinical complications have typically been lumped together). The costs of care for SCD patients are high and increasing.10 For the year 2010–2011, it was estimated that the total costs of hospital admissions for a SCD crisis (as a primary diagnosis) added up to more than £18,000,000 in England.11 In London, the highest hospital admission rates are seen among males in their forties, a demographic group in which rates increased from 7.6 to 26.8 per 100,000 between 2001 and 2009.12 The vast majority of patients with SCD in the UK and in France live in capital cities (68% in London, 70% in the Paris area).13 Identifying environmental factors triggering clinical complications in urban settings could therefore lead to better patient care, which could result in improved quality of life for patients with SCD and their relatives, as well as in reductions in hospital admissions and healthcare costs. We investigated the associations between weather, air quality, and daily hospital admissions for pain, fever and acute chest syndrome (ACS) of young patients known to have SCD, over a 5-year period in London and Paris using generalized additive models (GAM) and distributed lag non-linear models (DLNM), adjusted for long-term trends and day of the week. We then compared our results with those of previous studies and discuss the direct impact that these results could have on the prevention of hospital admissions for SCD.

Methods Data sources We extracted anonymized daily hospital admission records from 1st January, 2008 to 31st December, 2012 for patients with SCD under the age of 18 years old at the time of admission living within a radius of ten kilometers from each of the following hospitals: King’s College Hospital (Camberwell), Evelina Children’s Hospital (Lambeth) and Royal London Hospital (Whitechapel) in London; and the Necker Hospital for Sick Children (15th arrondissement) in Paris (Online Supplementary Figure S1). Recorded reasons for hospital admissions were pain, fever, ACS and other. Information on the patients’ genotype, either HbSS or HbSC, was available for individuals admitted to the three hospitals in London, but not the one in Paris. Outcome data were collected by inspection of specific databases of SCD patients and admissions at each hospital to optimize accuracy. Too few admissions of patients with HbS β-thalassemia, a third common form of SCD, were haematologica | 2017; 102(4)

available to be included in the study. Official meteorological data on daily rainfall (mm); air temperature (°C), relative humidity (%), wind speed (m/s) and atmospheric pressure (hPa) were extracted from the British Atmospheric Data Centre (BADC, http://badc.nerc.ac.uk/ view/badc.nerc.ac.uk__ATOM__dataent_ukmo-midas) for several monitoring stations, including Heathrow (51°28’13”N, 0°27’02”W, code 708, the reference station in London) and St. James Park (51°30’17”N, 0°07’52”W, code 697, the nearest station to the three hospitals). Measurements were highly consistent across different monitoring stations and, as a result, only data from Heathrow were included in the final analyses. Based on these preliminary analyses (not shown), data were purchased from Météo France only for one meteorological station, Paris Montsouris (48°49'18"N, 2°20'12"E). A composite index of temperature and relative humidity was calculated as a measure of apparent (or “feels like”) temperature using the following equations:14,15 Equation 1: DT ≈ T – Equation 2: AT = – 2.653 +(0.994×T)+(0.0153×(DT2) where DT is the dew point temperature in °C; RH is the relative humidity; AT is the apparent temperature in °C; and T is the ambient temperature in °C. Lawrence’s simple approximation is fairly accurate for relative humidity values above 50%, which match conditions in London and Paris. In addition, a “wind chill” index was also included as a composite index of temperature and wind speed:16 Equation 3: Twc=13.12+0.6215×Ta-11.37×V0.16+0.3965×Ta×V0.16 where Twc is the wind chill index; Ta is the air temperature in °C; and V is the wind speed at standard anemometer height (10 meters), in km/h. Daily mean concentrations of carbon monoxide (CO, mg/m3), nitrogen oxides (NOX = NO + NO2; μg/m3), sulfur dioxide (SO2, μg/m3), and ozone (O3, μg/m3), particle matter in two size ranges (<10 μm or PM10; and <2.5 μm, both expressed in μg/m3), black carbon (μg/m3) and particle number (N/cm3) were extracted from the London Air Quality Network (http://www.londonair.org.uk/); the DEFRA Black Carbon (https://uk-air.defra.gov.uk/networks/ network-info?view=ukbsn) and the DEFRA Particle Numbers and Concentrations Networks (https://uk-air.defra.gov.uk/networks/ network-info?view=particle) for London; and from the AirParif Network (http://www.airparif.asso.fr/en/) for Paris. Because not all the monitoring stations recorded all the above pollutants for the entire period of time of the study, some records are missing from the time-series. We therefore kept only data from the most complete monitoring stations (i.e. records available for at least 80% of days during the study period) and filled the gaps using an expectation–maximization imputation algorithm for multivariate normal time-series implemented in the mnimput function of the mtsdi R package. Missing values were therefore estimated by accounting for both correlation between time-series (i.e. from other monitoring stations) and time structure of the series itself (Online Supplementary Code S1). Air pollutant concentrations were normalized using a log transformation. To assess the error in imputed values, cross-validation based on a left-out sample of 100 daily records was conducted and the root mean squared error and normalized root mean square error were calculated (Online Supplementary Table S1). Separate analyses were run for monitoring stations categorized as “background” and “roadside” sites in London in order to identify potential associations with specific pollution caused by traffic. Descriptive statistics of hospital admissions (outcome) and envi667


F.B. Piel et al.

ronmental variables (exposure) in each of the study settings are shown in Table 1 and Table 2, respectively. Standardized z-score meteorological and air pollution data (Equation 4) were used in the time-series analyses in order to generate relative risks per one standard deviation increase. Statistical differences between admission rates per year, season, month and day of the week were identified by analyses of variance with the Tukey honestly significant difference test. Equation 4: where x is the exposure record, μ is the mean of the exposure records over the study period, and σ is the standard deviation of the exposure records over the study period.

Data analysis We first explored the relationships between the different outcomes and standardized exposure variables using a quasi-Poisson GAM. We used flexible thin-plate regression splines with shrinkage for long-term trends, seasonality, effects of the year, month and day of the week, and tested for weekend effects. Secondly, we implemented two standard methods commonly used to assess the relationship between an exposure variable and a health outcome in time-series analyses: the DLNM and an aggregated case-crossover study. DLNM is a flexible modelling framework to describe potential associations with non-linear and delayed effects in time-series data.17 Aggregated case-crossover studies provide an efficient framework for evaluating associations

between transient exposures and the onset of rare acute events, when exposure measurements are not available for each individual18 In a DLNM, seasonality, long-term trends and confounding by other time-varying factors (e.g. temperature) are typically corrected by fitting flexible spline functions of the different covariates. While delayed exposure effects can be explored for a specific lag, the DLNM offers the advantage of considering all lags under investigation together. While various maximum lags (up to 3 weeks) were tested, a lag of 1 week was considered the most relevant, biologically. The standard analysis of aggregated casecrossover studies is by conditional logistic regression on a timeseries dataset, in which each case day (a day with at least one hospital admission for SCD) is matched to all the other days within a given time window (e.g. 1 month). Relatively short time windows avoid long-term or seasonal effects, accounted for by strata, Fourier series or splines in DLNM. Various levels of constraints can be added by matching case and control days for a given covariate (e.g. temperature within 1°C) or a combination of covariates (e.g. temperature within 1°C and day of the week). While both methods have been previously used individually to assess environmental influences on SCD hospital admissions,8,9 the consistency of results between them has not been previously investigated. Thirdly, based on the results from single-exposure models, we explored multiple-exposure GAM for combined lags of 0 and 1. The different combinations of exposure variables used are shown in Online Supplementary Table S5. Finally, sensitivity analyses were performed throughout the whole model selection by: (i) exploring a full range of measure-

Table 1. Summary statistics of sickle cell disease admission data (outcome) in London and Paris between 1st January, 2008 and 31st December, 2012.

City

Hospital

Average age (years)

Male/female ratio

King’s College Hospital

8

1.76

Royal London Hospital

10

1.00

Evelina London Hospital

6

0.78

Total

8

1.19

Necker Hospital

8

1.09

Reason for admission

Number of admissions % SC

SCD

SS

471

439 283 62 47 47 403 258 77 25 43 510 318 129 32 29 1352 859 268 104 119

93% 60% 13% 10% 10% 91% 58% 17% 6% 10% 91% 57% 23% 6% 5% 92% 58% 18% 7% 8%

347 201 51 12 83

100% 58% 15% 3% 24%

%

London All Pain Fever ACS Other All Pain Fever ACS Other All Pain Fever ACS Other All Pain Fever ACS Other

445

558

1474

32 16 6 3 7 42 30 12 0 0 48 29 14 1 4 122 75 32 4 11

7% 3% 1% 1% 1% 9% 7% 3% 0% 0% 9% 5% 3% 0% 1% 8% 5% 2% 0% 1%

Paris

668

All Pain Fever ACS Other

347

haematologica | 2017; 102(4)


Enviromental factors and admissions for SCD

ments (e.g. NO, NOX, NO2 in turn) from several individual monitoring stations and average values; (ii) checking the consistency of the results across different methods; and (iii) selecting the best-performing model based on objective criteria (generalized cross-validation, Bayesian information criteria or Akaike information criterion). All the analyses were performed with R 3.2.4 and full scripts of the code used are provided in the Online Supplementary Material and are available on request. The study was discussed with the local research ethics committees, and formal ethical approval was not deemed necessary. All analyzed data were fully anonymized. The research was conducted in accordance with the Helsinki Declaration of 1975, as revised in 2008.

Results Over the 5-year study period, 1,887 and 346 hospital admissions for SCD (HbSS and HbSC only) were recorded in London and Paris, respectively. The proportions of HbSS and HbSC in London, and the reasons for admissions in London and Paris are shown in Table 1. The average daily number of SCD admissions was 1.03 in London (across the 3 hospitals) and 0.19 in Paris (1 hospital), with maximums of five and four per day, respectively. Although individual patients’ data were not analyzed as part of this study, it is worth noting that some patients may have been admitted several times over the study period. Average daily hospital admissions of SCD patients revealed differences in temporal patterns (yearly, monthly, daily and per season) between cities, genotypes and reasons for admissions (Table 3 and Figure 3). Average daily hospital admissions in London increased significantly from 0.52/day in 2008 to 0.98/day in 2009 (ANOVA, P=0.001) before stabilizing around 0.85 admissions per day, probably reflecting increasing numbers of patients

over that time period. No statistically significant difference was observed between years in Paris. In London, admission rates were significantly higher in autumn than in spring (P=0.045), while in Paris, there were fewest admissions in summer (P=0.029). Hospital admissions, particularly for pain, were less frequent during weekends, in both London (P<0.0001) and Paris (P=0.042). In London, there were significantly more admissions of SCD patients on Mondays than on Saturdays (P<0.0001), although the peak of admissions of HbSS patients for pain was observed on Tuesday. GAM testing for associations between environmental factors and hospital admissions for SCD in London revealed relative risks per one standard deviation increase of 1.06 with a 95% confidence interval (CI) of 1.01-1.12 for rainfall, and 0.93 (95% CI: 0.88-0.99) for maximum atmospheric pressure (Figure 1 and Online Supplementary Table S2). Specific GAM looking at genotypes and reasons for admissions suggested that the former association was mostly seen in HbSS patients admitted for pain (1.07; 95% CI: 1.01-1.14), whereas the latter association was strongest in SCD patients with fever (0.84; 95% CI: 0.750.95). Further associations were found between HbSS patients with pain and maximum wind speed (1.09; 95% CI: 1.02-1.16); with fever and CO (1.14; 95% CI: 1.011.30), and with other complications and PM2.5 (1.22; 95% CI: 1.02-1.46). Similar results were obtained when looking only at “background” or “roadside” monitoring stations (results not shown). No specific associations were identified for HbSC patients, possibly due to their smaller numbers. In Paris, we found a relative risk of 0.75 (95% CI: 0.57-0.99) for patients with pain in relation to minimum temperature. DLNM, which account for lag effects, only supported associations with rainfall (1.06; 95% CI: 1.01-1.12) at lag 0 in London (Figure 2 and Online Supplementary Table S3). An association with black carbon at lag 6 was also found

Table 2. Summary statistics of the meteorological and air quality parameters (exposure) in London and Paris between 1st January, 2008 and 31st December, 2012.

Rainfall (mm) Maximum temperature (˚C) Minimum temperature (˚C) Maximum wind speed (m/s) Maximum pressure (hPa) Maximum relative humidity (%) CO (μg/m3) NO2 (μg/m3) NOX (μg/m3) O3 (μg/m3) SO2 (μg/m3) PM10 (μg/m3) PM2.5 (μg/m3) Black carbon (μg/m3) Particle number (N/cm3)

London Minimum

Mean

SD

1.69 14.91 7.72 21.26 1,017.89 91.33 0.37 56.18 126.31 31.57 3.47 28.62 16.65 5.57 23,919.06

3.64 6.42 5.19 6.89 9.74 5.95 0.12 16.81 56.49 16.13 1.89 10.77 10.90 2.25 7,457.71

0.00 -0.70 -9.40 5.00 984.60 64.10 0.17 16.52 24.41 1.09 0.52 9.35 3.29 0.91 6,890.75

Paris Maximum

Mean

31.80 31.00 19.70 51.00 1,043.60 100.00 1.58 122.32 537.97 92.35 30.35 96.38 84.76 16.08 60,832.66

1.60 16.34 9.03 10.71 1,016.30 85.92 0.62 35.75 68.33 45.15 / 30.14 20.35 / /

SD

Minimum

Maximum

3.68 7.74 5.79 3.64 9.14 8.45 0.18 12.98 36.85 19.76 / 13.22 12.05 / /

0.00 -3.90 -8.90 3.10 974.60 49.00 0.25 8.46 18.24 1.33 / 7.13 3.92 / /

43.10 38.40 22.60 33.80 1,041.50 99.00 1.92 99.99 387.95 134.52 / 131.87 119.82 / /

CO: carbon monoxide; NO2: nitrogen dioxide; NOx: nitrogen oxides; SO2: sulfur dioxide; PM: particulate matter.

haematologica | 2017; 102(4)

669


F.B. Piel et al.

(1.08; 95% CI: 1.02-1.16). A similar effect of wind speed was found in Paris for a lag of 3 days (1.08; 95% CI: 1.021.13). In addition, an association was found with CO at lag 6 (1.14; 95% CI: 1.00-1.29). For HbSC, significant associations with maximum pressure (0.66; 95% CI: 0.52-0.83 at lag 0) and maximum relative humidity (0.91; 95% CI: 0.84-1.00 at lag 3; 0.88; 95% CI: 0.79-0.99 at lag 4) were found (Online Supplementary Figure S2). Statistically significant associations often differed when comparing the main reasons for hospital admission (Online Supplementary Figure S3). For example, maximum temperature was a risk factor at lags 1 and 2 for ACS but not for fever or pain, while maximum pressure appeared protective at lag 0 for pain and fever but not for ACS. In London, the results of multiple-exposure GAM support an association between admissions for pain for patients with HbSS, and rainfall and maximum wind speed, while maximum pressure appeared protective for HbSS patients with fever (Online Supplementary Table S5).

No statistically significant associations were found in multiple-exposure analyses for Paris. These results were consistent with the findings of single-exposure GAM. A summary of the convergence and divergence of associations identified in London and Paris is presented in Table 4.

Discussion A better understanding of the environmental factors triggering clinical complications in patients with SCD could allow healthcare professionals to give more accurate information to patients about the risks associated with certain conditions, facilitating behavioral changes to avoid clinical complications and hospital admissions. Evidence generated so far about the influence of meteorological factors and pollutants on symptoms in SCD has often provided discordant results, which are difficult to translate into health policies and advice for patients. This is partly because previous studies were mostly small, combining

Table 3. Effects of day of the week, weekend, season and year on admissions for sickle cell disease in Paris and London between 1st January, 2008 and 31st December, 2012, based on analyses of variance. Minimum and maximum values are highlighted in green and red, respectively, for columns in which a statistically significant difference was found (P<0.05).

N

All

Average daily admissions London SS Pain Fever ACS Other All

SC Pain

Fever

N

All

All

Paris SS Pain

Fever Other

Day of the week Monday Tuesday Wednesday Thursday Friday Saturday Sunday P-value

253 245 201 210 226 163 174

0.969 0.939 0.770 0.805 0.866 0.625 0.667 <0.0001

0.881 0.851 0.709 0.751 0.797 0.571 0.613 <0.0001

0.556 0.598 0.467 0.452 0.460 0.356 0.402 0.001

0.188 0.138 0.107 0.149 0.192 0.123 0.130 0.116

0.050 0.065 0.069 0.050 0.069 0.061 0.034 0.644

0.088 0.050 0.065 0.100 0.077 0.031 0.046 0.026

0.088 0.088 0.061 0.054 0.069 0.054 0.054 0.508

0.556 0.598 0.467 0.452 0.460 0.356 0.402 0.922

0.188 0.138 0.107 0.149 0.192 0.123 0.130 0.354

57 61 52 54 42 36 45

0.218 0.234 0.199 0.207 0.161 0.138 0.172 0.202

0.123 0.146 0.115 0.130 0.088 0.077 0.092 0.227

0.034 0.031 0.027 0.019 0.027 0.031 0.027 0.978

0.061 0.046 0.050 0.054 0.038 0.023 0.046 0.574

Weekend Working days Weekend P-value

1135 337

0.870 0.646 <0.0001

0.798 0.507 0.592 0.379 <0.0001 0.001

0.155 0.126 0.169

0.061 0.048 0.320

0.076 0.038 0.005

0.072 0.054 0.186

0.507 0.379 0.717

0.155 0.126 0.056

266 0.204 81 0.155 0.042

0.120 0.084 0.046

0.028 0.029 0.899

0.050 0.034 0.181

Season Spring Summer Autumn Winter P-value

328 360 402 382

0.713 0.783 0.884 0.845 0.045

0.659 0.713 0.811 0.774 0.062

0.450 0.404 0.488 0.540 0.029

0.126 0.172 0.171 0.117 0.063

0.043 0.048 0.066 0.071 0.254

0.039 0.089 0.086 0.046 0.003

0.054 0.070 0.073 0.071 0.719

0.450 0.404 0.488 0.540 0.812

0.126 0.172 0.171 0.117 0.017

82 69 109 87

0.178 0.150 0.240 0.192 0.029

0.117 0.070 0.143 0.111 0.015

0.022 0.026 0.035 0.029 0.700

0.035 0.043 0.053 0.051 0.599

Year 2008 2009 2010 2011 2012 P-value

190 353 320 297 312

0.519 0.967 0.877 0.814 0.852 0.001

0.475 0.901 0.808 0.759 0.751 0.007

0.328 0.545 0.501 0.474 0.503 0.017

0.074 0.208 0.167 0.151 0.134 0.339

0.044 0.071 0.058 0.060 0.052 0.893

0.030 0.077 0.082 0.074 0.063 0.140

0.044 0.066 0.068 0.055 0.101 0.019

0.328 0.545 0.501 0.474 0.503 0.331

0.074 0.208 0.167 0.151 0.134 0.235

84 80 47 69 67

0.230 0.219 0.129 0.189 0.183 0.107

0.128 0.121 0.077 0.110 0.115 0.506

0.033 0.033 0.008 0.030 0.036 0.925

0.060 0.049 0.036 0.049 0.033 0.135

670

haematologica | 2017; 102(4)


Enviromental factors and admissions for SCD

reasons for admissions and not distinguishing between different types of SCD, in addition to the variability of climate effects in different countries. Perhaps the most consistently quoted effect is the increase in episodes of acute pain associated with cold weather. Using high-quality hospital records from London and Paris, combined with rigorous time-series analysis methods, our results do not support strong associations between hospital admissions for SCD and temperature. This might be related to the urban environment in high-income countries, in which the effects of temperature changes may be countered by access to warm clothes and heated buildings. Environmental factors that consistently appeared significant throughout our analyses were rainfall, wind speed and atmospheric pressure. Wind speed has been identified

in several previous studies in urban settings, and is emerging as one of the most important meteorological factors.13 Rainfall has not been consistently linked to increased hospital admissions, but emerges as an important factor, particularly precipitating pain in children with HbSS. Both high wind speed and rainfall have the effect of causing rapid skin cooling, which has been implicated as a cause of vaso-occlusive pain in physiological experiments,19 and might be the mechanism of action in this case. While standard composite indices used in this study (i.e. apparent temperature and wind chill) did not reveal statistically significant associations, the development of a novel, specific composite index allowing the risk of hospitalization of SCD patients to be predicted based on the above results would warrant further investigation.

A

B

Figure 1. Forest plots of associations between environmental factors and admissions for sickle cell disease. The plots show associations at lags 0 and 1 between environmental factors, including weather and air pollution, and hospital admissions for sickle cell disease in London and Paris, based on generalized additive models corrected for longterm trends and weekend effect. Panel A shows variables with statistically significant associations, while panel B shows those with non-statistically significant associations.

haematologica | 2017; 102(4)

671


F.B. Piel et al. A

B

Figure 2. Lag plots of relative risks of hospital admissions for sickle cell disease according to exposure to environmental factors. The lag plots of relative risks (RR) and 95% confidence intervals (CI) per standard deviation (SD) increase in 17 exposure variables (8 for meteorological conditions and 9 for air quality) are based on distributed lag non-linear models with all lags (0-7 days) modeled together using a polynomial constraint for sickle cell anemia (HbSS) admissions in London (red) and Paris (blue) between 1st January, 2008 and 31st December, 2012. Panel A shows variables with statistically significant associations, while panel B shows those with non-statistically significant associations. Statistically significant risks are shown in a brighter red or blue for London and Paris, respectively. Data on black carbon and particle number were not available for Paris.

672

haematologica | 2017; 102(4)


Enviromental factors and admissions for SCD

A

B

C

D

Figure 3. Average daily hospital admissions for sickle cell anemia. The average daily admissions are shown per (A) day of the week, (B) weekday/weekend, (C) season and (D) year between 1st January, 2008 and 31st December, 2012 in three hospitals in London (red) and one hospital in Paris (blue). The number of * indicates the level of statistical significance (***P<0.001, *P<0.05).

We also identified a clear weekend effect, in both London and Paris, which is relevant in the broader context of healthcare provision.20 Lower admission rates during weekends, particularly for pain, may arise from many different social and logistical issues, but are unlikely to be primarily related to environmental factors. Perhaps the most plausible explanation is that parents were able to stay at home and look after their children at the weekends, whereas this becomes much harder during the working week. It does suggest that improved community support for families with sick children may be effective at reducing hospital admissions. Although patients included in this study were managed by hematologists familiar with SCD, delaying seeking healthcare could also reflect the distress of facing common misconceptions (e.g. lack of tolerance to pain, drug addiction) previously reported among medical staff and of longer waiting times compared to other complications (e.g. long bone fracture) previously reported in emergency departments.21 To the best of our knowledge, this study is the largest to use accurate hospital-based registers of patients with SCD with data specifically collected for this study. Other large studies have relied on coded data generated for routine administrative purposes, which are often associated with misclassification errors. We also analyzed the different types of SCD separately, as there is considerable evidence that the pathophysiology of HbSS and HbSC disease is significantly different.22 Furthermore, we focused on young children to avoid a series of confounding factors involved at older ages (e.g. smoking, occupational exposure, comorbidities) and used rigorous statistical methods, which revealed mostly consistent results for the main associations identified. Conducting separate analyses for haematologica | 2017; 102(4)

each genotype and reason for admission revealed important differences, which could partly explain the inconsistency of previous results. Despite focusing on a 5-year period, the number of admissions in Paris and the number of admissions of HbSC patients in London remained relatively limited, which led to large confidence intervals, potentially masking some associations. Extreme temperatures are believed to trigger acute vasoocclusive complications in patients with SCD. This is reflected in advice given to patients, but we could not find consistent support for such an association in our study, although increasing minimum temperature was associated with a significant reduction in admissions for acute pain in Paris (relative risk 0.75; 95% CI: 0.57 â&#x20AC;&#x201C; 0.99) (Online Supplementary Table S2). Instead, we found significant associations with maximum wind speed, which have previously been reported for London.6 In contrast to an earlier, smaller study in London, we did not find significant associations between increased numbers of SCD admissions and low concentrations of NOx, low concentrations of CO and high concentrations of O3.7 A recent study of 17,710 emergency hospital admissions of SCD patients in Paris concluded that most weather conditions and air pollutants assessed were correlated to each other and influenced the rate of such admissions in SCD patients over a lag period of 1 week.8 CO concentrations, day-to-day mean temperature drop and higher wind speed were associated with increased risks in a multipleexposure analysis. Contrary to our study, the authors did not find a weekend effect, which might be due to their focus on emergency admissions. Despite using a much larger number of admissions, their data were partly based on International Classification of Disease codes, included 673


F.B. Piel et al. Table 4. Summary of the congruence and divergence of statistically significant associations between environmental factors and hospital admissions for sickle cell anemia (SS) in London and Paris between 1st January, 2008 and 31st December, 2012. Red indicates risk factors, while green indicates protective factors for the following reasons of admissions: all, pain, fever, acute chest syndrome (ACS) or other. Main associations are shown in bold.

Factor Temporal patterns

Meteorological factors

Air quality factors

Day of the week Weekend Season Year Rainfall Min Temperature Max Temperature Wind speed Atmospheric pressure Relative humidity Carbon monoxide Ozone PM2.5 PM10 Black carbon Particle number

London (n=1,474)

Higher on Mondays, lower on Sundays No differences Lower during weekends Higher in Autumn Lower in Spring Lower in Summer Lower in 2008, higher in 2009 No differences All, Pain & Fever No differences No differences Pain ACS All & Pain All & Pain No differences ACS No differences Fever All Fever No differences Pain & Other No differences ACS No differences All, Pain No data Pain No data

all types of SCD and all reasons for admission, and covered a much broader age range (2 to 70 years old). These findings may differ from ours because risk factors for acute complications may be very different in children than adults; additionally, children are known to spend most of their time close to home, exposed to the same environment, whereas adults often work far from home and are potentially exposed to several different environments each day. Another recent study assessing the association between air pollution and emergency visits of children with SCD in Sao Paulo, Brazil, found remarkably high increases in relation to PM10, NO2, CO and O3.9 We could not find consistent risks associated with pollutants in this study, particularly of the magnitude described in Brazil. While both studies tested for lag effects, the Brazilian study looked at lags of up to 4 days while we tested for effects up to 1 week. The levels of exposure to pollution and environmental co-factors are very different in Brazil (e.g. NO2 = 104.59 μg/m3 ± 48.56; T°min = 15.23°C ± 3.40) compared to London (NO2 = 56.18 μg/m3 ± 16.84; T°min = 7.73°C ± 5.19) and Paris (NO2 = 35.76 μg/m3 ± 13.00; T°min = 9.04°C ± 5.79), and it is perhaps unsurprising that the findings are different. Environmental factors are important determinants of acute complications in children with SCD, but these effects are complex and differ significantly with geography and city design, even between apparently similar cities such as London and Paris. Better understanding of these factors in different geographic settings is important to allow patients and families to be given accurate information on how to reduce the risk of acute complications. This approach is particularly important in a chronic disease, such as SCD, for which there are few effective therapeutic options. Although the precise mechanism by which wind speed could trigger complications in SCD is not clear, it represents the environmental factor that is 674

Paris (n=347)

most consistently identified in association studies in European cities. Further studies are needed to accurately define environmental effects in SCD. These are particularly relevant in some cities in sub-Saharan Africa and India, where pollution levels and patient numbers are very high,23 and are likely to become more relevant as global warming and air pollution increase. Future studies need to consider the different types of SCD separately, and also consider how these may change with the age of the patient. Other important questions on environmental effects which need to be answered include the role of the home environment and the long-term effects of exposure to air pollutants.13 Due to the range of complications associated with SCD and to differences in exposures between patients living in urban environments, monitoring the exposure of large number of patients through personal devices (e.g mobile phone apps or personal monitors) might be particularly informative. Environmental factors in SCD are particularly important to understand because they can be manipulated relatively easily and cheaply with simple advice, unlike genetic causes of variation. Increased knowledge in this area will also be valuable for public health services, to understand when more patients will be admitted to hospital, and what housing requirements are important for families with SCD. Acknowledgments The authors acknowledge METEO FRANCE for supplying meteorological data for Paris, and Véronique Ghersi from AirParif for advice about air quality data for Paris. This study had no specific funding. The authors would like to thank the Stroke Association for supporting some of the work involved in this study, although this Association had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. haematologica | 2017; 102(4)


Enviromental factors and admissions for SCD

References 9. 1. Rees DC, Williams TN, Gladwin MT. Sickle-cell disease. Lancet. 2010;376 (9757):2018-2031. 2. Steinberg MH. Predicting clinical severity in sickle cell anaemia. Br J Haematol. 2005;129(4):465-481. 3. Sebastiani P, Solovieff N, Hartley SW, et al. Genetic modifiers of the severity of sickle cell anemia identified through a genomewide association study. Am J Hematol. 2010;85(1):29-35. 4. Serjeant GR. The natural history of sickle cell disease. Cold Spring Harb Perspect Med. 2013;3(10):a011783. 5. Redwood AM, Williams EM, Desal P, Serjeant GR. Climate and painful crisis of sickle-cell disease in Jamaica. BMJ. 1976;1(6001):66-68. 6. Jones S, Duncan ER, Thomas N, et al. Windy weather and low humidity are associated with an increased number of hospital admissions for acute pain and sickle cell disease in an urban environment with a maritime temperate climate. Br J Haematol. 2005;131(4):530-533. 7. Yallop D, Duncan ER, Norris E, et al. The associations between air quality and the number of hospital admissions for acute pain and sickle-cell disease in an urban environment. Br J Haematol. 2007;136(6): 844-848. 8. Mekontso Dessap A, Contou D, DandineRoulland C, et al. Environmental influences on daily emergency admissions in sickle-

haematologica | 2017; 102(4)

10.

11.

12.

13.

14.

15.

16.

cell disease patients. Medicine. 2014;93 (29):e280. Barbosa SM, Farhat SC, Martins LC, et al. Air pollution and children's health: sickle cell disease. Cad Saude Publica. 2015;31 (2):265-275. Kauf TL, Coates TD, Huazhi L, ModyPatel N, Hartzema AG. The cost of health care for children and adults with sickle cell disease. Am J Hematol. 2009;84(6):323327. Pizzo E, Laverty AA, Phekoo KJ, et al. A retrospective analysis of the cost of hospitalizations for sickle cell disease with crisis in England, 2010/11. J Public Health. 2015;37(3):529-539. AlJuburi G, Laverty AA, Green SA, et al. Trends in hospital admissions for sickle cell disease in England, 2001/02â&#x20AC;&#x201C;2009/10. J Public Health. 2012;34(4):570-576. Tewari S, Brousse V, Piel FB, Menzel S, Rees DC. Environmental determinants of severity in sickle cell disease. Haematologica. 2015;100(9):1108-1116. Lawrence MG. The relationship between relative humidity and the dewpoint temperature in moist air: a simple conversion and applications. B Am Meteorol Soc. 2005;86(2):225-233. Rodopoulou S, Samoli E, Analitis A, Atkinson RW, deâ&#x20AC;&#x2122;Donato FK, Katsouyanni K. Searching for the best modeling specification for assessing the effects of temperature and humidity on health: a time series analysis in three European cities. Int J Biometeorol. 2015;59(11):1585-1596. Blazejczyk K, Epstein Y, Jendritzky G,

17.

18.

19.

20.

21.

22.

23.

Staiger H, Tinz B. Comparison of UTCI to selected thermal indices. Int J Biometeorol. 2012;56(3):515-535. Bhaskaran K, Gasparrini A, Hajat S, Smeeth L, Armstrong B. Time series regression studies in environmental epidemiology. Int J Epidemiol. 2013;42(4):1187-1195. Armstrong BG, Gasparrini A, Tobias A. Conditional Poisson models: a flexible alternative to conditional logistic case cross-over analysis. BMC Med Res Methodol. 2014;14(122. Mohan J, Marshall JM, Reid HL, Thomas PW, Hambleton I, Serjeant GR. Peripheral vascular response to mild indirect cooling in patients with homozygous sickle cell (SS) disease and the frequency of painful crisis. Clin Sci. 1998;94(2):111-120. Freemantle N, Ray D, McNulty D, et al. Increased mortality associated with weekend hospital admission: a case for expanded seven day services? BMJ. 2015;351:h4596. Haywood C, Tanabe P, Naik R, Beach MC, Lanzkron S. The impact of race and disease on sickle cell patient wait times in the emergency department. Am J Emerg Med. 2013;31(4):651-656. Rees DC, Thein SL, Osei A, et al. The clinical significance of K-Cl cotransport activity in red cells of patients with HbSC disease. Haematologica. 2015;100(5):595-600. Piel FB, Hay SI, Gupta S, Weatherall DJ, Williams TN. Global burden of sickle cell anaemia in children under five, 2010-2050: modelling based on demographics, excess mortality, and interventions. PLoS Med. 2013;10(7):e1001484.

675


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Red Cell Biology & its Disorders

Ferrata Storti Foundation

Erythrocyte survival is controlled by microRNA-142

Natalia Rivkin,1 Elik Chapnik, 1 Alexander Mildner,2 Gregory Barshtein,3 Ziv Porat,4 Elena Kartvelishvily,5 Tali Dadosh,5 Yehudit Birger,6 Gail Amir,8 Saul Yedgar,3 Shai Izraeli,6,7 Steffen Jung2 and Eran Hornstein1

Department of Molecular Genetics, Weizmann Institute of Science, Rehovot; Department of Immunology Weizmann Institute of Science, Rehovot; 3Department of Biochemistry and Molecular Biology, Hebrew university, Hadassah Medical School, Jerusalem; 4Flow Cytometry Unit, Life Sciences Core Facilities, Weizmann Institute of Science, Rehovot; 5Department of Chemical Research Support, Weizmann Institute of Science, Rehovot; 6Functional Genomics and Leukemic Research, Cancer Research Center, Sheba Medical Center, Ramat Gan; 7Department of Human Molecular Genetics and Biochemistry, Tel Aviv University and 8Department of Pathology, Hadassah Medical Center, Jerusalem, Israel 1 2

Haematologica 2017 Volume 102(4):676-685

ABSTRACT

H Correspondence: eran.hornstein@weizmann.ac.il

Received: September 7, 2016. Accepted: November 22, 2016. Pre-published: December 1, 2016. doi:10.3324/haematol.2016.156109 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/676 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

676

ematopoietic–specific microRNA-142 is a critical regulator of various blood cell lineages, but its role in erythrocytes is unexplored. Herein, we characterize the impact of microRNA-142 on erythrocyte physiology and molecular cell biology, using a mouse loss-of-function allele. We report that microRNA-142 is required for maintaining the typical erythrocyte biconcave shape and structural resilience, for the normal metabolism of reactive oxygen species, and for overall lifespan. microRNA-142 further controls ACTIN filament homeostasis and membrane skeleton organization. The analyses presented reveal previously unappreciated functions of microRNA-142 and contribute to an emerging view of small RNAs as key players in erythropoiesis. Finally, the work herein demonstrates how a housekeeping network of cytoskeletal regulators can be reshaped by a single micro-RNA denominator in a cell type specific manner.

Introduction microRNAs (miRNAs) are genome encoded small non-coding RNAs that posttranscriptionally regulate gene expression. The hematopoietic–specific microRNA142 (miR-142) gene, and particularly its mature miR-142-3p product, has emerged as a critical regulator of various blood lineages. The miR-142 gene locus was historically associated with the (8;17) translocation, in B-cell leukemia,1 several years before miRNAs were discovered. Recent experimental evidence uncovered miR142 involvement in the differentiation and function of hemangioblasts,2,3 lymphocytes,4-7 neutrophils8 and macrophages.7,9,10 Furthermore, we have recently uncovered a key role for miR-142 in the maintenance of CD4+ dendritic cells11 by characterizing a novel mouse model with deletion of the miR-142 allele.12 Using this model, we also discovered that miR-142 is expressed in the bipotent MK-erythroid precursors (PreMegEs) and demonstrated that miR-142 function is essential for the differentiation of megakaryocytes (MKs) and that its loss results in pronounced thrombocytopenia.12 Several other groups independently developed miR-142 mouse alleles, including Kramer et al.7 and Shrestha et al.13 Erythrocytes are the most abundant cell type in the circulation. Erythrocyte numbers in peripheral blood are tightly controlled, and new red blood cells are generated to congregate and compensate for conditions of stress, such as in the case of traumatic blood loss.14,15 The maintenance of normal mechanical properties is essential to permit erythrocytes to withstand shear forces in the circulation. The flexibility and strength of the erythrocyte membrane is conferred by stereotyped pentagonal and hexagonal arrays of short ACTIN filament, which are cross-linked by long, flexible spectrin molecules.16 ANKYRIN, BAND 3 /SLC4A1 and PROTEIN haematologica | 2017; 102(4)


Erythrocyte survival is controlled by miR-142

4.1/EPB41A also play important roles in the cytoskeleton scaffold, and their dysfunction is associated with various types of hemolytic anemia.17,18 High oxygen carrying capacity, hemoglobin autoxidation and high polyunsaturated fatty acid content in the lipid bilayer make mature erythrocytes particularly susceptible to oxidative damage. Accordingly, oxidative damage inflicts cytoskeletal reorganization and shortens erythrocyte life span, which is ~40 days in Mus musculus, when adequate antioxidant response is manifested.19 miRNAs, and the effector protein cofactor ARGONAUTE,20 play important roles in erythropoiesis. For example, miR-451/144, whose biogenesis depends on ARGONAUTE and not on Dicer activity, promotes efficient erythropoiesis by inhibiting 14-3-3zeta, Rab14 and and miR-191 regulates erythrocyte Gata2,21-27 enucleation.28 In the work herein, we test the hypothesis that miR-142 activity is pivotal in controlling erythrocyte survival, cytoskeleton and function. miR-142 regulates the erythrocyte cytoskeleton morphology and biomechanical properties and its capacity to cope with oxidative stress, affecting the survival of red blood cells in vivo.

Methods Mouse genetics and cell biology The miR-142 null allele is descrcribed by Chapnik and colleagues.12 Mouse strains were housed and handled in accordance with protocols approved by the Institutional Animal Care and Use Committee of the Weizmann Institute of Science. Phenylhydrazine (PHZ) protocol follows that of Gutierrez and colleagues.29 Red blood cell (RBC) life span analysis was performed with EZ-Link Sulfo-NHS-Biotin (Thermo Scientific). Deformability and osmotic fragility was determined as in Relevy et al.,30 Barshtein et al.31 and Beutler and colleagues.32

Microscopy Erythrocyte ultrastructural analysis was perfomed with secondary electron (SE) detector in high resolution Ultra 55 (Zeiss) or ESEM (FEI) microscopes. Super resolution micrographs captured on Vutara SR200 STORM microscope with 647 nm laser excitation power density of about 7 kW/cm2. Z-stack image was measured by acquiring 700 frames at 50 Hz for each z position at 0.1 um steps, and 405-nm activation laser power was ramped slowly to maintain optimal single-molecule density. Single-molecule fitting was performed with Vutara software. Fluorescence-activated cell sorting (FACS) and high-speed cell imaging analysis in flow with ImageStreamX protocols are described by Chapnik et al.12 and George and colleagues.33 A detailed description of the Material and Methods, including primers and antibodies used, is provided in the Online Supplementary Materials.

Results In screening for hematopoietic phenotypes of the miR142 knockout model,12 we performed complete blood counts (Figure 1A-C) that revealed a reduction of erythrocyte numbers, hemoglobin levels and hematocrit concentration. Red blood cell distribution width (RDW) and the percentage of circulating reticulocytes increased significantly (Figure 1D,E). Microscopic analysis of methylene haematologica | 2017; 102(4)

blue stained smears demonstrated more reticulocytes in miR-142–/– peripheral blood than in controls. The fraction of irregular cell morphologies was also higher in miR-142deficient peripheral blood, relative to controls (Figure 1F). Because loss of miRNA gene phenotypes are often better characterized under sensitized stress conditions,34,35 we studied recovery after the introduction of a hemolytic challenge with PHZ, a chemical that induces hemolytic anemia.21 The administration of two PHZ injections on successive days is a standard protocol from which wildtype (WT) animals fully recovered. However, it resulted in the death of all miR-142–/– mice (Figure 2A). We then used a reduced PHZ dose by employing a single injection. Under this regimen, miR-142–/– mice exhibited anemia that was significantly more severe than in WT controls (Figure 2B-F). Recovery from the milder protocol of PHZ-induced anemia was delayed in miR-142–/– mice, however, all mutant mice survived. A complete blood count, 11 days after insult, revealed that mutant animals recover more slowly than controls. For example, hemoglobin levels after 11 days were 14 g / dl and 8 g / ld in WT and miR-142 null animals, respectively. These data demonstrate that hematopoietic-specific miR-142 activity is required for normal erythrocyte levels under steady state, and for adequately coping with the hemolytic stress induced by PHZ. To test if miR-142 knockout erythrocytes exhibit cellautonomous sensitivity to PHZ we performed an in vitro study by incubating circulating erythrocytes with different PHZ doses. Methylene blue staining revealed that miR142 null erythrocytes are more susceptible to PHZinduced hemolysis than control erythrocytes (Figure 2G). PHZ is known to induce oxidative stress, and reactive oxygen species (ROS) are reported to play a part in the toxic mechanism.36 Therefore, we measured a fluorescent indicator for ROS, CM-H2DCFDA, by flow cytometry in untreated WT and miR-142 null circulating erythrocytes. We observed a two-fold upregulation of ROS species in Ter-119HI erythrocytes (Figure 2H,I). Finally, to test miR142 null erythrocyte lifespan in vivo, we performed a pulsechase study of erythrocytes by Sulfo-NHS-Biotin labeling. miR-142 erythrocyte lifespan was approximately 10% shorter than control cells, when calculated on days 21-42 after PHZ treatment (Figure 2J). Therefore, it seems that miR-142-deficient erythrocytes exhibit a shorter than expected lifespan in vivo and synthesize more ROS, which plausibly contributes to their death, resulting in the observed anemia. Oxidative damage has been shown to cause changes to RBC morphology, biomechanical properties and cytoskeletal reorganization.37 We therefore sought to characterize the morphology of miR-142–/– erythrocytes by scanning electron microscopy (SEM). The study revealed that miR-142–/– erythrocytes were larger (radial dimension) on average than control cells and that abnormal forms, such as leptocytes and knizocytes, were more prevalent than in controls (Figure 3A-C). We also observed a higher proportion of miR-142–/– erythrocytes with prominent membrane defects. These deformations included hole-like structures that were observed only in miR-142–/– erythrocytes but not in control cells (Figure 3D). Taken together, these data demonstrate that miR-142 is required for normal erythrocyte morphology. Next, we explored ACTIN distribution, via ImageStreamX flow cytometry. miR-142–/– erythrocytes displayed decreased ACTIN polarity and were less circular 677


N. Rivkin et al.

A

D

B

C

E

F

Figure 1. Ineffective erythropoiesis in miR-142–/– animals, erythroid hyperplasia and anemia. Bar graphs are shown for measures of (A) red blood cell numbers (RBC), (B) hemoglobin (HGB), (C) hematocrit (HCT) and (D) red cell distribution width (RDW) values in sixmonth-old miR-142–/– mice and controls. (E) Reticulocytes (Retic) levels in three-month-old miR142–/– mice and controls. (F) Upper panels, micrographs of miR-142–/– and WT peripheral blood smears, stained with Methylene blue, demonstrating residual RNA in reticulocytes (black arrows) and of MayGrunwald-Giemsa staining (lower panels), which depict abnormal morphology of miR-142–/– erythrocytes. Scale bars, 5 μm. Representative results from one of two independent experiments are shown (mean ± SEM) with at least five animals in each group. *P<0.05; **P<0.005; ***P<0.0005.

than control RBCs (Figure 4A-D), consistent with changes in the organization of the erythrocyte cytoskeleton. Therefore, F-ACTIN assembly may be a key mechanism, whereby miR-142 controls erythroid cell development. The analysis was substantiated by super-resolution stochastic optical reconstruction microscopy (STORM), which demonstrated the typical dense submembranous ACTIN meshwork in WT erythrocytes. However, aberrant F-ACTIN polarity was observed in miR-142-deficient cells, in accordance with ImageStreamX measurements. These structural changes in ACTIN distribution are observed in erythrocytes deficient of miR-142, even when the cellular morphology was seemingly normal (Figure 4E). Therefore, in contrast to the symmetrically-rounded 678

WT red blood cell, the ACTIN meshwork of miR-142deficient erythrocytes drive irregularities to the cytoskeleton. Erythrocyte ability to pass through the microvasculature of the splenic sinuses is defined by their deformability, a characteristic that depends on cytoskeletal properties. The assessment of deformability may therefore expose changes in the ACTIN cytoskeleton functions. We quantified miR-142–/– erythrocyte deformability, under flowinduced shear stress (3.0 Pa), by measuring the major and minor axes elongation ratio of slide-adherent erythrocytes, as in the work of Relevy and colleagues.30 Intriguingly, miR-142–/– cells were more deformable than control erythrocytes (Figure 5A). In addition, an independent analysis of osmotic fragility revealed that miR-142–/– haematologica | 2017; 102(4)


Erythrocyte survival is controlled by miR-142

A

B

E

H

C

D

F

G

I

J

Figure 2. Vulnerability of miR-142–/– erythrocytes to oxidative challenge and increased ROS levels. (A) Kaplan-Meier survival plot of WT (n=12) and miR 142–/–(n=13) mice, treated by two intraperitoneal injections of PHZ (48 mg/kg) on the first two days of the study (arrows). Bar graph of erythrocyte measures in peripheral blood of animal subjected to reduced PHZ dosing (single injection of 48 mg/kg, n=4 per group), including (B) blood cell numbers (RBC), (C) hemoglobin (HGB), (D) hematocrit (HCT), (E) red cell distribution width (RDW) values and (F) percentage of circulating reticulocytes (Retic). (G) In vitro hemolysis, measured as percentage of live RBCs in glass hemocytometer, 4 hrs after incubation with increasing phenylhydrazine (PHZ) concentrations. (H) ROS quantified by FACS of cells treated with CMH2DCFDA and co-gated as Ter119HI and (I) bar graph depicting the percentage of erythroid cells with relatively high ROS, taken from top right plot quadrant and averaged for three independent experimental repeats from WT and miR-142–/– mice (n=3 per genotype). (J) In vivo eryhtorcyte lifespan, gated by streptavidin-Cy5 and Ter119HI antibodies in control and miR-142–/– mice, pulse-chased by Sulfo-NHS-Biotin labeling (n=4). *P<0.05; **P<0.005.

haematologica | 2017; 102(4)

679


N. Rivkin et al. A

B

C

D

Figure 3. miR-142–/– erythrocytes exhibit defective morphology. (A) Scanning electron microscopy reveals the stereotypic biconcave form of WT erythrocytes and atypical morphology of miR-142–/– erythrocytes (knizocytes, leptocytes). Micrograph scales noted in panels. (B) Quantification of cell surface area, (C) percentage of cells with abnormal morphology or (D) structural membrane changes in miR-142–/– (n=3) and WT cells (n=3). Representative results from one of two independent experiments (mean ± SEM) with >100 cells per group. *P<0.05; ***P<0.0005. AU: arbitrary unit.

cells are more resistant to osmotic lysis than WT control cells (Figure 5B). We concluded that miR-142 activity is necessary for maintaining normal membrane mechanical properties, and its deficiency leads to abnormal cellular deformability. The typical shape of the RBC and its elasticity are attributed to cell membrane proteins, surface charge and phospholipid composition, whose interactions reinforce the erythrocyte membrane with a deformable cytoskeleton network. Key steps in understanding the composition and function of the erythrocyte cytoskeleton were taken in the 1980’s with the revelation of a designated meshwork composed of ACTIN, SPECTRIN, ANKYRIN, BAND 3 /SLC4A1 and PROTEIN 4.1/EPB41.16-18,38 miR-142 deficiency downregulated SPECTRIN and upregulated ADDUCTIN and ANKYRIN levels, relative to controls (Figure 6A-I). Since the canonical erythrocyte cytoskeleton proteins do not harbor binding sites for miR-142, these effects were probably indirect, and may be related to the derepression of COFILIN, GRLF1 or other immediate targets of miR-142. K-562 is a human erythroleukemia line, derived from chronic myelogenous leukemia in terminal blast crises.39 K562 cells can be driven to erythrocyte differentiation with hemin treatment.40 To test if the main observations with miR-142 loss-of-function are extendable to human RBCs, we knocked down miR-142-3p by transfecting K-562 cells with a miRZip-142-3p plasmid and induced erythroid dif680

ferentiation with hemin. miR-142-3p was knocked down to 40% of the expression levels in K-562 cells that were transfected with miRZip-control vector (Figure 7A) and accordingly, miR-142-3p targets were derepressed, including CFL2, GRLF1 and WASL (Figure 7B). ImageStreamX cytometry revealed an increased average cell area, reduced circularity and decreased F-ACTIN polarity, consistent with observations in mouse miR-142 knockout erythrocytes (Figure 7C-F). Therefore, knockdown of human miR-1423p in K-562 cells recapitulates several phenotypes observed in vivo in the mouse miR-142 null allele. To test the relevance of miR-142 targets we next designed a rescue experiment, whereby we knocked down miR-142 in K-562 cells and then targeted derepressed targets by short hairpin RNA (shRNA). The shRNAs, expressed from lentiviral vectors, effectively reduced WASL, CFL2 and ARHGAP35/GRLF1 expression. ImageStreamX flow cytometry of miR-142-3p knockdown K-562 cells, in which WASL, CFL2 and ARHGAP35/GRLF1 were silenced, partially recovered the circularity phenotype defect imposed by miR-142 deficiency, relative to a lentivirus encoding a shRNA directed against red fluorescent protein, (shRFP), that was employed as control (Figure 7G,H). These data are consistent with the activity of ACTIN-binding proteins as effectors of miR-142-3p in control of the erythroid cell cytoskeleton. In summary, our study characterizes a new regulator of erythrocytes. miR-142 is an interesting regulahaematologica | 2017; 102(4)


Erythrocyte survival is controlled by miR-142

A

B

C

D

E

Figure 4. Disturbed cytoskeletal architecture in the absence of miR-142. (A) Representative single-cell images of RBCs, obtained by ImageStreamX flow cytometer. Bright-field (BR, gray), Ter119HI (yellow), Alexa Fluor 647-Phalloidin (F-actin, red), CD71Neg (green) and HoechstNeg (purple). Scale bar, 7 μm. (B) Decreased ACTIN polarity, (C) ACTIN position in the cell and (D) low circularity in miR-142−/− RBCs relative to WT controls revealed by ImageStreamX analysis. (E) Super-resolution microscopy images of F-ACTIN depth in RBCs, top and side view. Scale bar, 2 μm. Three animals per group (mean ± SEM). *P<0.05; **P<0.005; ***P<0.0005. AU: arbitrary unit.

tor as it controls the cytoskeleton in several hematopoietic cell lines. Furthermore, miR-142 regulates the erythrocytes capacity to cope with ROS. One of the intriguing conclusions emerging from the current study is that a housekeeping network of cytoskeletal regulators can be reshaped by a single miRNA denominator in a cell type specific manner.

Discussion Our study demonstrates that miR-142 controls critical facets of erythrocyte maturity, playing roles in the regulahaematologica | 2017; 102(4)

tion of RBC size, mechano-physical properties, lifespan and numbers, in vivo. Specifically, miR-142 functions help erythrocytes to cope with oxidative stress, and regulate the cytoskeleton. Accordingly, irregular morphology, abnormal mechanical properties and elevated ROS levels prevail in miR-142-deficient erythrocytes. Erythrocytes are sensitive to oxidative stress because of the high physiological hemoglobin level.41,42 Accordingly, the consequences of oxidative damage are often manifested as changes in erythrocyte deformability, splenic sequestration and a decreased lifespan.43-46 The sensitivity of miR142 null cells to ROS is consistent with hypersensitivity to 681


N. Rivkin et al. A

B

Figure 5. Disturbed RBC deformability and osmotic fragility of miR-142–/–. (A) RBC deformability is described by the elongation ratio (ER) distribution in miR-142−/− (n=4) and WT (n=4) cells under shear stress of 3.0 Pa. miR-142−/− curve is shifted to the right, meaning that mutant cells are more deformable than control erythrocytes. (B) miR-142−/− osmotic fragility curve is shifted to the left of a typical WT curve. miR-142−/− RBCs are less susceptible to decreased osmotic pressure and undergo lysis only after greater reduction in NaCl concentrations, relative to WT cells. n=4 WT and 4 miR-142−/− experimental repeats. *P<0.05. RBC: red blood cell; NaCI: sodium chloride.

PHZ-induced oxidative stress, as compared with WT littermates. However, the mechanisms by which miR-142 regulates oxidative stress are still unknown. Several proteins that are known to maintain the RBC structure are dysregulated in miR-142-deficient erythrocytes. However, since aductin, ankyrin and spectrin mRNAs do not harbor cis binding sites (miRNA recognition elements) for miR-142 in their 3ˈ untranslated regions, plausibly they are indirectly controlled by the miRNA. Direct targets, whose expression is derepressed and hence upregulated in miR-142-deficient adult erythrocytes, include Cofilin and Grlf1, which are regulated by the same miRNA also in MKs.12 Consistently, the knockdown of WASL, CFL2 and ARHGAP35 / GRLF1 in K-562 cells partially rescued the typical circular architecture of erythroid 682

cells. These data provide initial evidence that targets that were originally characterized in MKs also function downstream of miR-142-3p in the erythroid lineage. However the exact contribution of specific targets should be explored in future works, and it is also likely that additional miR-142-3p targets participate in the regulation of the ACTIN network and cytoskeleton organization. Abnormal erythrocyte structure or membrane deformability are observed in many clinical red cell disorders,47 therefore, the new link to miR-142 suggests that this miRNA gene might be an essential component in erythrocyte pathologies. Furthermore, based on the study of the human K-562 cell line, it may be that miR-142 functions are conserved to humans. In summary, our analysis suggests a critical role for miRhaematologica | 2017; 102(4)


Erythrocyte survival is controlled by miR-142

A

Figure 6. miR-142 activity is required for normal RBC protein expression. (A) Western blots of RBC proteins probed with the indicated antibodies. Each panel is representative of results from >5 blots. (B) Bar graph represents the ratio of SPECTRIN/GAPDH. (C) GRLF1/GAPDH. (D) ADUCTIN/GAPDH. (E) ACTIN/GAPDH. (F) BAND3/GAPDH. (G) ANKYRIN/GAPDH. (H) COFILLIN/GAPDH. (I) WASL/GAPDH. *P<0.05.

B

C

D

E

F

G

H

I

142 in controlling a network of erythrocyte proteins that are required for the unique cellular morphology, function, and for oxidative defense. More generally, the use of mouse genetics to uncover the functions of a lineage-specific miRNA helps to explain how ubiquitous regulatory networks can be reshaped to meet with the specialized requirements of a unique cell type. In the future, it will be important to investigate additional functions of miR-142, including its potential ability haematologica | 2017; 102(4)

to regulate erythropoiesis, the mechanisms by which 142 contributes to anti-oxidative activity, and whether the cytoskeleton network downstream of miR-142, that functions in megakaryocytes and erythrocytes, is present in other hematopoietic cells wherein miR-142 is expressed. Acknowledgments The authors would like to thank Ofira Higfa and Yehudah Melamed for veterinary services and husbandry and Dr. Joseph 683


N. Rivkin et al. A

B

C

D

G

E

F

H

Lotem for helpful discussions. The work is funded by the Minerva foundation and Minna-James-Heineman Stiftung through Minerva. Work at the Hornstein lab is further funded by an ERC consolidator program, Israel Science Foundation, the Legacy-heritage program, The Bruno and Ilse Frick Foundation for Research on ALS, the ALS Therapy Alliance, The Motor Neurone Disease Association (UK), the Thierry Latran Foundation for ALS research, the ERA-Net for Research Programmes on Rare Diseases (FP7), A. Alfred Taubman through IsrALS, Teva Pharmaceutical Industries Ltd as part of the Israeli National Network of Excellence in Neuroscience (NNE), Yeda-Sela, YedaCEO, Israel Ministry of Trade and Industry, Y. Leon Benoziyo

684

Figure 7. miR-142 regulates ACTIN cytoskeletal architecture and dynamics in K-562 cells. Expression of miR-142 (A) and its targets (B) in K562 cells transfected with a knockdown (KD) miR-ZIP-142 vector. Representative micrographs, obtained with ImageStreamX flow cytometer and stained with Alexa Fluor 647Phalloidin (F-actin, red), and Hoechst (blue, C). Scale bar, 7 Îźm. Decreased F-ACTIN polarity (D), increased cell area (E) and decreased circularity (F) in K-562 cells after miR-142 knockdown, relative to WT controls. K-562 transduced with miR-ZIP-142 knockdown vector and concomitantly with a lentivirus encoding a shRNA directed against RFP (shRFP) or a set of shRNA vectors as indicated. ImageStreamX flow cytometery that was performed 72 hrs later revealed that knockdown of targets partially restored F-ACTIN polarity (G) and circularity (H). Representative results from one of two independent experiments (mean Âą SEM), three experimental repeats per group. *P<0.05; **P<0.005; ***P<0.0005. AU: arbitrary unit; n.s: non significant; sh: short hairpin; RFP: red fluorescent protein.

Institute for Molecular Medicine, The Kekst Family Institute for Medical Genetics, The David and Fela Shapell Family Center for Genetic Disorders Research, The Crown Human Genome Center, the Nathan, Shirley, Philip and Charlene Vener New Scientist Fund, Julius and Ray Charlestein Foundation, The Fraida Foundation, The Wolfson Family Charitable Trust, The Adelis Foundation, MERCK (UK), Maria Halphen, and the Estates of Fannie Sherr, Lola Asseof and Lilly Fulop. Electron microscopy studies were partially founded by The Moskowitz Center for Nano and Bio-Nano Imaging at WIS. E.H. is Head of The Nella and Leon Benoziyo Center for Neurological Diseases, and the lab is further supported by Dr. Sydney Brenner.

haematologica | 2017; 102(4)


Running Title

References 1. Gauwerky CE, Huebner K, Isobe M, Nowell PC, Croce CM. Activation of MYC in a masked t(8;17) translocation results in an aggressive B-cell leukemia. Proc Natl Acad Sci USA. 1989;86(22):8867-8871. 2. Nimmo R, Ciau-Uitz A, Ruiz-Herguido C, et al. MiR-142-3p controls the specification of definitive hemangioblasts during ontogeny. Dev Cell. 2013;26(3):237-249. 3. Lu X, Li X, He Q, et al. miR-142-3p regulates the formation and differentiation of hematopoietic stem cells in vertebrates. Cell Res. 2013;23(12):1356-1368. 4. Chen CZ, Li L, Lodish HF, Bartel DP. MicroRNAs modulate hematopoietic lineage differentiation. Science. 2004;303 (5654):83-86. 5. Zhou Q, Haupt S, Prots I, et al. miR-142-3p is involved in CD25+ CD4 T cell proliferation by targeting the expression of glycoprotein A repetitions predominant. J Immunol. 2013;190(12):6579-6588. 6. Sun Y, Oravecz-Wilson K, Mathewson N, et al. Mature T cell responses are controlled by microRNA-142. J Clin Invest. 2015; 125(7):2825-2840. 7. Kramer NJ, Wang WL, Reyes EY, et al. Altered lymphopoiesis and immunodeficiency in miR-142 null mice. Blood. 2015; 125(24):3720-3730. 8. Fan HB, Liu YJ, Wang L, et al. miR-142-3p acts as an essential modulator of neutrophil development in zebrafish. Blood. 2014; 124(8):1320-1330. 9. Lagrange B, Martin RZ, Droin N, et al. A role for miR-142-3p in colony-stimulating factor 1-induced monocyte differentiation into macrophages. Biochim Biophys Acta. 2013;1833(8):1936-1946. 10. Sonda N, Simonato F, Peranzoni E, et al. miR-142-3p prevents macrophage differentiation during cancer-induced myelopoiesis. Immunity. 2013;38(6):1236-1249. 11. Mildner A, Chapnik E, Manor O, et al. Mononuclear phagocyte miRNome analysis identifies miR-142 as critical regulator of murine dendritic cell homeostasis. Blood. 2013;121(6):1016-1027. 12. Chapnik E, Rivkin N, Mildner A, et al. miR142 orchestrates a network of actin cytoskeleton regulators during megakaryopoiesis. Elife. 2014;3:e01964. 13. Shrestha A, Carraro G, El Agha E, et al. Generation and Validation of miR-142 Knock Out Mice. PloS one. 2015; 10(9):e0136913. 14. Broudy VC, Lin NL, Priestley GV, Nocka K, Wolf NS. Interaction of stem cell factor and its receptor c-kit mediates lodgment and acute expansion of hematopoietic cells in the murine spleen. Blood. 1996;88(1):75-81. 15. Bauer A, Tronche F, Wessely O, et al. The glucocorticoid receptor is required for stress erythropoiesis. Genes Dev. 1999;13(22): 2996-3002.

haematologica | 2017; 102(4)

16. Byers TJ, Branton D. Visualization of the protein associations in the erythrocyte membrane skeleton. Proc Natl Acad Sci USA. 1985;82(18):6153-6157. 17. Lazarides E, Woods C. Biogenesis of the red blood cell membrane-skeleton and the control of erythroid morphogenesis. Annu Rev Cell Biol. 1989;5:427-452. 18. Mohandas N, Evans E. Mechanical properties of the red cell membrane in relation to molecular structure and genetic defects. Annu Rev Biophys Biomol Struct. 1994;23:787-818. 19. Mohanty JG, Nagababu E, Rifkind JM. Red blood cell oxidative stress impairs oxygen delivery and induces red blood cell aging. Front Physiol. 2014;5:84. 20. O'Carroll D, Mecklenbrauker I, Das PP, et al. A Slicer-independent role for Argonaute 2 in hematopoiesis and the microRNA pathway. Genes Dev. 2007;21(16):1999-2004. 21. Rasmussen KD, Simmini S, Abreu-Goodger C, et al. The miR-144/451 locus is required for erythroid homeostasis. J Exp Med. 2010;207(7):1351-1358. 22. Kim M, Tan YS, Cheng WC, Kingsbury TJ, Heimfeld S, Civin CI. MIR144 and MIR451 regulate human erythropoiesis via RAB14. Br J Haematol. 2015;168(4):583-597. 23. Patrick DM, Zhang CC, Tao Y, et al. Defective erythroid differentiation in miR451 mutant mice mediated by 14-3-3zeta. Genes Dev. 2010;24(15):1614-1619. 24. Yu D, dos Santos CO, Zhao G, et al. miR451 protects against erythroid oxidant stress by repressing 14-3-3zeta. Genes Dev. 2010;24(15):1620-1633. 25. Dore LC, Amigo JD, Dos Santos CO, et al. A GATA-1-regulated microRNA locus essential for erythropoiesis. Proc Natl Acad Sci USA. 2008;105(9):3333-3338. 26. Pase L, Layton JE, Kloosterman WP, Carradice D, Waterhouse PM, Lieschke GJ. miR-451 regulates zebrafish erythroid maturation in vivo via its target gata2. Blood. 2009;113(8):1794-1804. 27. Cheloufi S, Dos Santos CO, Chong MM, Hannon GJ. A dicer-independent miRNA biogenesis pathway that requires Ago catalysis. Nature. 2010;465(7298):584-589. 28. Zhang L, Flygare J, Wong P, Lim B, Lodish HF. miR-191 regulates mouse erythroblast enucleation by down-regulating Riok3 and Mxi1. Genes Dev. 2011;25(2):119-124. 29. Gutierrez L, Tsukamoto S, Suzuki M, et al. Ablation of Gata1 in adult mice results in aplastic crisis, revealing its essential role in steady-state and stress erythropoiesis. Blood. 2008;111(8):4375-4385. 30. Relevy H, Koshkaryev A, Manny N, Yedgar S, Barshtein G. Blood banking-induced alteration of red blood cell flow properties. Transfusion. 2008;48(1):136-146. 31. Barshtein G, Gural A, Manny N, Zelig O, Yedgar S, Arbell D. Storage-induced damage to red blood cell mechanical properties can be only partially reversed by rejuvena-

32.

33.

34. 35. 36.

37.

38. 39.

40.

41.

42. 43.

44.

45.

46. 47.

tion. Transfus Med Hemother. 2014;41(3): 197-204. Beutler E, Kuhl W, West C. The osmotic fragility of erythrocytes after prolonged liquid storage and after reinfusion. Blood. 1982;59(6):1141-1147. George TC, Fanning SL, Fitzgerald-Bocarsly P, et al. Quantitative measurement of nuclear translocation events using similarity analysis of multispectral cellular images obtained in flow. J Immunol Methods. 2006;311(1-2):117-129. Hornstein E, Shomron N. Canalization of development by microRNAs. Nat Genet. 2006;38 Suppl:S20-24. Emde A, Hornstein E. miRNAs at the interface of cellular stress and disease. EMBO J. 2014;33(13):1428-1437. Kondo M, Itoh S, Kusaka T, Imai T, Isobe K, Onishi S. The ability of neonatal and maternal erythrocytes to produce reactive oxygen species in response to oxidative stress. Early Hum Dev. 2002;66(2):81-88. Sinha A, Chu TT, Dao M, Chandramohanadas R. Single-cell evaluation of red blood cell bio-mechanical and nano-structural alterations upon chemically induced oxidative stress. Sci Rep. 2015;5: 9768. Branton D, Cohen CM, Tyler J. Interaction of cytoskeletal proteins on the human erythrocyte membrane. Cell. 1981;24(1):24-32. Lozzio CB, Lozzio BB. Human chronic myelogenous leukemia cell-line with positive Philadelphia chromosome. Blood. 1975;45(3):321-334. Rutherford TR, Clegg JB, Weatherall DJ. K562 human leukaemic cells synthesise embryonic haemoglobin in response to haemin. Nature. 1979;280(5718):164-165. Hebbel RP, Leung A, Mohandas N. Oxidation-induced changes in microrheologic properties of the red blood cell membrane. Blood. 1990;76(5):1015-1020. Fibach E, Rachmilewitz E. The role of oxidative stress in hemolytic anemia. Curr Mol Med. 2008;8(7):609-619. Wang S, Dale GL, Song P, Viollet B, Zou MH. AMPKalpha1 deletion shortens erythrocyte life span in mice: role of oxidative stress. J Biol Chem. 2010;285(26):19976-19985. Snyder LM, Fortier NL, Trainor J, et al. Effect of hydrogen peroxide exposure on normal human erythrocyte deformability, morphology, surface characteristics, and spectrin-hemoglobin cross-linking. J Clin Invest. 1985;76(5):1971-1977. Mohandas N, Clark MR, Jacobs MS, Shohet SB. Analysis of factors regulating erythrocyte deformability. J Clin Invest. 1980;66(3):563-573. Cimen MY. Free radical metabolism in human erythrocytes. Clinica Chim Acta. 2008;390(1-2):1-11. An X, Mohandas N. Disorders of red cell membrane. Br J Haematol. 2008;141(3): 367-375.

685


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Hemostasis

Ferrata Storti Foundation

Factor VIII/V C-domain swaps reveal discrete C-domain roles in factor VIII function and intracellular trafficking

Eduard H.T.M. Ebberink,1* Eveline A.M. Bouwens,1* Esther Bloem,1 MariĂŤtte Boon-Spijker,1 Maartje van den Biggelaar,1 Jan Voorberg,1,3 Alexander B. Meijer1,2 and Koen Mertens1,2

Haematologica 2017 Volume 102(4):686-694

Department of Plasma Proteins, Sanquin Research, Amsterdam; 2Department of Pharmaceutics, Utrecht Institute for Pharmaceutical Sciences, Utrecht University and 3 Landsteiner Laboratory of AMC and Sanquin, University of Amsterdam, the Netherlands 1

*EHTM and EAMB contributed equally to this work

ABSTRACT

F

Correspondence: k.mertens@sanquin.nl

Received: August 2, 2016. Accepted: December 23, 2016. Pre-published: January 5, 2017. doi:10.3324/haematol.2016.153163 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/686 Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

686

actor VIII C-domains are believed to have specific functions in cofactor activity and in interactions with von Willebrand factor. We have previously shown that factor VIII is co-targeted with von Willebrand factor to the Weibel-Palade bodies in blood outgrowth endothelial cells, even when factor VIII carries mutations in the light chain that are associated with defective von Willebrand factor binding. In this study, we addressed the contribution of individual factor VIII Cdomains in intracellular targeting, von Willebrand factor binding and cofactor activity by factor VIII/V C-domain swapping. Blood outgrowth endothelial cells were transduced with lentivirus encoding factor V, factor VIII or YFP-tagged C-domain chimeras, and examined by confocal microscopy. The same chimeras were produced in HEK293-cells for in vitro characterization and chemical foot-printing by mass spectrometry. In contrast to factor VIII, factor V did not target to Weibel-Palade bodies. The chimeras showed reduced Weibel-Palade body targeting, suggesting that this requires the factor VIII C1-C2 region. The factor VIII/V-C1 chimera did not bind von Willebrand factor and had reduced affinity for activated factor IX, whereas the factor VIII/V-C2 chimera showed a minor reduction in von Willebrand factor binding and normal interaction with activated factor IX. This suggests that mainly the C1-domain carries factor VIII-specific features in assembly with von Willebrand factor and activated factor IX. Foot-printing analysis of the chimeras revealed increased exposure of lysine residues in the A1/C2- and C1/C2-domain interface, suggesting increased C2-domain mobility and disruption of the natural C-domain tandem pair orientation. Apparently, this affects intracellular trafficking, but not extracellular function.

Introduction Factor VIII (FVIII) serves as a co-factor for activated factor IX (FIXa) in the factor X (FX) activating complex. It consists of 2332 amino acids with a distinct domain structure: A1-a1-A2-a2-B-a3-A3-C1-C2.1 Intracellular processing of the B-domain yields a heterodimeric FVIII protein with a 90-220 kDa heavy chain (A1-a1-A2-a2) non-covalently associated with a 80 kDa light chain (a3-A3-C1-C2).2 FVIII circulates in complex with the multimeric glycoprotein von Willebrand factor (VWF) that protects FVIII from premature clearance and proteolytic degradation. Complex assembly occurs over an extended surface on FVIII, spanning the entire light chain.3-5 The sulfated tyrosine on position 1680 is essential for binding to VWF and mutation of this tyrosine results in impaired complex formation with VWF.3 Recently it has been shown that FVIII is expressed in endothelial cells.6-8 Previous work showed that FVIII overexpressed in endothelial cells co-sorts with VWF to the haematologica | 2017; 102(4)


Discrete role of C-domains of FVIII

secretory organelles designated Weibel-Palade bodies (WPB).9-11 The precise interaction mediating sorting to WPB has not been clarified, although it has been generally assumed that VWF plays a key role as a sorting chaperone. In contrast to this view, we have shown that FVIII sorting to WPB does not require the high-affinity interaction via the sulfated tyrosine on position 1680 in the a3domain.11,12 Moreover, endothelial cells with mutations in the FVIII C1- and C2-domains leading to impaired extracellular VWF/FVIII complex assembly show apparently normal expression of FVIII and storage in WPB.12 Like FVIII, coagulation factor V (FV) comprises two lipid-binding C-domains that form a similar side-by-side pair.13-17 FV shares ~40% sequence homology with FVIII and has a similar domain structure (A1-A2-a2-B-A3-a3C1-C2).18 FV functions as a cofactor for FXa in the prothrombinase complex, demonstrating that FVIII and FV serve a similar cofactor function. Unlike FVIII, however, FV neither circulates in complex with VWF, nor does it act as a cofactor for FIXa in the activation of FX. In the present study, we addressed the contribution of C-domains to FVIII intracellular targeting and extracellular function by constructing FVIII chimeras carrying FV C1- or C2domains and exploring the functional and structural implications of these C-domain swaps.

Methods Factor VIII constructs All constructs used in this study encoded B-domain-deleted FVIII (BDD-FVIII) variants in order to meet size restrictions in the lentiviral packaging system.10 For the same reason FV, too, was Bdomain-deleted (BDD-FV).19 In BDD-FVIII-YFP, yellow fluorescent protein (YFP) replaced the B-domain, as described for its green fluorescent protein (GFP) equivalent elsewhere.10,11 Construction of plasmids encoding the YFP-tagged BDD-FVIII/FV chimeras is described in the Online Supplementary Material. For functional studies, BDD-FVIII-YFP and the chimeras were constructed in the pcDNA3.1 vector for production in HEK293 cells. To simplify nomenclature, the term BDD-FVIII is replaced by FVIII throughout this paper. Consequently, the BDD-FVIII-YFP chimeras containing the FV-C1 or -C2 domain are referred to as FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2, respectively.

Immunofluorescence microscopy of lentiviral-transduced endothelial cells The isolation of blood outgrowth endothelial cells (BOEC) and their subsequent transduction with lentivirus have been described previously.10 A detailed description of the antibody staining of BDD-FV and BDD-FVIII can be found in the Online Supplementary Material. Z-stacks (0.4-Îźm intervals) were taken with confocal laser scanning microscopy using a Zeiss LSM510 equipped with Plan NeoFluar 63x/1.4 Oil objective (Carl Zeiss, Heidelberg, Germany). Images were processed with Zeiss LSM510 version 4.0 software and LSM image browser (Carl Zeiss, Heidelberg, Germany). Secretion of FVIII and FV was quantified by enzymelinked immunosorbent assay (ELISA) as described previously, with the exception that the FV ELISA used the monoclonal anti-light chain antibody CLB-FV-4, and purified FV as a reference.12,19

Purified factor VIII-yellow fluorescent protein variants FVIII-YFP/FV chimeras and FVIII-YFP were produced in stable cell lines (HEK293) and purified by immunoaffinity chromatography using a monoclonal antibody (VK34) followed by anion haematologica | 2017; 102(4)

exchange chromatography (Q Sepharose FF, GE Healthcare, Uppsala, Sweden) as described in detail elsewhere.20 Purified FVIIIYFP variants were homogeneous, and comprised a YFP-carrying heavy chain of approximately 110 kDa and an 80 kDa light chain (see Online Supplementary Figure S1). FVIII-YFP concentrations were determined by ELISA, employing the monoclonal anti-light chain antibody KM33 (anti-C1) or EL14 (anti-C2) for immobilization, and the anti-heavy chain antibody CLB-CAg-9 for detection.21 Normal human plasma served as the standard. FVIII activity was determined using a chromogenic assay (Chromogenix, Milan, Italy), and the activity/antigen ratios were 1.0 for FVIII-YFP, 0.9 for FVIII-YFP/FV-C2, and 0.4 for FVIII-YFP/FV-C1. In all functional studies FVIII concentrations were based on antigen concentrations, assuming that 1 U/mL corresponds to 0.3 nM.

Characterization of factor VIII-yellow fluorescent protein variants Interactions of purified FVIII-YFP variants with recombinant full-length VWF11 were assessed by surface plasmon resonance analysis using a BIAcore 3000 biosensor (Biacore AB, Uppsala, Sweden) as described previously.21 Details of the data analysis are provided in the Online Supplementary Material. Interactions of FVIII-YFP and the FVIII-YFP/FV chimeras with FIXa and phospholipids were inferred from FX activation studies, as described in detail elsewhere.21 Structural differences between FVIII-YFP, FV and the C-domain-swapped chimeras were probed by chemical foot-printing using lysine-reactive tandem-mass-tags (TMT) and mass spectrometry as described previously.22 A full description of the processing of labeled proteins into peptides and mass spectrometry analysis is given in the Online Supplementary Material.

Results Differential intracellular accumulation of factor V compared to factor VIII Because FV is structurally highly homologous to FVIII, we compared FVIII and FV with respect to WPB trafficking in BOEC expressing BDD-FVIII or BDD-FV via lentiviral transduction. Transduced BOEC secreted much larger amounts of FV (140 pmol/106 cells/72 h) than FVIII (typically 1-5 pmol/106 cells/72 h). Confocal microscopy revealed that, as expected, FVIII retained within the endothelial cells completely co-localized with VWF in WPB (Figure 1A). Consistent with co-localization, FVIIIcontaining WPB were round.10,23,24 In contrast, FV did not traffic to WPB (Figure 1B). The FV-transduced cells contained WPB which were negative for FV and retained their typical elongated morphology similar to those of the surrounding non-transduced cells. Instead, prominent background staining for FV could be detected in what might represent the endoplasmic reticulum (Figure 1B). Apparently, trafficking of FV is different from that of FVIII.

C-domains of factor VIII contribute to intracellular trafficking To study the contribution of individual C-domains of the C-domain pair to FVIII sorting, we exchanged single FVIII C-domains for those of FV. Because of potential difficulties in staining these variants with antibodies directed at the FVIII light chain, we expressed YFP-tagged FVIII (FVIII-YFP) in this experiment. To this end BOEC were transduced with lentivirus encoding for FVIII-YFP and YFP-tagged FVIII variants with swapped C1- or C2687


E.H.T.M. Ebberink et al.

domain (FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2). FVIIIYFP/FV chimera production, as assessed by antigen levels in the conditioned medium, ranged between 0.03 – 0.3 pmol/106 cells/72 h. The inherent variability in expression levels and transduction efficiency did not allow for quantification of the WPB-targeted fraction of the FVIII-YFP/FV chimeras. However, for transduced BOEC, trafficking could be studied by fluorescence microscopy. Confocal microscopy showed that YFP-tagged FVIII sorted exclusively to WPB (Figure 2A) and was retained in round WPB, similar to untagged FVIII (Figure 1A). This was to be expected because YFP tags as such do not affect FVIII trafficking to WPB.12,24 Substitution of the C1-domain resulted in total loss of co-localization of FVIII-YFP/FV-C1 with VWF in WPB (Figure 2B). Like FV, FVIII-YFP/FV-C1 was shown to accumulate intracellularly, without any apparent punctuated YFP signal as seen in Figure 2A (upper right side panels). The FVIII-YFP/FV-C2 variant displayed some residual sorting involving round WPB (Figure 2C). However, not all WPB were positive for YFP (see inset, Figure 2C), and a major part of the YFP signal was localized intracellularly, presumably in the endoplasmic reticulum. These results indicate that the C1-domain is the main driver of FVIII trafficking to WPB, although the C2domain contributes to this process as well.

Figure 1. Differential sorting of FVIII and FV in blood outgrowth endothelial cells. Confocal images of BOEC expressing (A) B-domain-deleted FVIII or (B) Bdomain-deleted FV. Both FVIII and FV are shown in green, and staining of VWF is shown in red. Yellow indicates co-localization of VWF with FVIII or FV. Images of the separate green and red channels are depicted on the right side. FVIII is solely visible co-stored with VWF in WPB. While FVIII-containing WPB have a round morphology, FVIII-negative WPB remain elongated. FV is not visible in WPB. The white scale bar represents 20 μm.

688

The factor VIII C1-domain is critical for von Willebrand factor-binding Given that the FVIII-YFP/FV chimeras displayed reduced co-localization with VWF in BOEC, we studied the ability of the chimeras to interact with VWF employing surface plasmon resonance analysis. This enables a time-resolved monitoring of the association and dissociation between two interactive proteins, by measuring mass increase and

Figure 2. Intracellular localization of FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2 chimeras in blood outgrowth endothelial cells. Confocal microscopy of BOEC expressing YFP-tagged B-domain-deleted FVIII variants. Merged signals of the YFP-tagged FVIII variants (green) and Alexa-633 stained VWF (red) (co-localization in yellow) are shown on the left side. On the right side image exports of the separate channels are displayed: YFP (green) top panel and Alexa-633 (red) lower panel. (A) YFP-tagged FVIII co-localizes with VWF in round WPB, while surrounding non-transduced BOEC, which are negative for YFP, contain elongated WPB. (B) FVIII-YFP/FV-C1 does not co-localize with VWF, and the WPB remain elongated. (C) Some of FVIII-YFP/FV-C2 is located in round WPB with VWF, but most of the FVIII-YFP/FV-C2 is dispersed throughout the cell. The white scale bar represents 20 μm.

haematologica | 2017; 102(4)


Discrete role of C-domains of FVIII

decrease due to the interaction of a soluble component (in this case, FVIII-YFP/FV chimeras) with an immobilized binding partner (in this case, VWF).12,21 FVIII-YFP variants, in varying concentrations, were passed over a chip with immobilized VWF and analyzed for surface-bound mass change, expressed in Resonance Units (RU). Figure 3 shows the maximal binding response as a function of the FVIII concentration, which can be used to derive an estimate of the dissociation constant Kd.21,25 The apparent Kd for FVIII-YFP was 8 nM, while FVIII-YFP/FV-C2 showed a slight reduction in VWF binding, with an apparent Kd of approximately 16 nM. In contrast, FVIII-YFP/FV-C1 showed no appreciable interaction with VWF at all (Figure 3A). Similar data were obtained at pH 5.5, which should resemble the pH in the mature secretory compartment.26 FVIII variants revealed complex binding kinetics, from which the affinity could not be directly inferred by standard curve fitting of the sensorgrams, due to pronounced heterogeneity in the dissociation phase at pH 5.5 (Online Supplementary Figure S2). Data obtained at pH 7.4 were

A

B

less complex, and revealed an apparent Kd of 3 nM for FVIII-YPF and 8 nM for FVIII-YFP/FV-C2. Irrespective of the data analysis used, these experiments demonstrate that FVIII-YFP/FV-C2 and FVIII-YFP display similar interactions with VWF, whereas FVIII-YFP/FV-C1 shows almost no VWF binding. Apparently, replacement of the FVIII C2-domain with that of FV conserves VWF-binding. In contrast, the FVIII C1-domain is irreplaceable for the interaction with VWF.

Substitution of the factor VIII C1-domain, but not C2-domain, affects cofactor activity Thus far, purified chimeras were characterized in the absence of lipids although FVIII/FV C-domains are known to assemble on (negatively charged) membranes.13-16 To study the function of the FVIII-YFP/FV chimeras further, we examined complex formation with FIXa on phospholipid vesicles and their capability to convert the physiological substrate, FX. Membrane-binding was assessed by varying the phospholipid vesicle concentration at a fixed concentration of FIXa (16 nM). When using phospholipid vesicles containing 15% phosphatidylserine, no reduction in FXa generation rates were observed with FVIII-YFP/FVC2 in comparison with FVIII-YFP (Figure 4A). A lower maximal FXa generation rate was seen with FVIII-YFP/FVC1. However, both chimeras showed the same apparent affinity to membranes as FVIII-YFP (half-maximal response ~0.5 μM). The reduction in maximal response of FVIIIYFP/FV-C1 is due to reduced binding of FIXa, as was assessed by varying the FIXa concentration at a fixed phospholipid concentration (Figure 4B). Therein, the FXa generation rates of FVIII-YFP/FV-C2 and FVIII-YFP are identical. FVIII-YFP/FV-C1, however, needed a concentration of at least 30 nM to reach the same level. Similar data were obtained in the presence of vesicles containing 5% phosphatidylserine, with the exception that the defect of the FVIII-YFP/FV-C1 variant proved more prominent (Figure 4D). At this low phosphatidylserine content the activity of FVIII-YFP/FV-C2 was also somewhat reduced at low lipid concentrations (Figure 4C). These data demonstrate that in terms of cofactor activity the C2-domain is interchangeable with that of FV. However, swapping the C1-domain introduces a defect that involves assembly with FIXa.

Distorted C-domain pairing in chimeras containing factor V C-domains

Figure 3. Interaction of FVIII-YFP and FVIII-YFP/FV chimeras with von Willebrand factor. Surface plasmon resonance (SPR) analysis was performed using the BIAcore 3000 system as described elsewhere.21 FVIII-YFP (closed squares), FVIIIYFP/FV-C1 (closed circles) and FVIII-YFP/FV-C2 (open triangles) were passed over a CM5 chip coated with recombinant VWF (7, 24 and 37 fmol/mm2) in a buffer containing 150 mM NaCl, 5 mM CaCl2, 2.4% glycerol (v/v), 0.005% Tween 20 (v/v), and 20 mM HEPES (pH 7.4, panel A) or 20 mM MES (pH 5.5, panel B) for 240 s at 20 μL/min at 25 °C. The signal of a non-coated CM5 channel was subtracted to correct for differences in buffer composition. Response upon the onset of dissociation was taken to represent maximal binding and was plotted against the FVIII concentration. Values represent mean resonance units (RU) ± SD. Data were analyzed by a non-linear regression using a single hyperbola. This revealed apparent Kd values of approximately 8 and 16 nM for FVIII-YFP and FVIII-YFP/FVC2, respectively. VWF binding of FVIII-YFP/FV-C1 was too low for quantitative analysis. Individual SPR sensorgrams are shown in Online Supplementary Figure S2.

haematologica | 2017; 102(4)

To examine the structural integrity of the chimeras, structural differences between FVIII-YFP and the chimeras were studied by lysine residue labeling with TMT. This approach consists of a pairwise comparison between proteins by labeling each with one of two isobaric labels (TMT-126 or TMT-127), which, upon tandem mass spectrometry, reveal their lysine-bound indicator mass of either 126 or 127 Da. This enables simultaneous identification and quantification of surface-exposed lysine residues.22 We used TMT-126 for FVIII-YFP and TMT-127 for the FVIII-YFP/FV chimeras and expressed relative incorporation of TMT labels in the ratio of TMT127/TMT-126 per individual peptide. A ratio above 1 indicates increased lysine exposure in the FVIII-YFP/FV chimera (TMT-127 labeled) compared to the FVIII-YFP (TMT-126 labeled), and vice versa for ratios below 1. Comparing the TMT-labeling incorporation of the FVIIIYFP/FV chimeras to that of FVIII-YFP revealed that for most of the lysine residues, TMT-127/TMT-126 ratios 689


E.H.T.M. Ebberink et al.

were around 1 (Online Supplementary Figure S3). This indicates that accessibility for most of the lysine residues remained unchanged. However, some peptides showed a TMT-127/TMT-126 ratio >1 in both chimeras (Figure 5A). Most of these peptides contain lysine residues situated at the C2/A1-domain interface (Lys107, Lys123 and Lys127) (Figure 5B). Peptides covering the C2-domain lysine residues Lys2239 and Lys2249 showed increased accessibility only in FVIII-YFP/FV-C1 (Figure 5A). Due to the swapping of the C2-domain and thus the absence of TMT-127 labeled counterparts, FVIII-YFP/FV-C2 showed a TMT-127/TMT-126 ratio of almost zero for these lysine residues. Apart from lysine residues with increased accessibility, FVIII-YFP/FV-C2 also displayed lysine residues in the C1-domain with decreased accessibility. These are Lys2020, Lys2110 and/or Lys2111. (Figure 5A, B and Online Supplementary Figure S3). Using TMT labeling, the lysine exposure of the swapped C-domains themselves could also be investigated in comparison with their native conformation in FV. This was done by labeling FV with TMT-126 and comparing it to the chimeras labeled with TMT-127. In this way only the swapped C-domains were examined. Again, most peptides displayed an unchanged lysine exposure (ratios were approximately 1). However, one C1-domain peptide containing two lysine residues directed towards the opposing C2-domain in the FV model had an average TMT-127/TMT-126 ratio of 3 (Lys1941 and Lys1954) (Figure 6A,C). Within the swapped C2-domain, two peptides indicated an increased accessibility of its lysine residues: one peptide with lysine residues directed towards to the C1-domain (Lys2157 and Lys2161) and one

690

A

B

C

D

peptide containing a lysine residue directed towards the A1-domain (Lys2137) (Figure 6B,C). Compared to FV, these lysine residues are more accessible for TMT labeling in the chimeras. Thus, taken together the comparisons between FVIII-YFP, FV and the chimeras, the C1/C2- and C2/A1-domain interfaces become exposed which suggests that pairing of the C-domains is altered in both FVIIIYFP/FV chimeras.

Discussion The site of FVIII biosynthesis has remained a matter of debate. It was previously established that FVIII overexpression in BOEC leads to co-storage of FVIII with VWF in WPB.9,10 The relevance of this model system has recently been established by findings that FVIII is expressed in endothelial cells.6-8 However, the mechanism underlying FVIII trafficking and secretion remains poorly understood. Here we studied the expression of FVIII, FV and FVIII/FV C-domain chimeras in BOEC. In contrast to FV, FVIII was solely found in WPB (Figure 1). This implies a specific FVIII mechanism, which appears to be affected once one of the two FVIII C-domains is swapped with FV (Figure 2). Particularly, C1-domain swapping appeared destructive, although FVIII-YFP/FV-C2 also displayed a trafficking defect. Thus, both C-domains appear to contribute. Remarkably, C-domains that bind tightly to phospholipids tend to occur in a tandem pair as observed in FVIII, FV, lactadherin and developmental-endothelial-locus 1 (del-1).27 Next to a tandem or multimeric organization, secretion of these strong lipid-binders tends to be regulated. For

Figure 4. FVIII cofactor activity of the FVIII-YFP chimeras. FVIII-YFP (closed squares), FVIII-YFP/FV-C1 (closed circles) and FVIII-YFP/FV-C2 (open triangles) were examined for their capability to function as cofactor in the activation of FX. FX activation studies were performed in the presence of phospholipid vesicles containing 15% phosphatidylserine, 20% phosphatidylethanolamine, and 65% phosphatidylcholine (panels A and B) or 5% phosphatidylserine, 20% phosphatidylethanolamine, and 75% phosphatidylcholine panels (C and D) in a buffer containing 40 mM Tris-HCl (pH 7.8), 150 mM NaCl and 0.2% (w/v) bovine serum albumin. FIXa titrations (panels B and D) were performed by incubation of 25 ÂľM of phospholipid vesicles mixed with 0-64 nM FIXa, 0.3 nM FVIII and 200 nM FX. Phospholipid titrations (panels A and C) were performed by incubation of 0-80 ÎźM phospholipid vesicles mixed with 16 nM FIXa, 0.3 nM FVIII and 200 nM FX. Reactions were initiated by addition of 1.5 mM CaCl2 and 1 nM thrombin. Subsamples of the reaction mixture were taken at 30 s intervals and analyzed for FXa using the substrate S2765 containing the thrombin inhibitor I-2581.21 Absorbance values were converted into molar concentrations using a standard curve of active-site titrated purified FXa.

haematologica | 2017; 102(4)


Discrete role of C-domains of FVIII

instance, lactadherin released by several cell types is associated with confined membrane vesicles (exosomes).28,29 As reported by others, deletion of one of the two lactadherin C-domains results in a loss of the exosome-mediated secretion.29 This suggests that trafficking of lactadherin requires an intact C-domain tandem pair, which might also be required for FVIII. Within both FVIII-YFP/FV chimeras, TMT foot-printing revealed more accessibility between the A1/C2- and C1/C2-domain interface, implying a loose C-domain tandem pair with increased C2-domain mobility (Figures 5 and 6). This extends the notion that the C2-domain in FVIII has limited interdomain contacts.30 In two available FVIII crystal structures,30,31 the C2-domain Lys2239 interaction with a nearby glutamic acid (Glu122 in the A1domain) seems to be the only evident electrostatic interaction (Online Supplementary Figure S4). A swap of the C2domain may affect this interaction even though Lys2293 is conserved in FV. The swapped Lys2239 could not be detected; however, Glu122 neighboring lysine residues

A

Lys123 and Lys127 could be measured and had an increased TMT-127/TMT-126 ratio (Figure 5). Strikingly, when swapping the C1-domain, Lys2239 could be resolved and also displayed increased labeling despite its remote location. Perhaps, introduction of a FV C-domain produces an unfavorable interaction in the C1/C2-domain interface which in turn affects the C2/A1-domain interface. By one-sided insertion of a FV C-domain in FVIII, interactions would be disrupted due to loss of the interacting counterparts. Indeed, Lys1941 and Lys2161 display increased accessibility in the chimeras while in FV they most likely interact with residues on their opposite Cdomain Phe2163 and Glu2034 (Online Supplementary Figure S4).32 The A3/C1-domain interface could hardly be probed because it contains a limited number of lysine residues. Nonetheless three of these resideues, at positions 2010, 2110 and 2111 in the C1-domain, displayed reduced surface exposure in the FVIII-YFP/FV-C2 chimera (Figure 5 and Online Supplementary Figure S3). The apparent protection of these lysine residues suggests that the top of the

B

Figure 5. Comparison of FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2 with FVIII-YFP by labeling with tandem-mass tags. FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2 were labeled with TMT-127 and FVIII-YFP with TMT-126 to assess differences in lysine residue exposure. Labeled FVIII variants were then processed into peptides and measured for TMT incorporation. The lysine residues with a TMT-127/TMT-126 ratio different from 1 are: Lys47+Lys48, Lys107, Lys123+Lys127, Lys2020, Lys2110+Lys2111 and Lys2239+2249. (A) The peptides containing the same lysine residue(s) and their TMT-127/TMT-126 ratio per chimera are shown in the different panels. The mean TMT127/TMT-126 ratio Âą SD is given by the black lines; peptides and their resolved TMT-127/TMT-126 ratio are represented by gray dots. A TMT-127/TMT-126 ratio above or below 1 (dashed line) indicates increased or decreased incorporation of TMT-127 and therefore altered accessibility of those lysine residues in the chimera compared to FVIII-YFP. Peptides covering lysine residues 2239 and 2249 show a TMT-127/TMT-126 ratio > 1 for chimera FVIII-YFP/FV-C1. In the FVIII-YFP/FV-C2 chimera these lysine residues have a near zero ratio because the FVIII C2-domain peptides are not present in the FVIII-YFP/FV-C2 chimera (only TMT-126 labeled FVIII-YFP lysine residues could be detected). Software analysis considers a TMT-127/TMT-126 ratio of zero erroneous and therefore may produce ratios with a minimum value of 0.01. (B) Lysine residues with increased accessibility for TMT labeling are indicated in the FVIII crystal structure (2R7E)30 as red spheres, whereas lysine residues with reduced reactivity are indicated in blue.

haematologica | 2017; 102(4)

691


E.H.T.M. Ebberink et al.

A

C

B

Figure 6. Comparison of FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2 with FV by labeling with tandem-mass tags. FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2 were labeled with TMT127, and FV with TMT-126. The mean TMT-127/TMT-126 ratio Âą SD (black lines) is given for peptides containing the same lysine residues from (A) the C1-domain in blue and (B) the C2-domain in red of FVIII-YFP/FV-C1 and FVIII-YFP/FV-C2, respectively. Discrete peptides and their measured TMT-127/TMT-126 ratios are represented by dots. Lysine residues with a TMT-127/TMT-126 ratio above 1 (dashed line) are Lys1941+Lys1954, Lys2137 and Lys2157+Lys2161. Such a TMT-127/TMT-126 ratio indicates increased incorporation of TMT-127 and therefore increased accessibility of those lysine residues in the chimera compared to FV wild-type. (C) The lysine residues with TMT-127/TMT-126 ratios above 1 are indicated in a model of FVa (1Y61)32 with spheres. The C1-domain is indicated in blue and the C2-domain in red.

C1-domain associates more tightly with the A3-domain in this chimera, thus maintaining the A3/C1-domain interface undisrupted. An overall limited disruption of the domain organization is reflected in the activity of the FVIII/FV-C2 chimera. The FVIII/FV-C2 chimera, despite its more accessible C1/C2- and C2/A1-domain interfaces, displays normal FIXa binding and FX activation. Moreover, swapping the entire C2-domain within FVIII does not result in any functional loss (Figure 4B, D). Interestingly, as reported by others, deletion of the C2-domain results in a 4-fold reduction in FIXa affinity.33 Apparently, the presence of the C1domain alone is insufficient for full cofactor function. However, the C2-domain of FV fully compensates for this defect (Figure 4B). This apparent paradox can be explained by a complementary role for the C2-domain in its FIXabinding conformation. Unlike the C2-domain, the C1domain contributes directly to FIXa binding. This is in agreement with the observation of Wakabayashi et al., who noted that a FVIII variant with the C1-domain replaced by a second C2-domain exhibits an approximately 9-fold reduced affinity for FIXa.34 The FVIII/FV-C1 chimera has a VWF-binding defect that 692

is more severe than that of FVIII lacking the sulfated tyrosine on position 1680 (FVIII-Y1680F).11,12 This is even more prominent at pH 5.5 (Figure 3 and Online Supplementary Figure S2) and implies that apart from the sulfated Tyr1680, the FVIII C1-domain is essential for VWF/FVIII complex formation. This is in agreement with previous reports that mutations in the C1-domain can interfere with VWF binding,5,12 although this does not exclude an additional contribution of the C2-domain.4 A direct C1domain interaction with VWF is further supported by recent studies using hydrogen-deuterium exchange and electron microscopy,35,36 Previously, we analyzed a variety of FVIII mutations that are associated with reduced interaction with VWF, but normal trafficking to WBP.12 The fact that these variants retained some residual VWF binding at pH 5.5 led us to speculate that this might be sufficient for trafficking to WPB.12 In this respect the FVIII/FV-C1 chimera is more defective than the mutations causing hemophilia A which we studied. If FVIII storage with VWF is driven exclusively by VWF binding, this might explain the trafficking defect for FVIII/FV-C1 (Figure 2B). Surprisingly, although the VWF binding of FVIII/FV-C2 is close to normal, this haematologica | 2017; 102(4)


Discrete role of C-domains of FVIII

chimera also displays reduced WPB storage and intracellular accumulation in BOEC (Figure 2C). This supports the conclusion that FVIII trafficking to WPB requires the tandem C-domain pair. This would be compatible with our previous data, because none of the hemophilia A-causing mutations that we analyzed can be expected to disrupt the C-domain pair.12 One limitation of lentiviral BOEC transduction is the variability in expression which precludes obtaining quantitative information for direct comparison of the extent of WPB trafficking with VWF-binding affinity. As we showed previously, this issue can be addressed by subcellular fractionation, provided that expression and transduction efficiency are in the same order of magnitude.10,12 However, this proved unfeasible for the present set of FVIII/FV chimeras. Nevertheless, it seems evident that the intracellular YFP-staining for the FVIII/FV-C2 chimera, despite its high affinity for VWF, is different from that of FVIII-YFP or BDD-FVIII (Figures 1A and 2A,C). Likewise, the FVIII/FV-C1 chimera, which displays no appreciable VWF interaction, could not be visualized intracellularly (Figure 2B). Despite these trafficking defects in BOEC, the chimeras could be expressed in HEK293 cells, and after purification displayed structural integrity and functionality (Figures 4-6 and Online

References 1. Lenting PJ, van Mourik JA, Mertens K. The life cycle of coagulation factor VIII in view of its structure and function. Blood. 1998;92(11):3983-3996. 2. Thompson AR. Structure and function of the factor VIII gene and protein. Semin Thromb Hemost. 2003;29(1):11-22. 3. Leyte A, van Schijndel HB, Niehrs C, et al. Sulfation of Tyr1680 of human blood coagulation factor VIII is essential for the interaction of factor VIII with von Willebrand factor. J Biol Chem. 1991;266(2):740-746. 4. Saenko EL, Scandella D. The acidic region of the factor VIII light chain and the C2 domain together form the high affinity binding site for von Willebrand factor. J Biol Chem. 1997;272(29):18007-18014. 5. Jacquemin M, Lavend'homme R, Benhida A, et al. A novel cause of mild/moderate hemophilia A: mutations scattered in the factor VIII C1 domain reduce factor VIII binding to on Willebrand factor. Blood. 2000;96(3):958-965. 6. Fahs SA, Hille MT, Shi Q, Weiler H, Montgomery RR. A conditional knockout mouse model reveals endothelial cells as the principal and possibly exclusive source of plasma factor VIII. Blood. 2014;123(24):3706-3713. 7. Everett LA, Cleuren AC, Khoriaty RN, Ginsburg D. Murine coagulation factor VIII is synthesized in endothelial cells. Blood. 2014;123(24):3697-3705. 8. Shahani T, Covens K, Lavend'homme R, et al. Human liver sinusoidal endothelial cells but not hepatocytes contain factor VIII. J Thromb Haemost. 2014;12(1):36-42. 9. Rosenberg JB, Foster PA, Kaufman RJ, et al. Intracellular trafficking of factor VIII to von Willebrand factor storage granules. J Clin

haematologica | 2017; 102(4)

Supplementary Figure S3). This might suggest that besides VWF binding another, as yet unknown mechanism could play a role in endothelial FVIII storage and secretion. A role for the A1-domain could not be excluded because the chimeras, compared to FVIII-YFP, display a structural change in the A1-domain as well (Lys47+Lys48, Figure 5). Whether or not these A1-domain residues contribute to maintaining the C-domain tandem organization remains an open question. Our finding that the C2-domain of FVIII can be replaced by that of FV without compromising FVIII activity may have translational implications. It has been well established that hemophilia A patients with FVIII inhibitors often have antibodies against the C2-domain.37 It seems conceivable that such inhibitors may be bypassed by FVIII containing the FV C2-domain. Because swapping the C1 domain eliminates VWF binding and affects the interaction with FIXa, the potential of FVIII containing the FV C1-domain for bypassing C1-domain-directed inhibitors seems less evident. Acknowledgments The authors thank J.M. Koornneef for providing purified factor V. This study was supported by the Landsteiner Foundation for Blood Transfusion Research (LSBR).

Invest. 1998;101(3):613-624. 10. Van den Biggelaar M, Bouwens EAM, Kootstra NA, Hebbel RP, Voorberg J, Mertens K. Storage and regulated secretion of factor VIII in blood outgrowth endothelial cells. Haematologica. 2009;94(5):670678. 11. Van den Biggelaar M, Bierings R, Storm G, Voorberg J, Mertens K. Requirements for cellular co-trafficking of factor VIII and von Willebrand factor to Weibel-Palade bodies. J Thromb Haemost. 2007;5(11): 2235-2242. 12. Van den Biggelaar M, Bouwens EA, Voorberg J, Mertens K. Storage of factor VIII variants with impaired von Willebrand factor binding in Weibel-Palade bodies in endothelial cells. PLoS One. 2011;6(8):e24163. 13. Lu J, Pipe SW, Miao H, Jacquemin M, Gilbert GE. A membrane-interactive surface on the factor VIII C1 domain cooperates with the C2 domain for cofactor function. Blood. 2011;117(11):3181-3189. 14. Gilbert GE, Novakovic VA, Kaufman RJ, Miao H, Pipe SW. Conservative mutations in the C2 domains of factor VIII and factor V alter phospholipid binding and cofactor activity. Blood. 2012;120(9):1923-1932. 15. Gilbert GE, Kaufman RJ, Arena AA, Miao H, Pipe SW. Four hydrophobic amino acids of the factor VIII C2 domain are constituents of both the membrane-binding and von Willebrand factor-binding motifs. J Biol Chem. 2002;277(8):6374-6381. 16. Peng W, Quinn-Allen MA, Kane WH. Mutation of hydrophobic residues in the factor Va C1 and C2 domains blocks membrane-dependent prothrombin activation. J Thromb Haemost. 2005;3(2):351-354. 17. Adams TE, Hockin MF, Mann KG, Everse SJ. The crystal structure of activated protein C-inactivated bovine factor Va: Implications for cofactor function. Proc

18.

19.

20.

21.

22.

23.

24.

Natl Acad Sci USA. 2004;101(24):89188923. Bos MH, Camire RM. Blood coagulation factors V and VIII: molecular mechanisms of procofactor activation. J Coagul Disord. 2010;2(2):19-27. Bos MHA, Meijerman DWE, van der Zwaan C, Mertens K. Does activated protein C-resistant factor V contribute to thrombin generation in hemophilic plasma? J Thromb Haemost. 2005;3(3):522530. Meems H, van den Biggelaar M, Rondaij M, van der Zwaan C, Mertens K, Meijer AB. C1 domain residues Lys 2092 and Phe 2093 are of major importance for the endocytic uptake of coagulation factor VIII. Int J Biochem Cell Biol. 2011;43(8): 1114-1121. Bloem E, van den Biggelaar M, Wroblewska A, Voorberg J, et al. Factor VIII C1 domain spikes 2092-2093 and 21582159 comprise regions that modulate cofactor function and cellular uptake. J Biol Chem. 2013;288(41):29670-29679. Bloem E, Ebberink EH, van den Biggelaar M, van der Zwaan C, Mertens K, Meijer AB. A novel chemical footprinting approach identifies critical lysine residues involved in the binding of receptor-associated protein to cluster II of LDL receptorrelated protein. Biochem J. 2015;468(1):6572 Bouwens EA, Mourik MJ, van den Biggelaar M, et al. Factor VIII alters tubular organization and functional properties of von Willebrand factor stored in WeibelPalade bodies. Blood. 2011;118(22):59475956. Van den Biggelaar M, Meijer AB, Voorberg J, Mertens K. Intracellular cotrafficking of factor VIII and von Willebrand factor type 2N variants to storage organelles. Blood. 2009;113(13):3102-3109.

693


E.H.T.M. Ebberink et al. 25. Van den Biggelaar M, Madsen JJ, Faber JH, et al. Factor VIII interacts with the endocytic receptor low-density lipoprotein receptor-related protein 1 via an extended surface comprising "hot-spot" lysine residues. J Biol Chem. 2015;290(27):16463-16476. 26. Wu MM, Grabe M, Adams S, Tsien RY, Moore HP, Machen TE. Mechanisms of pH regulation in the regulated secretory pathway. J Biol Chem. 2001;276(35):3302733035. 27. Vogel W. Discoidin domain receptors: structural relations and functional implications. FASEB J. 1999;13 (Suppl):S77-82. 28. VĂŠron P, Segura E, Sugano G, Amigorena S, ThĂŠry C. Accumulation of MFG-E8/lactadherin on exosomes from immature dendritic cells. Blood Cells Mol Dis. 2005;35(2):81-88. 29. Oshima K, Aoki N, Kato T, Kitajima K, Matsuda T. Secretion of a peripheral membrane protein, MFG-E8, as a complex with

694

30.

31.

32.

33.

membrane vesicles. Eur J Biochem. 2002;269(4):1209-1218. Shen BW, Spiegel PC, Chang C-H, et al. The tertiary structure and domain organization of coagulation factor VIII. Blood. 2008;111(3):1240-1247. Ngo JC, Huang M, Roth DA, Furie BC, Furie B. Crystal structure of human factor VIII: implications for theformation of the factor IXa-factor VIIIa complex. Structure. 2008;16(4):597-606. Orban T, Kalafatis M, Gogonea V. Completed three-dimensional model of human coagulation factor vs. molecular dynamics simulations and structural analyses. Biochemistry. 2005:44(39):1308213090. Wakabayashi H, Griffiths AE, Fay PJ. Factor VIII lacking the C2 domain retains cofactor activity in vitro. J Biol Chem. 2010;285(33):25176-25184.

34. Wakabayashi H, Fay PJ. Replacing the factor VIII C1 domain with a second C2 domain reduces factor VIII stability and affinity for factor IXa. J Biol Chem. 2013;288(43):31289-31297. 35. Chiu PL, Bou-Assaf GM, Chhabra ES, et al. Mapping the interaction between factor VIII and von Willebrand factor by electron microscopy and mass spectrometry. Blood. 2015;126(8):935-938. 36. Yee A, Oleskie AN, Dosey AM, et al. Visualization of an N-terminal fragment of von Willebrand factor in complex with factor VIII. Blood. 2015;126(8):939-942. 37. Meeks SL, Healey JF, Parker ET, Barrow RT, Lollar P. Antihuman factor VIII C2 domain antibodies in hemophilia A mice recognize a functionally complex continuous spectrum of epitopes dominated by inhibitors of factor VIII activation. Blood. 2007;110 (13):4234-4242.

haematologica | 2017; 102(4)


ARTICLE

Platelet Biology & its Disorders

The transcription factor GATA1 regulates NBEAL2 expression through a long-distance enhancer

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Anouck Wijgaerts,1 Christine Wittevrongel,1 Chantal Thys,1 Timothy Devos,2 Kathelijne Peerlinck,1 Marloes R. Tijssen,3,4 Chris Van Geet1,5 and Kathleen Freson 1

Department of Cardiovascular Sciences, Center for Molecular and Vascular Biology, KULeuven, Belgium; 2Department of Haematology, University Hospitals Leuven, Belgium; 3NHS Blood and Transplant, Cambridge Biomedical Campus, UK; 4Department of Haematology, University of Cambridge, Cambridge Biomedical Campus, UK and 5 Department of Pediatrics, University Hospitals Leuven, Belgium 1

Haematologica 2017 Volume 102(4):695-706

ABSTRACT

G

ray platelet syndrome is named after the gray appearance of platelets due to the absence of α-granules. It is caused by recessive mutations in NBEAL2, resulting in macrothrombocytopenia and myelofibrosis. Though using the term gray platelets for GATA1 deficiency has been debated, a reduced number of α-granules has been described for macrothrombocytopenia due to GATA1 mutations. We compared platelet size and number of α-granules for two NBEAL2 and two GATA1-deficient patients and found reduced numbers of α-granules for all, with the defect being more pronounced for NBEAL2 deficiency. We further hypothesized that the granule defect for GATA1 is due to a defective control of NBEAL2 expression. Remarkably, platelets from two patients, and Gata1-deficient mice, expressed almost no NBEAL2. The differentiation of GATA1 patient-derived CD34+ stem cells to megakaryocytes showed defective proplatelet and α-granule formation with strongly reduced NBEAL2 protein and ribonucleic acid expression. Chromatin immunoprecipitation sequencing revealed 5 GATA binding sites in a regulatory region 31 kb upstream of NBEAL2 covered by a H3K4Me1 mark indicative of an enhancer locus. Luciferase reporter constructs containing this region confirmed its enhancer activity in K562 cells, and mutagenesis of the GATA1 binding sites resulted in significantly reduced enhancer activity. Moreover, DNA binding studies showed that GATA1 and GATA2 physically interact with this enhancer region. GATA1 depletion using small interfering ribonucleic acid in K562 cells also resulted in reduced NBEAL2 expression. In conclusion, we herein show a long-distance regulatory region with GATA1 binding sites as being a strong enhancer for NBEAL2 expression.

Introduction Platelets play a critical role in hemostasis and contain secretory granules that are essential to maintain their normal function. Of the three types of granules (dense, α and lysosomes), α-granules are the most abundant type that store essential proteins for platelet adhesion and blood coagulation.1 Platelets are shed into the blood stream after the formation of long cytoplasmic extensions from differentiated megakaryocytes (MKs), a process called proplatelet formation. The α-granules are formed in the early MKs and are actively transported along microtubules from the MK cell body towards the proplatelets.1 In circulating platelets, these α-granules mature further and proteins will be actively or passively taken up from the plasma by receptor-mediated endocytosis, in addition to the proteins loaded in these granules during their biosynthesis in the MK.2 Recently, it has become clear that haematologica | 2017; 102(4)

Correspondence: kathleen.freson@med.kuleuven.be

Received: July 16, 2016. Accepted: January 10, 2017. Pre-published: January 12, 2017. doi:10.3324/haematol.2016.152777 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/695 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

695


A. Wijgaerts et al. α-granule proteins play important roles in nonhemostatic events, such as wound healing, cancer, inflammation and innate immunity.3 Gray platelet syndrome (GPS, MIM139090) is a rare inherited platelet disorder characterized by mild to moderate bleeding, macrothrombocytopenia and markedly reduced or absent α-granules that typically results in a grayish appearance of platelets under a light microscope.4 Using three independent next-generation sequencing approaches, recessive mutations in NBEAL2 were found to cause GPS.5-7 NBEAL2 encodes a 2.754 amino acid polypeptide, neurobeachin-like-2, that contains BEACH (named after Beige and Chediak-Higashi), ARM (Armadillo), Con A-like lectin (Concanavalin A-like lectin domain), PH (Pleckstrin Homology-domain like) and WD40 domains. The exact role of BEACH domain-containing proteins remains largely unknown, but they are generally large proteins that are able to control diverse cellular mechanisms such as vesicular transport, apoptosis, membrane dynamics and receptor signaling.8 Interestingly, LYST and NBEA, two other BEACH domain-containing proteins, have been shown to be implicated in platelet dense granule defects.9-10 Regarding these proteins, their exact contribution in granule formation and trafficking during MK and platelet formation also remains unknown. In addition, the term GPS has been used for macrothrombocytopenia patients with defects in X-linked GATA111 (MIM305371), or having an autosomal dominant loss-of-function variant in GFI1B12 (MIM187900). Both genes encode for transcription factors that regulate megakaryopoiesis, and defects result in enlarged platelets with fewer (though not absent) α-granules, as studies in additional GATA1 and GFI1B patients illustrated.13-18 Based

on a recent detailed comparison of clinical and laboratory data for NBEAL2, GATA1 and GFI1B patients,3 it was suggested not to use the term GPS for other macrothrombocytopenia disorders that comprise abnormal α-granule numbers, except for patients with NBEAL2 defects. Important similarities indeed do exist (e.g., large platelets with paucity of α-granules, myelofibrosis, platelet dysfunction) but also important differences, such as red cell abnormalities that are only described for GATA1 and GFI1B defects and the presence of CD34-positive platelets for GFI1B defects. In the study herein, we hypothesized that the α-granule defect found for GATA1 macrothrombocytopenia might be due to reduced NBEAL2 expression. We compared platelet morphology and NBEAL2 protein expression using samples from two NBEAL2 and two GATA1-deficient macrothrombocytopenia patients. In vitro megakaryopoiesis from hematopoietic stem cells (HSCs) from the GATA1 defective patients was studied with a focus on proplatelet formation, α-granule formation and NBEAL2 expression. Finally, The Encyclopedia of DNA Elements (ENCODE) and GATA1/GATA2 Chromatin immunoprecipitation (ChIP) sequencing data were used to predict putative enhancer elements upstream of the NBEAL2 gene. A long-distance (31 kb upstream) region was identified as a potential enhancer that was characterized using reporter and DNA-binding assays.

Methods Patient studies Table 1 contains clinical and laboratory data for two previously described GATA1-13,14 and two NBEAL2-deficient patients. Genetic screening of NBEAL2 was performed using a high-

Table 1. Comparison of clinical and laboratory data obtained for patients with defects in NBEAL2 and GATA1.

NBEAL2 MIM13909 Mode of inheritance Genetic defect Protein defect Dyserythropoiesis Myelofibrosis Macrothrombocytopenia α-granule defect as determined by EM Megakaryocytic emperipolesis Plt count, x 109/L* MPV, fL Bleeding severity Defective platelet aggregation

Autosomal recessive Homozygous Homozygous c.7440G>A c.5721-1G>C W2480X Predicted M1908X No Yes No Yes Yes Yes ND 40 55 > Max value > Max value Mild to moderate ND Performed at low plt count in PRP: Impaired ADP, collagen, arachidonic acid, and ristocetin-induced plt aggregration

GATA1 MIM300367/314050 X-linked c.653A>G

c.652G>T

D218G

D218Y

Yes No Mild TP

Yes + anemia Yes Severe TP Yes Yes

53 > Max value Mild Performed at low plt count in PRP: Impaired collagen, and ristocetin- induced plt aggregation

8 > Max value Severe ND

Normal platelet count should be between 150-400 x 109/L. ND: not determined; EM: electron microscopy; Plt: platelet; MPV: mean platelet volume; PRP: platelet rich plasma; ADP: adenosine diphosphate; TP: thrombocytopenia.

696

haematologica | 2017; 102(4)


GATA1 controls NBEAL2 expression

throughput sequencing test.19 Platelet testing and electron microscopy (EM) analysis was performed as described.13 The platelet surface area (μm2) and number of α-granules/μm2 were measured with ImageJ software. Patients, or their legal represen-

tatives, signed informed consent to participate in the enhanced clinical and laboratory phenotyping studies. The Ethics Committee of the University Hospitals Leuven approved the study (ref. ML3580).

A

B

C

D

Figure 1. Variants in NBEAL2 and GATA1 result in macrothrombocytopenia with a paucity of α-granules. (A) Schematic diagram displaying protein domains in the NBEAL2 and GATA1 protein. The NBEAL2 protein contains an ARM-type fold, Con A-like lectin, PH, BEACH and WD40 domain. The NBEAL2 predicted M1908X and NBEAL2 W2480X variants are indicated. The GATA1 protein contains a transactivation domain (TAD), N-terminal finger domain (NF), and C-terminal finger domain (CF). Locations of the GATA1 D218Y and D218G variants are indicated. Protein drawing is not to scale. (B) EM images of platelets for the 4 patients with different variants in comparison to an unrelated control show enlarged platelets that are more round instead of discoid, with only a few α-granules and more open canalicular system forming vacuoles (Scale bars = 1μm). (C,D) The platelet area (μm2) and number of α-granules/μm2 were quantified. Values are the means ± SEM as quantified for two GATA1 patients, two NBEAL2 patients and three unrelated controls. The number of α-granules/μm2 showed a significant reduction for GATA1 patients and an even more pronounced reduction for the patients with NBEAL2 variants, which was significant compared to the patients with GATA1 variants. ***P<0.001, ****P<0.0001, one-way analysis of variance (ANOVA) with Bonferroni’s multiple test. AA; amino acids; ARM: armadillo.

haematologica | 2017; 102(4)

697


A. Wijgaerts et al. Institute of Molecular Medicine).22 Blots were performed as described in the Online Supplementary Methods.

Megakaryocyte cultures, proplatelet formation and immunohistochemistry CD34+ HSCs were isolated from the peripheral blood of patients with the GATA1 D218Y and D218G variant and unrelated healthy controls. The MK differentiation assays and immunostainings are described in the Online Supplementary Methods. Proplatelet-forming MKs were analyzed as described.20,21

Immunoblot analysis Proteins were isolated from platelets and in vitro differentiated MKs as described.13 Platelet extracts were also obtained from C57BL/6/Gata1-deficient mice (ΔneoΔHS), which have been described previously (Dr. Paresh Vyas, Department of Haematology and MRC Molecular Haematology Unit, Weatherall

B

A

Control vs D218Y

Control vs D218Y

C

Integrin

D

698

E

Luciferase reporter assays to measure NBEAL2 enhancer activity The NBEAL2 enhancer region is located about 31 kb upstream of the NBEAL2 gene and within an intron of the CCDC12 gene (GRCh37/hg19) and was cloned as two fragments, these being Chr3:46988756-46989716 (comprising 3 potential GATA sites BS1-2-3) and Chr3:46989733-46990332 (comprising 2 potential GATA sites BS-4-5) in the pGL3 promotor (enhancer-less) vector (Promega). Details of cloning, mutagenesis, transfection of K562 cells and luciferase measurements are described in the Online Supplementary Methods.23

Control vs D218G

Control vs D218G

Figure 2. Effect of GATA1 D218Y and D218G variants on megakaryopoiesis. MKs were differentiated from peripheral blood-derived CD34+ HSCs from GATA1 patients D218Y, D218G and unrelated healthy controls. The MK differentiation assay was performed twice with different controls. (A) (Top) Proplatelet formation was quantified at day 11 or day 13 of differentiation in duplicate liquid cultures for the control and the GATA1 patients. ****P<0.0001, t-test. (Bottom) Proplatelet-forming MKs were present in the control and to a lesser extent in the GATA1 D218G (arrow). Though large MKs were present for GATA1 patient D218Y, proplatelet formation was not seen. (B) PCR data performed on RNA collected at day 6, and day 11 or 13, respectively, from the D218Y and D218G defective MKs. Expression levels of NBEAL2 and ITGB3 were normalized to GADPH expression. ****P<0.0001, t-test. (C) Immunoblot analysis for NBEAL2 with a rabbit polyclonal antibody and integrin β3 were performed (left) using total protein MK lysates at days 8 and 11 for the control and day 11 for GATA1 D218Y (right) using total protein MK lysates at day 13 for the control and GATA1 D218G. Uncropped blots are shown in the Online Supplementary Figure S3. (D, E) (Top) Representative immunofluorescence confocal microscopy images of differentiated MKs at days 8 (D) and 11 (E) of culture. The MKs were stained for VWF (green) and actin (red). The control shows VWF present inside the α-granules at day 8 with a further increase of VWF at day 11. The MKs of GATA1 patient D218Y were only weakly positive for VWF both at day 8 and 11 (Scale bar =10μm). (Bottom) Quantification of VWF staining showed a significant reduction of VWF in GATA1 D218Y patient at days 8 (D) and 11 (E) of culture in comparison to the control. Values are means ± SEM as quantified for 10 randomly selected images. Panel D: day 8, *P=0.0465, t-test; panel E: *P=0.0322, t-test. RNA: ribonucleic acid; GADPH: glyceraldehyde 3-phosphate dehydrogenase; VWF: von Willebrand factor; MK: megakaryocytes.

haematologica | 2017; 102(4)


GATA1 controls NBEAL2 expression

Quantitative (q)RT-PCR to quantify GATA1, GATA2, NBEAL2, and ITGB3 expression Total ribonucleic acid (RNA) from K562 cells was extracted with TRIzol (Invitrogen). Expression of GATA1, GATA2, NBEAL2 and ITGB3 were measured with Sybr Green quantitative real-time polymerase chain reaction (qRT-PCR) using the ABI 7000 real-time PCR machine (Life Technologies). Expression was quantified via the ΔΔCt method in arbitrary units24 (see the Online Supplementary Methods for primer details).

DNA binding assay Biotinylated PCR fragments for BS-1-2-3 and BS-4-5 were bound to Superparamagnetic streptavidin beads (Hyglos) and added to nuclear extracts (NE) isolated from HEK293 cells transfected with GATA1 or GATA2. The cloning, binding reaction, and blots are described in the Online Supplementary Methods.

GATA1 knockdown in K562 cells using siRNA GATA1 depletion in K562 cells was obtained after transfection with SMARTpool GATA1 small interfering RNA (siRNA)25 or negative control SMARTpool siRNA (Dharmacon) (see Online Supplementary Methods).

Glutathione S-Transferase-pull down assay with GATA1-NF Glutathione S-Transferase (GST)-coupled GATA1-NF or GST-

A

only bound to glutathione sepharose beads (GE Healthcare) was prepared as described.13 Immunoprecipitation with GFI1B26 was performed as described in the Online Supplementary Methods.

Results Clinical description and platelet morphology studies in macrothrombocytopenia patients with NBEAL2 and GATA1 variants Table 1 describes clinical and laboratory phenotypes for two GPS patients with recessive NBEAL2 variants and two boys with GATA1 variants. One GPS patient is homozygous for W2480X, that is predicted to result in a non-functional protein without WD40 domain (Figure 1A). The other GPS patient has a homozygous splice variant c.57211G>C (Figure 1A), predicted to result in an early stop codon M1908X that would delete the PH, BEACH and WD40 domains. Indeed, a similar splice variant, c.5720+1G>A, in another GPS patient has been demonstrated to result in M1908X, based on expression studies.27 The macrothrombocytopenia patients with GATA1 variants D218G13 and D218Y6 were previously described (Figure 1A). Both mutations are located in the N-terminal finger domain, which is required for association with the coactivator friend-of-GATA1 (FOG1) that contributes to

B

Integrin

Integrin

C

D

E

Integrin

Integrin

Figure 3. NBEAL2 expression in GATA1 and NBEAL2-deficient platelets. (A) Immunoblot analysis for NBEAL2 was performed with a rabbit polyclonal NBEAL2 antibody (epitope against amino acids 1865-1939) for platelet lysates from patients with NBEAL2 M1908X, NBEAL2 W2480X, GATA1 D218Y and GATA1 D218G. Integrin β3 was used as loading control. (B) Replicated immunoblots performed with a rabbit monoclonal NBEAL2 antibody (ab187162; epitope against amino acids 1-100) for platelet lysates from patients with NBEAL2 W2480X and GATA1 D218Y defects. Integrin β3 was used as loading control. (C) Quantification of NBEAL2 expression. **P=0.0077, ***P=0.0006, t-test. (D, E) Expression of NBEAL2 in platelets from hemizygous male (D) and homozygous female (E) Gata1-deficient mice in comparison to sex-matched wild-type mice. For each lane platelets were pooled from 10 mice for Gata1 and 5 mice for control. Uncropped blots are shown in the Online Supplementary Figure S4. WT: wild-type.

haematologica | 2017; 102(4)

699


A. Wijgaerts et al.

the stability of DNA binding to a palindromic GATA recognition sequence.28 GATA1 D218Y resulted in a severe clinical phenotype with deep macrothrombocytopenia, while patients with GATA1 D218G only have mild macrothrombocytopenia (Table 1). All patients have macrothrombocytopenia, and blood smears and EM studies of their platelets clearly show a gray appearance of large platelets due to a paucity of Îą-granules. Myelofibrosis was present in the patient with NBEAL2 W2840X and the patient with GATA1 D218Y. Bone marrow studies observed the presence of megakaryocyte emperipolesis in all patients except the patient with the

NBEAL2 M1908X variant. The platelet aggregation defect for the patient with the NBEAL2 M1908X variant was similar to the defect found for the patient with GATA1 D218G. The major difference between these cases is that the GPS patients had normal red blood cells, while dyserythropoiesis was clearly present in GATA1 D218G13 and severe anemia with dyserythropoiesis in GATA1 D218Y.14 Detailed morphological examination of platelets by EM for all patients was performed and compared. All had enlarged platelets that are more round instead of discoid with fewer Îą-granules and a pronounced open canalicular system forming large vacuoles in some platelets (Figure

A

B

Figure 4. Visualization of NBEAL2 enhancer with GATA binding peaks discovered by ChIP sequencing. (A) Part of the CCDC12 gene is shown upstream of the NBEAL2 gene using the UCSC genome browser (chr3:4697520147051622 using GRCh37/hg19). The upper layer shows the position of two NBEAL2 enhancer peaks within an intron of the CCDC12 gene. H3K4Me1, H3K4Me3 and H3K27Ac profiles for K562 (blue) and NHEK (purple) cell lines are displayed (data produced by Bernstein lab at Broad Institute as part of the ENCODE database).34 DNase Hypersensitive 1 regions are displayed as gray to black boxes (from less to more open chromatin conformation). Transcription factor ChIP sequencing data are shown for GATA1 and GATA2.36 A gray box encloses each peak cluster of transcription factors occupancy (with darkness of the box being proportional to the maximum signal strength). The ChIA-PET track shows the chromatin interaction in K562 cells determined by RNA polymerase II. Physical interaction is shown between the NBEAL2 enhancer and its promotor (ENCODE/GIS-Ruan). Conservation across vertebrates is displayed as blue histograms at the bottom of the figure using the Vertebrate Multiz Alignment & Conservation (100 Species) UCSC track. (B) ChIP sequencing data (chr3:46988420-47051185) of primary human MKs are shown in density plots and displayed in the UCSC genome browser beneath the tracks for gene structure. H3K4me3 and H3K9Ac are visualized in the upper tracks (blue). GATA1 (purple) and GATA2 (pink) binding peaks are visualized in the bottom tracks. The NBEAL2 enhancer is located (chr3:46988970-46989880). UCSC: University of California, Santa Cruz; DNase: deoxyribonuclease; Vent. Cons; 100 vertebrates conservation.

700

haematologica | 2017; 102(4)


GATA1 controls NBEAL2 expression

1B). Quantification of the platelet size for three unrelated controls, two GATA1 and two NBEAL2 patients, showed significantly larger platelets in both patient groups. There was no difference between the patient groups (Figure 1C). The number of α-granules corrected for platelet size showed a pronounced reduction for GATA1 patients and even more for the NBEAL2 patients, that were now significantly different from GATA1 patients (Figure 1D).

Abnormal NBEAL2 expression in megakaryocytes from GATA1 patients Recent in vitro megakaryopoiesis studies using HSCs from GPS patients with NBEAL2 mutations showed normal MK differentiation with defective proplatelet formation and reduced α-granule proteins such as von Willebrand factor (VWF), thrombospondin and P-

A

selectin.29 Peripheral blood-derived HSCs from GATA1 patients D218Y and D218G were differentiated to MK to quantify proplatelet formation during two independent culture experiments with each compared to another unrelated control. Similar as for NBEAL2 deficiency, large MKs were present, but the MKs for both patients showed a significantly decreased formation of proplatelets that was more pronounced for D218Y (Figure 2A). Quantitative real-time polymerase chain reaction (QRT-PCR) was performed using RNA from MKs at differentiation days 6 and 11 (for D218Y) or days 6 and 13 (for D218G) (Figure 2B). NBEAL2 and ITGB3 were significantly decreased throughout the MK differentiation for the GATA1 patients, but again the defect was more pronounced for the D218Y patient. We performed immunoblot analysis using total protein lysates from differentiated MKs to quantify

B

C

Figure 5. NBEAL2 enhancer activity using luciferase reporter assays. (A) The binding peaks for GATA1 (purple) and GATA2 (pink) discovered by ChIP sequencing of primary human MKs are visualized (ch3:4698683046991970). Schematic overview of the location of the 5 potential GATA binding sites. BS-1-2-3 are located in the first peak and BS4-5 in the second peak. The two fragments are cloned in a luciferase construct 3 prime to the luciferase reporter gene to generate pGL3-BS-1-2-3 and pGL3-BS-4-5. Binding site mutants are pGL3-BS-M1-2-3, pGL3-BS-1M2-3, pGL3-BS-1-2-M3 and pGL3-BS-M4-M5 with the mutations indicated. (B) qRT-PCR for RNA from K562 was performed to determine the expression of GATA1 and GATA2 relative to GADPH. Expression was quantified via the ΔΔCt method in arbitrary units. (C) The luciferase expression after transfection with different constructs as indicated in K562 cells. The pGL3 promotor vector without enhancer sequence was used as control construct. Each plasmid was assayed in triplicate in six separate transfection experiments. ****P<0.0001, one-way analysis of variance (ANOVA) with Bonferroni’s multiple test. RNA: ribonucleic acid; BS: binding site.

haematologica | 2017; 102(4)

701


A. Wijgaerts et al. NBEAL2 and integrin β3 expression. Remarkably, NBEAL2 expression was absent in GATA1 D218Y compared to control MKs at differentiation day 11 (Figure 2C). As integrin β3 levels were also lower for GATA1 D218Y, we also loaded day 8 differentiated MKs from the control, and these cells expressed comparable integrin β3 levels while NBEAL2 was clearly detectable. NBEAL2 expression was also studied at a later time point in MK differentiation (day 13) for GATA D218G, but its expression remained low compared to control MKs (Figure 2C). To exclude NBEAL2 degradation in GATA1 mutant MKs, the expression of filamin (FLN), a large protein sensitive to calpain-mediated cleavage, was found to be comparable to control MKs (Online Supplementary Figure S1). MKs were stained at days 8 and 11 for VWF and actin to visualize α-granule formation (Figure 2D,E). During these final stages of megakaryopoiesis, VWF staining was significantly reduced for GATA1 D218Y MKs. These data could indicate that part of the similar MK and platelet phenotypes between NBEAL2 and GATA1 deficiency are due to the fact that GATA1 is a transcription factor that drives NBEAL2 expression.

Strongly reduced NBEAL2 expression in GATA1-deficient platelets Immunoblot analyses were performed to compare NBEAL2 protein expression levels in platelets of patients with variants in GATA1 or NBEAL2 (Figure 3). Complete protein extracts of control platelets showed a large protein that corresponds to full-length NBEAL2 (302 kDa) (Figure 3A,B). Platelets of the patient with the homozygous splice mutation that would result in the predicted M1908X showed a strongly reduced, but not absent expression of full-length NBEAL2. There was also the presence of a band with a lower molecular weight of about 150 kDa, that could represent a cleaved truncated product, though we did not confirm this by mass spectrometry (Figure 3A,B). The homozygous W2480X NBEAL2 variant is predicted to result in a truncated protein of 27 kDa shorter than wild-type NBEAL2, a small difference that is probably not detectable by gel electrophoresis of such a large protein. However, platelets from this patient clearly expressed very low NBEAL2 levels (Figure 3A,B). It was surprising to observe that platelets from the patients with the D218Y or D218G GATA1 variants expressed no NBEAL2 (Figure 3A,B). An antibody for integrin β3 was used as loading control, as we previously found that integrin β3 expression was relatively normal for GATA1-deficient platelets using flow cytometry.13 NBEAL2 blots were performed with a rabbit polyclonal (Figure 3A; epitope against amino acids 1865-1939) and a rabbit monoclonal (Figure 3B; epitope against amino acids 1-100) NBEAL2 antibody with comparable results. Only for the patients with NBEAL2 W2480X and GATA1 D218Y was sufficient platelet extract available to triplicate the blots and quantify expression (Figure 3C). Quantification showed a marked decrease in NBEAL2 expression for platelets with the NBEAL2 W2480X variant, while NBEAL2 was not detected in GATA1 D218Y deficient platelets. We also performed immunoblot analysis using total platelet extracts from hemizygous male and homozygous female Gata1-deficient mice and compared them to wild-type mice (Figure 3D,E). Equal amounts of integrin β3 positive platelet extracts were loaded. NBEAL2 was again strongly reduced for the Gata1-deficient mice that were previously 702

described by others to display macrothrombocytopenia, defective megakaryopoiesis and reduced α-granule proteins in MKs.22,30,31 Female carriers of GATA1 variants are asymptomatic, and our earlier studies revealed skewed X inactivation and no mutant GATA1 RNA in platelets from the D218Y carrier, while the D218G carrier had mild skewing of X inactivation with the presence of the D218G mutation in platelet RNA.14 We also performed immunoblot analysis to compare NBEAL2 protein expression in platelets of female carriers of GATA1 D218G and D218Y. Interestingly, no difference in NBEAL2 expression was found for D218Y, while D218G carriers had slightly lower levels of NBEAL2 (Online Supplementary Figure S2). This decrease in NBEAL2 expression is not enough to cause a phenotype as these carriers have a normal platelet count and volume.13,14 To support the absence of protein degradation in these samples, blots were performed for FLN and GPIbα (Online Supplementary Figure S2).

Identification of a long-distance NBEAL2 enhancer region GATA1 is an important transcription factor that binds the sequence WGATAR in regulatory elements of many genes important for erythropoiesis and megakaryopoiesis.32 Our data clearly showed that GATA1 could regulate NBEAL2 expression and therefore, we analyzed the NBEAL2 chromosomal region using ENCODE data.33 Figure 4A shows potential GATA1 and GATA2 binding sites (BS) that are present in the chromosomal NBEAL2 region. Interestingly, some of these binding sites are clustered in a region that is located 31 kb upstream of the NBEAL2 gene, within an intronic region of the nearby gene CCDC12 (Chr3:46988970-46989880). This region is covered by a H3K4Me1 histone mark that is typically associated with an active enhancer. We only subtracted ChIP sequencing data from ENCODE that were determined for the myelogenous leukemia cell line K562 (Figure 4A, in blue), known to express GATA1 and GATA234 and epidermal keratinocytes (Normal Human Epidermal Keratinocytes (NHEK), Figure 4A, in purple), included as non-blood cells. A H3K4Me3 mark covers the NBEAL2 promotor region that contains multiple GATA1 and GATA2 binding sites. The layered H3K27Ac track shows levels of enrichment of the H3K27Ac histone mark, which is often found near active regulatory elements, and this peak overlaps the potential NBEAL2 enhancer. Chromatin interaction analysis with paired-end tag sequencing (or ChIA-PET) data for K562 cells further shows a physical interaction between this enhancer region and the NBEAL2 promotor (Figure 4A). The final lane displays the 100 vertebrates basewise conservation by PhyloP. In addition, data were subtracted from a recent study that determined the genome-wide binding sites for the 5 key hematopoietic transcription factors, GATA1, GATA2, RUNX1, FLI1, and TAL1/SCL, in primary human MKs.35 Interestingly, the NBEAL2 enhancer region is covered with ChIP sequencing peaks that are specific for GATA1 and GATA2 (Figure 4B). The exact location of these 2 peaks was also marked in Figure 4A at the top panel. The first binding peak contains 3 potential GATA binding sites of which 2 have the reverted sequence (referred to as BS-1, BS-2, BS-3) while the 2nd peak contains 2 potential GATA binding sites (referred to as BS-4 and BS-5) (Figure 5A). haematologica | 2017; 102(4)


GATA1 controls NBEAL2 expression

NBEAL2 enhancer activity using a luciferase reporter assay The two enhancer peaks were cloned as overlapping fragments (BS-1-2-3 and BS-4-5) in the pGL3 promotor vector to obtain luciferase reporter constructs (Figure 5A). Polymerase chain reaction (PCR) mutagenesis was used to obtain the following mutant luciferase reporter constructs: BS-M1, BS-M2, BS-M3 and BS-M4-M5 that lack the GATA binding site and will not bind to GATA1/2 (Figure 5A). We compared the enhancer activity of the different luciferase constructs in K562 that express high GATA1 and lower GATA2 levels (Figure 5B). Both enhancer fragments BS-12-3 and BS-4-5 were able to significantly increase luciferase expression and mutagenesis of only GATA BS-2, and BS-3 blocked this enhancing activity. The mutants BS-

1, BS-4 and BS-5 did not significantly change the enhancing activity. The combination of BS-1-2-3 and BS-4-5 showed a cumulative effect on the enhancer activity and the combination of BS-1-M2-3 and BS-M4-M5 blocked this enhancer activity (Figure 5C).

DNA binding assay confirms binding of GATA to NBEAL2 enhancer We next performed a DNA binding assay using biotinlabelled DNA fragments that cover the two NBEAL2 enhancer peaks (again referred to as BS-1-2-3 and BS-4-5). These biotinylated DNA fragments, bound to superparamagnetic streptavidin beads, were incubated with NE from GATA1 or GATA2 overexpressing HEK293 cells. The enhancer fragment BS-1-2-3 interacts with GATA1 and

A

B

C

D

haematologica | 2017; 102(4)

Figure 6. DNA binding assays for NBEAL2 enhancer region with GATA1 and GATA2 and the effect of GATA1 depletion on NBEAL2 expression. (A) DNA fragments BS1-2-3 and BS-4-5 were labeled with biotin and bound to superparamagnetic streptavidin beads. NE were used from GATA1 or GATA2 overexpression HEK293 cells. (left) Immunoblots for GATA1, GATA2 and STAT5 were performed on proteins bound to BS-1-2-3. STAT5 served as negative control. The unbound fraction is also shown. (Right) Immunoblots for GATA1 and GATA2 were performed on proteins bound to BS-4-5 and BS-M4-M5. GATA1 and GATA2 bind the wild-type, but not the mutant DNA fragments. (B) qRT-PCR was performed on K562 cells with a GATA1 siRNA depletion to determine the expression of GATA1 (left) and NBEAL2 (right) relative to GADPH after 24 and 48 hours. (left) **P=0.0073, t-test; (right) ***P=0.0004, t-test. (C) (left) Immunoblots analysis for NBEAL2 and integrin β3 on protein lysates collected after 48h. (Right) Quantification of NBEAL2 expression. **P=0.0050, t-test. (D) Immunoprecipitation using GST control beads and GST-GATA1 NF beads and protein lysates from HEK293 cells transfected with the long and short GFI1B isoforms. These results are representative for three independent experiments. Control GST-beads (only GST sequence) and non-transfected (NT) HEK293 cells were used as negative controls. SiRNA: small interfering RNA; NE: nuclear extracts; RNA: ribonucleic acid; GADPH: glyceraldehyde 3-phosphate dehydrogenase; GST: Glutathione S-Transferase; BS: binding sites; NF: N-terminal finger.

703


A. Wijgaerts et al.

GATA2, but not with STAT5 for which there is no recognition site in the enhancer region (Figure 6A). Also the 2nd peak BS-4-5 interacts strongly with GATA1 and GATA2 while mutagenesis of both GATA BS-4 and BS-5 in this fragment completely inhibits these interactions (Figure 6B). However, as mutagenesis of BS-4 and BS-5 did not change the enhancer activity (Figure 5C), these GATA sites might not be important for this enhancer region.

siRNA mediated GATA1 depletion in K562 cells results in decreased NBEAL2 We performed a GATA1 knockdown in K562 cells using SMARTpool GATA1 siRNA and compared it to cells transfected with SMARTpool control siRNA (Figure 6B). After 24h, qRT-PCR showed a significant reduction in GATA1 expression while NBEAL2 expression remained normal. However, after 48h a significantly decreased NBEAL2 expression was observed. This was confirmed by using immunoblot analysis, with strongly decreased NBEAL2 levels 48h after transfection (Figure 6C).

GATA1 binding to GFI1B FOG1 is known as a transcriptional co-regulator of GATA1 via interaction with its zinc fingers and the N-terminal GATA1 zinc finger (NF).13,14 We hypothesized that the zinc fingers of GFI1B might also interact with this NF GATA1 domain, as a strong phenotype homology was noticed between GATA1 and GFI1B macrothrombocytopenia. Therefore, a pull-down assay was performed that showed a physical interaction between the GATA1 NF and the large and short GFI1B isoform (Figure 6D).

Discussion Important similarities, but also differences exist between clinical phenotypes caused by NBEAL2, GATA1 and GFI1B germline defects in humans.36 They share the presence of large platelets with α-granule defects, bone marrow fibrosis with the presence of emperipolesis, platelet dysfunction and different degrees of bleeding. However, these diseases have a different mode of inheritance (recessive, X-linked and dominant for NBEAL2, GATA1 and GFI1B defects, respectively). More importantly, GATA1 and GFI1B, but not NBEAL2 defects are characterized by red blood cell abnormalities, and the severity of their α-granule defect is also different. Quantification of αgranule numbers for two NBEAL2 and two GATA1-deficient patients using EM indeed showed decreased α-granule numbers for all, but also a significant difference between both groups with almost no α-granules for NBEAL2 defects. Nevertheless, because of the important resemblance in the α-granule defect, we hypothesized that the transcription factor GATA1 could regulate NBEAL2 expression during megakaryopoiesis. Analogous to the recent findings from Di Buduo et al.29 who studied megakaryopoiesis using blood stem cells from GPS patients, stem cells from GATA1-deficient patients showed normal MK differentiation, but a severe reduction in proplatelet formation with decreased αgranules. Abnormal MK maturation with severely impaired cytoplasmic maturation, including fewer platelet-specific granules, has been described for Gata1deficient mice that lack an upstream regulatory region that controls Gata1 expression in MKs.22,37 Moreover, 704

reduced proplatelet and α-granule formation have also been described for V205G Gata1 transgenic mice that phenocopy patients with X-linked macrothrombocytopenia, as this mutation blocks binding of the GATA1 NF to FOG1.38 The exact pathogenic mechanism of missense variants in GATA1 NF is not really understood, but our previous in vitro binding assays with D218G and D218Y showed normal direct interaction with DNA and decreased interactions with FOG1.13,14 However, more recent murine erythroid cell-based studies could only confirm the defective interaction between FOG1 and GATA1 D218Y, while D218G was shown to interfere with the GATA1-TAL1 cofactor complex.39 If GATA1 D218G and D218Y are predicted to have different effects on DNA binding, they both resulted in abnormal NBEAL2 expression. Moreover, GATA1 depletion via deletion of a MK-specific enhancer in mice or using siRNA also resulted in decreased NBEAL2 expression. Therefore, our studies indicate that NBEAL2 expression can be regulated by changes in GATA1 levels or via functional changes in GATA1 NF. However, more studies are needed to evaluate how these GATA1 NF variants could change direct or indirect DNA binding in gene promoters and enhancer regions. In addition, for GFI1B patients, bone marrow MKs showed extensive peripheral cytoplasm with irregular proplatelets and were largely devoid of cell organelles,12 while more recent in vitro differentiation assays with mutant GFI1B constructs in stem cells showed abnormal proplatelet formation.40 Further studies are needed to discover if GATA1-GFI1B co-transcriptional complexes could regulate NBEAL2 expression during megakaryopoiesis. To our knowledge, it was not yet known that NBEAL2 is a GATA1 target gene. GATA1 drives many important target genes for red blood cell and MK formation such as the globin genes, EPO, EPOR, GATA2, NFE2, GP1BA, GP1BB, PF4, MPL and others.37,41 Interestingly, a BernardSoulier syndrome patient with macrothrombocytopenia was described with a variant in the promoter of the GP1BB gene altering GATA1 binding together with a 22q11 deletion (involving the GP1BB gene) on the other allele.42 This could mean that large platelets in GATA1 patients are due to defective GP1BB expression. This also implicates that the macrothrombocytopenia phenotype caused by GATA1 deficiency is complex and can be the result of many genes that show aberrant GATA1-drive expression, but part of the overall phenotype and especially the α-granule defect could be due to NBEAL2 loss. Moreover, a long-distance enhancer region 31 kb upstream of the NBEAL2 gene was discovered in data from ENCODE for K562 cells that express high levels of GATA1 and whole genome ChIP sequencing data for GATA1 and GATA2 binding sites using primary MKs (Figure 4).35 Typical WGATAR recognition sequences (GATA1/2 BS) are present in this region, and a physical interaction between this enhancer region and the NBEAL2 promoter was predicted based on ChIA-PET experiments using K562 cells.43 It is known that long-distance enhancers come into close proximity of target promoters through looping.44 Transcription factors and their complexes mediate such enhancer-promoter loop formation. The enhancer-gene loop is necessary for transcriptional upregulation, but how the looping changes the transcriptional output is still unclear. Chromatin looping has been shown to be relevant for the reactivation of globin genes,45 haematologica | 2017; 102(4)


GATA1 controls NBEAL2 expression

but examples of such complex transcriptional regulation via interactions between enhancers and promoters important for gene expression during megakaryopoiesis remains largely unstudied. Insights into such mechanisms would be highly relevant to understand genome-wide association study (GWAS) findings that showed associations of mostly noncoding variants with changes in platelet count and volume.46 It might therefore be that the noncoding variant influences expression of a long-distance gene rather than the closest gene. In the study herein, the NBEAL2 enhancer is actually located within an intron of another gene, CCDC12 (Figure 4). We confirmed the binding of GATA1 and GATA2 to this NBEAL2 enhancer region with a DNA binding assay, and showed that in luciferase reporter assays it functions as an enhancer element that depends mainly on GATA BS-2 and BS-3. Such GATA binding sites can of course be occupied by GATA1 or GATA2. GATA2 is essential for the maintenance of HSCs and progenitor cells, whereas GATA1 drives the differentiation of HSCs into megakaryopoiesis. GATA1 represses GATA2 transcription, and this involves GATA1-mediated displacement of GATA2 from chromatin, a process that is

References 1. Blair P, Flaumenhaft R. Platelet alpha-granules: basic biology and clinical correlates. Blood Rev. 2009;23(4):177-189. 2. Italiano JE Jr, Battinelli EM. Selective sorting of alpha-granule proteins. J Thromb Haemost. 2009;7(1):173-176. 3. Nurden AT, Nurden P. Should any genetic defect affecting -granules in platelets be classified as gray platelet syndrome? Am J Hematol. 2016;91(7):714-718. 4. Raccuglia G. Gray platelet syndrome. A variety of qualitative platelet disorders. Am J Med. 1971;51(6):818-828. 5. Albers CA, Cvejic A, Favier R, et al. Exome sequencing identifies NBEAL2 as the causative gene for gray platelet syndrome. Nat Genet. 2011;43(8):735-737. 6. Gunay-Aygun M, Falik-Zaccai TC, Vilboux T, et al. NBEAL2 is mutated in gray platelet syndrome and is required for biogenesis of platelet -granules. Nat Genet. 2011; 43(8):732-734. 7. Kahr WH, Hinckley J, Li L, et al. Mutations in NBEAL2, encoding a BEACH protein, cause gray platelet syndrome. Nat Genet. 2011;43(8):738-740. 8. Cullinane AR, Schäffer AA, Huizing M. The BEACH is hot: a LYST of emerging roles for BEACH-domain containing proteins in human disease. Traffic. 2013; 14(7):749-766. 9. Introne W, Boissy RE, Gahl WA. Clinical, molecular, and cell biological aspects of Chediak-Higashi syndrome. Mol Genet Metab. 1999;68(2):283-303. 10. Castermans D, Volders K, Crepel A, et al. SCAMP5, NBEA and AMISYN: three candidate genes for autism involved in secretion of large dense-core vesicles. Hum Mol Genet. 2010;19(7):1368-1378. 11. Tubman VN, Levine JE, Campagna DR, et al. X-linked gray platelet syndrome due to a GATA1 Arg216Gln mutation. Blood. 2007;109(8):3297-3299. 12. Monteferrario D, Bolar NA, Marneth AE, et

haematologica | 2017; 102(4)

13.

14.

15.

16.

17.

18. 19.

20.

21.

called a GATA switch.47 It is known that GATA switches play an important role in the differentiation of blood stem cells.48 Such a GATA switch can occur for the GATA building sites in the enhancer of the NBEAL2 gene; further studies are needed to unravel this exact mechanism as this might identify the precise moment that NBEAL2 is expressed during megakaryopoiesis. In conclusion, the study herein provides an explanation for the phenotypic resemblance between NBEAL2 and GATA1 defective macrothrombocytopenia with paucity of α-granules, as GATA1 is important for NBEAL2 expression during megakaryopoiesis via a long-distance enhancer. Funding KF is supported by the Fund for Scientific Research-Flanders (FWO-Vlaanderen, Belgium, G.0B17.13N] and by the Research Council of the University of Leuven (BOF KU Leuven‚ Belgium, OT/14/098]. CVG is holder of the Bayer and Norbert Heimburger (CSL Behring) Chairs. MRT was supported by a Fellowship from the European Haematology Association and the British Heart Foundation (PG/13/77/30375).

al. A dominant-negative GFI1B mutation in the gray platelet syndrome. N Engl J Med. 2014;370(3):245-253. Freson K, Devriendt K, Matthijs G, et al. Platelet characteristics in patients with Xlinked macrothrombocytopenia because of a novel GATA1 mutation. Blood. 2001;98(1):85-92. Freson K, Matthijs G, Thys C, et al. Different substitutions at residue D218 of the X-linked transcription factor GATA1 lead to altered clinical severity of macrothrombocytopenia and anemia and are associated with variable skewed X inactivation. Hum Mol Genet. 2002;11(2):147152. Balduini CL, Pecci A, Loffredo G, et al. Effects of the R216Q mutation of GATA-1 on erythropoiesis and megakaryocytopoiesis. Thromb Haemost. 2004;91(1): 129-140. Balduini CL, De Candia E, Savoia A. Why the disorder induced by GATA1 Arg216Gln mutation should be called "X-linked thrombocytopenia with thalassemia" rather than "X-linked gray platelet syndrome". Blood. 2007;110(7):2770-2771. Stevenson WS, Morel-Kopp MC, Chen Q, et al. GFI1B mutation causes a bleeding disorder with abnormal platelet function. J Thromb Haemost. 2013;11(11):2039-2047. Stevenson WS, Morel-Kopp MC, Ward CM. Platelets are not all gray in GFI1B disease. Clin Genet. 2015;87(3):299. Simeoni I, Stephens JC, Hu F, et al. A highthroughput sequencing test for diagnosing inherited bleeding, thrombotic, and platelet disorders. Blood. 2016;127(23):2791-2803. Freson K, Peeters K, De Vos R, et al. PACAP and its receptor VPAC1 regulate megakaryocyte maturation: therapeutic implications. Blood. 2008;111(4):1885-1893. Turro E, Greene D, Wijgaerts A, et al. A dominant gain-of-function mutation in universal tyrosine kinase SRC causes thrombocytopenia, myelofibrosis, bleeding, and bone pathologies. Sci Transl Med. 2016;8(328):328-330.

22. Shivdasani RA, Fujiwara Y, McDevitt MA, Orkin SH. A lineage-selective knockout establishes the critical role of transcription factor GATA-1 in megakaryocyte growth and platelet development. EMBO J. 1997;16(13):3965-3973. 23. Albers CA, Paul DS, Schulze H, et al. Compound inheritance of a low-frequency regulatory SNP and a rare null mutation in exon-junction complex subunit RBM8A causes TAR syndrome. Nat Genet. 2012;44(4):435-439. 24. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(- Ct)) method. Methods. 2001;25:402–408. 25. Lan X, Witt H, Katsumura K, et al. Integration of Hi-C and ChIP-seq data reveals distinct types of chromatin linkages. Nucleic Acids Res. 2012;40(16):7690-7704. 26. Chen L, Kostadima M, Martens JH, et al. Transcriptional diversity during lineage commitment of human blood progenitors. Science. 2014;345(6204):12511033. 27. Bottega R, Pecci A, De Candia E, et al. Correlation between platelet phenotype and NBEAL2 genotype in patients with congenital thrombocytopenia and -granule deficiency. Haematologica. 2013;98(6):868874. 28. Trainor CD, Omichinski JG, Vandergon TL, Gronenborn AM, Clore GM, Felsenfeld G. A palindromic regulatory site within vertebrate GATA-1 promoters requires both zinc fingers of the GATA-1 DNA-binding domain for high-affinity interaction. Mol Cell Biol. 1996;16(5):2238-2247. 29. Di Buduo CA, Alberelli MA, Glembostky AC, et al. Abnormal proplatelet formation and emperipolesis in cultured human megakaryocytes from gray platelet syndrome patients. Sci Rep. 2016;6:23213. 30. Zingariello M, Fabucci ME, Bosco D, et al. Differential localization of P-selectin and von Willebrand factor during megakaryocyte maturation. Biotech Histochem. 2010;85(3):157-170. 31. Zetterberg E, Verrucci M, Martelli F, et al.

705


A. Wijgaerts et al.

32. 33.

34.

35.

36.

706

Abnormal P-selectin localization during megakaryocyte development determines thrombosis in the gata1low model of myelofibrosis. Platelets. 2014;25(7):539547. Harigae H. GATA transcription factors and hematological diseases. Tohoku J Exp Med. 2006;210(1):1-9. ENCODE Project Consortium. An integrated encyclopedia of DNA elements in the human genome. Nature. 2012; 489(7414):57-74. Fujiwara T, O'Geen H, Keles S, Blahnik K, Linnemann AK, Kang YA, Choi K, Farnham PJ, Bresnick EH. Discovering hematopoietic mechanisms through genome-wide analysis of GATA factor chromatin occupancy. Mol Cell. 2009;36(4):667-681. Tijssen MR, Cvejic A, Joshi A, et al. Genome-wide analysis of simultaneous GATA1/2, RUNX1, FLI1, and SCL binding in megakaryocytes identifies hematopoietic regulators. Dev Cell. 2011;20(5):597-609. Kies C, Harms JM. Copper absorption as affected by supplemental calcium, magnesium, manganese, selenium and potassium. Adv Exp Med Biol. 1989;258:45-58.

37. Vyas P, Ault K, Jackson CW, Orkin SH, Shivdasani RA. Consequences of GATA-1 deficiency in megakaryocytes and platelets. Blood. 1999;93(9)2867-2875. 38. Shimizu R, Ohneda K, Engel JD, Trainor CD, Yamamoto M. Transgenic rescue of GATA-1-deficient mice with GATA-1 lacking a FOG-1 association site phenocopies patients with X-linked thrombocytopenia. Blood. 2004;103(7):2560-2567. 39. Campbell AE, Wilkinson-White L, Mackay JP, Matthews JM, Blobel GA. Analysis of disease-causing GATA1 mutations in murine gene complementation systems. Blood. 2013;121(26):5218-5227. 40. Kitamura K, Okuno Y, Yoshida K, et al. Functional characterization of a novel GFI1B mutation causing congenital macrothrombocytopenia. J Thromb Haemost. 2016 April 28. [Epub ahead of print] 41. Ferreira R, Ohneda K, Yamamoto M, Philipsen S. GATA1 function, a paradigm for transcription factors in hematopoiesis. Mol Cell Biol. 2005;25(4):1215-1227. 42. Ludlow LB, Schick BP, Budarf ML, et al. Identification of a mutation in a GATA

43.

44. 45.

46.

47.

48.

binding site of the platelet glycoprotein Ibbeta promoter resulting in the BernardSoulier syndrome. J Biol Chem. 1996; 271(36):22076-22080. Heidari N, Phanstiel DH, He C, et al. Genome-wide map of regulatory interactions in the human genome. Genome Res. 2014;24(12):1905-1917. Krivega I, Dean A. Enhancer and promoter interactions-long distance calls. Curr Opin Genet Dev. 2012;22(2):79-85. Deng W, Rupon JW, Krivega I, et al. Reactivation of developmentally silenced globin genes by forced chromatin looping. Cell. 2014;158(4):849-860. Gieger C, Radhakrishnan A, Cvejic A, et al. New gene functions in megakaryopoiesis and platelet formation. Nature. 480(7376):201-208. Bresnick EH, Lee HY, Fujiwara T, Johnson KD, Keles S. GATA switches as developmental drivers. J Biol Chem. 2010; 285(41):31087-31093. Tian T, Smith-Miles K. Mathematical modeling GATA-switching for regulating the differentiation of hematopoietic stem cell. BMC Syst Biol. 2014;8(1):S8.

haematologica | 2017; 102(4)


ARTICLE

Acute Myeloid Leukemia

CD244 maintains the proliferation ability of leukemia initiating cells through SHP-2/p27kip1 signaling

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Feifei Zhang,1* Xiaoye Liu,1* Chiqi Chen,1 Jun Zhu,2 Zhuo Yu,1 Jingjing Xie,3 Li Xie,1 Haitao Bai,2 Yaping Zhang,1 Xia Fang,4 Hao Gu,1 Chun Wang,2 Wei Weng,1 Cheng Cheng Zhang,5 Guo-Qiang Chen,1 Aibing Liang4 and Junke Zheng1

Hongqiao International Institute of Medicine, Shanghai Tongren Hospital, Key Laboratory of Cell Differentiation and Apoptosis of Chinese Ministry of Education, Faculty of Basic Medicine, Shanghai Jiao Tong University School of Medicine, China; 2 Department of Hematology, The 1st Peopleâ&#x20AC;&#x2122;s Hospital, Shanghai Jiao Tong University School of Medicine, China; 3Binzhou Medical University, Taishan Scholar Immunology Program, Yantai, China; 4Department of Hematology, Tongji Hospital of Tongji University School of Medicine, Shanghai, China and 5Departments of Physiology and Developmental Biology, UT Southwestern Medical Center, Dallas, TX, USA 1

Haematologica 2017 Volume 102(4):707-718

*FZ and XL contributed equally to this work

ABSTRACT

T

argeting leukemia initiating cells is considered to be an effective way to cure leukemia, for which it is critical to identify novel therapeutic targets. Herein, we demonstrate that CD244, which was initially reported as a key regulator for natural killer cells, is highly expressed on both mouse and human leukemia initiating cells. Upon CD244 knockdown, human leukemia cell lines and primary leukemia cells have markedly impaired proliferation abilities both in vitro and in vivo. Interestingly, the repopulation ability of both mouse and human hematopoietic stem cells is not impaired upon CD244 knockdown. Using an MLL-AF9-induced murine acute myeloid leukemia model, we show that leukemogenesis is dramatically delayed upon CD244 deletion, together with remarkably reduced Mac1+/c-Kit+ leukemia cells (enriched for leukemia initiating cells). Mechanistically, we reveal that CD244 is associated with c-Kit and p27 except for SHP-2 as previously reported. CD244 co-operates with c-Kit to activate SHP-2 signaling to dephosphorylate p27 and maintain its stability to promote leukemia development. Collectively, we provide intriguing evidence that the surface immune molecule CD244 plays an important role in the maintenance of stemness of leukemia initiating cells, but not in hematopoietic stem cells. CD244 may represent a novel therapeutic target for the treatment of acute myeloid leukemia.

Correspondence: chengq@shsmu.edu.cn/ zhengjunke@shsmu.edu.cn/ lab7182@tongji.edu.cn Received: June 27, 2016. Accepted: January 13, 2017. Pre-published: January 25, 2017. doi:10.3324/haematol.2016.151555 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/707 Š2017 Ferrata Storti Foundation

Introduction It was suggested that leukemia initiating cells (LICs) or leukemia stem cells (LSCs) are responsible for initiation, development and relapse of leukemia. Identification of novel theraputic targets specific to LICs is the key for the eradication of leukemia. We previously identified that an immune inhibitory receptor, LILRB2, played an essential role in the stemness maintenance of LICs and hematopoietic stem cells (HSCs).1 Studies from other groups also suggest that targeting certain surface immune molecules of LICs may be an attractive way to block leukemogenesis.2,3 However, an ideal therapeutic target for leukemia treatment should only be essential for maintaining the pool of LICs, but not for HSCs. We hypothesized that other surface immune molecules, including both inhibitory and stimulatory recephaematologica | 2017; 102(4)

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

707


F. Zhang et al. tors, may function differentially between HSCs and LICs.4 It is urgent to understand such molecules for the eradication of LICs. CD244 is one of the immune molecules initially identified on natural killer (NK) cells, a subset of CD8 T cells and monocytes. It has been reported that CD244 is also highly expressed on human HSCs, but may not (or may rarely) be expressed on mouse HSCs,5-7 although its function is still not clear. CD244 is a member of the CD2 subset of the immunoglobulin superfamily with four immune-receptor based tyrosine switch motifs (ITSM)8 which can function as an activating or inhibitory receptor by alternative adaptor recruitments including SAP, EAT-2, SHP-2 and SHP-1. Human CD244 interacts with SAP to activate NK cells.9,10 However, in the absence of SAP, CD244 associates with the inhibitory phosphatase SHP-2 or SHP-1 to function as an inhibitory receptor.11,12 CD244 is also known to associate with PI3K and SHIP1 as this receptor also signals through the inositol phospholipid signaling pathway.13-15 Nevertheless, the functions of CD244 in leukemia development and the potential targets involved in CD244/SHP2 signaling remain largely unknown. In this study, we showed that knockdown of CD244 by shRNAs induced a marked decrease of the proliferation abilities in both human AML cell lines and primary LICs, but not in mouse and human HSCs. Furthermore, CD244 deletion remarkably delayed leukemogenesis and depleted LICs in a murine MLL-AF9-transduced AML model. CD244 co-operates with c-Kit and initiates downstream SHP-2/p27 signaling to manipulate the activities of LICs. Therefore, CD244 may be an ideal target for the eradication of acute myeloid LICs.

(1000 rad) C57BL/6 mice by retro-orbital injection. Indicated YFP+ bone marrow cells from primary transplanted mice were further infused into recipient mice for secondary transplantation or limiting dilution analysis. Flow cytometry and cell cycle analyses were performed as we described previously.1 For analysis of lineages and LICs, either peripheral blood or bone marrow cells were stained with anti-mouse Mac-1-APC, anti-mouse Gr-1-PE, antimouse CD3-APC, anti-mouse B220-PE, or anti-mouse c-Kit-PE antibodies (eBioscience). CD244 expression on human LICs or HSCs were labeled by the antibodies against human CD45 (FITC), Lineage marker (APC), CD34 (eFluor 450), CD38 (PE-Cy7), CD90 (PE-Cy5.5), CD45RA (PE), CD123 (PE) and CD244 (Biotin). Expression of CD244 and c-Kit on mouse or human LICs (or cell lines) were detected by anti-mouse CD244-PE or human CD244APC and anti-human c-Kit-PE antibodies (eBioscience). Cell cycle status was measured with Ki-67/7-AAD (or Hoechst 33342) staining (BD Pharmingen) according to the manufacturer’s instructions.

Study approval Bone marrow mononuclear cells were obtained from the patients following diagnostic work at the Department of Hematology at Xinhua Hospital or the 1st People’s Hospital; human cord blood were obtained from the Department of Gynaecology and Obstetrics at the 6th People’s Hospital, Shanghai Jiao Tong University School of Medicine. Written informed consent was obtained from all of the patients and all the procedures were approved by the Ethics Committee for Medical Research (IRB) at Shanghai Jiao Tong University School of Medicine. Methods related to mouse information, western blotting and co-immunoprecipitation, quantitative RT-PCR, colony forming unit assays, bone marrow transplantation and statistical analyses are available in the Online Supplementary Appendix.

Results Methods Lentivirus construction, infection and in vivo xenograft The lentiviral vector PLL3.7 was used to express shRNAs designed to target CD244 (sequences are listed in Online Supplementary Table S1). Using a calcium phosphate transfection method, lentivirus constructs together with the packaging plasmids of pSPAX2 and pMD2G were mixed and transfected into 293T cells. Lentiviruses were used for the following infection on human leukemia cell lines, human cord blood CD34+ HSCs or acute myeloid leukemia (AML) samples. Human cord blood CD34+ HSCs were purified by using CD34 enrichment Kit (Miltenyi). Details of the clinical sample information are listed in Online Supplementary Table S2. Total AML cells from Samples #12 to #16, #19 and #20 (subtype, M5) were used for infection, in vitro colony assay and in vivo transplantation. MV4-11 cells (2.5x106), human cord blood CD34+ HSCs (2x103) or primary AML cells (2x106) were resuspended in 200 μL or 50 μL (for primary AML cells) PBS and transplanted into sublethally irradiated (250 rad) NOD-SCID mice by either retro-orbital or intra-tibial injection. All the mice were sacrificed for determination of engraftment at 2-3 months post transplantation.

Retroviral infection, transplantation and flow cytometric analysis MLL-AF9-expressing retroviruses were produced in 293T cells with an MSCV-MLL-AF9-IRES-YFP encoding plasmid.16 Lin– fetal liver cells were isolated from wild-type (WT) and CD244 knockout (KO) mice and infected with MLL-AF9 retroviruses by two rounds of spinoculation in the presence of 4 μg/mL polybrene. Infected cells (200,000) were transplanted into lethally irradiated 708

CD244 is required for the proliferation of both human leukemia cell lines and acute myeloid LICs To identify novel surface immune molecules (SIMs) that regulate the stemness of HSCs or LICs, we screened approximately 30 potential candidates that were known to be expressed on immune cells.17 As shown in Online Supplementary Figure S1A (mean fluorescence intensity) and Figure 1A (frequencies of SIM+ cells), we found that several immune receptors, including IREM-1, CD244 and JAM3, were highly expressed on both human CD34+ hematopoietic stem/progenitor cells (HSPCs) and AML cells transduced with MLL-AF9 oncogene in human CD34+ HSCs, MA9 cells.18 Intriguingly, one of these molecules, CD244, which is critical for functions of NK cells, was expressed at the highest level on both HSPCs and AML cells. To elucidate the roles of CD244 in human AML, we first examined the protein levels of CD244 on different human AML cell lines. Most of the AML cell lines expressed CD244, including Kasumi-1 (M2), NB4 (M3), HL-60 (M3), THP-1 (M5), U937 (M5), MV4-11 (M5) and HEL (M6) (Figure 1B). We then constructed several shRNAs to specifically knockdown CD244 to evaluate its roles in cell proliferation. shCD244#1 and shCD244#2 efficiently reduced the CD244 levels as compared with that in scrambled cells, respectively, measured by either flow cytometric analysis or quantitative RT-PCR (Figure 1C and D). Since shCD244#2 had the highest knockdown efficiency, it was mainly used for the following experiments to unravel the functions of CD244 in human leukemia cells. haematologica | 2017; 102(4)


CD244 in leukemogenesis regulation

We then down-regulated the expression of CD244 in MV4-11 cells with shCD244#2 and found that these cells had notable decreased proliferation ability in vitro even two days upon CD244 knockdown (Figure 1E). Meanwhile, shCD244#1 had a similar effect when it was used to knock down CD244 in MV4-11 cells (Online Supplementary Figure S1B). We also observed similar effects

in other CD244-knockdown leukemia cell lines including HL-60, U937 and THP-1 (Online Supplementary Figure S1CE). We then injected CD244-knockdown MV4-11 cells into NOD-SCID mice and demonstrated that loss of CD244 led to a significantly decreased engraftment compared to that in the control (6.29±4.35% vs. 63.90±1.21%) (Figure 1F).

A

C

B

D

H

E

F

I

G

J

K

Figure 1. CD244 is required for the proliferation of both human leukemia cell lines and acute myeloid leukemia initiating cells (LICs). (A) Frequencies of representative surface immune molecules on human cord blood hematopoietic stem/progenitor cells (HSPCs) and MLL-AF9-transduced human MA9 cells were measured by flow cytometric analysis. Cord blood mononuclear cells (MNCs) were used for serving as the control of total cell population. (B) Representative flow cytometric analysis of CD244 expression on different leukemia cell lines including Kasumi-1 (M2), NB4 (M3), HL-60 (M3), THP-1 (M5), U937 (M5), MV4-11 (M5) and HEL (M6) (isotype control, red line). (C and D) Knockdown efficiency of CD244 targeted by scrambled shRNA (Scr), shCD244#1 and shCD244#2 was evaluated by either flow cytometric analysis (C) or quantitative real-time RT-PCR (qRT-PCR) as in (D). (E) Cell numbers were counted at indicated days after infection with CD244-targeting shCD244#2 or scrambled shRNA in a representative experiment (n=3). (F) MV4-11 cells infected with CD244-targeting shCD244#2 and control cells were injected into NOD-SCID mice. Three months later, engrafted leukemia cells in the bone marrow were detected by anti-human CD45 antibodies (n=6-8). (G and H) Representative images or colony numbers of patients’ CD34+ AML cells were examined after infection with CD244-targeting shCD244#2 or scrambled shRNA in a representative experiment (n=3). (I) Human primary leukemia cells infected with CD244-targeting shCD244#2 and scrambled shRNA were injected into NOD-SCID mice. Two months later, engrafted leukemia cells in the bone marrow were evaluated by flow cytometeric analysis in a representative experiment. (J) Repopulation was evaluated in human cord blood CD34+ cells after infection with CD244-targeting shCD244#2 or scrambled shRNA 12 weeks after transplantation (n=5). (K) Knockdown efficiency of CD244 targeted by scrambled shRNA (Scr) and shCD244#2 was evaluated in the stably engrafted GFP+ CB cells by flow cytometric analysis 12 weeks after transplantation (*P<0.05).

haematologica | 2017; 102(4)

709


F. Zhang et al.

To address the function of CD244 in human AML CD34+ cells, we analyzed the levels of CD244 in different types of human AML samples (M2, M3, M4 and M5) (Online Supplementary Table S2) by flow cytometry and found that most of the AML CD34+ cells (enriched for LICs) were positive for CD244. In contrast, most of the tested B-cell acute lymphoblastic leukemia (B-cell ALL) CD34+ cells expressed at a much lower level of CD244 (Online Supplementary Figure S2A and B). Interestingly,

A

CD244 also expressed on some CD34– cells including blasts. Because it has been reported that leukemia blasts have differentially expressed surface markers compared to the residual HPSCs, such as CD123 or CD45RA,19-21 we further examined the CD244 level on the immunophenotypic CD45dimSSClow Lin–CD34+CD38–CD90–CD45RA+ or Lin–CD34+CD38–CD90–CD123+ CD34+ blasts (CD34+ LICs) and CD45dimSSClow Lin–CD34–CD38–CD90–CD45RA+ or CD45dimSSClow Lin–CD34–CD38–CD90–CD123+ CD34–

B

C

D

E

F

G

H

Figure 2. CD244 promotes leukemogenesis in a murine acute myeloid leukemia (AML) model. (A) Representative flow cytometric analysis of CD244 expression in mouse MLL-AF9-transduced wild-type (WT) leukemia cells compared to isotype control. (B) Representative flow cytometric analysis for Mac-1+/c-Kit+ LICs in the bone marrow of primary recipient mice (left). Quantification of the percentage of Mac-1+/c-Kit+ cells from different mice (n=5, right). (C) Representative flow cytometric analysis for detection of YFP+ leukemia cells in peripheral blood of the recipients receiving WT or CD244-null leukemia cells from primary transplantation (left). Quantification of the percentage of leukemia cells (n=5, right). (D) Secondary transplantation of 100 leukemia cells displayed significantly delayed onset of leukemogenesis by CD244-null cells compared to WT cells (n=5, log-rank test). (E) Comparison of the sizes of spleens and livers of the mice transplanted with WT or CD244-null leukemia cells upon secondary transplantation (n=5-7). (F) Histological hematoxylin & eosin staining of AML infiltration in the livers and spleens of mice, as shown in (E). (G) Repopulation was evaluated with WT and CD244-null donor bone marrow cells at indicated time points post transplantation (n=5). (H) Multi-lineages were analyzed in WT and CD244-null donor cells 20 weeks after transplantation (n=5).

710

haematologica | 2017; 102(4)


CD244 in leukemogenesis regulation

blasts (CD34– LICs). The flow cytometric analysis showed that almost all the CD34+ blasts expressed CD244, while CD34– blasts had a significantly lower level of CD244 compared to that of CD34+ ones (Online Supplementary Figure S2C). Interestingly, CD34+ blasts, but not CD34– blasts, had similar CD244 expression level with Lin–CD34+CD38–CD90+CD45RA– or Lin–CD34+CD38CD90+CD123– residual HSCs (Online Supplementary Figure S2D-F). We then knocked down the CD244 levels in several human AML samples with shCD244#1 or shCD244#2. Consistently, the deletion of CD244 led to a notable delay in the growth of CD34+ LICs by shCD244#2 or shCD244#1 (Online Supplementary Figure S3A). Moreover, the in vitro functional colony forming assay revealed a significant declination in colony numbers (Figure 1G and H). We further performed serial re-plating experiments with human AML cells upon the knockdown of CD244 (Online Supplementary Figure S3B and C), which showed that the colony numbers were significantly reduced after serial replating (Online Supplementary Figure S3D). Importantly, when human AML cells were knocked down by shCD244#2 targeting CD244 and transplanted into NODSCID mice, the engraftment was dramatically decreased compared to controls (Figure 1I and Online Supplementary Figure S3E). Since CD244 is also expressed on Lin–CD34+CD38–CD90+CD45RA+ immunophenotypic human HSCs (Online Supplementary Figure S3F), but may not be expressed or may be expressed only in some fractions of mouse HSCs,5-7 we decided to investigate its role in human HSCs as well. Interestingly, no obvious proliferation defects were observed in CD244-knockdown

A

human cord blood HSCs (Online Supplementary Figure S3G). Then CD244-knockdown HSCs were infused into NOD-SCID mice and the repopulation was analyzed at 12 weeks post transplantation, showing that CD244 was not required for the repopulation ability of HSCs (Figure 1J and Online Supplementary Figure S3H and I). We also examined the knockdown efficiency of CD244 in human HSCs after transplantation, where the stably engrafted GFP+ CB cells infected either by shCD244#2 or Scramble shRNA were gated for the analysis of CD244 expression, which showed that CD244 was decreased to around 67.5% of that in the control (Figure 1K). This may also indicate that the lack of phenotype might not be due to an escape mechanism by CD244-expressing HSCs.

CD244 promotes leukemogenesis in a murine AML model We further used a transplantable murine AML model driven by the MLL-AF9 oncogene22 to extensively explore the function of CD244 in leukemia development. We first evaluated the protein levels of CD244 on WT leukemia cells (tagged with YFP) and found that more than 92.9% of them were positive for CD244 as compared to isotype control by flow cytometry (Figure 2A). We further monitored leukemia progression by determining WT and CD244-null (hereafter referred to as KO) circulating leukemia cells in peripheral blood (Online Supplementary Figure S4A), which only expressed myeloid cell markers (Mac-1 and Gr-1), but not lymphoid cell markers (CD3 and B220) (Online Supplementary Figure S4B and C). Surprisingly, we did not find significant differences either

B

D

C

Figure 3. CD244 regulates the proliferation ability of leukemia initiating cells (LICs). (A) Representative flow cytometric analysis for the distribution of wild-type (WT) or CD244-null Mac-1+/c-Kit+ cells (enriched for LICs) in the bone marrow (BM) and spleens (SP) of the recipients upon secondary transplantation (left). Quantification of the percentage of Mac-1+/c-Kit+ cells (n=5-7 for BM and n=3-4 for SP, respectively, right). (B) Representative images for colony forming units for WT and CD244null acute myeloid leukemia (AML) cells in 1st and 2nd plating (left). Colony numbers were compared in WT and CD244-null AML cells of the secondary recipients after serial plating (n=3, right). (C) Stages of cell cycle were determined by using Ki-67 and Hoechst 33342 staining in WT and CD244-null LICs of the recipients upon secondary transplantation (left). Quantitative analysis of the cell cycle in a representative experiment (n=4, right). (D) Representative flow cytometric analysis for apoptosis of WT or CD244-null Mac-1+/c-Kit+ cells in the bone marrow of the recipients upon secondary transplantation (left). Quantification of the percentages of apoptotic Mac-1+/c-Kit+ cells (n=3, right).

haematologica | 2017; 102(4)

711


F. Zhang et al.

for the frequencies of YFP+ leukemia cells or the differentiation status as compared by the percentages of differentiated Mac-1+/Gr-1+ leukemia cells (Online Supplementary Figure S4). There was no significant change in survival between WT and CD244-null leukemic mice as well (Online Supplementary Figure S4D). Nevertheless, we observed a notable decrease of Mac-1+/c-Kit+ LICs in the bone marrow of CD244-null primary recipients than WT ones (35.23±3.52% vs. 45.88±2.45%) (Figure 2B). Thereby, we performed a secondary transplantation to further reveal the effect of CD244 in LICs. We found that CD244-null YFP+ leukemia cells in peripheral blood were dramatically reduced compared to WT counterparts at three weeks post transplantation (23.21±5.98% vs. 64.26±4.42%) (Figure 2C). Most importantly, recipients of MLL-AF9-transduced CD244-null cells had a markedly extended survival (57 vs. 33) (Figure 2D and Online Supplementary Figure S4E). We further performed the limiting dilution assay to evaluate the LIC frequency in CD244-null leukemia cells, which was markedly decreased compared to WT controls (1 in 724 vs. 1 in 59)

A

(Online Supplementary Table S3). Consistently, we also found that the sizes of livers and spleens were much smaller in the mice transplanted with CD244-null AML cells (Figure 2E). This was further confirmed by histological hematoxylin & eosin staining, showing much less infiltration in the mice injected with CD244-null leukemia cells (Figure 2F). These results also indicate the loss of selfrenewal ability in CD244-null LICs. Consistent with our observations in human HSCs, CD244-null mouse HSCs also had normal repopulation ability 20 weeks after transplantation (Figure 2G and H). Because we did not uncover drastic changes in primary recipients, we then mainly focused on the phenotypes in secondary recipients.

CD244 promotes the proliferation of LICs To further characterize the CD244-null AML, we first examined the phenotypic LIC frequencies in the bone marrow and spleens of mice receiving leukemia cells from primary recipients, which was significantly reduced to 60.9% of that in WT counterparts in the bone marrow (27.54±1.60% vs. 45.21±1.47%) (Figure 3A). We also

B

C

D

E F

G

H

Figure 4. CD244 manipulates p27 levels to maintain the proliferation ability of leukemia-initiating cells (LICs). (A) Potential candidates in wild-type (WT) and CD244null LICs examined by qRT-PCR. (B) p27 levels were compared between WT and CD244-null LICs by western blotting analysis (left). Quantification of p27 levels between WT and CD244-null LICs (right). (C) p27 levels were compared between WT and CD244-null LICs by immunofluorescence staining in different sets of experiments. (D) Cytoplasmic and nuclear form of p27 of WT and CD244-null LICs were determined by western blotting. (E and F) p27 was over-expressed in both WT and CD244-null AML cells and transplanted into the recipient mice. Survival (E) and p27 levels (F) were analyzed among the mice receiving WT, CD244-null, p27-overexpressed WT and CD244-null AML cells (n=5-10, log-rank test). (G) Colony forming assay were performed in vitro and representative images for colonies (left) and colony forming unit (CFU) numbers (right) derived from WT, CD244-null, p27-over-expressed WT and CD244-null AML cells were calculated (n=3). (H) Stages of cell cycle were determined by using Ki-67 and Hoechst 33342 staining in panel (G). A representative FACS plot (left) and quantification data were shown (n=3, right).

712

haematologica | 2017; 102(4)


CD244 in leukemogenesis regulation

observed a similar trend in the leukemic spleens although it is not statistically different (Figure 3A). Consistently, we also found that there was a much lower frequency of CD244-null Lin–IL7-R–Sca-1–c-Kit+CD34+FcγR+ L-GMPs (enriched in LICs) compared to that of WT controls (Online Supplementary Figure 4F and G), which indicates

A

CD244 may play a potential role in the maintenance of LICs. An in vitro colony-forming assay also illustrated that the clonogenic potential of CD244-null leukemia cells isolated from secondary recipients was greatly declined to 62.8%, 26.7% and 42.5% of that in WT counterparts upon first-third plating, respectively (Figure 3B), indicating

B

C

E

F

G

H

D

Figure 5. CD244 regulates the p27 stability through SHP-2 signaling. (A) Fc-tagged CD244 and StrepII-tagged p27 were over-expressed in 293T cells, and their lysates were immunoprecipitated by protein A beads, followed by western blotting analysis for p27 and SHP-2. (B) Fc-tagged CD244 and StrepII-tagged p27 were over-expressed in 293T cells, and their lysates were immunoprecipitated by StrepII beads, followed by western blotting analysis for Fc (CD244) and SHP-2. (C) StrepIItagged p27 were over-expressed in 293T cells, and their lysates were immunoprecipitated by SHP-2 or IgG control antibodies, followed by western blotting analysis for p27. (D) The lysates of primary mouse AML cells were immunoprecipitated by SHP-2 or IgG control antibodies, followed by western blotting analysis for the levels of endogenous CD244 and p27. (E) Fc-tagged CD244 and StrepII-tagged p27 were over-expressed in 293T cells, and their lysates were immunoprecipitated by StrepII beads, followed by western blotting analysis for p-SHP-2, SHP-2 and 4G10 (for p-p27). (F) Fc-tagged CD244, StrepII-tagged p27 and shRNA targeting SHP-2 (shSHP2#3) were transfected into 293T cells, and their lysates were immunoprecipitated by StrepII beads, followed by western blotting analysis for p-SHP-2, SHP-2 and 4G10 (for p-p27). The backbone empty vector (EV) was used as the control. (G) Human primary leukemia cells were infected with CD244-targeting shCD244#2, SHP2-targeting shSHP-2#3 and scrambled shRNA. And their lysates were immunoprecipitated by p27 antibody, followed by western blotting analysis for the levels of endogenous p-SHP-2, SHP-2, 4G10 (for p-p27), p27 and CD244. (H) HEL cells were infected with scrambled shRNA, CD244-targeting shCD244#2, or CD244-targeting shCD244#2 and followed by the overexpression of SHP-2 and their lysates were immunoprecipitated by p27 antibody, followed by western blotting analysis for the levels of p-SHP-2, SHP-2, 4G10 (for p-p27), p27 and CD244.

haematologica | 2017; 102(4)

713


F. Zhang et al.

a loss of self-renewal ability upon CD244 deletion. We then further performed a cell cycle analysis and found there was only slightly decreased G0 frequencies (representing a quiescent status) in CD244-null LICs than WT ones (9.65Âą0.63% vs. 12.75Âą0.79%) (Figure 3C). Although there is some evidence to show the loss of quiescence may lead to the exhaustion of LICs,23 the connection between quiescence and stemness in LICs is still controversial since several lines of studies also imply the existence of a subset of actively cycling, non-quiescent AML cells enriched for LIC activities.24 In addition, we also did not observe significant changes in apoptosis between WT and CD244-null LICs (Figure 3D). These results suggest that CD244 may play important roles in the proliferation (or self-renewal) rather than sustaining the quiescence or apoptosis of LICs.

CD244 manipulates p27 levels to maintain the proliferation ability of LICs To understand the underlying mechanisms that affect the proliferation (or self-renewal) ability in CD244-null LICs, we performed quantitative RT-PCR with WT and CD244-null Mac-1+/c-Kit+ LICs. Taking into account the phenotypes in mouse CD244-null and human CD244-knockdown LICs, we mainly focused on the changes of molecules related to cell cycle and self-renewal, which showed that p27 was significantly down-regulated in CD244-null LICs (Figure 4A). We further evaluated p27 levels by western blotting or immunofluorescence staining

A

E

B

C

F

and demonstrated that p27 levels were strikingly reduced in CD244-null LICs (Figure 4B and C). In addition, we also isolated cyptoplasmic proteins from WT and CD244-null LICs and examined the p27 levels by western blot, which showed that there was a remarkable decrease of cytoplasmic p27, but not nuclear form of p27 (Figure 4D). Although the mRNA level of p16 was up-regulated almost 2-fold, the protein level of p16 was not significantly changed in CD244-null LICs (Online Supplementary Figure S5A), which indicates that p16 does not play an important role in the CD244-mediated leukemogenesis. Although it has been reported that p27 acts as a cell cycle inhibitor by interacting with CDK2 or CDK4, and may participate in sustaining the quiescent state in the MLL-leukemia cells,25,26 p27 has also been reported to serve as cytoplasmic oncoprotein in BCR-ABL1 transformed chronic myeloid leukemia to promote cell growth and block apoptosis, rather than its effects on cyclin-dependent kinase (CDKs).27 We also noticed that the p27 was mainly located in cytoplasm and notably reduced in CD244-null Mac1+/c-Kit+ LICs (Figure 4C and D), indicating p27 may exert its additional effects (may enhance proliferation or selfrenewal) on leukemia development. To confirm whether p27 is a direct downstream target for CD244, we overexpressed mouse p27 in both WT and CD244-null leukemia cells and transplanted them into the recipient mice. The mice transplanted with the p27-over-expressed CD244-null AML cells developed leukemia significantly faster than CD244-null control cells, which was comparable to WT counterparts (Figure 4E). Interestingly, the over-

D

G

H

Figure 6. CD244 interacts with c-Kit to maintain the stability of p27. (A) Fc-tagged CD244 and c-Kit were over-expressed in 293T cells and their lysates were immunoprecipitated by protein A beads, followed by western blotting analysis for c-Kit. (B) Fc-tagged CD244 and c-Kit were over-expressed in 293T cells and their lysates were immunoprecipitated by c-Kit antibodies, followed by western blotting analysis for Fc (CD244). (C) StrepII-tagged p27 and c-Kit were over-expressed in 293T cells and their lysates were immunoprecipitated by StrepII beads, followed by western blotting analysis for c-Kit. (D) The lysates of HEL cells were immunoprecipitated by p27 or IgG control antibodies, followed by western blotting analysis for the levels of endogenous CD244, SHP-2 and c-Kit. (E) The lysates of HEL cells were immunoprecipitated by SHP-2 or IgG control antibodies, followed by western blotting analysis for the levels of endogenous CD244, p27 and c-Kit. (F) HEL cells transfected with shKit#3 targeting c-Kit or scrambled shRNA were used for the detection of c-Kit, p-SHP-2, SHP-2, p-p27 and p27 by western blotting analysis. (G) FACS sorted c-Kit+ and c-Kit- mouse AML cells were evaluated for the expression levels of c-Kit, p-SHP-2, SHP-2, p-p27 and p27 by western blotting analysis. (H) The lysates of WT and CD244-null primary AML cells were immunoprecipitated by p27 antibodies, followed by western blotting analysis for the levels of endogenous p-SHP-2, SHP2, p-p27, c-Kit and CD244. IgG antibodies or the backbone empty vector (EV) was used as the control.

714

haematologica | 2017; 102(4)


CD244 in leukemogenesis regulation

expression of p27 in WT leukemia cells slightly suppressed their proliferation (Figure 4E). These data also suggested that p27 may have a dose-dependent effect on leukemogenesis. The p27 levels in WT, CD244-null, p27over-expressed WT and CD244-null AML cells after transplantation were further confirmed by western blotting (Figure 4F). In addition, we have performed a functional assay by testing the colony forming ability in vitro to directly evaluate the p27 effect on the proliferation of LICs. Interestingly, the ectopic expression of p27 enhanced the proliferation of CD244-null LICs in vitro, but not WT ones

B

C

CD 24 sh

Sc r

4#

2

A

(Figure 4G). However, high level of p27 had no impact on the cell cycle status of both WT and CD244-null leukemic blasts (Figure 4H). In contrast, we also noticed that the colonies derived from CD244-null leukemia cells appeared much smaller and more diffused compared to WT ones, which could be rescued by the p27 OE (Figure 4G), suggesting the cell division of CD244-null LICs was slowed down upon p27 deletion. In addition, we also revealed that apoptosis was not altered in CD244-null LICs (Figure 3D). Therefore, these results suggest that CD244 maintains leukemogenic potential of AML cells through p27, which may mainly contribute to the proliferation (or self-

D

F

E

G

Figure 7. CD244 controls the activities of human leukemia cells through SHP-2/p27 signaling. (A) CD244 was knocked down with shCD244#2 in HEL cells, followed by western blotting analysis for p27. (B) CD244 was stably over-expressed in K562 cells, followed by western blotting analysis for p27. (C) Representative flow cytometric analysis for the expression of CD244 and c-Kit in cord blood mononuclear cells (MNC) and patientsâ&#x20AC;&#x2122; acute myeloid leukemia (AML) cells. CD34+ cells in MNC and AML samples were gated and quantified for different fractions (n=5 for MNC and n=7 for AML). (D) p-SHP-2 and p27 expression levels were examined in CD244low and CD244-high fractions by immunofluorescence staining. (E) The lysates of human primary AML cells (type M5) were immunoprecipitated by p27 or IgG control antibodies, followed by western blotting analysis for the levels of endogenous CD244, SHP-2 and c-Kit. (F) Three human leukemia samples (type M5) were crosslinked with anti-CD244 antibodies for 24 h, and their lysates were used for the detection of p-SHP-2, SHP-2, p-p27 and p27 by western blotting analysis. (G) Working model for the roles of CD244 in leukemia initiating cells.

haematologica | 2017; 102(4)

715


F. Zhang et al.

renewal ability), but less likely to the quiescence or apoptosis of LICs.

CD244 regulates the p27 stability through SHP-2 signaling In consideration of the significant decrease of p27 expression levels in the CD244-null LICs, we decided to delineate how p27 is regulated by CD244. We first performed the co-immunoprecipitation assay to pull down CD244 in 293T cells and demonstrated that CD244 was associated with p27 (Figure 5A). We further examined whether the phosphatase, SHP-2, served as a mediator between CD244 and p27 since SHP-2 was known to be involved in CD244 signaling. Conceivably, SHP-2 was found to be interacted with CD244 (Figure 5A). Conversely, both CD244 and SHP-2 could also be coimmunoprecipitated by p27 (Figure 5B). p27 could also be detected when SHP-2 was pulled down from 293T cells (Figure 5C). Moreover, importantly, endogenous CD244 and p27 were associated with SHP-2 in primary mouse AML cells as determined by the co-immunoprecipitation experiment (Figure 5D). Several studies show that tyrosine phosphorylation at Y88/Y89 of p27 leads to its conversion into a non-inhibitor of cyclin-CDK complex and is required for efficient phosphorylation on T187 by CDK2, followed by ubiquitination mediated degradation.7,13,22 Recently, SHP-2 is reported to be a phosphatase for tyrosine dephosphorylation of p27 and can be translocated into nucleus upon stimulation and stabilizes the p27 protein levels in the cytoplasm of AML cells.26,28 Indeed, overexpression of CD244 could dramatically up-regulate the phosphorylation of SHP-2 which substantially dephosphorylated p27 (Figure 5E). To reveal whether SHP-2 is a direct player for CD244 to dephosphorylate p27, we knocked down SHP-2 with a validated shSHP-2#3 (Online Supplementary Figure S5B) in CD244over-expressed 293T cells and examined the phospho-p27 levels upon co-immunoprecipitation. Indeed, phoshpop27 was highly up-regulated after SHP-2 deletion (Figure 5F). Furthermore, we have performed the experiment in human primary AML cells to confirm the regulatory role of SHP-2. We demonstrated that the p-p27 was significantly elevated upon SHP-2 or CD244 knockdown in human primary AML cells (Figure 5G and Online Supplementary Figure S5C). Then we continued to knock down CD244 in HEL cells to examine the regulatory role of SHP-2. Interestingly, we did find that p-SHP-2 and p27 were notably reduced, which could be rescued by overexpressing SHP-2 in CD244-knockdown HEL cells (Figure 5H). Taken together, CD244 is associated with SHP-2 to down-regulate the phosphorylation of p27, which further sustains its stability.

CD244 interacts with c-Kit to maintain the stability of p27 c-Kit has been reported as a critical regulator for both HSCs and AML initiation through SHP-2 or SCL pathway. Constitutively, activation of c-Kit may lead to the marked acceleration of leukemia development.29-34 Regarding the significant decrease of c-Kit+ LICs in CD244-null leukemic mice, we speculate that there is a connection between CD244, c-Kit and p27. Therefore, we conducted a coimmunoprecipitation experiment with overexpression of both Fc-tagged CD244 and c-Kit in 293T cells, followed by pull-down with either protein A beads or anti-c-Kit antibodies. CD244 was indeed associated with c-Kit (Figure 716

6A and B). We also provided further evidence to show that p27 was directly interacted with c-Kit when p27 was pulled down from 293T cells (Figure 6C). Consistently, we further confirmed the interactions of CD244, SHP-2 and cKit in leukemia cell line, HEL cells, which is an erythroid leukemia cell line expressing both CD244 and c-Kit (Figure 6D and E). Recently, Marcelo et al. reported that c-Kit promoted the expression of p27 in hemogenic endothelial cells,35 which led us to further evaluate the relationship of c-Kit and p27 in leukemia cells. By using shKit#3 specifically targeting c-Kit (Online Supplementary Figure S5D), we knocked down c-Kit expression in HEL cells and demonstrated that knockdown of c-Kit could efficiently downregulate phosphorylation of SHP-2 and total p27 expression levels, but enhance the p-p27 level (Figure 6F). Because HEL cells were JAK2 V617F positive and may exert artificial effect on SHP-2/p27 signaling, we have further performed the experiment related to the role of c-Kit on primary mouse AML cells. As shown in Figure 6G, we examined the c-Kit/SHP-2/p27 pathways in c-Kit+ or c-Kit– mouse AML cells and demonstrated that c-Kit level indeed enhances the SHP-2 signaling to decrease the phosphorylation level of p27 to maintain the stability of total p27 expression. Next, we pulled down p27 with the antibody specifically against p27 in WT and CD244-null AML cells, and revealed c-Kit, SHP-2 and CD244 were indeed directly associated with p27 in physiological conditions (Figure 6H and Online Supplementary Figure S5E). Phospho-SHP-2 level was also significantly decreased, which was concordant with increased level of phospho-p27 (Figure 6H and Online Supplementary Figure S5E). Importantly, the interactions between these three proteins were CD244-dependent since there were much lower levels of p-SHP-2 and c-Kit, or higher levels of p-p27 in CD244-null leukemia cells (Figure 6H and Online Supplementary Figure S5E). Although we found no significant difference in the phospho-SHP-2 level in the total lysate of WT and CD244-null AML cells, there was a remarkable decreased phospho-SHP-2 level in CD244-null AML cells when p27 was immunoprecipitated, indicating that only the p27-interacting form of SHP-2 was affected upon CD244 deletion (Figure 6H and Online Supplementary Figure S5E). These results suggest that CD244 might co-ordinate with c-Kit to regulate the levels of p27 through SHP-2 pathway to maintain the proliferation ability of LICs.

CD244 controls the proliferation of human leukemia cells through SHP-2/p27 signaling To tease apart the CD244/SHP-2/p27 signaling in human leukemia cells, we examined the p27 levels in either CD244-silenced HEL cells or CD244-over-expressed K562 cells (expressing very low levels of CD244) by immunoblotting, which showed that p27 levels were also consistently down-regulated or up-regulated, respectively (Figure 7A and B). We then measured the expression levels of CD244 and c-Kit in CD34+ cells (enriched for LICs) from several AML patients and demonstrated that 61.3% of CD34+/CD244+ leukemia cells were co-expressed with c-Kit, which was much higher than cord blood CD34+/CD244+ mononuclear cells (MNC) (61.31±14.42% vs. 8.61±2.25%) (Figure 7C). These results indicate that CD244 and c-Kit may also function synergistically in human LICs. We further isolated CD244-low and CD244haematologica | 2017; 102(4)


CD244 in leukemogenesis regulation

high leukemia cells and examined the phospho-SHP-2 and p27 level by immunofluorescence staining. The phosphoSHP-2 and p27 levels were much higher in the cytoplasm of CD244-high cells (Figure 7D). We next examined the CD244/c-Kit/SHP-2/p27 signaling in human primary AML cells and revealed similar interactions among these molecules as in a mouse leukemia model, as analyzed by the co-immunoprecipitation experiments (Figure 7E). Furthermore, CD244 receptors of human CD34+ LICs were cross-linked with the functional anti-CD244 antibodies to test whether the downstream SHP-2/p27 signaling could be efficiently induced. In line with the findings in mouse leukemia cells, phospho-SHP-2 and p27 levels were indeed increased, while p-p27 was consistently decreased upon the antibody stimulation in all the 3 tested human samples (Figure 7F). In summary, a working model is shown in Figure 7G, and our findings show that CD244 collaborates with c-Kit to regulate SHP-2 phosphatase activity to dephosphorylate and stabilize p27 in cytoplasm for maintaining the proliferation (self-renewal) ability of LICs.

Discussion In this study, we characterized CD244 function in AML initiating cells and provided solid evidence that CD244 has a differential effect on LICs compared to normal HSCs. We further revealed CD244 might co-operate with c-Kit to mediate its downstream signaling through SHP2/p27 to regulate the proliferation (self-renewal) of LICs. To our knowledge, this is the first body of evidence showing that CD244 is critical for leukemia development. These results consolidated our previous hypothesis that immune inhibitory receptors are involved in maintenance of the stemness of HSCs and LICs. Intriguingly, our study also indicated that both ITIM and ITSM containing immune molecules may be essential for different types of leukemia. Our previous findings suggested that many ITIM-containing immune molecules recruit SHP-2, SHP-1 or SHIP1 to initiate numerous downstream signaling pathways. Our current study also indicates that CD244 is also associated with SHP-2 as well as p27 to control leukemia development. Nonetheless, how CD244 exerts its effect on SHP-2 and p27 needs further clarification. Our data showed that CD244 could up-regulate p27 levels either through transactivation or post-translational modification. However, it is still not clear which pathway plays the dominant roles in leukemogenesis and how these dual functions are integrated. Meanwhile, how CD244 transactivates p27 is still a mystery. p27 is known to be important to maintain the repopulation ability and quiescence of HSCs by co-operating with p57.36,37 p27 has also been found to play differential roles in different types of cells,

References 1. Zheng J, Umikawa M, Cui C, et al. Inhibitory receptors bind ANGPTLs and support blood stem cells and leukaemia development. Nature. 2012;485(7400):656660. 2. Majeti R, Chao MP, Alizadeh AA, et al. CD47 is an adverse prognostic factor and

haematologica | 2017; 102(4)

including acting as a cell cycle inhibitor to maintain the quiescent status of LICs25 and enhancing the migration of mouse embryonic fibroblasts.38 Nevertheless, several studies claimed that LICs existed in a non-quiescent population of cells controlled by certain cyclins, such as cyclin D2.39 Iwasaki et al. also demonstrated that CD93 marks a non-quiescent human LIC population and regulates LIC self-renewal predominantly by silencing CDKN2B.24 Herein, we also found CD244 may sustain the proliferation ability of LICs by stabilizing p27 levels independent of cell cycle inhibition, suggesting cell cycle regulators play differential roles in normal or cancer stem cells, which is critical for the development of new strategies for cancer treatment. Furthermore, our findings indicate that c-Kit may interact with CD244 to regulate the expression levels of p27. However, there are several intriguing questions that remain to be addressed. 1) Which domain of c-Kit interplays with CD244: the extra-cellular or intra-cellular domain? 2) How does c-Kit stabilize p27 levels through CD244? 3) Does CD244 up-regulate the expression of cKit? 4) Although it is well known that c-Kit has an important function in mouse LIC maintenance, its role in human LICs and related mechanisms remains elusive. We are currently in the process of investigating these issues to pinpoint the CD244/c-Kit/SHP-2/p27 signaling in LICs. Our study indicates that CD244 is highly expressed on different types of AML, but may not be expressed on Bcell ALL, which suggests that the expression of CD244 may serve as a threshold for myeloid differentiation. Whether there are any other specific regulators controlling CD244 expression in LICs needs to be clarified. Determining the novel mechanisms controlling the expression and activation of CD244 will open a new avenue for treatment of AML. In conclusion, our results provide strong evidence that CD244 co-operates with cKit to regulate leukemogenesis through SHP-2/p27 signaling. Acknowledgments We thank Dr. Colin Stewart of the Frederick National Laboratory for Cancer Research for providing the CD244-/mice. We appreciate the kindness of Dr. Yueying Wang at Ruijing Hospital, Shanghai, providing us with the plasmid c-Kit. Funding This work was supported by grants from National Basic Research Program of China (973Program, 2014CB965000; NO2015CB910403), National Natural Science Foundation of China (81370654, 81422001, 81570093), the 1000-Youth Elite Program, the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning, the Innovation Program of Shanghai Municipal Education Commission (13G20), 1R01CA172268, CPRIT RP140402, and Taishan Scholar Immunology Program.

therapeutic antibody target on human acute myeloid leukemia stem cells. Cell. 2009;138(2):286-299. 3. Chao MP, Alizadeh AA, Tang C, et al. Therapeutic antibody targeting of CD47 eliminates human acute lymphoblastic leukemia. Cancer Res. 2011;71(4):13741384. 4. Zheng J, Song C, Zhang CC. A new chapter: hematopoietic stem cells are direct

players in immunity. Cell Biosci. 2011;1:33. 5. Kiel MJ, Yilmaz OH, Iwashita T, Terhorst C, Morrison SJ. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell. 2005;121(7):1109-1121. 6. Sintes J, Romero X, Marin P, Terhorst C, Engel P. Differential expression of CD150 (SLAM) family receptors by human hematopoietic stem and progenitor cells.

717


F. Zhang et al. Exp Hematol. 2008;36(9):1199-1204. 7. Oguro H, Ding L, Morrison SJ. SLAM family markers resolve functionally distinct subpopulations of hematopoietic stem cells and multipotent progenitors. Cell Stem Cell. 2013;13(1):102-116. 8. Mathew PA, Garni-Wagner BA, Land K, et al. Cloning and characterization of the 2B4 gene encoding a molecule associated with non-MHC-restricted killing mediated by activated natural killer cells and T cells. J Immunol. 1993;151(10):5328-5337. 9. Tangye SG, Lazetic S, Woollatt E, Sutherland GR, Lanier LL, Phillips JH. Cutting edge: human 2B4, an activating NK cell receptor, recruits the protein tyrosine phosphatase SHP-2 and the adaptor signaling protein SAP. J Immunol. 1999; 162(12):6981-6985. 10. Sayos J, Nguyen KB, Wu C, et al. Potential pathways for regulation of NK and T cell responses: differential X-linked lymphoproliferative syndrome gene product SAP interactions with SLAM and 2B4. Int Immunol. 2000;12(12):1749-1757. 11. Parolini S, Bottino C, Falco M, et al. Xlinked lymphoproliferative disease. 2B4 molecules displaying inhibitory rather than activating function are responsible for the inability of natural killer cells to kill Epstein-Barr virus-infected cells. J Exp Med. 2000;192(3):337-346. 12. Tangye SG, Phillips JH, Lanier LL, Nichols KE. Functional requirement for SAP in 2B4mediated activation of human natural killer cells as revealed by the X-linked lymphoproliferative syndrome. J Immunol. 2000;165(6):2932-2936. 13. Eissmann P, Beauchamp L, Wooters J, Tilton JC, Long EO, Watzl C. Molecular basis for positive and negative signaling by the natural killer cell receptor 2B4 (CD244). Blood. 2005;105(12):4722-4729. 14. Wahle JA, Paraiso KH, Costello AL, Goll EL, Sentman CL, Kerr WG. Cutting edge: dominance by an MHC-independent inhibitory receptor compromises NK killing of complex targets. J Immunol. 2006;176(12):71657169. 15. Wahle JA, Paraiso KH, Kendig RD, et al. Inappropriate recruitment and activity by the Src homology region 2 domain-containing phosphatase 1 (SHP1) is responsible for receptor dominance in the SHIP-deficient NK cell. J Immunol. 2007; 179(12):8009-8015. 16. Krivtsov AV, Twomey D, Feng Z, et al.

718

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Transformation from committed progenitor to leukaemia stem cell initiated by MLLAF9. Nature. 2006;442(7104):818-822. Zhu Y, Yao S, Chen L. Cell surface signaling molecules in the control of immune responses: a tide model. Immunity. 2011;34(4):466-478. Wei J, Wunderlich M, Fox C, et al. Microenvironment determines lineage fate in a human model of MLL-AF9 leukemia. Cancer Cell. 2008;13(6):483-495. Goardon N, Marchi E, Atzberger A, et al. Coexistence of LMPP-like and GMP-like leukemia stem cells in acute myeloid leukemia. Cancer Cell. 2011;19(1):138-152. Jin L, Lee EM, Ramshaw HS, et al. Monoclonal antibody-mediated targeting of CD123, IL-3 receptor alpha chain, eliminates human acute myeloid leukemic stem cells. Cell Stem Cell. 2009;5(1):31-42. Jordan CT, Upchurch D, Szilvassy SJ, et al. The interleukin-3 receptor alpha chain is a unique marker for human acute myelogenous leukemia stem cells. Leukemia. 2000;14(10):1777-1784. Somervaille TC, Cleary ML. Identification and characterization of leukemia stem cells in murine MLL-AF9 acute myeloid leukemia. Cancer Cell. 2006;10(4):257-268. Lechman ER, Gentner B, Ng SW, et al. miR126 Regulates Distinct Self-Renewal Outcomes in Normal and Malignant Hematopoietic Stem Cells. Cancer Cell. 2016;29(2):214-228. Iwasaki M, Liedtke M, Gentles AJ, Cleary ML. CD93 Marks a Non-Quiescent Human Leukemia Stem Cell Population and Is Required for Development of MLLRearranged Acute Myeloid Leukemia. Cell Stem Cell. 2015;17(4):412-421. Zhang J, Seet CS, Sun C, et al. p27kip1 maintains a subset of leukemia stem cells in the quiescent state in murine MLLleukemia. Mol Oncol. 2013;7(6):1069-1082. Kardinal C, Dangers M, Kardinal A, et al. Tyrosine phosphorylation modulates binding preference to cyclin-dependent kinases and subcellular localization of p27Kip1 in the acute promyelocytic leukemia cell line NB4. Blood. 2006;107(3):1133-1140. Agarwal A, Mackenzie RJ, Besson A, et al. BCR-ABL1 promotes leukemia by converting p27 into a cytoplasmic oncoprotein. Blood. 2014;124(22):3260-3273. Tossidou I, Dangers M, Koch A, Brandt DT, Schiffer M, Kardinal C. Tyrosine phos-

29.

30.

31.

32.

33. 34.

35.

36.

37.

38.

39.

phatase SHP-2 is a regulator of p27(Kip1) tyrosine phosphorylation. Cell Cycle. 2008;7(24):3858-3868. Scolnik MP, Morilla R, de Bracco MM, Catovsky D, Matutes E. CD34 and CD117 are overexpressed in AML and may be valuable to detect minimal residual disease. Leuk Res. 2002;26(7):615-619. Boissel N, Leroy H, Brethon B, et al. Incidence and prognostic impact of c-Kit, FLT3, and Ras gene mutations in core binding factor acute myeloid leukemia (CBFAML). Leukemia. 2006;20(6):965-70. Hassan HT. c-Kit expression in human normal and malignant stem cells prognostic and therapeutic implications. Leuk Res. 2009;33(1):5-10. Chavez-Gonzalez A, Dorantes-Acosta E, Moreno-Lorenzana D, Alvarado-Moreno A, Arriaga-Pizano L, Mayani H. Expression of CD90, CD96, CD117, and CD123 on different hematopoietic cell populations from pediatric patients with acute myeloid leukemia. Arch Med Res. 2014;45(4):343350. Lacombe J, Krosl G, Tremblay M, et al. Genetic interaction between Kit and Scl. Blood. 2013;122(7):1150-1161. Zhu HH, Ji K, Alderson N, et al. Kit-Shp2Kit signaling acts to maintain a functional hematopoietic stem and progenitor cell pool. Blood. 2011;117(20):5350-5361. Marcelo KL, Sills TM, Coskun S, et al. Hemogenic endothelial cell specification requires c-Kit, Notch signaling, and p27mediated cell-cycle control. Dev Cell. 2013; 27(5):504-515. Cheng T, Rodrigues N, Dombkowski D, Stier S, Scadden DT. Stem cell repopulation efficiency but not pool size is governed by p27(kip1). Nat Med. 2000;6(11):1235-1240. Zou P, Yoshihara H, Hosokawa K, et al. p57(Kip2) and p27(Kip1) cooperate to maintain hematopoietic stem cell quiescence through interactions with Hsc70. Cell Stem Cell. 2011;9(3):247-261. Zhang D, Wang Y, Liang Y, et al. Loss of p27 upregulates MnSOD in a STAT3dependent manner, disrupts intracellular redox activity and enhances cell migration. J Cell Sci. 2014;127(Pt 13):2920-2933. Chen BB, Glasser JR, Coon TA, et al. F-box protein FBXL2 targets cyclin D2 for ubiquitination and degradation to inhibit leukemic cell proliferation. Blood. 2012; 119(13):3132-3134.

haematologica | 2017; 102(4)


ARTICLE

Acute Myeloid Leukemia

Phase I study of the aurora A kinase inhibitor alisertib with induction chemotherapy in patients with acute myeloid leukemia

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Amir T. Fathi,1* Seth A. Wander,1 Traci M. Blonquist,2 Andrew M. Brunner,1 Philip C. Amrein,1 Jeffrey Supko,1 Nicole M. Hermance,3 Amity L. Manning,3 Hossein Sadrzadeh,1 Karen K. Ballen,1 Eyal C. Attar,1 Timothy A. Graubert,1 Gabriela Hobbs,1 Christelle Joseph,1 Ashley M. Perry,1 Meghan Burke,1 Regina Silver,1 Julia Foster,1 Meghan Bergeron,1 Aura Y. Ramos,1 Tina T. Som,1 Kaitlyn M. Fishman,1 Kristin L. McGregor,1 Christine Connolly,1 Donna S. Neuberg2 and Yi-Bin Chen1 Massachusetts General Hospital Cancer Center, Harvard Medical School, Boston; Dana-Farber Cancer Institute, Harvard Medical School, Boston and 3Worcester Polytechnic Institute, Department of Biology, Worcester, MA, USA

1 2

Haematologica 2017 Volume 102(4):719-727

ABSTRACT

A

berrant expression of aurora kinase A is implicated in the genesis of various neoplasms, including acute myeloid leukemia. Alisertib, an aurora A kinase inhibitor, has demonstrated efficacy as monotherapy in trials of myeloid malignancy, and this efficacy appears enhanced in combination with conventional chemotherapies. In this phase I, dose-escalation study, newly diagnosed patients received conventional induction with cytarabine and idarubicin, after which alisertib was administered for 7 days. Dose escalation occurred via cohorts. Patients could then receive up to four cycles of consolidation, incorporating alisertib, and thereafter alisertib maintenance for up to 12 months. Twenty-two patients were enrolled. One dose limiting toxicity occurred at dose level 2 (prolonged thrombocytopenia), and the recommended phase 2 dose was established at 30mg twice daily. Common therapyrelated toxicities included cytopenias and mucositis. Only three (14%) patients had persistent disease at mid-cycle, requiring “5+2” reinduction. The composite remission rate (complete remission and complete remission with incomplete neutrophil recovery) was 86% (nineteen of twenty-two patients; 90% CI 68-96%). Among those over age 65 and those with high-risk disease (secondary acute leukemia or cytogenetically high-risk disease), the composite remission rate was 88% and 100%, respectively. The median follow up was 13.5 months. Of those treated at the recommended phase 2 dose, the 12-month overall survival and progression-free survival were 62% (90% CI 33-81%) and 42% (90% CI 1765%), respectively. Alisertib is well tolerated when combined with induction chemotherapy in acute myeloid leukemia, with a promising suggestion of efficacy. (clinicaltrials.gov Identifier:01779843).

Correspondence: afathi@partners.org

Received: October 15, 2016. Accepted: December 26, 2016. Pre-published: December 29, 2016. doi:10.3324/haematol.2016.158394 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/719 ©2017 Ferrata Storti Foundation

Introduction Acute myeloid leukemia (AML) is an aggressive hematologic malignancy associated with poor outcomes. Current standard treatment includes remission induction chemotherapy, typically consisting of cytarabine combined with an anthracycline, an approach which has remained unchanged for more than 30 years.1-3 Over the last decade, a series of novel therapeutics have been combined with this chemotherapy backbone, with most yielding negative or inconsistent results.4-7 More recently, some improvement in outcomes have been noted with certain FLT3 tyrosine kinase inhibitors and antibody-drug conjugates8-10 in combination with induction haematologica | 2017; 102(4)

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

719


A.T. Fathi et al.

chemotherapy; such benefits may be limited to certain subcategories of AML. For the majority of patients, new and effective approaches are still needed to enhance the current standard of care. Aurora kinases are a family of serine-threonine kinases that regulate multiple phases of the mitotic signaling cascade.11 Aberrant upregulation of aurora A kinase (AAK) has been demonstrated in multiple malignancies, and its inhibition suppresses proliferation of neoplastic cells11,12 by triggering mitotic errors, aneuploidy, senescence, and apoptosis.13,14 This process appears especially applicable to cells during or shortly following exposure to cytotoxic chemotherapy.15-18 Alisertib is a potent, orally available inhibitor of AAK. Preclinical studies in AML cell lines, patient samples, and animal models have demonstrated potent cytotoxicity, diminished clonal survival, and promotion of apoptosis.15 Alisertib has been evaluated for safety and efficacy in multiple clinical studies,19-22 including those regarding hematological malignancy,23-26 and has been associated with clinical response. Alisertib has also been studied in combination with other neoplastic agents, where it was also demonstrated to be safe and well-tolerated.15-18 There has been growing interest in targeting AAK as a therapeutic approach in myeloid neoplasms, and pre-clinical studies have suggested that targeting AAK expression with alisertib promotes tumor sensitivity to cytarabine.15 A subsequent phase II trial of alisertib monotherapy demonstrated efficacy in a subset of patients with AML and high-grade myelodysplastic syndromes (MDS). 27 Given the oncogenic role of aurora kinases, the activity of alisertib monotherapy in AML, and suggestion of synergy in combination with chemotherapies, we performed a phase I study to evaluate the safety and tolerability of alisertib combined with conventional induction chemotherapy for newly diagnosed AML.

Methods This study (clinicaltrials.gov Identifier:01779843) was approved by the local institutional review board, and conducted in accordance with the declaration of Helsinki. Patients were eligible for enrollment if they were age 18 or older with previously untreated AML based on WHO criteria (≥ 20% bone marrow blasts).28 They were required to have an Eastern Cooperative Oncology Group (ECOG) performance status of 0-2, cardiac ejection fraction of ≥50%, and intact organ function including aspartate aminotransferase (AST), alanine aminotransferase (ALT), and alkaline phosphatase of <5 times the upper limit of normal, direct bilirubin of <2.0 mg/dL, and creatinine clearance of ≥40 mL/min. Standard cytogenetic and molecular testing was performed at diagnosis. Those with acute promyelocytic leukemia or with core-binding factors alterations, t(8;21) or inv(16)/t(16;16), were excluded.

Treatment “7+3” induction chemotherapy included continuous cytarabine infusion at 100 mg/m2 on days 1 through 7, and idarubicin, dosed at 12 mg/m2 administered on days 1-3 by intravenous bolus. Starting on the day after the conclusion of cytarabine infusion, alisertib was administered orally according to dose level, twice daily (BID), for 7 days. A bone marrow biopsy was performed at mid-induction 720

(between days 13 and 16). If bone marrow cellularity was ≥20% and myeloblast involvement was >5%, “5+2” reinduction, with continuous infusion of cytarabine at 100mg/m2 on days 1 to 5 and idarubicin 12 mg/m2 intravenously on days 1 and 2, was administered. If the bone marrow cellularity was <20% but with >5% myeloblasts, “5+2” reinduction was per clinician discretion. Patients receiving reinduction did not receive additional doses of alisertib during the remission induction period. A bone marrow evaluation was performed at the time of peripheral hematologic recovery (absolute neutrophil count (ANC) >1000/µL and platelet count >100,000/μL) or by day 40 (range 35-42) or day 60 (if “5+2” was administered) in the absence of optimal hematologic recovery. A marrow biopsy was also performed on clinical suspicion of resistant disease. Response criteria29 were categorized as complete remission (CR), complete remission with incomplete platelet recovery (CRp), complete remission with incomplete neutrophil recovery (CRi), or refractory disease. Those achieving CR, CRi, or CRp, and eligible for allogeneic hematopoietic cell transplantation (HCT), could come off study treatment during follow up for that purpose; they were then followed for relapse and survival outcomes. Otherwise, responding patients were eligible for up to four cycles of consolidation therapy, at their clinician's discretion. Consolidation therapy consisted of cytarabine intravenously dosed at 3 gm/m2 every twelve hours on days 1, 3, and 5 in patients < age 60, or 2 gm/m2 daily on days 1-5 in those ≥ age 60. Starting on day 6 of each consolidation cycle, they received BID alisertib according to dose level for 7 days. Once patients completed the cycles of consolidation and achieved count recovery, they were eligible for alisertib maintenance. This was administered BID according to the patient’s dose level on days 1-7 of 21-day cycles, and continued until 12 months after the start of induction, or until disease progression. Patients were enrolled in one of three dose cohorts in a “3+3” dose-escalation design. The three alisertib dose levels were 10 mg, 20 mg, and 30 mg BID. Enrollment was stopped at dose levels until all three patients in a cohort were assessed for treatment-related dose-limiting toxicities (DLTs). The DLT period was from initiation up to day 40, or day 60 if “5+2” was administered. If no DLTs were experienced by the first three patients, three patients were treated at the next dose level. If one DLT was experienced, an additional three patients were enrolled at the same dose level. If fewer than 2 DLTs were experienced among the six, dose escalation was permitted. If 2 or more DLTs were experienced, the previous lower dose was deemed the maximum tolerated dose (MTD). Should DLTs not be encountered, the highest dose level (30 mg BID) within the protocol would be the recommended phase 2 dose (RP2D). Once the MTD or RP2D was identified, an additional six patients were to be treated at that dose level. Toxicities were graded according to Common Terminology Criteria for Adverse Events (CTCAE version 4.0). Patients who experienced grade 3 or 4 non-hematologic toxicity related to the study drug, and which persisted for longer than 48 hours without resolution to ≤ grade 2, stopped alisertib. Grade 3 or 4 myelosuppression did not lead to alisertib cessation. DLTs were defined as any grade 4 or 5 non-hematologic toxicity, but excluding toxicities such as infection related to neutropenia, grade 4 haematologica | 2017; 102(4)


Alisertib and induction chemotherapy in AML

fatigue or anorexia, and grade 4 nausea, vomiting, diarrhea, or electrolyte abnormalities which were reversible with the appropriate therapies. Grade 3 non-hematological toxicities that did not resolve to grade 2 by day 40 were also considered DLTs, unless they were attributed to persistent AML. Grade 4 neutropenia or thrombocytopenia at day 40 following induction, or day 60 if “5+2” reinduction was administered, were also considered DLTs, if the cytopenias were not thought to be related to the underlying leukemia.

Correlative Methods Culture of primary human AML cells Cryopreserved primary bone marrow samples, collected at baseline, were thawed and cultured for 36-48h on confluent irradiated (2,000 rads) mouse MS5 stromal cells in roswell park memorial institute (RPMI) medium supplemented with 20% fetal bovine serum (FBS), 1% penicillin/streptomycin, 50μM β-mercaptoethanol, and human cytokines including stem cell factor (SCF) (100ng/ml), interleukin (IL)-3 (10ng/ml), IL-6 (20ng/ml), thrombopoietin (TPO) (10ng/ml), and FLT3 ligand (FLT3L, 10ng/ml) (Peprotech). Once cells were proliferative, each sample was transferred to two new wells of irradiated stroma and left untreated, or treated with 50nM alisertib for 18 hrs.

Immunofluorescence Staining and Imaging Patient leukemic cells were then spun at 1,000rpm for 3 minutes onto poly-lysine coated coverslips and fixed in ice-cold methanol. Coverslips were blocked in trisbuffered saline (TBS)/ bovine serum albumin (BSA), stained for tubulin (dmlα: Sigma) and DNA (0.2 μg/mL 4′,6-diamidino-2-phenylindole (DAPI)), and mounted on coverslips using ProLong Antifade mounting medium (Molecular Probes). A Nikon Ti-E fluorescence microscope equipped with a 60x objective and a Zyla sCMOS camera was used to identify and capture images of mitotic cells in each sample. The identification and characterization of the mitotic spindle structure was performed on a minimum of 20 mitotic figures per condition.

Pharmacokinetics Samples to determine the steady state minimum concentration of alisertib in plasma (Cminss) were obtained from each patient just prior to dosing on days 9, 11, and 14 during the first cycle. Peripheral blood (6mL) was drawn into plastic tubes containing freeze-dried K2EDTA, mixed by inversion, placed over ice until centrifuged (1,300 g, 10min, 4°C), whereupon the plasma was removed and stored at -80ºC until assayed. The concentration of alisertib in human plasma was determined by reversed-phase high performance liquid chromatography with tandem mass spectrometric detection as previously reported, with minor modifications.22 The analytical method was validated as recommended by current FDA Guidance for Industry: Bioanalytical Method Validation, May 2001, to document selectivity, carryover, accuracy, precision, absolute recovery, and matrix effects. Alisertib was determined with an interday accuracy of 100.8% and a precision of 1.3% at the 5.0 ng/mL lower limit of quantitation. Cminss was calculated for each patient as the average of the assayed concentration of alisertib in predose samples obtained on days 11 and 14. The geometric mean and standard deviation were calculated for Cminss values for patients evaluated at each dose level. haematologica | 2017; 102(4)

Statistical Analysis The primary study objective was to determine the type, frequency, and severity of drug related toxicities and define the MTD/RP2D. The study utilized a 3+3 design, such that if the DLT rate exceeded 30% at a given level it was less likely that the alisertib dose would be escalated. The secondary objectives included assessment of the complete response rate (CR/CRi/CRp), and the rates of overall survival (OS) and progression-free survival (PFS) at one year. OS was defined as the time from diagnosis to death and censored at the last known date alive. PFS was defined as the time from diagnosis to the first occurrence of progression or death and censored at the last known date alive and disease-free. For the two patients who died before the end of the DLT monitoring period, the date when they were deemed off-treatment was used for survival estimates. Survival was estimated for the trial and at the RP2D (cohort 3 and dose expansion) with the KaplanMeier method. The proportion of patients achieving a CR or CRi was estimated along with a 90% exact binomial confidence interval. All grade 3 or higher toxicity was tabled regardless of attribution and dose cohort. The date for either count recovery or day 40 (D40), whichever was reported, were used to determine when toxicities occurred. Grade 3 or higher induction toxicities were tabled according to the dose cohort. Additionally, for patients proceeding to consolidation therapy, post-induction toxicities were tabled. Pharmacokinetic and correlative studies, as well as patient and disease characteristics, including the number of patients completing each stage of therapy (induction, consolidation, and maintenance) were presented descriptively.

Results Patient Characteristics Between May 2013 and January 2015, twenty-two patients were enrolled on study. Baseline patient characteristics are summarized in Table 1. The median age was 62.7 (range 41.7-75.1), and the majority were Caucasian and male. Eight patients (36%) were aged ≥ 65 years. Eight patients (36%) had secondary AML, six with preceding MDS, one with a preceding myeloproliferative neoplasm, and one with therapy-related AML. The majority of patients had intermediate-risk AML based upon cytogenetics (82%) or upon European LeukemiaNet (ELN) criteria (73%). The most frequent additional mutations identified were NRAS (27%), IDH2 R140 (23%), NPM1 (18%), IDH1 (14%), TP53 (14%), and FLT3-ITD (14%).

Treatment and Toxicity Profile A total of twenty-two patients were enrolled: three patients (10mg BID), seven patients (20mg BID), six patients (30mg BID), and an additional six at the determined RP2D of 30mg BID. All patients experienced at least one expected grade 3/4 toxicity of anemia, leukopenia, thrombocytopenia, and febrile neutropenia. A single DLT, prolonged grade 4 thrombocytopenia (beyond 40 days), thought to possibly be related to alisertib, occurred at dose level two (20mg BID), prompting the enrollment of six patients at that dose level. No additional DLTs were observed. An expansion cohort of six patients was then 721


A.T. Fathi et al.

added at 30mg BID. Two patient deaths occurred prior to the DLT assessment period. The first mortality, at dose level two, was caused by sepsis and thought to be unrelated to the study drug; this patient was replaced due to DLT assessment ineligibility. The second mortality was due to subarachnoid hemorrhage, again considered to be unrelated to the study drug; this patient was in the expansion cohort, and no replacement was performed. Alisertib was well tolerated. All grade 3 and higher nonhematologic toxicities, regardless of attribution, are summarized in Table 2. In addition to the DLT noted above and expected hematologic toxicities, other grade 4 toxicities included respiratory failure (one patient, 5%) and sepsis (two patients, 9%), neither of which were thought to be related to alisertib. The most frequent grade 3 toxicity was rash (four patients, 18%), which were also not felt to Table 1. Patient and disease characteristics.

Patient Characteristic N Dose Level 10 mg BID 20 mg BID 30 mg BID Expansion Age at Dx, median (range) Sex Male Female Race Asian Black/African American Other White Ethnicity Hispanic or Latino Non-Hispanic Ethnicity not known Disease Characteristics Treatment-Related AML Underlying myeloid neoplasm† Cytogenetic Risk Poor-Risk Intermediate Mutations* FLT3** FLT3-ITD FLT3-TKD NPM1 CEBPA IDH1 IDH2 NRAS KRAS TP53 ELN Prognostic Groups Favorable*** Intermediate-1 Intermediate-2 Adverse

N. (%) 22 3 (14) 7 (32) 6 (27) 6 (27) 62.7 (41.7-75.1) 15 (68) 7 (32) 1 (4) 1 (4) 3 (14) 17 (77)

Efficacy Treatment responses are summarized in Table 4, along with chromosomal and molecular features available at diagnosis for treated patients. Of the twenty-two patients treated, only three (14%) had persistent disease at mid-treatment, all of whom went on to receive “5+2” reinduction therapy per protocol. A remission was achieved in 86% (19/22; 90% CI 68-96%) of patients (64% CR, 23% CRi). Of the remaining three patients, two (9%) died during induction therapy and were inevaluable for marrow response, and one (5%) had refractory disease. The 12-month OS and PFS were 75% (90% CI 55-87%) and 54% (34-71%), respectively. Seven of the eight patients over age 65 years (87%) achieved a complete remission (six CR, one CRi, one death). Ten patients had high-risk AML, as defined as secondary/therapy-related AML and/or displaying poorrisk cytogenetic aberrations as per Medical Research Council (MRC) and ELN criteria; the overall remission rate for this group was 100% (80% CR, 20% CRi).

0 (0) 14 (64) 8 (36)

Table 2. Non-Hematologic Toxicities (grade 3 and higher) noted on study.

Toxicity Description

Grade 3 N. %

1 (4) 7 (32)

Colitis Febrile neutropenia Hypertension Hyponatremia Hypotension Hypoxia Infection Intracranial hemorrhage Lymph node pain Nausea Non-cardiac chest pain Pneumonia Rash (maculopapular) Respiratory failure Sepsis Sinusitis Skin infection Tooth infection Typhlitis

2 18 1 1 1 1 1

9 82 5 5 5 5 5

1 1 1 1 4

5 5 5 5 18

4 (18) 18 (82) 3 (14) 3 (14) 1 (4) 4 (18) 2 (9) 3 (14) 6 (27) 6 (27) 3 (14) 3 (14) 2 (9) 9 (41) 7 (32) 4 (18)

*Denominator out of all patients, some with unknown mutation status. †Six cases of antecedent MDS, and one case of antecedent MPN. **One patient had concurrent FLT3-ITD and FLT3-TKD mutations detected. ***One case of isolated NPM1 mutation, and one case of isolated CEBPA mutation. Dx: diagnosis; AML: Acute myeloid leukemia; ELN: European LeukemiaNet.

722

be related to alisertib. Non-hematologic toxicities possibly or probably attributable to alisertib are provided in Table 3. The large majority were deemed < grade 3, apart from two reversible episodes of grade 3 gastrointestinal toxicity. Mucositis was the most frequent toxicity with a possible relation to alisertib – this occurred in six patients (27%, all < grade 3). The median time to partial peripheral count recovery, as defined by an ANC of 500/mm3 and a platelet count of 50,000/mm3, was 33 days (range 26-48 days). The median time to full count recovery (defined as an ANC of 1000/mm3 and platelet count of 100,000/mm3) was 36 days (range 28-52 days).

1 1 1 1 1

5 5 5 5 5

Grade 4 N. % 1

1 2

Grade 5 N. %

5

5 9

1

5

1

5

haematologica | 2017; 102(4)


Alisertib and induction chemotherapy in AML

Ten patients (45%) received at least one cycle of cytarabine consolidation therapy on study; two patients (9%) received four cycles of consolidation therapy. Two patients were not able to proceed to consolidation, as per the discretion of their treating physician, due to ongoing infection. Four patients (18%) received alisertib as maintenance therapy following consolidation. At the time of manuscript preparation, six patients (27%) had experienced disease relapse and six patients (27%) had died (Figure 1). Among the four patients who received maintenance alisertib therapy, one experienced relapse and there were no instances of disease-related mortality. Ten patients (45%) went on to receive allogeneic HSCT after completing protocol-based therapy, four of whom relapsed with two ultimately dying from the disease in the protocol-determined follow-up period. A total of twelve patients were treated at the RP2D (six from dose level three and six from the expansion cohort) of which eleven (92%; 90% CI 66-99%) achieved a CR/CRi. The median follow up for those remaining alive was 13.5 months. Figure 2 displays the OS and PFS for those treated at the RP2D. The 12-month OS for these patients was 62% (90% CI 33-81%) and the 12-month PFS was 42% (90% CI 17-65%).

Pharmacokinetics The mean (±SD) Cminss of alisertib was 229±89 nM, 552±317 nM and 652±304 nM for patients treated with doses of 10 mg, 20 mg, and 30 mg BID, respectively.

Mitotic Spindle Studies For a subset of patients, baseline samples of leukemic blasts were available for mitotic spindle study by immunofluorescence staining (Patients #4, #17, #20 and #22). Following 18 hours of treatment with 50nM of alisertib (a dose previously demonstrated to be optimal for studies in culture), all of these samples exhibited response (P<0.001). Treated samples demonstrated profound defects in mitotic spindle assembly, manifested as increases in mitotic cells exhibiting monopolar spindles and corresponding decreases in bipolar spindles (Figure 3), while untreated mitotic cells exhibited primarily bipolar or multipolar spindles.

Discussion The current therapeutic paradigm for AML is suboptimal for the majority of patients, and has remained unchanged for decades. Outcomes are particularly poor for certain higher risk subgroups, including older patients, those with secondary disease, and those with poor-risk cytogenetics. Effective novel approaches to the treatment of AML are necessary, and appear to be gradually emerging. Recent data has suggested that AAK inhibition with the targeted inhibitor alisertib may be an effective therapy across a range of cancers,19-24,26 including advanced myeloid malignancies.25 Others have found that the unique mitotic toxicity of alisertib may synergize with cytotoxic

Table 3. Non-hematologic toxicities on study by dose cohort, possibly or probably attributed to alisertib.

Toxicity Description Abdominal pain Anorexia Chills Constipation Diarrhea Dizziness Dysgeusia Dyspepsia Dyspnea Edema limbs Elevated alk phos Fatigue Headache Hyperbilirubinemia Hyperglycemia Hypertension Hypocalcemia Oral mucositis Nausea Skin changes / Rash Transaminitis

Cohort 1 (n=3) Gr1 Gr2 Gr3 Gr4

Cohort 2 (n=7) Gr1 Gr2 Gr3 Gr4 1 2 1 2 1 1

Cohort 3 (n=6) Gr1 Gr2 Gr3 Gr4

1

1

Expansion (n=6) Gr2 Gr3 Gr4 1 1

1 1

Gr1

1 1 1 1

1

1 1 1

2

1 4 4 1 1 1

1 1

2

1 1 2 3 1

3

2 1 1

1 1

1

2

2

alk phos: alkaline phosphatese; Gr: grade.

haematologica | 2017; 102(4)

723


A.T. Fathi et al. chemotherapy to enhance efficacy.15-18 Herein we report the results of a phase I study of alisertib combined with conventional induction chemotherapy for newly diagnosed AML. The addition of alisertib to induction was well-tolerated

without significant toxicity. Although limited by size in this phase I study, our data also suggests that the combination was promising, with higher response rates than those which have been historically reported.2,9,30,31 The composite rate of remission (CR and CRi) across all

Figure 1. Summary plot of patient clinical course, sorted by follow up and treatment arm. CR: complete remission; CRi: remission with incomplete neutrophil recovery.

Table 4. Course and outcome, by dose cohort, for patients on trial. Nineteen of twenty-two patients (86%) achieved a CR or CRi.

Case

Cohort Sex

Age

1 2 3

1 1 1

M F M

58 51 75

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

2 2 2 2 2 2 2 3 3 3 3 3 3 Exp Exp Exp Exp Exp Exp

M M F M F M M M M F F M M M F M F M M

62 70 43 48 74 71 61 69 61 61 68 69 63 42 43 64 59 64 67

Karyotype

Detected Mutations

Reinduction Response Consolidation Stem cell Status at (5+2) transplant 12-months

t(3;12;21) t(2;3) FLT3-ITD isochromosome (17q); monosomy 17p monosomy 7; trisomy 8 CEBPA, IDH2, NRAS normal IDH2 normal FLT3-ITD, FLT3-TKD, NPM1, NRAS normal NRAS, KRAS normal FLT3-ITD, NPM1 normal CEBPA, IDH1 trisomy 8 IDH2 trisomy 9; del(20q) JAK2 isochromosome (22q) complex TP53 normal IDH2 del (Y) NPM1, NRAS, APC normal IDH1, IDH2, CDH1, SMAD4 complex TP53, NRAS, KRAS normal IDH1 normal normal NPM1 normal IDH2 complex TP53, NRAS, KRAS

N N N

CR CR CRi

N Y N

Y Y N

Alive Alive Aliveâ&#x20AC;

N N N N N N N N N N N N N N Y N N Y Y

CRi CR CR Refractory CR CR CR CRi CR CR CR CR CR CRi CR CRi CR

N Y Y N N Y Y Y Y Y N Y N N Y N N

N N N N N N Y Y N Y Y Y Y N N Y Y

Dead Alive Alive Alive Alive Dead Alive Alive Alive Dead Alive Alive Alive Dead Dead Alive Alive* Alive Dead

*Remains on study. â&#x20AC; Lost to follow up. Patients 9 and 18 died prior to response assessment. Exp: Expansion; CR: complete remission; CRi: remission with incomplete neutrophil recovery.

724

haematologica | 2017; 102(4)


Alisertib and induction chemotherapy in AML

patients was 86%, and among those evaluable for response assessment, all but one achieved remission. More than one third of patients on study had high-risk AML, as established by karyotype or antecedent marrow process, and all of these patients achieved remission. All seven evaluable patients over age 65 also achieved remission. Mid-treatment bone marrow biopsies were performed according to the established approach for â&#x20AC;&#x153;7+3â&#x20AC;? induction, and only three of the twenty-two patients on study (14%) required reinduction for residual disease, a rate much lower than that previously reported with conventional induction.1,2,31 Among the high-risk patients studied, three patients had mutations impacting TP53, all with concurrent complex karyotypes. All achieved complete remission on study, and two went on to have stem cell transplantation. Nevertheless, none of the three were alive at 12 months, which is, unfortunately, consistent with the grim prognosis associated with the TP53 alteration.32-35 The pharmacokinetics of oral alisertib monotherapy has been characterized in phase I clinical trials.22,23 The drug has an apparent biological half-life of 19-23 hours, and steady-state pharmacokinetics is achieved within 7 days of repeated daily or twice daily dosing. The mean Cminss exhibited a relatively high degree of variability without a clear dose-dependent difference or even a reversal in magnitude between successive dose levels in some cases. Comparative data for the same dose levels evaluated in the present investigation have not previously been reported for the twice daily dosing schedule. Approximate values of the mean Cminss for the 50 mg BID were reported as 1,000 nM and 2,400 nM in the two single agent phase I studies. The Cminss of alisertib achieved with doses of 10, 20, and 30 mg BID in AML patients following the "7+3" remission induction regimen was in good general agreement with the expected drug levels based upon these prior early phase studies. Aurora kinases are essential regulators of chromosomal alignment and separation during mitosis. Each of the enzymes, aurora A, B and C kinases, have key and coordinated functions in the mitotic process,36-38 including centrosome maturation, spindle assembly, chromosomal separation, and mitotic checkpoint regulation.36,39-42 AAK amplification has been detected across a range of malignancies, although its role in malignant transformation remains unclear.43-45 Overexpression of AAK may be insufficient,

and some studies have suggested that additional concurrent activating mutations, in genes such as RAS, may be necessary to transform cells.46 A key contribution appears to be the role of AAK in chromosomal segregation, with overexpression leading to errors in this process. Aneuploidy, aberrant spindle formation, abortive mitoses, and other defects promote genetic instability and likely contribute to oncogenic transformation.38,43,47,48 In addition, an effect on tumor suppressor function may be important. AAK directly phosphorylates TP53, and amplification of the former may cause degradation of the latter.49-51 Nevertheless, given the broad impact on cell cycling and survival, the efficacy of AAK inhibitors in malignancy may be more generally related to anti-mitotic effects rather than to suppressing addiction to AAK activity.38 As a selective inhibitor of AAK, alisertib triggers the development of chromosomal defects and aneuploidy, with resulting cellular senescence and apoptosis in malignant cell lines.13,15 In multiple phase I and phase II clinical studies across solid tumor and hematologic malignancies,19-26 alisertib demonstrated a broad range of efficacy. In myeloid malignancies, a single-arm phase II study evaluated the efficacy of alisertib monotherapy in fifty-seven

Figure 2. Overall (OS) and progression-free survival (PFS) estimates for those treated in cohort 3 and the expansion cohort. OS is noted in the solid black line, while PFS is the red dashed line.

Figure 3. Spindle defects following alisertib treatment of baseline primary AML samples. Proliferative patient samples were left untreated, or treated with 50nM alisertib, and assessed for defects in mitotic spindle formation. The percentage of mitotic primary cells demonstrating defective monopolar spindle structures from untreated patient samples (blue bars) and those treated with alisertib (red bars) are demonstrated. All samples displayed an increase in monopolar spindle formation during mitosis following alisertib treatment. Representative images of bipolar and monopolar mitotic spindles in AML samples are shown in the panel on the right.

haematologica | 2017; 102(4)

725


A.T. Fathi et al. patients with advanced AML and high-grade MDS.25 Alisertib was administered at 50 mg twice daily for 7 consecutive days in 21-day cycles. The majority of patients had received prior therapies and had relapsed or refractory disease. Alisertib was well-tolerated, with the predominant grade 3/4 adverse events being febrile neutropenia, thrombocytopenia, anemia, and fatigue. Six responses were observed, all in AML patients, with five patients experiencing a partial remission, and one patient experiencing a prolonged CR lasting longer than a year. 49% of patients were also reported to have achieved stable disease.25 These findings suggested clinical activity with alisertib, now further supported by our data. Of twenty patients evaluable for response assessment, nineteen achieved remission. Only three of the twenty-two enrolled patients had persistent disease at mid-treatment with "7+3". We also investigated available baseline samples from patients on study. In this small subgroup of four patients, spindle formation was clearly impacted by alisertib, with a marked increase in aberrant monopolar spindle formation. Intriguingly, all four of these patients went on to achieve remission on study, despite three of the four having highrisk karyotypic abnormalities and/or MDS-derived secondary AML. Additional study of primary samples from AML patients is necessary in order to establish trends, and these are currently ongoing. The addition of alisertib was well-tolerated. Common toxicities on study were neutropenic fever, thrombocytopenia, and anemia (all expected with induction chemotherapy), and less common toxicities were rash and oral mucositis. Combining cytotoxic agents with induction chemotherapy raises concern for prolonged cytopenias; only one DLT, a case of prolonged thrombocytopenia, was noted on the study. Patients are frequently discharged following a degree of hematologic recovery rendering them clinically safe outside of the hospital, which at our institution typically includes an ANC of 500/mm3 and platelet transfusion independence. Encouragingly, among our patients, the median time to such a threshold was not

References 1. Rai KR, Holland JF, Glidewell OJ, et al. Treatment of acute myelocytic leukemia: a study by cancer and leukemia group B. Blood. 1981;58(6):1203-1212. 2. Yates J, Glidewell O, Wiernik P, et al. Cytosine arabinoside with daunorubicin or adriamycin for therapy of acute myelocytic leukemia: a CALGB study. Blood. 1982;60(2):454-462. 3. Yates JW, Wallace HJ, Jr., Ellison RR, Holland JF. Cytosine arabinoside (NSC-63878) and daunorubicin (NSC-83142) therapy in acute nonlymphocytic leukemia. Cancer Chemother Rep. 1973;57(4):485-488. 4. Garcia-Manero G, Tambaro FP, Bekele NB, et al. Phase II trial of vorinostat with idarubicin and cytarabine for patients with newly diagnosed acute myelogenous leukemia or myelodysplastic syndrome. J Clin Oncol. 2012;30(18):2204-2210. 5. Petersdorf SH, Kopecky KJ, Slovak M, et al. A phase 3 study of gemtuzumab ozogam-

726

6.

7.

8.

9.

greatly impacted, at 33 days. Nevertheless, we do acknowledge that there was a mild prolongation of time to peripheral count recovery, when compared to the approximate hospitalization duration expected for standard induction. However, this modest prolongation among our population did not contribute to morbidity or mortality in this study. The two deaths seen during induction, though related to cytopenias, occurred prior to day 30. Our study was limited by its small size, being a phase I study with the key aim of determining tolerability and an MTD/RP2D. It was also limited by the relative heterogeneity of its population, with a mixture of high- and intermediate-risk AML cases. The mutational profile of patients in this small phase I study may also not fully reflect the larger population of AML patients. Additionally, a sizeable proportion of our patients went on to consolidative stem cell transplantation, and did not receive ongoing therapy on study. These trends are important since they can impact the interpretation of efficacy endpoints, a consideration which reinforces the need for further clinical study. Despite this, the overall outcomes for our cohort are favorable relative to historical outcomes, particularly given their age (median 63 years) and disease risk (36% with secondary AML and 18% with poor-risk cytogenetics). We demonstrated that alisertib combined with induction chemotherapy is safe and welltolerated, and established an RP2D of 30mg BID in this setting. We also found that alisertib is safe among higher risk patients, with a promising suggestion of efficacy. All patients with high-risk AML (as defined by karyotype or antecedent marrow process), and all evaluable patients above age 65 achieved remission on study. Based on this promising data, we are currently conducting a phase II study of alisertib combined with “7+3” induction chemotherapy, specifically for newly diagnosed patients with higher risk AML (clinicaltrials.gov Identifier:02560025). Acknowledgments Takeda Pharmaceuticals kindly provided funding and a supply of alisertib for the conduct of this study.

icin during induction and postconsolidation therapy in younger patients with acute myeloid leukemia. Blood. 2013;121(24): 4854-4860. Ravandi F, Cortes JE, Jones D, et al. Phase I/II study of combination therapy with sorafenib, idarubicin, and cytarabine in younger patients with acute myeloid leukemia. J Clin Oncol. 2010;28(11):18561862. Serve H, Krug U, Wagner R, et al. Sorafenib in combination with intensive chemotherapy in elderly patients with acute myeloid leukemia: results from a randomized, placebo-controlled trial. J Clin Oncol. 2013;31 (25):3110-3118. Röllig C, Müller-Tidow C, Hüttmann A, et al. Sorafenib versus placebo in addition to standard therapy in younger patients with newly diagnosed acute myeloid leukemia: results from 267 patients treated in the randomized placebo-controlled SAL-Soraml trial. Blood (ASH Annual Meeting Abstracts). 2014;124 (Abstract 6). Stone RM, Mandrekar S, Sanford BL, et al.

The multi-kinase inhibitor midostaurin (M) prolongs survival compared with placebo (P) in combination with daunorubicin (D)/cytarabine (C) induction (ind), high-dose C consolidation (consol), and as maintenance (maint) therapy in newly diagnosed acute myeloid leukemia (AML) patients (pts) age 18-60 with FLT3 mutations (muts): An international prospective randomized (rand) P-controlled double-blind trial (CALGB 10603/RATIFY [Alliance]). Blood (ASH Annual Meeting Abstracts). 2015;126 (Abstract 6). 10. 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. 11. Malumbres M, Perez de Castro I. Aurora kinase A inhibitors: promising agents in antitumoral therapy. Expert Opin Ther Targets. 2014;18(12):1377-1393. 12. Keen N, Taylor S. Aurora-kinase inhibitors as anticancer agents. Nat Rev Cancer.

haematologica | 2017; 102(4)


Alisertib and induction chemotherapy in AML

2004;4(12):927-936. 13. Hoar K, Chakravarty A, Rabino C, et al. MLN8054, a small-molecule inhibitor of Aurora A, causes spindle pole and chromosome congression defects leading to aneuploidy. Mol Cell Biol. 2007;27(12):45134525. 14. Honda R, Korner R, Nigg EA. Exploring the functional interactions between Aurora B, INCENP, and survivin in mitosis. Mol Biol Cell. 2003;14(8):3325-3341. 15. Kelly KR, Nawrocki ST, Espitia CM, et al. Targeting Aurora A kinase activity with the investigational agent alisertib increases the efficacy of cytarabine through a FOXOdependent mechanism. Int J Cancer. 2012;131(11):2693-2703. 16. Sehdev V, Katsha A, Ecsedy J, Zaika A, Belkhiri A, El-Rifai W. The combination of alisertib, an investigational Aurora kinase A inhibitor, and docetaxel promotes cell death and reduces tumor growth in preclinical cell models of upper gastrointestinal adenocarcinomas. Cancer. 2013;119(4):904-914. 17. Sehdev V, Peng D, Soutto M, et al. The aurora kinase A inhibitor MLN8237 enhances cisplatin-induced cell death in esophageal adenocarcinoma cells. Mol Cancer Ther. 2012;11(3):763-774. 18. Huck JJ, Zhang M, Mettetal J, et al. Translational exposure-efficacy modeling to optimize the dose and schedule of taxanes combined with the investigational Aurora A kinase inhibitor MLN8237 (alisertib). Mol Cancer Ther. 2014;13(9):2170-2183. 19. Cervantes A, Elez E, Roda D, et al. Phase I pharmacokinetic/pharmacodynamic study of MLN8237, an investigational, oral, selective aurora a kinase inhibitor, in patients with advanced solid tumors. Clin Cancer Res. 2012;18(17):4764-4774. 20. Dees EC, Cohen RB, von Mehren M, et al. Phase I study of aurora A kinase inhibitor MLN8237 in advanced solid tumors: safety, pharmacokinetics, pharmacodynamics, and bioavailability of two oral formulations. Clin Cancer Res. 2012;18(17):4775-4784. 21. Melichar B, Adenis A, Lockhart AC, et al. Safety and activity of alisertib, an investigational aurora kinase A inhibitor, in patients with breast cancer, small-cell lung cancer, non-small-cell lung cancer, head and neck squamous-cell carcinoma, and gastrooesophageal adenocarcinoma: a five-arm phase 2 study. Lancet Oncol. 2015;16(4):395405. 22. Matulonis UA, Sharma S, Ghamande S, et al. Phase II study of MLN8237 (alisertib), an investigational Aurora A kinase inhibitor, in patients with platinum-resistant or -refractory epithelial ovarian, fallopian tube, or primary peritoneal carcinoma. Gynecol Oncol. 2012;127(1):63-69. 23. Barr PM, Li H, Spier C, et al. Phase II Intergroup Trial of Alisertib in Relapsed and Refractory Peripheral T-Cell Lymphoma and Transformed Mycosis Fungoides: SWOG 1108. J Clin Oncol. 2015;33(21):2399-2404.

haematologica | 2017; 102(4)

24. Friedberg JW, Mahadevan D, Cebula E, et al. Phase II study of alisertib, a selective Aurora A kinase inhibitor, in relapsed and refractory aggressive B- and T-cell non-Hodgkin lymphomas. J Clin Oncol. 2014;32(1):44-50. 25. Goldberg SL, Fenaux P, Craig MD, et al. An exploratory phase 2 study of investigational Aurora A kinase inhibitor alisertib (MLN8237) in acute myelogenous leukemia and myelodysplastic syndromes. Leuk Res Rep. 2014;3(2):58-61. 26. Kelly KR, Shea TC, Goy A, et al. Phase I study of MLN8237--investigational Aurora A kinase inhibitor--in relapsed/refractory multiple myeloma, non-Hodgkin lymphoma and chronic lymphocytic leukemia. Invest New Drugs. 2014;32(3):489-499. 27. Goldberg SL, Fenaux P, Craig MD, et al. Phase 2 study of MLN8237, an investigational aurora A Kinase (AAK) inhibitor in patients with acute myelogenous leukemia (AML) or myelodysplastic syndromes (MDS). Blood (ASH Annual Meeting Abstracts). 2010;116 (Abstract 3273). 28. Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood. 2009;114(5):937-951. 29. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21(24):4642-4649. 30. Attar EC, Johnson JL, Amrein PC, et al. Bortezomib added to daunorubicin and cytarabine during induction therapy and to intermediate-dose cytarabine for consolidation in patients with previously untreated acute myeloid leukemia age 60 to 75 years: CALGB (Alliance) study 10502. J Clin Oncol. 2013;31(7):923-929. 31. Fernandez HF, Sun Z, Yao X, et al. Anthracycline dose intensification in acute myeloid leukemia. N Engl J Med. 2009;361(13):1249-1259. 32. Bejar R, Stevenson K, Abdel-Wahab O, et al. Clinical effect of point mutations in myelodysplastic syndromes. N Engl J Med. 2011;364(26):2496-2506. 33. Middeke JM, Herold S, Rucker-Braun E, et al. TP53 mutation in patients with high-risk acute myeloid leukaemia treated with allogeneic haematopoietic stem cell transplantation. Br J Haematol. 2016;172(6):914-922. 34. Wattel E, Preudhomme C, Hecquet B, et al. p53 mutations are associated with resistance to chemotherapy and short survival in hematologic malignancies. Blood. 1994;84(9):3148-3157. 35. Wong TN, Ramsingh G, Young AL, et al. Role of TP53 mutations in the origin and evolution of therapy-related acute myeloid leukaemia. Nature. 2015;518(7540):552-555.

36. Barr AR, Gergely F. Aurora-A: the maker and breaker of spindle poles. J Cell Sci. 2007;120:2987-2996. 37. Vader G, Lens SM. The Aurora kinase family in cell division and cancer. Biochim Biophys Acta. 2008;1786(1):60-72. 38. Goldenson B, Crispino JD. The aurora kinases in cell cycle and leukemia. Oncogene. 2015;34(5):537-545. 39. Hegarat N, Smith E, Nayak G, Takeda S, Eyers PA, Hochegger H. Aurora A and Aurora B jointly coordinate chromosome segregation and anaphase microtubule dynamics. J Cell Biol. 2011;195(7):11031113. 40. Marumoto T, Hirota T, Morisaki T, et al. Roles of aurora-A kinase in mitotic entry and G2 checkpoint in mammalian cells. Genes Cells. 2002;7(11):1173-1182. 41. Marumoto T, Honda S, Hara T, et al. AuroraA kinase maintains the fidelity of early and late mitotic events in HeLa cells. J Biol Chem. 2003;278(51):51786-51795. 42. Mori D, Yano Y, Toyo-oka K, et al. NDEL1 phosphorylation by Aurora-A kinase is essential for centrosomal maturation, separation, and TACC3 recruitment. Mol Cell Biol. 2007;27(1):352-367. 43. Anand S, Penrhyn-Lowe S, Venkitaraman AR. AURORA-A amplification overrides the mitotic spindle assembly checkpoint, inducing resistance to Taxol. Cancer Cell. 2003;3(1):51-62. 44. Gu J, Gong Y, Huang M, Lu C, Spitz MR, Wu X. Polymorphisms of STK15 (Aurora-A) gene and lung cancer risk in Caucasians. Carcinogenesis. 2007;28(2):350-355. 45. Sakakura C, Hagiwara A, Yasuoka R, et al. Tumour-amplified kinase BTAK is amplified and overexpressed in gastric cancers with possible involvement in aneuploid formation. Br J Cancer. 2001;84(6):824-831. 46. Tatsuka M, Sato S, Kitajima S, et al. Overexpression of Aurora-A potentiates HRAS-mediated oncogenic transformation and is implicated in oral carcinogenesis. Oncogene. 2005;24(6):1122-1127. 47. Nigg EA. Mitotic kinases as regulators of cell division and its checkpoints. Nat Rev Mol Cell Biol. 2001;2(1):21-32. 48. Zhou H, Kuang J, Zhong L, et al. Tumour amplified kinase STK15/BTAK induces centrosome amplification, aneuploidy and transformation. Nat Genet. 1998;20(2):189193. 49. Katayama H, Sasai K, Kawai H, et al. Phosphorylation by aurora kinase A induces Mdm2-mediated destabilization and inhibition of p53. Nat Genet. 2004;36(1):55-62. 50. Mao JH, Wu D, Perez-Losada J, et al. Crosstalk between Aurora-A and p53: frequent deletion or downregulation of AuroraA in tumors from p53 null mice. Cancer Cell. 2007;11(2):161-173. 51. Nair JS, Ho AL, Schwartz GK. The induction of polyploidy or apoptosis by the Aurora A kinase inhibitor MK8745 is p53-dependent. Cell cycle. 2012;11(4):807-817.

727


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Myeloid Leukemia

Ferrata Storti Foundation

Lenalidomide combined with intensive chemotherapy in acute myeloid leukemia and higher-risk myelodysplastic syndrome with 5q deletion. Results of a phase II study by the Groupe Francophone Des Myélodysplasies

Lionel Ades,1 Thomas Prebet,2 Aspasia Stamatoullas,3 Christian Recher,4 Romain Guieze,5 Emmanuel Raffoux,6 Krimo Bouabdallah,7 Mathilde Hunault,8 Eric Wattel,9 Laure Stalnikiewicz,10 Andrea Toma,11 Hervé Dombret,6 Norbert Vey,2 Marie Sebert,1 Claude Gardin,12 Cendrine Chaffaut,13 Sylvie Chevret13 and Pierre Fenaux1 Hôpital St Louis, Assistance Publique-Hôpitaux de Paris and Paris 7 University; 2Institut Paoli-Calmettes, Marseille; 3Service d’Hématologie, Rouen; 4Hématologie, CHU de Toulouse ; 5Service d’Hématologie, Clermont-Ferrand; 6Hôpital Saint Louis, Paris; 7 Service d’Hématologie Clinique et de Thérapie Cellulaire, CHU de Bordeaux Hôpital du Haut-Lévêque, Pessac; 8Service d’Hématologie, Angers; 9Service d’Hématologie, Lyon; 10 Service d’Hématologie, Lens; 11Hôpital Henri Mondor, Créteil; 12Hopital Avicenne and Paris 13 University and 13SBIM, Hôpital Saint Louis, Paris, France 1

Haematologica 2017 Volume 102(4):728-735

ABSTRACT

Correspondence: pierre.fenaux@aphp.fr

Received: July 2, 2016. Accepted: December 16, 2016. Pre-published: December 29, 2016. doi:10.3324/haematol.2016.151894 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/728 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

728

P

atients with acute myeloblastic leukemia or higher risk myelodysplastic syndromes with 5q deletion (generally within a complex karyotype) respond poorly to intensive chemotherapy and have very poor survival. In this population, we evaluated escalating doses of lenalidomide combined with intensive chemotherapy in a phase II study. Treatment consisted of daunorubicin (45 mg/m2/day, days 1-3 in cohort 1, escalated to 60 mg/m2/day, days 1-3 in cohorts 2 and 3) combined with cytosine arabinoside (200 mg/m2/day, days 1-7) and lenalidomide (10 mg/day, days 1-21 in cohorts 1 and 2, escalated to 25 mg/day, days 1-21 in cohort 3). Eighty-two patients with 5q deletion were enrolled, including 62 with acute myeloblastic leukemia, 62/79 (78%) of whom had a complex karyotype (median 7 cytogenetic abnormalities, all but 2 of them monosomal) and three had unknown karyotypes. Thirty-eight patients (46%) achieved complete remission and the overall response rate was 58.5%. Among the 62 patients with a complex karyotype, 27 achieved complete remission (44%) and 21 had cytogenetic responses. A lower response rate was observed in patients with acute myeloblastic leukemia but other pretreatment factors, including cytogenetic complexity and treatment cohort, did not significantly influence response. Fifteen patients underwent allogeneic stem cell transplantation, including 11 patients in first remission. The 1-year cumulative incidence of relapse was 64.6% and the median overall survival was 8.2 months. By comparison with conventional intensive chemotherapy, the treatment protocol we used appeared to produce higher hematologic and cytogenetic complete remission rates in patients with very poor cytogenetics, but response duration was short in this very poor risk population, highlighting the need for better post-induction strategies. Clinical trial registry number: NCT00885508 haematologica | 2017; 102(4)


Lenalidomide and chemotherapy in higher-risk MDS/AML

Introduction Deletion 5q [del(5q)] is the most common cytogenetic abnormality in myelodysplastic syndromes (MDS).1 It can occur as a single abnormality, especially in MDS without excess of bone marrow blasts and is associated in that case with a relatively favorable prognosis. On the other hand, the prognosis of MDS with del(5q) and increased bone marrow blast percentage and/or additional cytogenetic abnormalities, and of acute myeloblastic leukemia (AML) with del(5q) (isolated or complex) is poor.2 Furthermore, del(5q) is found in 40–50% of higher-risk MDS and AML with complex karyotypes (sometimes detectable only using fluorescence in situ hybridization techniques).3,4 Such patients respond poorly to intensive anthracycline-cytosine arabinoside chemotherapy with only 20–30% obtaining a complete remission, of short duration.5–8 In these patients, azacytidine may possibly produce a somewhat higher response rate, but responses are of very short duration.9 Most complex karyotypes that include del(5q) are monosomal karyotypes, and at least 50% have a TP53 mutation,10 which is also associated with a very low complete response (CR) rate; long-term survival with conventional intensive chemotherapy is very rare and there is a high relapse rate after allogeneic stem cell transplantation.11 In patients with MDS with del (5q) who had low- or intermediate-1-risk disease according to the International Prognostic Scoring System (IPSS) lenalidomide produced an erythroid response in 65% to 70% of patients, and cytogenetic partial or complete response in 50% to 75% of responders.12,13 More recently, lenalidomide was tested in IPSS intermediate-2- and high-risk MDS with del(5q), with a lower 28% overall response in our experience, and 36% in the Nordic MDS group’s experience.14,15 On the other hand, six of the nine patients with isolated del(5q) in our series reached hematologic complete remission, compared to one of the 38 patients with additional cytogenetic abnormalities. This suggests a specific effect of lenalidomide on the del(5q) clone, which was possibly sufficient to induce a response in the case of isolated del(5q), but not if other chromosomal abnormalities were present.14 Those results prompted us to combine intensive chemotherapy and lenalidomide in higher risk MDS and AML with del(5q), generally as part of a complex monosomal karyotype.

Methods Trial design This was a phase II clinical trial (NCT00885508) using the combination of anthracycline-cytosine arabinoside (AraC) chemotherapy and lenalidomide in IPSS high- and intermediate-2 (“higher”)risk MDS and AML with 5q(del). We used the Simon two-stage phase II design in order to assess whether the response rate to lenalidomide combined with escalating dose chemotherapy, compared to that observed with chemotherapy alone or lenalidomide alone, was particularly promising (at least 50% responses) or not promising (less than 30% responses), controlling for type I and II error rates of 0.025 and 0.10, respectively. A first cohort of 31 patients was included at the first dose level (daunorubicin 45 mg/m2/day for 3 days), combined with lenalidomide 10 mg/day for 21 days, in order to estimate the dose-limiting toxicity, defined by more than three of ten haematologica | 2017; 102(4)

patients recovering from aplasia after more than 40 days, or the occurrence of unexpected grade III-IV non-hematologic toxicity. Efficacy was defined as a response rate, including CR, CR with incomplete recovery (CRi) or marrow CR, of at least 50%. An interim analysis was planned following inclusion of the first 31 patients, in order to implement a reduction of the anthracycline dose in the subsequent cohort in the case of dose-limiting toxicity, or on the contrary an increase in the anthracycline dose if toxicity was considered acceptable. After review of the first cohort by the Data Safety Monitoring Board, toxicity was considered acceptable and the dose of daunorubicin increased to 60 mg/m2/day for 3 days in the second cohort, keeping the same daily dose of lenalidomide. Finally, the protocol was extended in August 2011 to allow a dose escalation of lenalidomide to 25 mg/day in 20 additional patients, but the final dose escalation (to lenalidomide 50 mg/day) was denied due to dose-limiting toxicity.

Patients Inclusion criteria were as follows: (i) age 18 years or older; (ii) documented diagnosis of MDS or AML according to the FrenchAmerican-British classification16 and World Health Organization (WHO) 2008 criteria,17 with IPSS intermediate-2- or high-risk MDS,18 including cases of chronic myelomonocytic leukemia with a white blood cell count less than 13x109/L and refractory anemia with excess blasts in transformation (AML/RAEB-t); (iii) del(5q) by conventional cytogenetics or by fluorescence in situ hybridization in the case of cytogenetic failure, with or without additional chromosomal changes. Conventional cytogenetic analysis was performed by analyzing G- and R-banded metaphase chromosomes in at least 20 mitoses, and results were interpreted using International System Cytogenetic Nomenclature; (iv) no contraindication to anthracycline-based intensive chemotherapy; (v) written informed consent; and (vi) negative serum or urine pregnancy test in women of childbearing potential. The trial was approved by the Comité de Protection des Personnes Paris—Ile de France (ethical committee whose approval is valid for all participating French institutions). The Groupe Francophone des Myélodysplasies sponsored the trial, and Celgene (Paris, France) provided lenalidomide and a scientific grant, but was not involved in analyzing the results of the study or writing the manuscript.

Treatment Patients in the first cohort received induction treatment with daunorubicin (45 mg/m2/day, days 1-3, by IV push) + AraC (200 mg/m2/day, days 1-7, continuous infusion) and lenalidomide (10 mg/day, days 1-21, orally) and granulocyte colony-stimulating factor (from day 8 to the end of aplasia). The dosing and schedule of lenalidomide was similar to the dosing used in our previous phase II trial evaluating lenalidomide as a single agent in the same population.14 Responders [patients who achieved CR, CRi or marrow CR according to the International Working Group (IWG) AML criteria19 for AML, and IWG 2006 criteria for MDS20] received six consolidation courses of daunorubicin (45 mg/m2, day 1), AraC (120 mg/m2/day, days 1-5, subcutaneously) and lenalidomide 10 mg/day, days 1-15, followed by maintenance lenalidomide 10 mg/day as a continuous schedule until progression or toxicity. After daunorubicin at a dose of 45 mg/m2/day had proven safe in the first cohort (n=31), escalation to daunorubicin 60 mg/m2/day during induction (3 days) and consolidation (1 day) was made in an additional cohort of 33 patients. Finally, after this dose had also been proven safe in the second cohort, a third cohort of patients was given 25 mg/day of lenalidomide while the daunorubicin dose remained unchanged at 60 mg/m2/day. 729


L. Ades et al.

Endpoints The primary trial endpoint was hematologic response to induction treatment (including CR, CRi and marrow CR), according to IWG AML criteria19 for AML, and IWG 2006 criteria for MDS.20 Secondary endpoints included cumulative incidence of relapse, event-free survival, overall survival and safety. All patients who, after induction treatment, achieved a CR, CRi or marrow CR were considered responders and were to continue treatment until relapse. In agreement with MDS and AML response criteria, complete cytogenetic response was defined by the disappearance of all chromosomal abnormalities, including del(5q) and other additional abnormalities, without appearance of new ones. A partial cytogenetic response was defined by at least a 50% reduction of the number of mitoses with any chromosomal abnormality. In agreement with IWG 2006 recommendations, the response of patients with 20% to 30% marrow blasts (AML/RAEB-t patients) was evaluated according to criteria that apply to MDS.

Statistical analysis The statistical analysis was performed on a modified intent-totreat principle, excluding diagnostic errors and consent withdrawals. Medians with interquartile ranges (IQR) and numbers with percentages are given as summary statistics for quantitative and qualitative variables, respectively. Exact 95% confidence intervals (95% CI) were computed for response rates. Censored endpoints (overall survival and eventfree survival) were estimated by the non-parametric Kaplan-Meier method, and compared between randomized groups by the logrank test. Analyses were stratified on treatment cohorts. Prognostic factors for achieving CR were assessed by the Wilcoxon rank sum test or Fisher exact test. Multivariable logistic regression modeling was used to summarize prognostic information. All statistical tests were two-sided, with P values of 0.05 or less denoting statistical significance. Statistical computations were performed on SAS 9.3 (SAS Inc., Cary, NC; USA) and R 2.13.0 (http://www.R-project.org/), at the reference date of 1 January, 2015.

ty in nine (11%) patients and associated with more than one additional abnormality (complex karyotype) in 62 (78%) patients. In patients with a complex karyotype, the median number of cytogenetic abnormalities, in addition to del(5q), was seven (range, 3–17): 28 had chromosome 17p deletion (generally associated with TP53 mutation in MDS and AML) and all but two of the patients with additional abnormalities had a monosomal karyotype. Median baseline white blood cell count, platelet count and hemoglobin level were 2.6x109/L (IQR: 1.7-5.2), 46.5x109/L (IQR: 28–93) and 8.8 g/dL (IQR: 8.2-9.6), respectively. During induction treatment, 80% of the patients received the planned schedule of lenalidomide (21 days), whereas 20% received lenalidomide for only 5 to 20 days.

Treatment outcomes Forty-eight (58.5%) of the 82 patients responded, including 38 (46%) who achieved a CR, four (4.8%) a CRi, and six (7%) a marrow CR according to IWG 2006/AML criteria, while 34 had progression or induction failure. The response rate did not differ between the three cohorts (58%, 59%, 58% in the first, second and third cohorts, respectively; P=1.00, by the Fisher exact test). Overall, 28 of the 38 patients who achieved a CR also achieved a

Results Patients’ baseline characteristics Figure 1 shows the flow chart for the trial. Between February 2009 and May 2012, 85 patients from 13 centers were included, of whom 82 were evaluable [2 patients withdrew consent and did not receive treatment, and 1 was excluded because of the absence of del(5q)]. Among the 82 evaluable patients, who constituted the modified intent-to-treat population, 31 patients were included in the first cohort, 32 in the second cohort and 19 in the third cohort. Table 1 summarizes the patients’ main characteristics at inclusion, showing no obvious difference between the three treatment cohorts. The 82 patients included 42 males, with a median age of 66 years (IQR: 58–72; range, 30–79). At inclusion, according to the WHO 2008 classification, 20 patients had RAEB2, and 62 had AML (including 22 AML/RAEB-t, with 20%30% marrow blasts). Among the 20 RAEB-2 patients, 16 had IPSS high-risk disease and four had IPSS intermediate2-risk. Among the 79 available karyotypes [conventional cytogenetic analysis failed in the other 3 patients and del(5q) was only identified by fluorescence in situ hybridization analysis], del(5q) was isolated in eight (10%) patients, associated with one additional abnormali730

Figure 1. Flow chart of the study. LEN: lenalidomide (dose in mg/day); DNR: daunorubicin (dose in mg/m2/day).

haematologica | 2017; 102(4)


Lenalidomide and chemotherapy in higher-risk MDS/AML

cytogenetic response, comprising 18 complete cytogenetic responses and ten partial cytogenetic responses. Overall, 130 consolidation courses were administered to 41 of the patients who achieved a hematologic response after induction treatment. The seven other responders did not receive the planned consolidation courses, as two underwent allogeneic stem cell transplantation, two relapsed before consolidation and three received alternative consolidation treatment [azacitidine (n=2) and clofarabine (n=1)]. Fifteen patients, corresponding to 29% of the patients aged less than 70 years, defined by age as eligible for transplantation, underwent allogeneic stem cell transplantation, 11 in first remission (9 in complete remission, 2 in partial remission), and four after induction failure.

Thirty-five of the 48 responders subsequently relapsed, including 29 of the 38 patients who had achieved a CR, with a 1-year cumulative incidence of relapse of 64.6% (95% CI: 50.7-78.4). The median duration of response was 6 months, while that of CR was 6.2 months. Among the 11 patients allografted in first remission, six relapsed, three had an early death and two were alive in complete remission after 12 and 14 months. Three of the four patients allografted after induction failure died, two from relapse and one from graft-versus-host disease, and one was still alive after 8 months. Seventy-seven patients died, 16 of them within 90 days of treatment onset (early death) leading to an early death rate of 20.8% (95% CI: 12.4-31.5%). The median overall survival was 8.2 months (95% CI: 7.15-10.5), and 1- and 2-

Table 1. Main baseline characteristics of the patients in the three cohorts.

Variables

Overall N=82 n.

Age [IQR] Gender Female Male WHO 2008 diagnosis RAEB-2 AML FAB diagnosis RAEB RAEB-t AML Performance Status 0 1 2 Cytogenetics Failure (5q by FISH) Isolated del(5q) del(5q) +1 abnormality Complex

%

Cohort 1 LEN 10/DNR 45 N=31 n. %

Cohort 2 LEN 10/DNR 60 N=32 n. %

Cohort 3 LEN 25/DNR 60 N=19 n. %

66.1 [58.2;71.8]

65.2 [59.4;73.0]

66.7 [54.6;71.9]

66.1 [61;68.8]

40 42

49% 51%

15 16

48% 52%

17 15

53% 47%

8 11

42% 58%

21 61

26% 74%

11 20

35% 65%

6 26

19% 81%

4 15

21% 79%

20 22 40

24% 27% 49%

11 7 13

35% 23% 42%

6 5 21

19% 16% 65%

3 10 6

16% 53% 31%

20 36 10

30% 55% 15%

13 11 5

45% 38% 17%

5 16 3

21% 67% 12%

2 9 3

14% 64% 22%

3 8 9 62

4% 10% 11% 75%

2 2 2 25

6% 6% 6% 82%

1 2 5 24

3% 6% 16% 75%

0 4 2 13

0% 21% 10% 69%

LEN: lenalidomide (dose in mg/day); DNR: daunorubicin (dose in mg/m2/day); WHO: World Health Organization; RAEB: refractory anemia with excess blasts; FAB: French American British; RAEB-t: refractory anemia with excess blasts in transformation; AML: acute myeloid leukemia; FISH: fluorescence in situ hybridization.

A

B

C

Figure 2. Outcome in the different treatment cohorts. (A) Overall survival, (B) event-free survival, (C) cumulative incidence of relapse.

haematologica | 2017; 102(4)

731


L. Ades et al.

year overall survival rates were 30.5% (95% CI: 22â&#x20AC;&#x201C;42.3) and 13.4% (95% CI: 7.7-23.2), respectively. The median event-free survival was 5.7 months (95% CI: 4.4-7.2). No difference in outcome was observed between treatment cohorts for cumulative incidence of relapse, event-free survival or overall survival (Figure 2). When patients were censored at the time of transplantation, the median overall survival was 7.8 months (95% CI:

Table 2. Prognostic factors for complete response to induction treatment (univariate analysis)

Variables Age, years Gender Female Male WHO diagnosis RAEB AML FAB diagnosis RAEB RAEB-t AML Karyotype Complex Isolated del5q Del5q+1 abnormality WBC count (109/L) Hemoglobin level (g/dL) Platelet count (109/L) % circulating blasts % bone marrow blasts Serum albumin level Treatment cohort DNR 45, LEN 10 DNR 60, LEN 10 DNR 60, LEN 25

Complete remission (n.) and % or median [IQR]

P-value

66.6 [58.9 ;71.2]

0.57

(18/40) 45% (20/42) 48%

0.83

(14/21) 68% (24/61) 40%

0.037

(13/20) 65% (9/22) 41% (16/40) 40%

0.46

(27/62) 44% (2/8) 25% (6/9) 67% 2.5 [1.9;3.9] 8.95[8.5;10] 49.5 [31;92] 5 [0.75;15.5] 14[10;30] 39[32;41]

0.99

0.72 0.003 0.39 0.017 0.03 0.003

(14/31) 45% (16/32) 50% (8/19) 42%

0.88

IQR: interquartile range; WHO: World Health Organization; RAEB: refractory anemia with excess blasts; AML: acute myeloid leukemia; FAB: French American British; RAEBt: refractory anemia with excess blasts in transformation; WBC: white blood cell; DNR: daunorubicin (dose in mg/m2/day); LEN: lenalidomide (dose in mg/day).

A

B

6.7-9.4), and 1- and 2-year overall survival rates were 27.6% (95% CI: 18.6-41) and 8.6% (95% CI: 3.8-19.7), respectively. The median event-free survival was 4.8 months (95% CI: 4.3-6.6) (Figure 3). Among patients who did not receive an allograft, only 16 were alive at 1 year, and five at 2 years.

Prognostic factors for response and overall survival Tables 2 and 3 summarize prognostic factors for CR and overall survival, respectively. The CR rate was significantly lower among AML patients (including those with AML/RAEB-t) (40%) than among those with MDS (68%; P=0.037), with higher blast percentage (P=0.03), higher circulating blast percentage (P=0.017) and lower hemoglobin level (P=0.003). Among the eight patients with isolated del(5q), two achieved a CR (25%) and both of these patients also achieved a complete cytogenetic response. Among the nine patients with del(5q) in association with one other abnormality, six (88%) achieved a CR, five of whom also had a cytogenetic response (3 complete and 2 partial). Finally, among the 62 patients with a complex karyotype, 27 achieved a CR (44%) and 21 (34%) had a cytogenetic response (13 complete and 8 partial). The CR rate was 25%, 67% and 44% in patients with isolated del(5q), del(5q) with one additional abnormality, and del(5q) in a complex karyotype, respectively (P=0.24); the presence of chromosome 17p abnormalities did not have a significant influence on achievement of CR. Likewise, other factors, including the revised IPSS classification, cytogenetic complexity and treatment cohort, did not influence response achievement. By multivariate analysis, WHO 2008 diagnosis [odds ratio (OR)=0.3 (0.1;0.9); P=0.03)], percentage of circulating blasts [OR=0.95 (0.94-1), P=0.035] and baseline hemoglobin level [OR=1.73 (1.14;2.62); P=0.01] retained prognostic significance for achievement of CR. Prognostic factors associated with a shorter overall survival were a higher white blood cell count (P=0.003), higher percentage of circulating blasts (P=0.009), and higher platelet count (P=0.009), while cytogenetic complexity (HR= 1.45, 95% CI: 0.82-2.58; P=0.21) and treatment cohort had no significant influence. In a multivariate Cox model, only platelet count remained of prognostic value for survival. Finally, among responders, achieving a cytogenetic

C

Figure 3. Outcome in the different treatment cohorts with transplanted patients censored. (A) Overall survival, (B) event-free survival, (C) cumulative incidence of relapse.

732

haematologica | 2017; 102(4)


Lenalidomide and chemotherapy in higher-risk MDS/AML

response was not associated with a survival advantage (median survival: 11.6 versus 11.6 months, P=0.46, logrank test).

Toxicity The median duration of hospitalization during induction treatment was 30 days (IQR: 26–35; range, 7–70). In the 48 responders, the median time to an absolute neutrophil count >1x109/L and a platelet count >50x109/L was 23 days (IQR: 17–28) and 21 days (IQR: 16–26) respectively. The median number of red blood cell and platelet units transfused during induction treatment was ten (IQR: 8–12) and seven (IQR: 6–11), respectively. No obvious differences were observed between the treatment cohorts. Grade III-IV non-hematologic toxicities (Table 4) included transient liver toxicity with an increase in transaminases (n=7), increase in creatinine level (n=2), and lung disease (n=17) related mainly to sepsis. No other clinically relevant toxicities were observed during the induction course. Of note, the grade III-IV increases in transaminases were mainly observed in the 25 mg/day lenalidomide cohort: 6/19 (31%) compared to 1/63 (2%) in patients who received 10 mg/day (P=0.0004), suggesting that dose-limiting toxicity was reached at this dose level. Due to this hepatic dose-limiting toxicity, the escalating dose process planned (to lenalidomide 50 mg/day) was stopped, and the trial closed for inclusion. During consolidation cycles, 11 patients had to be hospitalized, in all cases due to sepsis, including ten (91%) who were admitted during the first consolidation course.

Table 3. Prognostic factors for overall survival (univariate analysis).

Variable

HR

95% CI

P value

Age Gender Female Male WHO diagnosis AML RAEB FAB diagnosis RAEB RAEB-t AML Karyotype Complex Isolated del(5q) Del(5q) + 1 abnormality Complex karyotype WBC count Platelets count % circulating blasts Treatment cohort DNR 45, LEN 10 DNR 60, LEN 10 DNR 60, LEN 25

1.01

(0.98-1.03)

0.61

1.00 0.75

(0.48-1.18)

1.00 0.57

(0.33-1.01)

0.053

1.00 1.66 1.57

(0.88-3.14) (0.88-2.79)

0.12 0.12

1.00 0.89 0.54 1.45 1.02 0.99 1.01

(0.42-1.89) (0.25-1.21) (0.82-2.58) (1.01-1.04) (0.99-1) (1.00-1.04)

0.77 0.14 0.21 0.003 0.009 0.009

1.00 0.81 0.95

(0.48-1.35) (0.52-1.73)

0.41 0.87

0.22

Statistically significant variables are shown in bold. HR: hazard ratio; 95% confidence interval; LEN: lenalidomide (dose in mg/day); DNR: daunorubicin (dose in mg/m2/day); WHO: World Health Organization; RAEB: refractory anemia with excess blasts; FAB: French American British; RAEB-t: refractory anemia with excess blasts in transformation; AML: acute myeloid leukemia; FISH: fluorescence in situ hybridization.

haematologica | 2017; 102(4)

Discussion The higher-risk MDS and AML population treated in the present study was characterized by a highly complex karyotype in most cases, monosomal in 73% of the patients, with 17p deletion (generally associated with TP53 mutation) in 34% of the patients. Using a combination of classical daunorubicin-AraC chemotherapy and lenalidomide, we report a response rate of 58.5% (CR rate, 46%). A cytogenetic CR was obtained in 18/38 (47%) of the patients who achieved a CR, including 80% of those with a complex karyotype. In AML, the 3+7 regimen (with either daunorubicin 60– 90 mg/m2 or idarubicin 10–12 mg/m2), remains the standard induction therapy, yielding CR rates of 60% to 80% in younger adults and 50% to 60% in older patients.7,21 However, in patients with a complex karyotype, CR rates ranging from 25% to 30% have been reported. Patients with a monosomal karyotype and/or TP53 mutations (which are generally associated with a complex karyotype) have even lower CR rates, usually below 20%.22–24 Karyotype also influences long-term survival and in elderly patients with high-risk karyotype, the overall survival after intensive chemotherapy has been reported to be 4 months. Thus the hematologic and cytogenetic response rate observed in our patients, most of whom had had a complex and monosomal karyotype, with conventional doses of anthracyclines and AraC plus lenalidomide, appears encouraging. In particular, 34% of our patients had 17p deletion, generally associated with a complex karyotype. Moreover, the cytogenetic response rate in patients with unfavorable karyotype achieving hematologic CR with intensive chemotherapy is reportedly low (28%),25 whereas in our series the cytogenetic response rate in the patients with a complex karyotype who achieved hematologic CR was 77%. The CR rate in the present study was nevertheless significantly lower in patients with AML than in those with MDS (40% versus 68%, P=0.037) while karyotype (complex or not) had no influence. This is an unexpected finding because the prognosis of these patients is usually correlated with a complex karyotype rather than with the WHO 2008 classification. Lenalidomide, in lower-risk MDS with del(5q), appears to act largely by targeting the malignant del(5q) clone, leading to a high incidence of complete cytogenetic response.12,13 It was recently reported that the mechanism of action of lenalidomide is mediated by the degradation

Table 4. Grade III-IV non-hematologic toxicities during induction treatment.

Patients Cardiovascular Lung toxicity Transaminases Gut toxicity Creatinine level Neurological

Lenalidomide 10 mg cohorts n. %

Lenalidomide 25 mg cohort n. %

63 2 10 1 2 2 2

19 4 7 6 2 0 1

3% 16% 2% 3% 3% 3%

21% 37% 32% 11% 5% 733


L. Ades et al.

of casein kinase 1A1 (CK1α), and that the heterozygous deletion of CSNK1A1 in del(5q) MDS allowed lenalidomide to target the malignant clone selectively.26 In higherrisk MDS and AML with del(5q), however, three phase 2 studies and two reports on using lenalidomide as a single agent showed response rates of only 25% to 35%.14,15,27–29 This lower efficacy could result from cytogenetic complexity and/or to the fact that deleted segments on chromosome 5 are often different in higher-risk MDS or AML with del(5q) and lower-risk MDS with del(5q).30 In our series of higher-risk MDS and AML patients with del(5q), however, six of the nine patients with isolated del(5q) achieved CR, compared to only 1/38 of patients with additional cytogenetic abnormalities, pointing to cytogenetic complexity as a major factor of resistance. This prompted us to add conventional 3+7 chemotherapy to lenalidomide in those patients, with higher response rates than using either chemotherapy or lenalidomide alone. The fact that 80% of hematologic responders also achieved a cytogenetic response may suggest an additive effect of lenalidomide and chemotherapy on del(5q) cells. Given that the del(5q) in MDS and AML appears to be an early genetic event, even in the case of complex karyotype, such an effect on early clonal cells may be particularly important. In a recent report on lenalidomide monotherapy, followed by lenalidomide (10 mg/day for 10 days) combined with intensive chemotherapy (cytarabine: 200 mg/m2 for 10 days, daunorubicin: 50 mg/m2 for 3 days and etoposide (100 mg/m2 for 5 days) in nine patients with higher-risk MDS or AML with chromosome 5 abnormalities, four achieved a response, including two CR.31 Lenalidomide may also have different mechanisms of action in AML and MDS, which we cannot exclude in the present case. Indeed, in AML patients without del(5q), a CR/CRi rate of 30% was obtained with lenalidomide as a single agent (50 mg/day).32,33 Similarly, in MDS patients without del(5q), lenalidomide induced an erythroid response with transfusion independence in 25% to 30% of the cases, suggesting a different mechanism of action than that in patients with del(5q).

References 1. Schanz J, Steidl C, Fonatsch C, et al. Coalesced multicentric analysis of 2,351 patients with myelodysplastic syndromes indicates an underestimation of poor-risk cytogenetics of myelodysplastic syndromes in the International Prognostic Scoring System. J Clin Oncol. 2011;29(15): 1963– 1970. 2. Giagounidis AA, Aul C. The 5q- syndrome. Cancer Treat Res. 2008;142:133–148. 3. Komrokji RS, Padron E, Ebert BL, List AF. Deletion 5q MDS: molecular and therapeutic implications. Best Pract Res Clin Haematol. 2013;26(4):365–375. 4. Martínez-Ramírez A, Urioste M, Melchor L, et al. Analysis of myelodysplastic syndromes with complex karyotypes by highresolution comparative genomic hybridization and subtelomeric CGH array. Genes Chromosomes Cancer. 2005;42(3):287– 298. 5. Fenaux P, Preudhomme C, Laï JL, Morel P,

734

6.

7.

8.

9.

Despite these results, most of our patients relapsed within a few months, and the median overall survival was only 8.2 months. In AML patients aged less than 60 years with del(5q) treated with intensive chemotherapy, the 4year overall survival rate was reported to be limited (23%) and even lower (2%) when associated with a monosomal karyotype.23 Similarly, in patients with higher-risk MDS and monosomal karyotype treated with a hypomethylating agent, the CR rate was low (17%) and overall survival remained short (7 months).34 The fact that lenalidomide does not act on stem cells but only on progenitors may provide a possible explanation for the early relapses observed despite the achievement of a CR.35 The postremission strategy with low-intensity chemotherapy combined with lenalidomide that we used was possibly suboptimal in this situation, suggesting that a high or intermediate dose of AraC combined with lenalidomide should be tested. It has also been suggested, based on a recent report that haploinsufficiency of Rps14 is associated with activation of S100A8-S100A9, that inhibition of S100A8-S100A9 with pharmaceutical agents could be of potential clinical interest in this situation.36,37 In terms of toxicity, our strategy was well tolerated, without any additional hematologic toxicities compared to those following similar dose intensive chemotherapy. The dose-limiting toxicity was reached with a daily dose of 25 mg of lenalidomide with transient grade III–IV increases in transaminases in 31% of the patients, preventing an increase of the lenalidomide dose to 50 mg/day. In a UK experience with a 10 mg/day dose of lenalidomide, two of the nine patients treated had a grade III increase in transaminases.31 In conclusion, in patients with a very unfavorable karyotype including del(5q), we report a hematologic response rate of 58.5% after induction treatment combining 3+7 chemotherapy and lenalidomide. This outcome is of potential clinical interest if consolidation strategies and pre-emptive therapy after transplantation can be found to avoid the very high relapse rate still observed in this very poor-risk population.

Beuscart R, Bauters F. Cytogenetics and their prognostic value in de novo acute myeloid leukaemia: a report on 283 cases. Br J Haematol. 1989;73(1):61–67. Estey EH, Pierce S, Keating MJ. Identification of a group of AML/MDS patients with a relatively favorable prognosis who have chromosome 5 and/or 7 abnormalities. Haematologica. 2000;85(3): 246–249. Gardin C, Turlure P, Fagot T, et al. Postremission treatment of elderly patients with acute myeloid leukemia in first complete remission after intensive induction chemotherapy: results of the multicenter randomized Acute Leukemia French Association (ALFA) 9803 trial. Blood. 2007;109(12):5129–5135. Grimwade D, Walker H, Harrison G, et al. The predictive value of hierarchical cytogenetic classification in older adults with acute myeloid leukemia (AML): analysis of 1065 patients entered into the United Kingdom Medical Research Council AML11 trial. Blood. 2001;98(5):1312–1320. Itzykson R, Thépot S, Quesnel B, et al.

10.

11.

12.

13.

Prognostic factors for response and overall survival in 282 patients with higher-risk myelodysplastic syndromes treated with aazacitidine. Blood. 2011;117(2):403–411. Sebaa A, Ades L, Baran-Marzack F, et al. Incidence of 17p deletions and TP53 mutation in myelodysplastic syndrome and acute myeloid leukemia with 5q deletion. Genes Chromosomes Cancer. 2012;51(12): 1086– 1092. Bejar R, Stevenson KE, Caughey B, et al. Somatic mutations predict poor outcome in patients with myelodysplastic syndrome after hematopoietic stem-cell transplantation. J Clin Oncol. 2014;32(25): 2691–2698. List A, Dewald G, Bennett J, et al. Lenalidomide in the myelodysplastic syndrome with chromosome 5q deletion. N Engl J Med. 2006;355(14):1456–1465. Fenaux P, Giagounidis A, Selleslag D, et al. A randomized phase 3 study of lenalidomide versus placebo in RBC transfusion-dependent patients with low-/intermediate-1-risk myelodysplastic syndromes with del5q. Blood. 2011;118(14):3765–3776.

haematologica | 2017; 102(4)


Lenalidomide and chemotherapy in higher-risk MDS/AML

14. Adès L, Boehrer S, Prebet T, et al. Efficacy and safety of lenalidomide in intermediate-2 or high-risk myelodysplastic syndromes with 5q deletion: results of a phase 2 study. Blood. 2009;113(17):3947–3952. 15. Möllgård L, Saft L, Treppendahl MB, Dybedal I, Nørgaard JM, Astermark J, et al. Clinical effect of increasing doses of lenalidomide in high-risk myelodysplastic syndrome and acute myeloid leukemia with chromosome 5 abnormalities. Haematologica. 2011;96(7):963–971. 16. Bennett JM, Catovsky D, Daniel MT, et al. Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French-American-British Cooperative Group. Ann Intern Med. 1985;103(4): 620– 625. 17. Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World Health Organization (WHO) classification of myeloid neoplasms and acute leukemia: rationale and important changes. Blood. 2009;114(5):937–951. 18. Greenberg P, Cox C, LeBeau MM, et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood. 1997;89(6):2079–2088. 19. Cheson BD, Bennett JM, Kopecky KJ, et al. Revised recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. J Clin Oncol. 2003;21 (24):4642–4649. 20. Cheson BD, Greenberg PL, Bennett JM, et al. Clinical application and proposal for modification of the International Working Group (IWG) response criteria in myelodysplasia. Blood. 2006;108(2):419–425. 21. Goldstone AH, Burnett AK, Wheatley K, et al. Attempts to improve treatment outcomes in acute myeloid leukemia (AML) in older

haematologica | 2017; 102(4)

22.

23.

24.

25.

26.

27.

28.

29.

patients: the results of the United Kingdom Medical Research Council AML11 trial. Blood. 2001;98(5):1302–1311. Medeiros BC, Othus M, Fang M, Roulston D, Appelbaum FR. Prognostic impact of monosomal karyotype in young adult and elderly acute myeloid leukemia: the Southwest Oncology Group (SWOG) experience. Blood. 2010;116(13):2224–2228. Breems DA, Van Putten WLJ, De Greef GE, et al. Monosomal karyotype in acute myeloid leukemia: a better indicator of poor prognosis than a complex karyotype. J Clin Oncol. 2008;26(29):4791–4797. Perrot A, Luquet I, Pigneux A, et al. Dismal prognostic value of monosomal karyotype in elderly patients with acute myeloid leukemia: a GOELAMS study of 186 patients with unfavorable cytogenetic abnormalities. Blood. 2011;118(3):679–685. Balleisen S, Kuendgen A, Hildebrandt B, Haas R, Germing U. Prognostic relevance of achieving cytogenetic remission in patients with acute myelogenous leukemia or highrisk myelodysplastic syndrome following induction chemotherapy. Leuk Res. 2009;33(9):1189–1193. Krönke J, Fink EC, Hollenbach PW, et al. Lenalidomide induces ubiquitination and degradation of CK1α in del(5q) MDS. Nature. 2015 ;523(7559):183–188. Sekeres MA, Gundacker H, Lancet J, et al. A phase 2 study of lenalidomide monotherapy in patients with deletion 5q acute myeloid leukemia: Southwest Oncology Group Study S0605. Blood. 2011;118(3):523–528. Melchert M, Williams C, List A. Remitting activity of lenalidomide in treatmentinduced myelodysplastic syndrome. Leukemia. 2007;21(7):1576–1578. Lancet JE, List AF, Moscinski LC. Treatment of deletion 5q acute myeloid leukemia with lenalidomide. Leukemia. 2007;21(3): 586–588.

30. Pedersen B, Jensen IM. Clinical and prognostic implications of chromosome 5q deletions: 96 high resolution studied patients. Leukemia. 1991;5(7):566–573. 31. Dennis M, Culligan D, Karamitros D, et al. Lenalidomide monotherapy and in combination with cytarabine, daunorubicin and etoposide for high-risk myelodysplasia and acute myeloid leukaemia with chromosome 5 abnormalities. Leuk Res Rep. 2013;2(2): 70–74. 32. Fehniger TA, Uy GL, Trinkaus K, et al. A phase 2 study of high-dose lenalidomide as initial therapy for older patients with acute myeloid leukemia. Blood. 2011;117(6): 1828–1833. 33. Blum W, Klisovic RB, Becker H, et al. Dose escalation of lenalidomide in relapsed or refractory acute leukemias. J Clin Oncol. 2010;28(33):4919–4925. 34. Lübbert M, Suciu S, Hagemeijer A, et al. Decitabine improves progression-free survival in older high-risk MDS patients with multiple autosomal monosomies: results of a subgroup analysis of the randomized phase III study 06011 of the EORTC Leukemia Cooperative Group and German MDS Study Group. Ann Hematol. 2016;95(2):191–199. 35. Tehranchi R, Woll PS, Anderson K, et al. Persistent malignant stem cells in del(5q) myelodysplasia in remission. N Engl J Med. 2010 ;363(11):1025–1037. 36. Schneider RK, Schenone M, Ferreira MV, et al. Rps14 haploinsufficiency causes a block in erythroid differentiation mediated by S100A8 and S100A9. Nat Med. 2016;22 (3):288–297. 37. Knipp S, Hildebrand B, Kündgen A, et al. Intensive chemotherapy is not recommended for patients aged >60 years who have myelodysplastic syndromes or acute myeloid leukemia with high-risk karyotypes. Cancer. 2007;110(2):345–352.

735


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Acute Lymphoblastic Leukemia

Ferrata Storti Foundation

CREBBP knockdown enhances RAS/RAF/MEK/ERK signaling in Ras pathway mutated acute lymphoblastic leukemia but does not modulate chemotherapeutic response

Zach A. Dixon,1 Lindsay Nicholson,2 Martin Zeppetzauer,3 Elizabeth Matheson,1 Paul Sinclair,1 Christine J. Harrison1 and Julie A. E Irving1 Newcastle Cancer Centre at the Northern Institute for Cancer Research, Newcastle University, Newcastle upon Tyne, UK; 2Academic Haematology, Newcastle University, Newcastle upon Tyne, UK and 3Children's Cancer Research Institute (CCRI), Leukemia Biology Group, Vienna, Austria

1

Haematologica 2017 Volume 102(4):736-745

ABSTRACT

R

doi:10.3324/haematol.2016.145177

elapsed acute lymphoblastic leukemia is the most common cause of cancer-related mortality in young people and new therapeutic strategies are needed to improve outcome. Recent studies have shown that heterozygous inactivating mutations in the histone acetyl transferase, CREBBP, are particularly frequent in relapsed childhood acute lymphoblastic leukemia and associated with a hyperdiploid karyotype and KRAS mutations. To study the functional impact of CREBBP haploinsufficiency in acute lymphoblastic leukemia, RNA interference was used to knock down expression of CREBBP in acute lymphoblastic leukemia cell lines and various primagraft acute lymphoblastic leukemia cells. We demonstrate that attenuation of CREBBP results in reduced acetylation of histone 3 lysine 18, but has no significant impact on cAMP-dependent target gene expression. Impaired induction of glucocorticoid receptor targets was only seen in 1 of 4 CREBBP knockdown models, and there was no significant difference in glucocorticoid-induced apoptosis, sensitivity to other acute lymphoblastic leukemia chemotherapeutics or histone deacetylase inhibitors. Importantly, we show that CREBBP directly acetylates KRAS and that CREBBP knockdown enhances signaling of the RAS/RAF/MEK/ERK pathway in Ras pathway mutated acute lymphoblastic leukemia cells, which are still sensitive to MEK inhibitors. Thus, CREBBP mutations might assist in enhancing oncogenic RAS signaling in acute lymphoblastic leukemia but do not alter response to MEK inhibitors.

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/736

Introduction

Correspondence: julie.irving@newcastle.ac.uk

Received: April 15, 2016. Accepted: December 13, 2016. Pre-published: December 15, 2016.

Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

736

Childhood acute lymphoblastic leukemia (ALL) is the most common form of childhood malignancy and cause of cancer-related death.1 Following many years of continually improving treatment protocols, incorporating risk stratification, the cure rate of children has reached excellent levels, with sustained remission approaching 90%.2 Still, relapse following therapy remains a major clinical problem, with 5-year survival rates of only 25% for children classified as high-risk.3,4 Understanding the mechanisms of relapse and targeting relapse-associated mutations may lead to improved therapies that are clearly necessary for these children.5 One gene implicated in ALL relapse encodes cyclic adenosine monophosphate (cAMP) response element binding protein (CREB) binding protein (CREBBP/CBP), a member of the KAT3 family of histone acetyltransferases (HAT) along with its paralog, EP300. CREBBP is involved in a wide range of processes, including cAMPdependent signaling, histone acetylation, acetylation-mediated activation or inactivation of non-histone proteins, Wnt signaling, cell cycle control, ubiquitination, haematologica | 2017; 102(4)


Impact of CREBBP haploinsufficiency

DNA damage repair and antigen presentation.6-12 Germline mutations in CREBBP cause Rubinstein-Taybi Syndrome, which is characterized by developmental defects and an increased susceptibility to malignancies.13,14 A study by Mullighan et al. identified that 18% of relapsed childhood ALL cases were CREBBP mutant,15 and further studies showed enrichment in the high hyperdiploid (HHD) (5168 chromosomes) and hypodiploid cytogenetic subgroups, seen in approximately 30% of cases.16-18 CREBBP is most commonly affected by heterozygous alterations, mainly point mutations, and less frequently by deletions. CREBBP mutations affect primarily the HAT domain leading to attenuation or loss of function of the mutant protein, but without altering the activity of the remaining wild-type allele.15 Thus, the ensuing functional outcome is haploinsufficiency. Biallelic alterations only occur in approximately 6% of cases.15,16 In mouse embryonic fibroblast cell models, CREBBP mutations were shown to cause reduced acetylation of CREBBP target residues, as well as reduced expression of cAMP-dependent and glucocorticoid (GC) responsive genes.15 These results, coupled with the observation that CREBBP mutations appear to be enriched at relapse, suggest that CREBBP mutations may be a determinant of drug resistance, increasing the risk of relapse. CREBBP mutations also frequently co-occur with Ras pathway activating mutations, particularly KRAS,15,17-19 suggestive of a possible link between CREBBP attenuation and oncogenic signaling. From a therapeutic view point, the global reduction in HAT activity in CREBBP mutated cells may be reversed by the use of histone deacetylase (HDAC) inhibitors and sensitivity to the HDAC inhibitor (HDACi), vorinostat, has been previously shown.15 Thus HDACi were proposed as potential therapies for CREBBP mutant ALL cases. In this study, we are the first to assess the functional effects of CREBBP haploinsufficiency in ALL cell lines and primary-derived (primagraft) ALL cells. Our data do not support a role of CREBBP mutations in modulating response to GC, other ALL chemotherapeutic drugs or HDACi. We show, however, that KRAS is directly acetylated by CREBBP and that knockdown of CREBBP is associated with enhanced signaling of the RAS/RAF/MEK/ERK pathway in Ras pathway mutant ALL cells. Importantly, sensitivity to MEK inhibition was preserved.

Methods Cell culture Two B-cell precursor ALL (BCP-ALL) cell lines lacking CREBBP alterations (as determined by Sanger Sequencing and COSMIC database), derived from pediatric samples, were used in this study. PreB 697 (recently re-named EU-3 by the original author20 and also referred to as “697” in cell line repositories) was a kind gift from Reinhard Kofler, Austria. These cells were cultured in RPMI-1640 (Sigma-Aldrich, Dorset, UK) supplemented with 10% fetal bovine serum (FBS) (Gibco, Rugby, UK). The near-haploid childhood BCP-ALL cell line, MHH-CALL-2,21,22 was purchased from DMSZ (Braunschweig, Germany) and was maintained in RPMI-1640, supplemented with 20% FBS. All cell lines were cultured at 37°C in 5% (v/v) carbon dioxide and were routinely tested for mycoplasma contamination using MycoAlert® (Lonza, Basel, Switzerland). Primagraft ALL cells were maintained in short-term culture in RPMI-1640 supplemented with 10% FBS. To produce a maximal intracellular cAMP response, cells were treated with 100 haematologica | 2017; 102(4)

μM 3-Isobutyl-1-methylxanthine (IBMX) (Sigma-Aldrich) and 10 μM forskolin (Sigma-Aldrich) for 90 minutes.

RNAi PreB 697 cells were transduced with GFP-tagged pGIPZ lentiviral small hairpin RNA (shRNA) (Thermo Scientific, Rugby, UK) targeted to CREBBP (shCBP) or a vector containing a non-silencing sequence (shNEG) as control. ShCBP and shNEG cells were cultured in selection medium consisting of RF10 containing 10 μg/mL puromycin (Gibco) for three weeks and monitored by standard flow cytometry until 100% GFP positivity was attained. Small interfering RNA (siRNA) pool transfection targeting CREBBP (sc-29244) or non-targeting control (sc-37007) (Santa Cruz Biotechnology, Wembley, UK) was carried out, using concentrations between 120 nM and 500 nM of siRNA pool by electroporation using an EPI 2500 Elektroporation impulse generator at 350 volts.

Generation of ALL primagrafts Primagrafts were originally created by intrafemoral injection of ALL cells into NOD SCID γ null mice and were characterized for Ras pathway mutation as previously described (Online Supplementary Methods and Online Supplementary Table S1).23 CREBBP alterations were assessed by SNP 6.0 arrays and genotyping console software (Affymetrix, High Wycombe, UK) and/or Sanger sequencing (primer sequences available on request). All in vivo work was carried out in compliance with a Home Office licence (60/4552).

Real-time quantitative PCR Real-time quantitative PCR (RQ-PCR) was carried out using the TaqMan® method on the 7500 Real-Time PCR System (Applied Biosystems, Warrington, UK). TaqMan® Universal PCR MasterMix (Applied Biosystems) and TaqMan® assay probes (Applied Biosystems) were used (Online Supplementary Table S2). Data analysis was conducted using the Comparative Ct Threshold Method (ΔΔCT) with TBP, a universally expressed housekeeping gene, as the reference gene.

Western blotting Membranes were probed with either anti-CBP (sc-369) (Santa Cruz), anti-AcH3K18 (ab1191) (Abcam, Cambridge, UK), anti-pERK (9101) (Cell Signalling, Hitchin, UK), anti-ERK2 (sc-153) (Santa Cruz), anti-KRAS (F234) (Santa Cruz), anti-acetyl lysine (New England Biolabs, Frankfurt am Main, Germany) or anti-αtubulin (T6074) (Sigma-Aldrich) primary antibodies. Horseradish peroxidase (HRP) conjugated secondary anti-rabbit or anti-mouse antibodies (Dako, Glostrup, Denmark) were used and membranes were developed using an ECL prime detection kit (Amersham Biosciences, Buckinghamshire, UK) followed by exposure to X-ray film. Densitometry was carried out using AIDA image analysis software (Raytest, Straubenhardt, Germany).

In vitro acetylation assay In vitro acetylation assays were performed to detect acetylation of KRAS by CREBBP using recombinant CREBBP and KRAS proteins (both Sigma-Aldrich, Vienna, Austria). Reaction was carried out for 1 hour at 30°C in non-denaturing lysis buffer with the addition of 10 μM sodium butyrate and 25 μM acetyl-CoA. Enzymatic reaction was terminated by addition of 6xSDS loading buffer and heating at 90°C for 10 minutes. Acetylation levels were assessed by Western blot analysis.

Alamar blue drug sensitivity assay Cells were treated with dexamethasone (Sigma-Aldrich), vorinostat (suberanilohydroxamic acid) (Selleckchem, Suffolk, UK) 737


Z.A. Dixon et al.

or U0126 (Calbiochem, Nottingham, UK), alone, or dexamethasone and vorinostat in combination, and incubated for 96 hours. Alamar blue reagent (Invitrogen, Rugby, UK) was added to each well and cells were incubated for an additional 4-8 hours. Cell viability was determined by fluorescence intensity output using a FLUOstar Omega plate reader and Omega data analysis software (BMG Labtech, Aylesbury, UK). Fluorescence readings were expressed as a percentage of the control vehicle (CV) and plotted as a survival curve using GraphPad Prism software (GraphPad Software Inc., San Diego, CA, USA).

Gene expression microarray Gene expression profiling (GEP) was carried out using Affymetrix technology as detailed in the Online Supplementary Methods.

Statistical analysis Statistical significance was determined using the unpaired Student t-test on GraphPad Prism software (GraphPad Software Inc.).

A

Results Stable CREBBP knockdown in BCP-ALL cells is associated with reduced acetylation of H3K18 and an expression profile linked to the GR The BCP-ALL cell line, PreB 697, was selected for lentiviral transduction as it was shown to have two CREBBP wild-type alleles (data not shown). It was also resolved at the correct protein size by western blotting (Figure 1A). Cells were transduced with a pGIPZ vector, targeting either CREBBP (shCBP) or a non-silencing control (shNEG). After selection, shCBP cells had an approximately 40% reduction in CREBBP mRNA transcript levels compared to shNEG (Figure 1B), which translated into a greater knockdown at the protein level (Figure 1A). Protein amounts were similar to those seen in primagraft ALL samples with mono-allelic CREBBP deletions (Figure 1C and D). To elucidate the functional effect of CREBBP knockdown, acetylation of CREBBP target residue, histone 3 lysine 18

B

D C

E

F

Figure 1. Generation and functional characterization of a stable CREBBP knockdown PreB-acute lymphoblastic leukemia (ALL) cell line. (A) Western analysis of CREBBP in PreB 697 shCBP and shNEG cells with α-tubulin serving as a loading control. (B) Histogram shows mean±SD (n=3) of CREBBP mRNA expression relative to TBP by RQ-PCR in PreB 697 shCBP and shNEG cells. (C) CREBBP protein expression relative to α-tubulin analyzed by western blotting in a panel of ALL primagrafts carrying wild-type (Wt 1, 2 and 3) or mono-allelic deletion in CREBBP (Del 1 and 2). (D) Histogram of the relative expression of CREBBP in PreB 697 cell lines (left two bars) and primagraft samples normalized to α-tubulin after densitometric analyses. (E) Western analyses of AcH3K18 relative to α-tubulin in PreB 697 shCBP and shNEG cells. (F) Growth curves of PreB 697 shCBP and shNEG cells at various seeding densities. Each point represents the mean cell number±SD (n=3).

738

haematologica | 2017; 102(4)


Impact of CREBBP haploinsufficiency

(H3K18), was assessed by western blotting and was significantly reduced compared to shNEG cells (Figure 1E). There was no effect of CREBBP knockdown on cell growth or viability (mean±SD; shCBP 32.4 hours±5.4; shNEG 30.7 hours±4.8; P=0.7 by Student t-test) (Figure 1F). Given the important role of CREBBP in cAMP-dependent signaling, CREBBP knock-down cells were next studied for expression of cAMP-dependent target genes in response to increased intracellular cAMP. Cells were treated with 100 μM IBMX and 10 μM forskolin to produce a maximal intracellular cAMP response, and analyzed by RQ-PCR for expression of a selection of cAMP-dependent genes. These genes included those previously identified in mouse embryonic fibroblasts (MEF)15 and also those determined by GEP of PreB 697 cells treated with IBMX and forskolin (Online Supplementary Table S3). While induction of gene expression was shown for all selected genes, there was no significant difference in the level of induced gene expression between shCBP and shNEG cells (P>0.3) (Figure 2). To investigate the effect of CREBBP knockdown on global gene expression, GEP was performed and the expression of 28 genes was found to be significantly altered in shCBP cells compared to shNEG (Online Supplementary Table S4). Ingenuity pathway analysis of the data set showed that transcription of NR3C1, the gene

encoding the glucocorticoid receptor (GR) may be affected in CREBBP knockdown cells, based on differential expression of the upstream genes SCG2, OAT and DNAJC15 (P=0.04), indicative of a link between CREBBP and GC response.

Stable CREBBP knockdown impairs expression of GR target genes but does not alter response to GC To assess the effect of CREBBP knockdown on GC response, shCBP cells were analyzed for induced expression of the classical GR transcriptional targets, GILZ (TSC22D3), FKBP5, NR3C1 (GR) and ITGA9, following 24-hour exposure to either 17 nM (GI50 for PreB 697 cells) dexamethasone or CV. ShCBP cells showed a significant impairment in the GC-induced expression of GILZ (P=0.009), FKBP5 (P=0.03) and NR3C1 (P=0.0003), with expression levels approximately half of that seen in shNEG cells, but no such effect was seen for ITGA9 (P=0.9) (Figure 3A). However, there was no significant effect on sensitivity to dexamethasone, assessed by viability assay following 96-hour exposure (GI50 values, mean±SD; shCBP 16.6nM±5.5 and shNEG 16.7nM±3.1; P=0.9) (Figure 3B). There was also no difference in drug response to the class I/II HDAC inhibitor, vorinostat (GI50 values, mean±SD; shCBP 485 nM±27.8 and shNEG

Figure 2. Stable CREBBP knockdown in PreB 697 cells and induced expression of cAMP-dependent genes. Histogram of RQ-PCR data showing mean and SD of cAMP-dependent gene expression relative to the reference gene TBP in PreB 697 shCBP and shNEG cells following 90-minute dosing with control vehicle (CV) or IBMX/forskolin (I&F) treatment (n=3).

haematologica | 2017; 102(4)

739


Z.A. Dixon et al.

471 nM±48.75; P=0.9) or to a combination of dexamethasone and vorinostat (P=0.8) (Figure 3B). Further to this, there was no significant differential sensitivity to the common ALL therapeutics, daunorubicin, vincristine, methotrexate and 6-thioguanine (P>0.07) (Online Supplementary Figure S1).

Transient knockdown of CREBBP in PreB 697 and MHH-CALL-2 cells has no impact on GR target gene induction or GC response To confirm these data, we used a pool of small interfering RNA (siRNA) targeting CREBBP transfected into PreB 697, as well as MHH-CALL-2 cells. Results were similar for both cell lines. For PreB 697 cells, CREBBP knockdown was associated with reduced acetyl H3K18 (AcH3K18) levels (Figure 4B) but there was no significant impairment

in cAMP-dependent target expression (Figure 4C). Unlike in the stably transduced PreB 697 cells, siRNA did not show a significant alteration of GC-induced expression of GR targets; GILZ, FKBP5 and ITGA9 (PreB 697 GILZ; P=0.6, FKBP5; P=0.6, ITGA9; P=0.2). There was, however, a statistically significant increase in the expression of NR3C1 (P=0.01) (Figure 4D) and an accompanying significant sensitization to dexamethasone in PreB 697 siCBP cells (GI50 values, mean±SD; siCBP 17 nM±1.6, control 34.7 nM±2.7; P=0.0006) (Figure 4E). In MHH-CALL-2 cells, transient CREBBP knockdown was associated with reduced AcH3K18 expression, but neither impairment in induced cAMP-dependent nor GR target gene expression (CXCR4, MKNK2, DUSP10 and RGS16; P>0.2 and GILZ, FKBP5, NR3C1 and ITGA9; P>0.7, respectively). In line with this, there was no difference in sensitivity to dexam-

A

B

Figure 3. Stable CREBBP knockdown in PreB 697 cells leads to impaired expression of glucocorticoid receptor (GR) target genes, but does not influence glucocorticoid (GC) sensitivity. (A) Histogram of RQ-PCR data showing mean and SD of mRNA expression of GR target genes relative to TBP in PreB 697 shCBP and shNEG cells following 24-hour exposure to 17 nM dexamethasone (n=3). (B) Viability curves of PreB 697 shCBP and shNEG cells after dosing with dexamethasone, vorinostat or a combination of both, in which the concentration of vorinostat was varied and that of dexamethasone kept constant at 17 nM. Viability curves show mean±SD (n=3).

740

haematologica | 2017; 102(4)


Impact of CREBBP haploinsufficiency

ethasone (GI50 values, mean±SD; siCBP 2.8 nM±0.59; control 3.8 nM±0.7; P=0.1) (Figure 5A-E).

CREBBP knockdown in primary HHD ALL cells does not alter response to GC To add further biological relevance to the study, CREBBP knockdown using siRNA pool transfection was carried out in HHD ALL primagrafts. Knockdown was confirmed by western analyses (Figure 6A) and in 2 of 3 primagraft samples was associated with a reduction in AcH3K18 levels (a representative example is shown in Figure 6B). Overall, CREBBP knockdown had a minimal effect on the induced expression of cAMP-dependent targets or of GR targets (Figure 6C and D). Importantly, as shown in all other models, CREBBP knockdown in all 3 HHD primagraft samples showed no significant effect on sensitivity to dexamethasone compared to vehicle-treated control cells (GI50 values, mean±SD; L779 siCBP 65.2 nM±38.3 vs. control 90.8 nM±18.2; P=0.4, L829R >10 μM for both; L914 siCBP 6.1 nM±2.2 vs. control 5.6 nM±0.41; P=0.7) (Figure 6E).

KRAS is a substrate of CREBBP and knockdown of CREBBP is associated with increased Ras signaling Recent studies have shown that CREBBP mutations frequently co-occur with RAS mutations in childhood HHD ALL.18,24 To test for acetylation of KRAS by CREBBP, we performed in vitro acetylation assays using recombinant

A

B

proteins. Direct acetylation of KRAS was shown in the presence of CREBBP, identifying KRAS as a substrate for CREBBP (Figure 7A). The functional impact of CREBBP attenuation on Ras pathway signaling was investigated by RNAi-mediated knockdown of CREBBP in PreB 697 cells which harbor NRAS (G12D) mutation and in 3 Ras pathway mutant HHD ALL primagrafts (L779, L829R and L914) with NRAS (Q61R), KRAS (G13D) and CBL/FLT3 (Δ836) mutations (Online Supplementary Table S1). Western blotting showed an increase in p-ERK levels in PreB 697 shCBP cells compared to shNEG cells and in L829R and L914 primagraft ALL samples after siCBP knockdown, suggesting that CREBBP reduction enhances Ras pathway activation (Figure 7B and C). To assess the influence on sensitivity to MEK inhibition (MEKi), PreB shCBP and shNEG cells were dosed with the benchmark MEKi, U0126. CREBBP knockdown had no effect on sensitivity to U0126 (GI50 values, mean±SD; shCBP 9.8 μM±3.5 vs. shNEG 10.4 μM±SD 1.7; P=0.8) (Figure 7D). CREBBP knockdown in MHH-CALL-2 cells, which lack RAS pathway mutations, did not alter p-ERK levels or MEKi sensitivity (Online Supplementary Figure S2).

Discussion While CREBBP mutations were initially identified in lymphoid malignancies, including relapsed childhood

D

C

E

Figure 4. Transient CREBBP knockdown in PreB 697 cells does not affect induced expression of cAMP-dependent or glucocorticoid receptor (GR) target genes and has marginal effect on glucocorticoid (GC) sensitivity. Representative western blot of CREBBP (A) and AcH3K18 (B) in siCBP and control treated PreB 697 cells 24 hours post transfection, with α-tubulin used as loading control. (C) Histograms of RQ-PCR data showing mean and SD of cAMP-dependent gene expression relative to the reference gene TBP in siCBP and control treated PreB 697 cells, following 90-minute dosing with control vehicle (CV) or IBMX/forskolin (I&F) treatment (n=3). (D) Histograms of RQ-PCR data showing mean and SD of mRNA expression of GR target genes relative to TBP in siCBP and control treated PreB 697 cells (n=3). (E) Viability curves of siCBP and control treated PreB 697 cells after dosing with dexamethasone. Values plotted represent the mean % viability±SD (n=3).

haematologica | 2017; 102(4)

741


Z.A. Dixon et al.

ALL, diffuse large B-cell lymphoma and follicular lymphoma,15,25-27 it is becoming increasingly apparent that CREBBP is inactivated in a range of solid tumor types, including bladder, medulloblastoma, lung, adenoid carcinoma, esophageal cancer and thalamic glioma.28-36 To date, experimental data on the functional consequences of CREBBP inactivation in oncogenesis and disease progression are sparse, with this study being one of the first in any tumor type. CREBBP knockdown was performed in ALL cell lines both stably and transiently. Knockdown was also achieved in primagrafts from ALL patients within the HHD cytogenetic subtype, known to be commonly associated with CREBBP mutations. In the majority of cell models, induction of cAMP-dependent gene expression was unaffected by CREBBP knockdown, suggesting that remnant CREBBP protein levels were sufficient to maintain cAMP-dependent signaling. While GEP of PreB 697 shCBP cells showed differential expression of genes upstream of the GR (i.e. SCG2, OAT, DNAJC15) and significant impairment of GR target gene expression in response to GC treatment, no impact was seen on GCinduced apoptosis. In fact, CREBBP knockdown in both ALL cell lines and all ALL primagrafts tested did not impact on GC sensitivity. However, expanding the range of cell lines and primary samples to include wild-type and other RAS pathway activating mutations, may clarify some of the variability we observed on the transcription of cAMP and GR targets. Nevertheless, our data are

A

B

in line with a recent clinical study of HHD ALL cases, which showed that patients with CREBBP mutation did not have a poor response to prednisone in BFM protocols.18 Prompted by the observation that CREBBP mutations frequently co-occur with Ras pathway activating mutations, including KRAS, NRAS, PTPN11 and FLT3,18 we investigated the influence of CREBBP knockdown on Ras signaling. Both cell lines and primagraft ALL cells, with different Ras pathway mutations, showed increased levels of p-ERK after CREBBP knockdown, suggesting that CREBBP attenuation enhances Ras signaling. We show that KRAS is directly acetylated by CREBBP, a secondary modification shown to have a negative regulatory effect on RAS activity by altering conformational stability of the Switch II domain and thus interaction with guanine nucleotide exchange factors.37 The link we have identified between mutations in epigenetic modifiers such as CREBBP and enhanced oncogenic signaling is supported by a recent study in early thymocyte precursor ALL which found that inactivation of the methyltransferase, EZH2, co-operates with mutant RAS to induce hyperactive cytokine signaling, principally through STAT3, and was associated with a reduced sensitivity to JAK inhibitors.38 Further to this, inactivation of another methyltransferase, SUZ12, has also been shown to co-operate with Ras signaling in a number of solid tumors, acting here to enhance downstream oncogenic Ras transcriptional signatures.39

D

C

E

Figure 5. Transient CREBBP knockdown in MHH-CALL-2 cells does not affect induced expression of cAMP-dependent or glucocorticoid receptor (GR) target genes and has no effect on glucocorticoid (GC) sensitivity. Representative western blot of CREBBP (A) and AcH3K18 (B) in siCBP and control treated MHH-CALL-2 cells 24 hours post transfection, with Îą-tubulin used as loading control. (C) Histograms of RQ-PCR data showing mean and SD of cAMP-dependent gene expression relative to the reference gene TBP in siCBP and control treated MHH-CALL-2 cells, following 90-minute dosing with control vehicle (CV) or IBMX/forskolin (I&F) treatment (n=3). (D) Histograms of RQ-PCR data showing mean and SD of mRNA expression of GR target genes relative to TBP in siCBP and control treated MHH-CALL-2 cells following 24-hour exposure to 17 nM dexamethasone (n=3). (E) Viability curves of siCBP and control treated MHH-CALL-2 cells after dosing with dexamethasone. Values plotted represent the mean % viabilityÂąSD (n=3).

742

haematologica | 2017; 102(4)


Impact of CREBBP haploinsufficiency

A

Figure 6. CREBBP knockdown in HHD primagraft ALL cells does not significantly impair the induced expression of cAMP-dependent or glucocorticoid receptor (GR) target genes and does not impact on glucocorticoid (GC) sensitivity. Representative western blot of CREBBP (A) and AcH3K18 (B) in siCBP and control treated HHD primagraft ALL cells 24 hours post transplantation, with α-tubulin used as loading control. (C) Histograms of RQ-PCR data showing mean and SD of the fold change relative to CV in cAMP responsive genes in siCBP and control treated HHD primagraft ALL cells, following 90-minute dosing with CV or IBMX/forskolin (I&F) treatment. TBP was used as reference gene. Histograms show means of triplicate wells±SD for L779 and L829R and triplicate primagraft samples for L914 (n=3). (D) Histograms of RQ-PCR data showing mean and SD of the fold upregulation of GR target genes relative to CV in siCBP and control treated HHD primagraft ALL cells following 24 hours exposure to 1 μM dexamethasone. Histograms show intra-assay for mean±SD L779 and L829R and interassay for L914 (n=3). (E) Viability curves of siCBP and control treated HHD primagraft ALL cells after dosing with dexamethasone. Values plotted represent the mean % viability±SD of triplicate wells for L779 and L829R and triplicate experiments for L914.

C

B

D E

A

B

C

D

Figure 7. CREBBP acetylates KRAS and attenuation of CREBBP increased Ras pathway signaling without altering sensitivity to MEK inhibitors. (A) Western blot showing the acetylation of the recombinant KRAS by CREBBP as well as the autoacetylation of CREBBP, using acetylated lysine antibody (top) and KRAS antibody (bottom) to confirm equal loading. (B) Western blot for p-ERK, ERK and α-tubulin in CREBBP knockdown PreB 697 (shCBP and shNEG), L779, L829R and L914 cells. (C) Histogram of the relative expression of p-ERK in PreB 697 (shCBP and shNEG) and CREBBP knockdown primagraft samples normalized to ERK and α-tubulin after densitometric analyses. (D) Viability curves of PreB 697 shCBP and shNEG cells after 96-hour dosing with U0126. Values plotted represent the mean±SD of 3 independent experiments.

haematologica | 2017; 102(4)

743


Z.A. Dixon et al.

Here, we show that, despite an enhanced Ras pathway activation in CREBBP attenuated ALL cells, sensitivity to MEKi was retained, which may have important clinical implications. Our group has recently reported on the preclinical evaluation of the MEKi, selumetinib, and demonstrated significant differential sensitivity in Ras pathwaymutated ALL compared to ALL without RAS mutations, both in vitro and in an orthotopic xenograft model engrafted with primary ALL cells.23,40 Clinical trials of selumetinib for relapsed Ras pathway mutated ALL cases are underway, and the data presented here suggest that patients with co-occurring CREBBP and RAS mutations might still be good candidates for MEKi therapy. CREBBP plays a complex role in a wide array of cellular functions through its acetylation of numerous histone and non-histone proteins, which now includes KRAS. Additional substrates include the tumor suppressor p53 and the oncogene BCL6, with acetylation leading to activation and inactivation, respectively.7,8 In follicular lymphoma, aberrant acetylation of both p53 and BCL6 was shown in CREBBP mutated samples, suggesting an effect on tumor suppressor/oncogene balance.25 Another study implicated CREBBP mutation as an early event in follicular lymphoma, contributing to immune evasion due to

References 1. Inaba H, Greaves M, Mullighan CG. Acute lymphoblastic leukaemia. Lancet. 2013;381(9881):1943-1955. 2. Pui CH, Yang JJ, Hunger SP, et al. Childhood Acute Lymphoblastic Leukemia: Progress Through Collaboration. J Clin Oncol. 2015; 33(27):2938-2948. 3. Bhojwani D, Pui CH. Relapsed childhood acute lymphoblastic leukaemia. Lancet Oncol. 2013;14(6):e205-217. 4. Parker C, Waters R, Leighton C, et al. Effect of mitoxantrone on outcome of children with first relapse of acute lymphoblastic leukaemia (ALL R3): an open-label randomised trial. Lancet. 2010;376(9757):20092017. 5. Irving JA. Towards an understanding of the biology and targeted treatment of paediatric relapsed acute lymphoblastic leukaemia. Br J Haematol. 2016;172(5):655-666. 6. Kalkhoven E. CBP and p300: HATs for different occasions. Biochem Pharmacol. 2004;68(6):1145-1155. 7. Tang Y, Zhao W, Chen Y, Zhao Y, Gu W. Acetylation is indispensable for p53 activation. Cell. 2008;133(4):612-626. 8. Bereshchenko OR, Gu W, Dalla-Favera R. Acetylation inactivates the transcriptional repressor BCL6. Nat Genet. 2002;32(4):606613. 9. Hecht A, Vleminckx K, Stemmler MP, van Roy F, Kemler R. The p300/CBP acetyltransferases function as transcriptional coactivators of beta-catenin in vertebrates. EMBO J. 2000;19(8):1839-1850. 10. Shi D, Pop MS, Kulikov R, Love IM, Kung AL, Grossman SR. CBP and p300 are cytoplasmic E4 polyubiquitin ligases for p53. Proc Natl Acad Sci USA. 2009;106(38): 16275-16280. 11. Zimmer SN, Lemieux ME, Karia BP, et al. Mice heterozygous for CREB binding

744

12.

13. 14. 15.

16.

17.

18.

19.

20.

decreased antigen presentation.41 Thus, CREBBP mutations may have wide-ranging but varying effects across tumor types. Understanding the oncogenic mechanism of CREBBP is paramount as it may serve as a therapeutic target. Rebalancing of the physiological acetylation levels resulting from loss of HAT activity with HDAC inhibitors is an obvious therapeutic strategy, particularly since vorinostat and romidepsin are approved for the treatment of recurrent/refractory cutaneous T-cell lymphoma.42,43 A recent report in lung cancer showed that CREBBP was synthetically lethal with its paralog EP300, and depletion of either gene led to apoptosis.44 However, no differential sensitivity to vorinostat was evident in our CREBBP knockdown BCP-ALL models. Thus, these data suggest that MEK inhibitors might still be exploited in RAS pathway mutant ALL cases, irrespective of the presence of CREBBP alterations. Acknowledgments The authors would like to gratefully acknowledge Cancer Research UK (PhD studentship to JAEI and LN) and Bloodwise and European Research Council (AdV grant to CJH) for supporting this study. The authors would also like to thank Professor Renate Panzer-GrĂźmayer for critical review of the manuscript.

protein are hypersensitive to gamma-radiation and invariably develop myelodysplastic/myeloproliferative neoplasm. Exp Hematol. 2012;40(4):295-306 e295. Kretsovali A, Agalioti T, Spilianakis C, Tzortzakaki E, Merika M, Papamatheakis J. Involvement of CREB binding protein in expression of major histocompatibility complex class II genes via interaction with the class II transactivator. Mol Cell Biol. 1998; 18(11):6777-6783. Roelfsema JH, Peters DJ. Rubinstein-Taybi syndrome: clinical and molecular overview. Expert Rev Mol Med. 2007;9(23):1-16. Miller RW, Rubinstein JH. Tumors in Rubinstein-Taybi syndrome. Am J Med Genet. 1995;56(1):112-115. Mullighan CG, Zhang J, Kasper LH, et al. CREBBP mutations in relapsed acute lymphoblastic leukaemia. Nature. 2011;471 (7337):235-239. Inthal A, Zeitlhofer P, Zeginigg M, et al. CREBBP HAT domain mutations prevail in relapse cases of high hyperdiploid childhood acute lymphoblastic leukemia. Leukemia. 2012;26(8):1797-1803. Holmfeldt L, Wei L, Diaz-Flores E, et al. The genomic landscape of hypodiploid acute lymphoblastic leukemia. Nat Genet. 2013; 45(3):242-252. Malinowska-Ozdowy K, Frech C, Schonegger A, et al. KRAS and CREBBP mutations: a relapse-linked malicious liaison in childhood high hyperdiploid acute lymphoblastic leukemia. Leukemia 2015; 29(8):1656-1667. Ma X, Edmonson M, Yergeau D, et al. Rise and fall of subclones from diagnosis to relapse in pediatric B-acute lymphoblastic leukaemia. Nat Commun. 2015;6:6604. Zhou M, Gu L, Li F, Zhu Y, Woods WG, Findley HW. DNA damage induces a novel p53-survivin signaling pathway regulating cell cycle and apoptosis in acute lymphoblastic leukemia cells. J Pharmacol Exp

Ther. 2002;303(1):124-131. 21. Tomeczkowski J, Yakisan E, Wieland B, Reiter A, Welte K, Sykora KW. Absence of G-CSF receptors and absent response to GCSF in childhood Burkitt's lymphoma and B-ALL cells. Br J Haematol. 1995;89(4):771779. 22. Aburawi HE, Biloglav A, Johansson B, Paulsson K. Cytogenetic and molecular genetic characterization of the 'high hyperdiploid' B-cell precursor acute lymphoblastic leukaemia cell line MHH-CALL-2 reveals a near-haploid origin. Br J Haematol. 2011;154(2):275-277. 23. Irving J, Matheson E, Minto L, et al. Ras pathway mutations are prevalent in relapsed childhood acute lymphoblastic leukemia and confer sensitivity to MEK inhibition. Blood. 2014;124(23):3420-3430. 24. Paulsson K, Lilljebjorn H, Biloglav A, et al. The genomic landscape of high hyperdiploid childhood acute lymphoblastic leukemia. Nat Genet. 2015;47(6):672-676. 25. Pasqualucci L, Dominguez-Sola D, Chiarenza A, et al. Inactivating mutations of acetyltransferase genes in B-cell lymphoma. Nature. 2011;471(7337):189-195. 26. Green MR, Gentles AJ, Nair RV, et al. Hierarchy in somatic mutations arising during genomic evolution and progression of follicular lymphoma. Blood. 2013;121 (9):1604-1611. 27. Pastore A, Jurinovic V, Kridel R, et al. Integration of gene mutations in risk prognostication for patients receiving first-line immunochemotherapy for follicular lymphoma: a retrospective analysis of a prospective clinical trial and validation in a population-based registry. Lancet Oncol. 2015;16(9):1111-1122. 28. Gui Y, Guo G, Huang Y, et al. Frequent mutations of chromatin remodeling genes in transitional cell carcinoma of the bladder. Nat Genet. 2011;43(9):875-878. 29. Robinson G, Parker M, Kranenburg TA, et

haematologica | 2017; 102(4)


Impact of CREBBP haploinsufficiency

30.

31.

32.

33. 34. 35.

al. Novel mutations target distinct subgroups of medulloblastoma. Nature. 2012; 488(7409):43-48. Peifer M, Fernandez-Cuesta L, Sos ML, et al. Integrative genome analyses identify key somatic driver mutations of small-cell lung cancer. Nat Genet. 2012;44(10):1104-1110. Han JY, Lee YS, Kim BC, et al. Wholegenome analysis of a patient with early-stage small-cell lung cancer. Pharmacogenomics J. 2014;14(6):503-508. Gao Y, Geng J, Hong X, et al. Expression of p300 and CBP is associated with poor prognosis in small cell lung cancer. Int J Clin Exp Pathol. 2014;7(2):760-767. Ho AS, Kannan K, Roy DM, et al. The mutational landscape of adenoid cystic carcinoma. Nat Genet. 2013;45(7):791-798. Gao YB, Chen ZL, Li JG, et al. Genetic landscape of esophageal squamous cell carcinoma. Nat Genet. 2014;46(10)1097-1102. Song Y, Li L, Ou Y, et al. Identification of

haematologica | 2017; 102(4)

36.

37.

38.

39.

40.

genomic alterations in oesophageal squamous cell cancer. Nature. 2014; 509(7498):91-95. Mukasa A, Aihara K, Gotoh K, et al. Frequent h3f3a k27m mutations in thalamic gliomas from young adult patients. Neuro Oncol. 2014;16(1):140-146. Yang MH, Nickerson S, Kim ET, et al. Regulation of RAS oncogenicity by acetylation. Proc Natl Acad Sci USA. 2012;109 (27):10843-10848. Danis E, Yamauchi T, Echanique K, et al. Ezh2 Controls an Early Hematopoietic Program and Growth and Survival Signaling in Early T Cell Precursor Acute Lymphoblastic Leukemia. Cell Rep. 2016;14(8):1953-1965. De Raedt T, Beert E, Pasmant E, et al. PRC2 loss amplifies Ras-driven transcription and confers sensitivity to BRD4-based therapies. Nature. 2014;514(7521):247-251. Ryan SL, Matheson E, Grossmann V, et al.

41.

42.

43.

44.

The role of the RAS pathway in iAMP21ALL. Leukemia. 2016;30(9):1824-1831. Green MR, Kihira S, Liu CL, et al. Mutations in early follicular lymphoma progenitors are associated with suppressed antigen presentation. Proc Natl Acad Sci USA. 2015; 112(10):E1116-1125. Thurn KT, Thomas S, Moore A, Munster PN. Rational therapeutic combinations with histone deacetylase inhibitors for the treatment of cancer. Future Oncol. 2011;7(2):263-283. Mann BS, Johnson JR, Cohen MH, Justice R, Pazdur R. FDA approval summary: vorinostat for treatment of advanced primary cutaneous T-cell lymphoma. Oncologist. 2007; 12(10):1247-1252. Ogiwara H, Sasaki M, Mitachi T, et al. Targeting p300 addiction in CBP-deficient cancers causes synthetic lethality via apoptotic cell death due to abrogation of MYC expression. Cancer Discov. 2015;6(4):430445.

745


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Chronic Lymphocytic Leukemia

Ferrata Storti Foundation

Haematologica 2017 Volume 102(4):746-754

miR-125b and miR-532-3p predict the efficiency of rituximab-mediated lymphodepletion in chronic lymphocytic leukemia patients. A French Innovative Leukemia Organization study

Anne-Laure Gagez,1,2* Isabelle Duroux-Richard,3* Stéphane Leprêtre,4 Frédérique Orsini-Piocelle,5 Rémi Letestu,6 Sophie De Guibert,7 Edouard Tuaillon,8 Véronique Leblond,9 Olfa Khalifa,3 Valérie Gouilleux-Gruart,10 Anne Banos,11 Olivier Tournilhac,12 Jehan Dupuis,13 Christian Jorgensen,3,14 Guillaume Cartron1,2 and Florence Apparailly3,14 1 CNRS UMR 5235, University of Montpellier; 2Department of Clinical Hematology, University Hospital Montpellier; 3INSERM, U1183, Institute of Regenerative Medicine and Biotherapy, University Hospital Montpellier; 4Henri Becquerel Center, Rouen; 5 Department of Clinical Hematology, Hospital Center of Annecy, Pringy; 6Department of Biological Hematology, APHP, GHUPSSD, Avicenne Hospital, Bobigny; 7Department of Clinical Hematology, Pontchaillou Hospital, Rennes; 8Department of BacteriologyVirology, University Hospital Montpellier; 9Department of Hematology, La Pitié Salpétrière Hospital, Paris; 10CNRS UMR 7292, François Rabelais University, University Hospital Tours; 11Department of Hematology, Cote Basque Hospital, Bayonne; 12 Department of Clinical Hematology, University Hospital Estaing, Clermont-Ferrand; 13 Unit of Lymphoid Hematologic Malignancies, Henri Mondor Hospital, Créteil and 14 Clinical department for Osteoarticular Diseases, University Hospital Lapeyronie, Montpellier, France

*A-LG and ID-R contributed equally to this work

ABSTRACT

Correspondence: g-cartron@chu-montpellier.fr

Received: July 20, 2016. Accepted: January 23, 2017. Pre-published: January 25, 2017. doi:10.3324/haematol.2016.153189 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/746 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

746

T

he underlying in vivo mechanisms of rituximab action remain incompletely understood in chronic lymphocytic leukemia. Recent data suggest that circulating micro-ribonucleic acids correlate with chronic lymphocytic leukemia progression and response to rituximab. Our study aimed at identifying circulating micro-ribonucleic acids that predict response to rituximab monotherapy in chronic lymphocytic leukemia patients. Using a hierarchical clustering of microribonucleic acid expression profiles discriminating 10 untreated patients with low or high lymphocyte counts, we found 26 micro-ribonucleic acids significantly deregulated. Using individual real-time reverse transcription polymerase chain reaction, the expression levels of microribonucleic acids representative of these two clusters were further validated in a larger cohort (n=61). MiR-125b and miR-532-3p were inversely correlated with rituximab-induced lymphodepletion (P=0.020 and P=0.001, respectively) and with the CD20 expression on CD19+ cells (P=0.0007 and P<0.0001, respectively). In silico analyses of genes putatively targeted by both micro-ribonucleic acids revealed a central role of the interleukin-10 pathway and CD20 (MS4A1) family members. Interestingly, both micro-ribonucleic acids were negatively correlated with MS4A1 expression, while they were positively correlated with MS4A3 and MSA47. Our results identify novel circulating predictive biomarkers for rituximab-mediated lymphodepletion efficacy in chronic lymphocytic leukemia, and suggest a novel molecular mechanism responsible for the rituximab mode of action that bridges miR-125b and miR-532-3p and CD20 family members. (clinicaltrials.gov Identifier: 01370772).

haematologica | 2017; 102(4)


miRNAs predicting rituximab efficiency

Introduction Micro-ribonucleic acids (miRNAs) are a class of small noncoding RNAs that regulate gene expression at the posttranscriptional level and play an important regulatory role in many cellular processes.1 Deregulated expression of miRNAs could play a critical oncogenic or tumor-suppressor role and has therefore been associated with cancer, including hematological malignancies and, especially, B-cell lymphomas.2-4 MiRNAs correlate with the clinical characteristics or outcome of chronic lymphocytic leukemia (CLL) patients, allowing the identification of CLL subgroups with worse outcomes.5-7 Some of these miRNAs contribute to the deregulation of pathways involved in CLL oncogenic processes, such as PI3K/Akt (miR-22), NFκB (miR-9 family), or toll-like receptor 9 (miR-17~92 family).8-10 The B-cell receptor (BCR) signaling pathway was recently shown to be directly regulated by miR-34, miR150, and miR-155 in CLL,11,12 in addition to BCL2 (miR15a/16), TCL1 (miR-29 and miR-181), P53 (miR-15a/miR16-1 cluster, miR-17-5p, miR-29c and miR-34a), or PTEN (miR-26a and miR-214).13-17 Response to a CLL treatment could also be regulated by miRNAs. Thus, patients refractory to fludarabine exhibit significantly higher expression levels of miR-21, miR-148a and miR-222 than fludarabine sensitive patients.18 The activation of P53-responsive genes was found only in fludarabine responsive patients, suggesting a possible link between abnormal miRNA expression and P53 pathway dysfunction in non-responder patients. Links between miRNAs, fludarabine-refractory CLL and genomic abnormalities were further demonstrated, underlying the crucial role of MYC and P53 regulatory networks in determining cell response to fludarabine in CLL.19 Finally, in a prospective clinical trial aiming at evaluating the contribution of 17p deletion or TP53 mutation in fludarabinerefractory CLL, miR-34a expression at baseline was lower than in a control cohort of CLL non-refractory patients.20 The detection of miRNAs in serum and other body fluids under physiological and pathological conditions raises the possibility of using them as diagnostic or prognostic biomarkers.21-23 Visone et al. found that blood expression levels of miR-181b decreased in progressive CLL patients but not in patients with stable disease.24 Recently, we have shown that high miR-125b blood concentration can predict clinical benefit of rituximab (MabThera®, Rituxan®) treatment in patients with rheumatoid arthritis, and preliminary results suggested a similar prognostic value of miR-125b in B-cell lymphoma patients.25 Interestingly, the expression of miR125b is high in hematopoietic stem cells (HSCs) and decreases in committed progenitors.26 In addition, overexpression of miR-125b in mice HSCs is associated with the development of lymphoproliferative disease.27 Finally, miR125b expression is reduced in CLL patients as compared with healthy donors.28 Altogether, these studies suggest that low expression of circulating miR-125b might be associated with lymphoproliferation and poor treatment response to rituximab. The in vivo mechanisms of rituximab action remain incompletely understood, and could differ depending on the subtype of B-lymphoproliferative disorders. In vitro data demonstrate that rituximab is able to induce apoptosis, complement-dependent cytotoxicity (CDC), antibody-dependent cellular cytotoxicity (ADCC) and antibody-dependent cellular phagocytosis (ADCP). Although the influence of FcγRIIIa-V158VV on rituximab response haematologica | 2017; 102(4)

in follicular lymphoma patients strongly suggests that ADCC occurs in vivo, there is a lack of evidence for the other immune mechanisms.29 One of the main barriers to improve our knowledge is the scarcity of clinical situations in which rituximab is used alone. Indeed, chemotherapy associated with rituximab pollutes any definitive conclusions in studies attempting to analyze in vivo rituximab mechanisms of action. Recently, we conducted a clinical phase II study testing a new approach of dose-dense rituximab pre-phase before immunochemotherapy.30 This study was based on increased rituximab elimination observed in CLL patients compared to lymphoma patients in different clinical trials.31 Thus, this study provided a unique opportunity to dissect the mechanisms of action of rituximab in CLL patients. The study herein aimed at identifying a blood-based miRNA signature at diagnosis, before rituximab monotherapy, that predicts rituximab efficacy in CLL patients. It also aimed at shedding light on the miRNA-mediated molecular mechanisms involved in rituximab's mode of action in vivo.

Methods CLL2010FMP protocol A prospective, randomized, open-label, phase II study (CLL2010FMP, clinicaltrials.gov Identifier: 01370772) included 140 treatment-naive patients (aged 18-65 years) diagnosed with confirmed chronic lymphocytic leukemia, according to the International Workshop on Chronic Lymphocytic Leukemia (IWCLL) 2008 criteria, and Binet stage C, or with active Binet stage A or B.32 An additional inclusion criteria was the absence of 17p deletion, assessed by fluorescence in situ hybridization (FISH) (<10% positive nuclei). Each patient provided written informed consent before enrolment. Participating centers are listed in the Online Supplementary Information. Patients were stratified according to IGHV mutational status, FISH analysis (11q deletion or not) and were randomly assigned to receive either 6 cycles of chemoimmunotherapy combining fludarabine, cyclophosphamide, and rituximab (FCR) (rituximab 375mg/m2 for the first course, Day (D)1 and 500mg/m2 for the others, fludarabine 40mg/m2/d D2-4, cyclophosphamide 250mg/m2/d D2-4) every 28 days, or DenseFCR with an intensified rituximab pre-phase (500mg on D0, and 2000mg on D1, D8 and D15) before initiating the standard FCR treatment. The main objective was to increase the complete response rate with undetectable minimal residual disease three months after treatment, as published previously.30 In the study herein we have explored miRNA signature in a cohort of 123 patients, 61 receiving rituximab pre-phase before immunochemotherapy (Dense-FCR arm) and 62 receiving the FCR chemotherapy (standard arm). The study was approved by the institutional ethics committee of each participating center according to the principles of the Declaration of Helsinki.

Gene expression analysis Details are described in the Online Supplementary Information.

FCGR3A genotyping Single-step multiplex allele-specific PCR assays were performed as initially described by Dall’Ozzo et al. introducing minor modifications (for details see the Online Supplementary Information).

IL-10 competent CLL cells identification Interleukin (IL)-10-competent CLL cell counts were determined by flow cytometry analysis of IL-10 production. Peripheral blood 747


A-L Gagez et al.

Table 1. Patients’ characteristics for the protocol CLL2010FMP.

Cohort (n=123) n (%) Median (IQR)

Age (years) Women Binet stage AB ECOG 0 IGHV unmutated Cytogenetic abnormalities Del(13q) Del(11q) Trisomy 12 β2 microglobulin (mg/L) IL-10-competent B cells FCGR3A V/V V/F F/F

FCR (n=62) n (%) Median (IQR)

58.45 (52.82-61-83) 33 (26.83) 16 (25.81) 91 (73.98) 44 (70.97) 86 (69.92) 41 (66.13) 75/119 (63.03) 39/59 (66.10) 54/96 (56.25) 24/120 (20.00) 9/82 (10.98) 114 (92.68) 44 (35.77) 14 (12.17) 55 (47.83) 46 (40.00)

59.95 (52.32-62.00) -

17 (27.87) 47 (77.05) 45 (73.77) 36/60 (60.00)

58.34 (53.22-61.36) -

3.22 (2.42-4.20) -

25/48 (52.08) 12/59 (20.34) 4/42 (9.52) 59 (96.72) 44 (72.13)

2.80 (2.32-3.68) 2.97 (0.98-9.58)

-

6 (10.34) 29 (50.00) 23 (39.66)

-

29/48 (60.42) 12/61 (19.67) 5/40 (12.50) 3.05 (2.35-4.00) 56 (90.32) 2.97 (0.98-9.58) -

Dense FCR (n=61) n (%) Median (IQR)

8 (14.04) 26 (45.61) 23 (40.35)

IQR: interquartile range; FCR: fludarabine, cyclophosphamide, and rituximab; ECOG: Eastern Cooperative Oncology Group; IL-10: interleukin-10.

mononuclear cells (PBMCs) were purified from peripheral blood samples of the Dense-FCR arm of the protocol using FicollHypaque density gradients (Eurobio, Courtaboeuf, France).33 Clonal CLL cells were identified as CD19+ CD5+ CD20int lymphocytes with a previously described protocol.34 Analyses were performed on a CyAnTM ADP flow cytometer (Beckman Coulter, Brea, CA, USA).

Cell surface CD20 expression analysis CD20 expression was quantified using CD20-PE QuantiBRITETM reagents (Ratio 1:1) according to manufacturer’s recommendations (BD Biosciences, Le Pont-de-Claix, France). Calibration and quantification were performed using a FACSCANTO II cytometer (BD, Biosciences, Le Pont-de-Claix, France), and details are in the Online Supplementary Information. Circulating CD20 antigen was evaluated by considering the lymphocyte count and blood volume for each patient.

Statistical analysis The distribution of data was tested with the Shapiro-Wilk test. A X2 or Fisher's test were used to compare categorical data. For numerical data, medians were compared using a Student's t-test or Mann-Whitney test. All variables with a P value <0.10 in univariate analysis were included in an intermediate model. The final model variables were determined by backward selection using a Student's t-test (P<0.05 as significant model). The Spearman‘s correlation test was used to assess the association between two numerical data. All statistical analyses were performed at the conventional two-tailed α level of 0.05 using R software version 3.0.2.10.

Results Patients’ characteristics Sixty-one CLL patients were allocated to receive rituximab pre-phase. Their clinical and biological characteristics are presented in Table 1. They did not differ from the entire cohort. Median age was 58 years (interquartile range (IQR): 53-61), 28% were females and 77% were 748

Table 2. List of the 26 miRNAs differentially expressed in CLL patients with low or high lymphocyte counts at D0.

miRNA ID

Cluster 1 hsa-miR-323-3p hsa-miR-99b hsa-miR-326 hsa-miR-365 hsa-miR-125b hsa-miR-494 hsa-miR-193b hsa-miR-211 hsa-miR-193a-5p hsa-miR-212 hsa-miR-184 Cluster 2 hsa-miR-328 hsa-miR-486 hsa-miR-532-3p hsa-miR-484 hsa-miR-324-3p hsa-miR-92a hsa-miR-339-5p hsa-miR-223 hsa-miR-423-5p hsa-miR-652 hsa-miR-30c hsa-miR-16 hsa-miR-26a hsa-miR-29c hsa-miR-29a

Ct value in high lymphocyte count (mean)

Fold Change low/high lymphocyte count (mean)

P

31.8 26.5 31.3 31.6 28.6 28.4 25.6 28.0 26.5 26.4 29.1

9.6 8.7 7.5 7.5 5.2 4.3 3.3 2.8 2.1 1.8 0.2

0.022 0.003 0.001 0.004 0.033 0.022 0.009 0.012 0.020 0.012 0.041

22.9 18.9 22.1 17.4 22.2 17.2 23.3 14.5 22.8 21.2 16.6 16.6 17.6 19.6 18.9

5.1 4.7 4.4 3.7 2.9 2.9 2.8 2.5 2.3 2.2 2.0 0.4 0.3 0.1 0.1

0.015 0.013 0.010 0.028 0.004 0.001 0.010 0.034 0.013 0.038 0.029 0.044 0.031 0.002 0.002

The cycle threshold (Ct) value is defined as the number of cycles required for the fluorescent signal to cross the background level. Fold Change is the fold ratio of the geometric means of miRNA expression from low and high lymphocyte count patients. The P value is determined by t-test and considered significant for P<0.05.

haematologica | 2017; 102(4)


miRNAs predicting rituximab efficiency

A

B

C

Figure 1. MicroRNAs expression profile discriminates CLL patients with low or high lymphocyte counts before treatment. (A) The profiles of 26 microRNAs significantly differently expressed (P<0.05) between high and low lymphocyte concentration samples isolated from CLL patients (n=5/group) were visualized using a supervised heatmap (average linkage and Pearson’s correlation). Relative miRNA expression was calculated using the comparative threshold cycle (Ct) method. For normalization, the mean Ct value of all miRNA targets was used. Dendrograms indicated the correlation between groups of samples and miRNAs. Samples are in columns and miRNAs in rows. Each column represents an individual sample and each row represents a single miRNA. The heat map shows relative levels of miRNA expression in a green (low relative expression) to red (high relative expression) scale. (B-C) Expression levels of 4 miRNAs representative of each cluster, miR-193 band miR-125b for cluster 1 (B), and miR-532-3p and miR-652 for cluster 2 (C), were measured for 123 CLL patients included in the CLL2010FMP protocol, using RT-qPCR. A significant inverse correlation was observed depending on the lymphocyte counts for miR-193b r=-0.19), miR-125b (r=-0.39), miR-652 (r=-0.30) and miR532-3p (r=-0.34) (Spearman's correlation test).

Binet stage A or B. Cytogenetic analysis demonstrated del11q in 20% of patients and median lymphocyte count was 89 G/L (IQR: 43-123).

Blood miRNA expression profile discriminates CLL patients with high and low lymphocytosis Because miR-125b expression is reduced in CLL patients compared with healthy donors and is a key regulator of lymphoproliferation,28 we performed real-time PCR-based high-throughput miRNA arrays comparing CLL patients with high lymphocyte counts (>93.93G/L, Q3) versus low lymphocyte counts (<11.67G/L, Q1) at baseline. Q1 and Q3 were interquartile values of lymphocyte counts at D0 for patients of the Dense-FCR arm analyzed by TaqMan low-density array (TLDA). After filtering (fold change ≥1.5 and cycle threshold (Ct) values <32) on the differentially expressed miRNAs, we found 5 miRNAs downregulated and 21 miRNAs upregulated in CLL patients with low lymphocyte counts compared with those with high lymphocyte counts (P<0.05) (Table 2). The heat map showed results of the unsupervised hierarchical clustering based on the significantly differentially expressed miRNAs (Figure 1A). Two patterns of miRNA expression profile named cluster 1 and cluster 2 were clearly identified according to lymphocyte counts. To confirm this finding, we selected four miRNAs (two from each cluster) based on technical criteria (21<Ct<29 and 2-fold difference between high and low lymphocyte count at D0 between two miRNAs in each cluster), and on the literature.28,35,36 We therefore assessed their expression in a cohort of 123 CLL patients (Figure 1B). Expression patterns of miR-193b, miR-125b, miR-652 and miR-532-3p were consistent with haematologica | 2017; 102(4)

the array data. Scatter plots confirmed that increased lymphocyte counts were inversely correlated with the expression levels of miR-193b and miR-125b (P=0.03 and P=0.0001, respectively) for cluster 1, and miR-652 and miR-532-3p (P=0.0017 and P=0.0001, respectively) for cluster 2 (Figure 1B). No significant correlation was found between individual miRNAs (miR-125b, miR-193b, miR532-3p, miR-652) and clinical (age, Binet stage, Eastern Cooperative Oncology Group (ECOG)) or biological (IGHV mutation, cytogenetic abnormalities (del11q, del13q, trisomy 12), B10 frequency, or FcγRIIIa-158V/F polymorphism) parameters (data not shown). Finally, our results demonstrated that all 4 miRNAs were markedly downregulated in the blood of CLL patients displaying high lymphocyte counts.

miR-125b and miR-532-3p expression levels correlate with lymphodepletion observed after rituximab treatment We hypothesized that lymphocyte depletion observed after rituximab infusions was related to in vivo rituximab activity, and then monitored the lymphocyte depletion in our CLL cohort following four rituximab infusions at D22. We thus determined whether a correlation existed between the miRNA expression profile that discriminates CLL patients with high lymphocytosis before treatment and the efficacy of lymphocyte depletion with rituximab. The median lymphocyte count was 89G/L (range: 4-351) before rituximab treatment and 3G/L (range: 0.1-189) after four rituximab infusions, bringing the median lymphocyte depletion after rituximab treatment up to 95.9% (range: 5.0-99.6). 749


A-L Gagez et al.

Only 2 out of the 4 validated miRNAs, namely miR125b and miR-532-3p, were significantly correlated with lymphodepletion, whereas miR-652 and miR-193b did not correlate with lymphodepletion. Thus, the lymphodepletion rate was inversely correlated with the expression levels of miR-125b and miR-532-3p (P=0.020 and P=0.001, respectively), as shown in the scatter plots in Figure 2. However, no correlation was found between miRNAs and the clinical response assessed 3 months after immunochemotherapy by FCR. Our group recently showed that the frequency of IL-10competent B cells adversely impacts lymphodepletion following rituximab treatment of CLL patients and also correlates with clinical response assessed 3 months after immuno-chemotherapy by FCR.34 Because we reported that the IL-10-competent regulatory B cells frequency predicts efficacy of rituximab-mediated lymphodepletion and clinical CLL outcome, we integrated this third variable in our analysis. Logistic regression analyses showed that only IL-10-competent B cells frequency and miR-532-3p were associated with 90% lymphodepletion after rituximab monotherapy (odds ratio (OR)=0.87; 95% confidence interval (CI)=0.76-0.97; P=0.014, and OR=0.0002; 95% CI=<10-4-0.34; P=0.029, respectively). Receiver operating characteristic curve (ROC) using IL-10-competent B cells frequency and miR-532-3p expression levels showed a highly discriminative power (area under the curve (AUC)=0.795; 95% CI=0.652-0.939), allowing one to predict patients who will have more than 90% of lymphodepletion after rituximab pre-phase.

Putative and validated target genes of miR-125b and miR-532-3p Using the miRWalk database, a tool that compares miRNAs binding sites resulting from 5 main existing miRNA-target prediction programs (DIANA, RNA22, PicTar, miRanda and TargetScan), we investigated putative target genes for the two miRNAs associated with rituximab-induced lymphodepletion.37 Two lists of putative target genes were obtained: 5053 genes for miR-125b and 6652 for miR-532-3p. The Venny program, an interactive tool for comparing lists, identified 3151 common genes targeted by both miR-125b and miR-532-3p.38 We then

A

compared them with transcriptomic datasets available for IL-10-competent B cells.39 Among the 104 genes differentially expressed in the study that compared IL-10+ and IL10- human regulatory B cells, 33 and 46 genes overlapped with miR-125b and miR-532-3p putative targeted genes, respectively.39 Importantly, 26 genes were common targets of both miRNAs (Figure 3A). Consequently, in the context of rituximab, which is known to target the pan-B-cell marker CD20/MS4A1, we wondered whether this gene could also be targeted by miR-125b and miR-532-3p. We found that both miRNAs were predicted to target MS4A1. Pathway enrichment analysis was performed using the web-based bioinformatics application Ingenuity Pathway Analysis (IPA Ingenuity Systems) based on the in silico 26 predicted target genes common to miR-125b, miR-532-3p and differentially regulated in human IL-10â&#x20AC;&#x201C;/IL-10+ regulatory B cells, as well as on MS4A1. A hierarchical layout was built with only miRNA/mRNA interactions displaying high predicted scores and for which the correlation was experimentally observed in humans (Figure 3B). All the 9 genes presented in this figure were associated with the IL-10 pathway (EGR3, IL1A, IL10, IL10RA, IRF4, IRF5, MS4A1, TLR7 and TSC22D3).

Expression of CD20 family members, miR-125b and miR-532-3p on CLL cells We analyzed the association between miR-125b and miR-532-3p expression levels and the CD20 surface expression on CD19+/CD5+ CLL cells. In both cases, a significant inverse correlation was observed between CD20 protein and miRNA expression levels (Figure 4). A high expression level of CD20 at the surface of CLL cells correlated with a low expression level of miR-125b and miR532-3p (P=0.0007 and P<0.0001, respectively). Since miRNAs are negative regulators of gene expression, we monitored the mRNA level of CD20 in the blood of CLL patients. CD20 (MS4A1) mRNA tended to be inversely correlated with both miR-125b and miR-532-3p (Figure 5A). Interestingly, the expression of two other members of the CD20 family, namely MS4A3 and MS4A7, might also be controlled by miR-125b and miR-532-3p. Indeed, both MS4A3 and MS4A7 mRNAs present putative binding sites for miR-125b and miR-532-3p, not only at the 3â&#x20AC;&#x2122;UTR

B

Figure 2. Efficacy of lymphodepletion after rituximab treatment inversely correlates with miR-125b and miR-532-3p expression levels. miR-125b (r=-0.42) (A) and miR-532-3p (r=-0.49) (B) expression levels were quantified by RT-qPCR. The lymphodepletion was measured 22 days after 4 doses of rituximab infusion and correlated with miRNAs (Spearman's correlation test) (n=61).

750

haematologica | 2017; 102(4)


miRNAs predicting rituximab efficiency

IL-10-competent B cell Figure 3. Target gene prediction for miR-125b and miR-532-3p. (A) To identify putative miR-125b and miR-532-3p target genes, we used miRWalk software. VENNY, an interactive tool for comparing lists with Venn Diagrams, was used to predict common genes between human regulatory B cells IL-10+ and IL-10â&#x20AC;&#x201C;,39 and putative miR-125b and miR-532-3p target genes. 26 genes putatively targeted by miR-125b and miR-532-3p, and specifically dysregulated in IL-10+ B cells transcriptome analysis are listed. (B) The layout of these 26 putative targets was built in the context of rituximab treatment (MS4A1 (CD20 gene)) using Ingenuity Pathway Analysis (IPA). 9 genes revealed a central role for the IL-10 pathway.

region, but also in the promoter, 5â&#x20AC;&#x2122;UTR and coding regions as shown in Figure 5B. MS4A7 mRNA levels were positively correlated with both miR-125b and miR-532-3p expression levels, whereas MS4A3 mRNA levels were positively correlated with miR-125b expression only. Overall, these results suggest that miR-125b and miR-5323p might differentially control the expression of the CD20 family members.

A

Discussion In the study herein, we investigated whether a bloodbased miRNA signature can predict the efficacy of rituximab-mediated lymphodepletion in CLL patients, and provide some clue as to the underlying molecular mechanisms. We showed that miR-125b, miR-193b, miR-652 and miR-532-3p expression levels were inversely correlated with lymphocyte counts in untreated patients, and that miR-125b and miR-532-3p negatively correlated with lymphocyte depletion after rituximab monotherapy. Finally, our data suggest that both miR-125b and miR-5323p expression levels might provide a link between the expression level of CD20 family members and the efficacy of rituximab-mediated lymphodepletion. Both miR-125b and miR-532-3p have been previously described in leukemia disorders. Recently, it has been shown that miR-125b was involved in specific subtypes of leukemia, either through IRF4 silencing, genetic deletion or chromosomal translocation as evidenced in B-cell leukemia or myeloid leukemia, respectively.40-42 One of the two genes encoding for the mature form of miR-125b, namely miR-125b-1, maps at 11q24, a chromosomal region that is close to the epicenter of 11q23 deletions found in CLL, and might explain why miR-125b expression is reduced in CLL patients compared to healthy donors.28 In the study herein, although the number of patients with a del11q was small, no correlation was haematologica | 2017; 102(4)

B

Figure 4. Inverse correlation between miR-125b and miR-532-3p expression levels and circulating CD20 surface antigen expression on CD19+ cells in CLL patients blood. CD20 expression levels on CD19+ lymphocytes were quantified using flow cytometry. PBMCs were collected and miR-125b (r=-0.37) (A) and miR-532-3p (r=-0.29) (B) were quantified using RT-qPCR. Circulating CD20 antigen was correlated to miRNA expression (Spearman's correlation test) (n=61).

751


A-L Gagez et al. A

P-value

rho

P-value

rho

P-value

rho

B

Figure 5. miR-125b and miR-532-3p expression levels correlate with the expression of three MS4A family members. (A) Spearman’s correlation between miR-125b or miR-532-3p and MS4A mRNA family expression levels in CLL patients blood (n=61). (B) Putative miR125b and miR-532-3p target binding sites on MS4A family members. CDS: coding DNA sequence; rho: correlation coefficient; UTR: untranslated region.

found with miR-125b expression levels. In a recent study investigating miRNA changes upon B-cell receptor stimulation in distinct subclasses of CLL patients, the expression of miR-532-3p was increased at 48 hours exclusively in CLL patients with stable disease.6 Like miR-125b, the role and implication of miR-532-3p in CLL are established, especially as it is strongly associated with progression-free survival in CLL.35 Our data thus identify a novel prognostic relevance for miRNAs, and specifically for miR-125b and miR-532-3p, which are able to predict the efficacy of rituximab-mediated lymphodepletion, independently of the clinical outcome. In the study herein, logistic regression analysis showed that IL-10-competent B cells frequency and miR532-3p were associated with lymphodepletion after rituximab pre-phase, whereas no correlation was found between miRNAs and the clinical response assessed 3 months after immunochemotherapy by FCR. Taking advantage of the signatures of these 2 miRNAs for predicting the effect of rituximab, we searched for putative target genes according to these variables. Using available software and databases, we identified 3151 common putative target genes for miR-125b and miR-532-3p, which represent over 62% and 50% of miR-125b and miR532-3p targets, respectively. The lack of sequence homology between miR-125b and miR-532-3p, neither for the seed sequence nor for the entire miRNA mature sequence, does not explain such a surprisingly large number of overlapping putative targets. None of the few common validated target genes reported so far for both miRNAs have been studied together in the same cellular 752

context. In silico analyses reveal that these two miRNAs rather target distinct sequences and/or regions of the same gene, suggesting a synergistic effect by collective target regulation by both miRNAs. Among the 104 genes differentially expressed between IL-10+ and IL-10− cells,39 we found over 50% of genes putatively targeted by miR-125b and/or miR-532-3p, among which 26 genes are common for both miRNAs. Pathway enrichment analysis identified 9 genes associated with the IL-10 pathway in the rituximab context. Some of these genes are already validated targets for miR-125b. Rossi et al. showed that miR-125b was involved in T-cell differentiation through the silencing of IL-10 receptor α (IL10RA). MiR-125b expression in CD4+ T cells could contribute to the maintenance of the naive state, while its downregulation is associated with the acquisition of an effector-memory phenotype.26 MiR125b inhibits the expression of IRF4 in B lymphocytes, diffuse large B-cell lymphomas and myeloma cell lines, and promotes myeloid and B-cell neoplasm by inducing tumorigenesis in mice hematopoietic progenitor cells.40,42,43 An indirect implication of miR-532-3p on TLR7 gene expression mediated by an upregulation of IL-4 was reported in peripheral blood samples from CLL patients.44,45 The dysregulation of miR-532-3p was also evidenced in Binet A stage CLL patients as compared with a normal B-cell subset population.35 Among the miRNAs tested in relation with clinical data, miR-532-3p is part of a miRNA-based signature strongly associated with progression-free survival.35 In humans, the MS4A gene includes CD20 and 18 other genes.46 Rituximab is a chimeric type I monoclonal antibody haematologica | 2017; 102(4)


miRNAs predicting rituximab efficiency

Figure 6. Schematic representation of the potential mechanisms of action of miR-125b and miR-532-3p on the modulation of rituximab activity in CLL cells. In CLL patients displaying low expression levels of miR-125b and miR-532-3p, the MS4A1 mRNA and CD20 surface receptor are upregulated, while the expression of MS4A3 and MS4A7 mRNA are downregulated (MS4A3 and MS4A7 protein expressions were unknown). Consequently, rituximab (RTX) efficacy is optimum for lymphodepletion (Lymphodepletion +++). In contrast, CLL patients with high expression levels of miR-125b and miR-532-3p display low levels of MS4A1 mRNA and CD20 surface receptor, and high levels of MS4A3 and MS4A7 mRNA (MS4A3 and MS4A7 protein expressions were unknown), which might form hetero-oligomers with CD20 and impede optimal lymphodepletion (Lymphodepletion +) by rituximab. Interrogation marks indicate data that were not experimentally confirmed in the present study.

that specifically binds to inter-tetramers of CD20.29 CD20 could form homo- or hetero-oligomeric complexes with CD20 family members.47 Herein, we show that miR-125b and miR-532-3p putatively target other members of the MS4A family, including MS4A1 (coding CD20), MS4A3 (alias Htm4) and MS4A7. MS4A3 is a cell surface signaling molecule involved in the cell cycle of hematopoietic and tumor cells, whereas MS4A7 is expressed in lymphoid tissues.47,48 However, their role in the mode of action of rituximab is unexplored. Herein, we showed a significant correlation between miR-125b and miR-532-3p expression levels and three MS4A family members in CLL patients; CD20 being negatively correlated and MS4A3 and MS4A7 being positively correlated.48,49 Negative regulation of gene expression by miRNAs through translational repression and deadenylation-dependent decay of messengers is widely described. Emerging evidence reveal that miRNAs and their associated multiprotein complexes can directly or indirectly stimulate gene expression.50 GW182 is an essential component of the repressive miRNA complex that interacts with Ago1/2, leading to mRNA degradation. In conditions of quiescence, the absence of GW182 favors the interaction of FXR1 with Ago1/2 and induces the translation of the targeted mRNA.50 In the context of CLL, which is a unique malignancy where quiescent B cells accumulate in the peripheral blood, we may hypothesize that miR-125b and miR-532-3p

References 1. Bartel DP. MicroRNAs: target recognition and regulatory functions. Cell. 2009;136(2):215-233. 2. Calin GA, Croce CM. MicroRNA signatures

haematologica | 2017; 102(4)

might differently affect the expression of the 3 MS4A family members and thus modulate the lymphodepletion outcome upon rituximab treatment. In Figure 6, we propose a schematic explanation on how miR-125b and miR-532-3p can act in CLL patient cells to modulate rituximab efficacy on lymphodepletion. Overall, our results suggest that miR-125b and miR-5323p are potential non-invasive biomarkers, detectable in the peripheral blood of CLL patients before treatment, which predict rituximab-mediated lymphodepletion efficacy. These miRNAs might also play a role in the molecular mechanisms involved in the rituximab-mediated mechanism of action, through their implication in the IL-10 pathway, including IL-10-competent B cells, and through the modulation of the MS4A family members’ expression. Further investigations in an independent cohort are needed to further explore these hypotheses. Funding This work is supported by a public grant overseen by the French National Research Agency (ANR) as part of the “Investissements d’Avenir” program (reference Labex MAbImprove: ANR-10-LABX-53-01), the INSERM and the university of Montpellier. This study was funded by the French Innovative Leukemia Organization (FILO) group and F. Hoffman-La Roche Ltd (Basel, Switzerland).

in human cancers. Nat Rev Cancer. 2006; 6(11):857-866. 3. Lawrie CH. MicroRNAs in hematological malignancies. Blood rev. 2013;27(3):143-154. 4. Croce CM. Causes and consequences of microRNA dysregulation in cancer. Nat Rev

Genet. 2009;10(10):704-714. 5. Li PP, Wang X. Role of signaling pathways and miRNAs in chronic lymphocytic leukemia. Chin Med J (Engl). 2013; 126(21): 4175-4182. 6. Tavolaro S, Colombo T, Chiaretti S, et al.

753


A-L Gagez et al.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

754

Increased chronic lymphocytic leukemia proliferation upon IgM stimulation is sustained by the upregulation of miR-132 and miR-212. Genes Chromosomes Cancer. 2015;54(4):222-234. Calin GA, Dumitru CD, Shimizu M, et al. Frequent deletions and down-regulation of micro- RNA genes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc Natl Acad Sci USA. 2002; 99(24):1552415529. Palacios F, Abreu C, Prieto D, et al. Activation of the PI3K/AKT pathway by microRNA-22 results in CLL B-cell proliferation. Leukemia. 2015;29(1):115-125. Rushworth SA, Murray MY, Barrera LN, Heasman SA, Zaitseva L, Macewan DJ. Understanding the role of miRNA in regulating NF-kappaB in blood cancer. Am J Cancer Res. 2012;2(1):65-74. Bomben R, Gobessi S, Dal Bo M, et al. The miR-17 approximately 92 family regulates the response to Toll-like receptor 9 triggering of CLL cells with unmutated IGHV genes. Leukemia. 2012;26(7):1584-1593. Cui B, Chen L, Zhang S, et al. MicroRNA155 influences B-cell receptor signaling and associates with aggressive disease in chronic lymphocytic leukemia. Blood. 2014; 124(4):546-554. Mraz M, Chen L, Rassenti LZ, et al. miR-150 influences B-cell receptor signaling in chronic lymphocytic leukemia by regulating expression of GAB1 and FOXP1. Blood. 2014;124(1):84-95. Cimmino A, Calin GA, Fabbri M, et al. miR15 and miR-16 induce apoptosis by targeting BCL2. Proc Natl Acad Sci USA. 2005;102(39):13944-13949. Pekarsky Y, Santanam U, Cimmino A, et al. Tcl1 expression in chronic lymphocytic leukemia is regulated by miR-29 and miR181. Cancer Res. 2006;66(24):11590-11593. Mraz M, Malinova K, Kotaskova J, et al. miR-34a, miR-29c and miR-17-5p are downregulated in CLL patients with TP53 abnormalities. Leukemia. 2009;23(6):1159-1163. Fabbri M, Bottoni A, Shimizu M, et al. Association of a microRNA/TP53 feedback circuitry with pathogenesis and outcome of B-cell chronic lymphocytic leukemia. JAMA. 2011;305(1):59-67. Zou ZJ, Fan L, Wang L, et al. miR-26a and miR-214 down-regulate expression of the PTEN gene in chronic lymphocytic leukemia, but not PTEN mutation or promoter methylation. Oncotarget. 2015; 6(2):1276-1285. Ferracin M, Zagatti B, Rizzotto L, et al. MicroRNAs involvement in fludarabine refractory chronic lymphocytic leukemia. Mol Cancer. 2010;9(123. Moussay E, Palissot V, Vallar L, et al. Determination of genes and microRNAs involved in the resistance to fludarabine in vivo in chronic lymphocytic leukemia. Mol Cancer. 2010;9:115. Zenz T, Habe S, Denzel T, et al. Detailed analysis of p53 pathway defects in fludarabine-refractory chronic lymphocytic leukemia (CLL): dissecting the contribution of 17p deletion, TP53 mutation, p53-p21 dysfunction, and miR34a in a prospective clinical trial. Blood. 2009;114(13):2589-2597. De Tullio G, De Fazio V, Sgherza N, et al. Challenges and opportunities of microRNAs in lymphomas. Molecules. 2014;19(9):

14723-14781. 22. Gilad S, Meiri E, Yogev Y, et al. Serum microRNAs are promising novel biomarkers. PloS One. 2008;3(9):e3148. 23. Mitchell PS, Parkin RK, Kroh EM, et al. Circulating microRNAs as stable bloodbased markers for cancer detection. Proc Natl Acad Sci USA. 2008;105(30):1051310518. 24. Visone R, Veronese A, Rassenti LZ, et al. miR-181b is a biomarker of disease progression in chronic lymphocytic leukemia. Blood. 2011;118(11):3072-3079. 25. Duroux-Richard I, Pers YM, Fabre S, et al. Circulating miRNA-125b is a potential biomarker predicting response to rituximab in rheumatoid arthritis. Mediators Inflamm. 2014;2014(342524. 26. Rossi RL, Rossetti G, Wenandy L, et al. Distinct microRNA signatures in human lymphocyte subsets and enforcement of the naive state in CD4+ T cells by the microRNA miR-125b. Nat Immunol. 2011; 12(8):796-803. 27. Ooi AG, Sahoo D, Adorno M, Wang Y, Weissman IL, Park CY. MicroRNA-125b expands hematopoietic stem cells and enriches for the lymphoid-balanced and lymphoid-biased subsets. Proc Natl Acad Sci USA. 2010;107(50):21505-21510. 28. Tili E, Michaille JJ, Luo Z, et al. The downregulation of miR-125b in chronic lymphocytic leukemias leads to metabolic adaptation of cells to a transformed state. Blood. 2012;120(13):2631-2638. 29. Cartron G, Trappe RU, Solal-Celigny P, Hallek M. Interindividual variability of response to rituximab: from biological origins to individualized therapies. Clin Cancer Res. 2011;17(1):19-30. 30. Lepretre S, Letestu R, Dartigeas C, et al. Results of a Phase II Randomizing Intensified Rituximab Pre-Phase Followed By Standard FCR Vs Standard FCR in Previously Untreated Patients with Active B-Chronic Lymphocytic Leukemia (B-CLL). CLL2010FMP (for fit medically patients): A Study of the french Cooperative Group on CLL and WM (FCGCLL/MW) and the “Groupe Ouest-Est d’Etudes Des Leucémies Aigües Et Autres Maladies Du sang” (GOELAMS). Blood. 2014;124(21): 3329. 31. Li J, Zhi J, Wenger M, et al. Population pharmacokinetics of rituximab in patients with chronic lymphocytic leukemia. J Clin Pharmacol. 2012;52(12):1918-1926. 32. Hallek M, Cheson BD, Catovsky D, et al. Guidelines for the diagnosis and treatment of chronic lymphocytic leukemia: a report from the International Workshop on Chronic Lymphocytic Leukemia updating the National Cancer Institute-Working Group 1996 guidelines. Blood. 2008; 111(12):5446-5456. 33. Iwata Y, Matsushita T, Horikawa M, et al. Characterization of a rare IL-10-competent B-cell subset in humans that parallels mouse regulatory B10 cells. Blood. 2011; 117(2):530-541. 34. Gagez AL, Tuaillon E, Cezar R, et al. Response to rituximab in B-CLL patients is adversely impacted by frequency of IL-10 competent B cells and FcgammaRIIIa polymorphism. A study of FCGCLL/WM and GOELAMS groups. Blood Cancer J. 2016; 6:e389.

35. Negrini M, Cutrona G, Bassi C, et al. microRNAome expression in chronic lymphocytic leukemia: comparison with normal B-cell subsets and correlations with prognostic and clinical parameters. Clin Cancer Res. 2014;20(15):4141-4153. 36. Moussay E, Wang K, Cho JH, et al. MicroRNA as biomarkers and regulators in B-cell chronic lymphocytic leukemia. Proc Natl Acad Sci USA. 2011;108(16):6573-6578. 37. Dweep H, Gretz N. miRWalk2.0: a comprehensive atlas of microRNA-target interactions. Nat Methods. 2015;12(8):697. 38. Oliveros, J.C. (2007) VENNY. An interactive tool for comparing lists with Venn Diagrams.http://bioinfogp.cnb.csic.es/tools/ venny/index.html. ' 39. Lin W, Cerny D, Chua E, et al. Human regulatory B cells combine phenotypic and genetic hallmarks with a distinct differentiation fate. J Immunol. 2014;193(5):22582266. 40. So AY, Sookram R, Chaudhuri AA, et al. Dual mechanisms by which miR-125b represses IRF4 to induce myeloid and B-cell leukemias. Blood. 2014;124(9):1502-1512. 41. Sonoki T, Iwanaga E, Mitsuya H, Asou N. Insertion of microRNA-125b-1, a human homologue of lin-4, into a rearranged immunoglobulin heavy chain gene locus in a patient with precursor B-cell acute lymphoblastic leukemia. Leukemia. 2005; 19(11):2009-2010. 42. Malumbres R, Sarosiek KA, Cubedo E, et al. Differentiation stage-specific expression of microRNAs in B lymphocytes and diffuse large B-cell lymphomas. Blood. 2009; 113(16):3754-3764. 43. Gururajan M, Haga CL, Das S, et al. MicroRNA 125b inhibition of B cell differentiation in germinal centers. Int Immunol. 2010;22(7):583-592. 44. Ruiz-Lafuente N, Alcaraz-Garcia MJ, Sebastian-Ruiz S, et al. The gene expression response of chronic lymphocytic leukemia cells to IL-4 is specific, depends on ZAP-70 status and is differentially affected by an NFkappaB inhibitor. PloS One. 2014; 9(10):e109533. 45. Ruiz-Lafuente N, Alcaraz-Garcia MJ, Sebastian-Ruiz S, et al. IL-4 up-regulates MiR-21 and the MiRNAs hosted in the CLCN5 gene in chronic lymphocytic leukemia. PloS One. 2015;10(4):e0124936. 46. Clark EA, Ledbetter JA. How does B cell depletion therapy work, and how can it be improved? Ann Rheum Dis. 2005;64 Suppl 4:iv77-80. 47. Liang Y, Tedder TF. Identification of a CD20, FcepsilonRIbeta-, and HTm4-related gene family: sixteen new MS4A family members expressed in human and mouse. Genomics. 2001;72(2):119-127. 48. Kutok JL, Yang X, Folkerth R, Adra CN. Characterization of the expression of HTm4 (MS4A3), a cell cycle regulator, in human peripheral blood cells and normal and malignant tissues. J Cell Mol Med. 2011;15(1):8693. 49. Gingras MC, Lapillonne H, Margolin JF. CFFM4: a new member of the CD20/FcepsilonRIbeta family. Immunogenetics. 2001;53(6):468-476. 50. Banzhaf-Strathmann J, Edbauer D. Good guy or bad guy: the opposing roles of microRNA 125b in cancer. Cell Commun Signal. 2014;12:30.

haematologica | 2017; 102(4)


ARTICLE

Non-Hodgkin Lymphoma

Inhibition of 4EBP phosphorylation mediates the cytotoxic effect of mechanistic target of rapamycin kinase inhibitors in aggressive B-cell lymphomas

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Chengfeng Bi,1 Xuan Zhang, 1 Ting Lu,1 Xiaoyan Zhang,1 Xianhuo Wang,1,3 Bin Meng,3 Huilai Zhang,3 Ping Wang,3 Julie M. Vose,1 Wing C. Chan,2 Timothy W. McKeithan2 and Kai Fu1,3

Departments of Pathology and Microbiology and Hematology Oncology, University of Nebraska Medical Center, Omaha, NE, USA; 2Department of Pathology, City of Hope Medical Center, Duarte, CA, USA, and 3The Sino-US Lymphoma Center, Tianjin Medical University Cancer Institute and Hospital, National Cancer Research Center, China

1

Haematologica 2017 Volume 102(4):755-764

ABSTRACT

M

echanistic target of rapamycin (mTOR) complex 1 is a central integrator of nutrient and growth factor inputs that controls cell growth in eukaryotes. The second generation of mTOR kinase inhibitors (TORKi), directly targeting the mTOR catalytic site, are more effective than rapamycin and its analogs in cancer treatment, particularly in inducing apoptosis. However, the mechanism underlying the cytotoxic effect of TORKi remains elusive. Herein, we demonstrate that TORKiinduced apoptosis is predominantly dependent on the loss of mTOR complex 1-mediated 4EBP activation. Knocking out RICTOR, a key component of mTOR complex 2, or inhibiting p70S6K has little effect on TORKi-induced apoptosis. Conversely, increasing the eIF4E:4EBP ratio by either overexpressing eIF4E or knocking out 4EBP1/2 protects lymphoma cells from TORKi-induced cytotoxicity. Furthermore, downregulation of MCL1 expression plays an important role in TORKi-induced apoptosis, whereas BCL-2 overexpression confers resistance to TORKi treatment. We further show that the therapeutic effect of TORKi in aggressive B-cell lymphomas can be predicted by BH3 profiling, and improved by combining it with pro-apoptotic drugs, especially BCL-2 inhibitors, both in vitro and in vivo. Taken together, the study herein provides mechanistic insight into TORKi cytotoxicity and identified a potential way to optimize its efficacy in the clinical treatment of aggressive B-cell lymphoma.

Correspondence: kfu@unmc.edu

Received: October 31, 2016. Accepted: January 18, 2017. Pre-published: January 19, 2017. doi:10.3324/haematol.2016.159160

Introduction Aggressive B-cell lymphomas are clinically and pathologically heterogeneous entities that consist of diffuse large B-cell lymphoma (DLBCL), Burkitt lymphoma (BL), mantle cell lymphoma (MCL), and rare variants, such as double hit lymphoma (DHL). Rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP)-based chemotherapy regimens have significantly improved the clinical outcome of patients with DLBCL and MCL, whereas intensive chemotherapy was commonly used for BL. DHL responds poorly to R-CHOP treatment and remains challenging in clinical practice. Toxicity and secondary cancer risk due to intensive therapy are also high, particularly in patients treated with high-dose regimens.1,2 Improvements in therapeutic strategies and target therapy are therefore urgently needed in aggressive B-cell lymphomas. With a central role in cell survival and growth, mechanistic target of rapamycin complex 1 (mTORC1) is one of the most important targets in cancer therapy, given that it is often deregulated in human cancers.3 The two best characterized targets of mTORC1 are p70S6K1 and 4EBP1.4 The 4EBP1-eIF4E pathway regulated translation is considered to be critically important in cancer cell survival and proliferation since it has a broad impact on haematologica | 2017; 102(4)

Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/755 Š2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

755


C. Bi et al. protein expression, including many oncogenes.5-8 Several mTORC1 inhibitors have been developed and investigated in the treatment of B-cell lymphomas. Rapamycin and its analogs (rapalogs) have been approved by the FDA for the treatment of relapsed MCL. However, many MCL patients respond poorly to rapalog treatment,9,10 and the efficacy of rapalogs in other B-cell lymphomas remains controversial.11,12 Rapalogs only block the effect of mTORC1 on low-affinity targets, such as S6KT389, but not on high-affinity targets, such as 4EBP1Thr37/46 or Grb10S150,13,14 and have no effect on the mechanistic target of rapamycin complex 2 (mTORC2).15,16 Despite the clinical application and multiple active clinical trials of rapalog, this type of mechanistic target of rapamycin (mTOR) inhibitor has been demonstrated to lack cytotoxic effects in lymphoma cells.17,18 Unlike rapalogs, TORKi function as ATP-competitive inhibitors and block all known mTORC1 targets, as well as mTORC2 activity, by directly targeting the mTOR kinase catalytic site.13,19,20 Several TORKi have been developed recently and appear to be promising in lymphoma treatment. Many of these studies addressed the effectiveness on the dual inhibition of mTORC1 and mTORC2,19,21-23 since mTORC2 directly phosphorylates and activates AKT, compromising the inhibitory effect on mTORC1.24 However, two additional feedback loops, through S6K and Grb10, respectively,25,26 may eventually lead to the activation of AKT when mTORC1 is inhibited, which makes the crosstalk between mTORC1 and AKT more elusive, especially upon the dual inhibition of mTORC1 and mTORC2 by TORKi. Because TORKi inhibits both mTORC2 and mTORC1, it is difficult to differentiate which effect plays a major role in the treatment of lymphoma. Although a previous study demonstrated that the block in proliferation of mouse embryonic fibroblasts by TORKi is predominantly through effective inhibition of mTORC1 and independent of mTORC2,13 it is still important to investigate the significance of mTORC2 inhibition in lymphoma cells, given that AKT signaling is one of the most commonly deregulated oncogenic pathways in B-cell malignancies. In addition, although TORKi is potent in inhibiting target of rapamycin (TOR) kinase, many clinical trials were halted prematurely, with the main problem being that cancer cells responded variably to this class of agent, which is probably because TORKi unequally inhibits the 5’cap-dependent translation of different transcripts.27 Therefore, fully assessing the effect of TORKi in aggressive B-cell lymphoma and understanding the molecular mechanism will help to optimize the utilization of these drugs in clinical practice. The study herein was designed to: (1) assess the effects of TORKi treatment, especially the cytotoxic effect in aggressive B-cell lymphomas, (2) determine the significance of 4EBP1 pathway inhibition in TORKi-induced cytotoxicity, and (3) identify the molecular basis for resistance to TORKi treatment in aggressive Bcell lymphomas.

Methods Ectopic gene expression with retrovirus transduction and gene knockout/knockdown with CRISPR-CAS9 system For retroviral transduction, cell lines were first engineered to express the ecotropic retroviral receptor as described previously.28 Retroviral constructs were packaged by co-transfecting HEK293T 756

cells with pCL-Eco and PMIP vector carrying GFP, eIF4E, MCL1, BCL-XL, BCL-2, or AKT. CRISPR-Cas9-mediated knockout was delivered by lentiCRISPR v2 vector (#52961, Addgene, Cambridge, MA, USA).29 For the double knockout experiment, 4EBP1 and 4EBP2 single guide RNAs (sgRNAs) were expressed by lentiCRISPR v2 and a lenti sgRNA expressing vector with a blasticidin resistance gene, respectively. Sequences of sgRNAs are listed in the Online Supplementary Table S1.

BH3 profiling BH3 profiling was performed as previously described.30 Briefly, lymphoma cells were suspended in dithioerythritol (DTE) buffer at a density of 2.5-3×106/ml and then mixed with an equal volume of dye solution containing 4 μM JC-1, 40 μg/ml oligomycin, 20 mM 2-mercaptoethanol, and 0.01% digitonin (w/v). BH3 peptides were dissolved in DTE buffer at a concentration of 160 μM. Equal volumes (15 μL) of cell/dye mix and peptide solution were added to a flat bottom 384 cell plate and incubated at 30°C in a plate reader. Fluorescent signals were detected at excitation of 545 nm and emission of 590 nm every 30 min. The maximum emission value was used for calculating mitochondrial depolarization. Dimethyl sulfoxide (DMSO) and carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) were used as negative and positive controls, respectively, with % depolarization = 1- (sample-FCCP) / (DMSO-FCCP).

Xenograft experiments Six- to eight-week-old female CB-17/SCID mice (The Jackson Laboratory, Bar Harbor, ME, USA) were subcutaneously inoculated on the flank with Ramos or Mino cells (5×106 cells/animal) suspended in 100 μL PBS. Tumor volume was calculated based on the formula V= L x S2 x 0.5 (L: the long axis; S: the short axis). When the xenograft reached a volume of ~200 mm3, mice were randomly assigned to the control, individual or combined treatment groups, with six mice for each group. Torin1 was first dissolved in 100% N-methyl-2-pyrrolidone and subsequently diluted with PEG400 and water at the ratio of 1:2:2, and intraperitoneally injected daily at a dose of 20 mg/kg. ABT-199 was prepared in 60% phosal 50 PG, 30% PEG400 and 10% ethanol, and given daily by gavage at the dose of 20 mg/kg. For the combined treatment, Torin1 was given 4 h before ABT-199. Animals were sacrificed when the control tumor reached ~2000 mm3 or after the loss of more than 10% of body weight. All animal studies were conducted in accordance with the NIH guidelines for animal care. All experimental procedures and protocols were approved by the Institutional Animal Care and Use Committee at University of Nebraska Medical Center (UNMC). For cell lines, inhibitors, and antibodies used, and for additional methods including cell proliferation and apoptosis assays, immunoblotting, luciferase assay and RT-PCR, please see the Online Supplementary Materials and Methods for details.

Results TORKi induces cytotoxicity in B-cell lymphoma cells To examine the effect of TORKi on the proliferation and survival of lymphoma cells, we selected two commonly used TORKi, Torin1 and AZD8055,13,19 to treat 17 aggressive B-cell lymphoma cell lines. Although these cell lines showed different sensitivity to the treatment, both drugs significantly inhibited cell proliferation in all tested cells, mostly in a dose-dependent manner (Figure 1A). There is no distinct correlation between the different types of lymphoma and the extent of inhibition. However, both drugs induced significant apoptosis in only a few lymphoma cell haematologica | 2017; 102(4)


mTOR kinase inhibitor treatment for B-cell lymphomas

A

B

Figure 1. TORKi inhibits cell proliferation and, to a varying extent, induces apoptosis in aggressive B-lymphoma cells. (A) Aggressive B-lymphoma cell lines were treated with TORKi Torin1 (upper panel) and AZD8055 (lower panel) at different doses. Viable cells were determined by MTS assay after 72 h of treatment with each drug. Data shown are the average of three experiments and are presented as mean ± SEM. (B) Apoptosis was quantified by flow cytometry with Annexin V and PI double staining after 48 h of treatment with 250 nM of Torin1 or 5 μM of AZD8055. Relative apoptosis was measured by calculating the fold change to apoptotic fractions of control cells. Data shown are the average of three experiments and are presented as mean ± SEM. MCL: mantle cell lymphoma; DLBCL: diffuse large B-cell lymphoma; DHL: double hit lymphoma; BL: burkitt lymphoma; GCB: germinal center B cell; ABC: activated B cell; Ctr: control.

lines. BL cell line Ramos exhibited the most significant apoptosis upon TORKi treatment, followed by DLBCL lines Tmd8, Su-dhl-6 and DHL line Dohh2; while among MCL lines, increased cell death was only observed in Mino cells (Figure 1B). Prolonged treatment with TORKi (96 h) did not induce significant apoptosis in resistant cell lines either (Online Supplementary Figure S1A). Since previous studies have shown that rapalogs induce cell stasis but barely affect cell survival,17,18 we proceeded to study the significance and mechanism of TORKi-induced apoptosis in lymphoma cells.

TORKi-induced apoptosis is independent of mTORC2 inhibition We first examined the downstream targets of mTORC1 and mTORC2 as well as the AKT activity in Mino cells upon TORKi treatment. As expected, phosphorylation of RPS6 and 4EBP1 were markedly inhibited upon AZD8055 treatment. Phosphorylation of AKTSer473, one of the mTORC2 targets, was also inhibited, confirming the effective inhibition of mTORC1 and mTORC2 by TORKi. Phosphorylation of AKTThr308, one of the targets of PI3Khaematologica | 2017; 102(4)

PDK, was increased upon AZD8055 treatment, suggesting inactivation of the S6K-PI3K negative feedback loop. Surprisingly, phosphorylation of GSK3β, one of the AKT downstream targets, was also increased, suggesting that the net AKT activity was, in fact, elevated (Figure 2A). To avoid unexpected effects secondary to an inadequate amount of AKT protein, we over-expressed AKT in Mino cells, which showed similar effects upon TORKi treatment (Figure 2A). As one of the most prominent features of TORKi, as compared to rapalogs, is their ability to inhibit both mTORC1 and mTORC2, we further examined whether mTORC2 inhibition plays an important role in TORKiinduced apoptosis. Given that the mTOR pathway crucially regulates protein translation, it would be difficult to evaluate the effects of knocking down one of the mTOR pathway components using post-transcriptional mechanisms, such as shRNA interference, partially due to subsequent alterations in protein synthesis. Therefore, we utilized the CRISPR-Cas9 system to knockout RICTOR from genome. Two sensitive cell lines, Ramos and Mino, were chosen for the study. Notably, the knocking out of RIC757


C. Bi et al.

TOR had little effect on cell survival in cells without treatment, while TORKi minimally increased apoptosis in Ramos cells (<10%) with RICTOR knockout but had virtually no additional effect in Mino cells (Figure 2B-D).

TORKi-induced apoptosis is independent of S6K inhibition To determine whether S6K inhibition plays a role in TORKi-induced apoptosis, we selected four cell lines, and treated them with either rapalog or TORKi. As expected, temsirolimus, a rapalog, blocked only the S6K pathway, as shown by decreased phosphorylation of S6K target RPS6S235/236, whereas TORKi blocked both S6K and 4EBP1 pathways in all tested cells (Online Supplementary Figure S1B). In addition, PF4708671, a selective inhibitor of S6K, had no significant impact on apoptosis in both Ramos and Mino cells (Online Supplementary Figures S1C and S1D), suggesting TORKi-induced apoptosis is independent of S6K inhibition.

Inhibition of 4EBP-eIF4E pathway plays an important role in TORKi-induced apoptosis When 4EBP1 binds to eIF4E, it prevents the formation of eIF4F by blocking the recruitment of eIF4G proteins. In fact, eIF4E binding proteins comprise not only 4EBP1 but also 4EBP2 and 4EBP3, all of which are regulated by mTORC1.8 The ratio of eIF4E/4EBP, rather than their individual protein levels, dictates the 5â&#x20AC;&#x2122;cap-dependent mRNA translation efficiency as well as the sensitivity to mTOR inhibition.8,31,32 To test whether the 4EBP-eIF4E pathway plays an important role in TORKi-induced apoptosis, we first depleted 4EBP1 in lymphoma cells. Ramos and Mino cells were transduced with CRISPR-Cas9 vectors which carry sgRNAs specifically targeting 4EBP1. After the establishment of stably expressing cells, sgRNA1 exhibit-

A

ed the highest efficiency (Figure 3A) and thus was used in subsequent experiments. For Mino cells, knockout of 4EBP1 almost blocked TORKi-induced apoptosis; the effect is limited in Ramos cells which showed higher sensitivity to TORKi treatment (Figure 3B,C). Since other 4EBPs may act similarly to 4EBP1, and the level of 4EBP3 is very low in leukocytes,33 we subsequently knocked out 4EBP2 using the CRISPR-Cas9 system. Of the examined sgRNAs, sgRNA2 showed the highest efficiency. Upon treatment, similar results were obtained in both Ramos and Mino 4EBP2 knockout cells, implying that a single 4EBP loss may be insufficient to completely rescue cells from apoptosis because of compensation from other 4EBPs (Figure 3D-F). Therefore, we knocked out both 4EBP1 and 4EBP2 by separate CRISPR-Cas9 constructs in Ramos cells (Figure 3G). Strikingly, the double knockout significantly abolished TORKi-induced apoptosis. Moreover, we found that MCL1 and BCL-XL were substantially upregulated in the double knockout Ramos cells (Figure 3H,I). Next, we sought to confirm the essential role of the 4EBP-eIF4E pathway in TORKi-induced apoptosis by overexpressing eIF4E in lymphoma cells (Figure 4A). Overexpression of eIF4E in both Ramos and Mino cells significantly attenuated TORKi-induced apoptosis (Figure 4B). We also investigated the level of MCL1, which is remarkably downregulated by TORKi treatment (Online Supplementary Figure S1A), and found that eIF4E-overexpressing cells exhibited a higher level of MCL1 than the control cells upon TORKi treatment (Figure 4A), suggesting that increased eIF4E attenuated the TORKi-induced apoptosis, at least partially, through upregulation of MCL1. We also performed similar experiments in Dohh2 cells, one of the DHL cell lines, and revealed similar findings (Figure 4A,B).

B

C

D

Figure 2. Knocking out Rictor has little effect on TORKi-induced apoptosis in aggressive B-lymphoma cells. (A) Mino cells were transduced with retroviral vectors expressing GFP and AKT. After puromycin selection, cells were treated with AZD8055 for 24 h. AKT and mTORC1 signaling were evaluated. (B) Ramos and Mino cells were transduced with a CRISPR-CAS9 vector that expresses a sgRNA targeting Rictor. After selection for 3 weeks, Rictor protein levels were examined by immunoblotting. (C and D) Cells transduced with Rictor-sgRNA1 were treated with AZD8055 (AZD) or Torin1 (Tor) for 48h. Apoptosis was quantified by flow cytometry with Annexin V and PI double staining. Shown is an average (Âą SEM) of two independent experiments. Ctr: control.

758

haematologica | 2017; 102(4)


mTOR kinase inhibitor treatment for B-cell lymphomas

TORKi treatment downregulates MCL1 and BCL-XL To further explore the molecular mechanisms of TORKi-induced apoptosis, we examined the effects of TORKi on three important anti-apoptotic proteins in Bcell lymphomas, BCL-2, MCL1 and BCL-XL. As expected, temsirolimus had little effect on the expression of these proteins, whereas TORKi substantially downregulated MCL1 expression, and moderately decreased the BCL-XL level. However, the BCL-2 level was barely affected by TORKi treatment. Interestingly, Ramos, the cell line most sensitive to TORKi treatment, expressed a very low level of BCL-2 (Online Supplementary Figure S1A). These findings imply that BCL-2 may play an important role in conferring resistance to mTORC1 inhibition. To confirm that downregulation of MCL1 and BCL-XL is responsible for TORKi-induced apoptosis, we retrovirally overexpressed MCL1 in Ramos, Mino and Dohh2 cells (Figure 5A-D). Overexpression of MCL1 protein partially rescued cells from TORKi-induced apoptosis in Ramos and was more effective in Mino and Dohh2 cells (Figure 5E-G). Similarly, overexpression of BCL-XL in Ramos cells only partially rescued cells, whereas co-expression of both MCL1 and BCL-XL resulted in a total resistance to TORKiinduced apoptosis. Furthermore, to confirm that high expression of BCL-2 is a major cause for TORKi resistance, we overexpressed BCL-2 in Ramos cells, which completely blocked apoptosis from TORKi treatment (Figure 5A,E).

A

B

D

E

G

Prediction of cell sensitivity to TORKi treatment by BH3 profiling The above findings suggested that reduced expression of MCL1 and BCL-XL secondary to mTORC1 inhibition plays an important role in TORKi-induced apoptosis, whereas elevated BCL-2 expression may correlate with the resistance. An important question is whether the expression of these proteins, especially BCL-2, can be used as a biomarker to predict the susceptibility of lymphoma patients to TORKi treatment. Because apoptosis is controlled by a complicated network consisting of apoptosis activators, sensitizers and anti-apoptotic proteins, quantifying one or a few of the anti-apoptotic proteins alone may be incapable of revealing the most critical proteins the cells rely on for survival. Therefore, we sought to use BH3 profiling, which exposes tumor cells to a series of peptides derived from BH3 proteins and directly determines the important anti-apoptotic proteins for cell survival.34 Because MCL1 is the main anti-apoptotic protein affected by TORKi treatment, we intended to focus on the effect of its specific binding peptide NOXA. BH3 profiling was performed on Ramos, Mino and Z138 cells, representing cells with high, moderate, and low sensitivity to TORKi treatment, respectively. Ramos cells overexpressing BCL-2 (Ramos-BCL-2), which is highly resistant to TORKI treatment (Figure 5E), were also included as control. Ramos cells clearly demonstrated high mitochondrial

C

F

H I

Figure 3. Knocking out of 4EBPs induces resistance to TORKi treatment. (A) Ramos and Mino cells were transduced with CRISPR-CAS9 vectors targeting 4EBP1 and immunoblotted with the indicated antibodies. (B and C) Ramos and Mino cells transduced with 4EBP1-sgRNA1 were treated with AZD8055 (AZD) or Torin1 (Tor) for 48 h, and apoptosis was evaluated using flow cytometry with Annexin V and PI double staining. (D) Cells were transduced with CRISPR-CAS9 vectors targeting 4EBP2 and immunoblotted with the indicated antibodies. (E and F) Ramos and Mino cells transduced with 4EBP2-sgRNA2 were treated with AZD or Tor for 48 h, and apoptosis was evaluated using flow cytometry with Annexin V and PI double staining. (G) Ramos was transduced with CRISPR-CAS9 vectors targeting both 4EPB1 and 4EBP2 and immunoblotted with the indicated antibodies. 48 h after treatment with AZD or Tor, 4EBP1/2 double knockout (Ramos-DKO) and control (RamosC) Ramos cells were (H) immunoblotted with antibodies against MCL1 and BCL-XL, and (I) analyzed by flow cytometry with Annexin V and PI double staining to evaluate apoptosis. All data (mean Âą SEM) shown are the average of two experiments. *P<0.05; **P<0.01; ***Pâ&#x2030;¤0.001; ****P<0.0001. Ctr: control.

haematologica | 2017; 102(4)

759


C. Bi et al.

permeabilization triggered by either the NOXA- or BCLXL-specific binding peptide HRK, whereas Ramos-BCL-2 showed no response to either of the peptides, highlighting the importance of MCL1 and BCL-XL for evading apoptosis in Ramos cells (Figure 6A). Consistent with the sensitivity to TORKi treatment, NOXA induced mild permeabilization in Mino but had no effect on Z138 cells (Figure 6A). We also performed BH3 profiling on a series of TORKi resistant lymphoma cell lines and found that the blocking of MCL1 by the NOXA peptide was unable to induce permeabilization in any of these cells (Online Supplementary Figure S2). It should also be noted that blocking BCL-XL by HRK activated the apoptotic pathway in both Mino and Z138 cells, yet they both showed a certain resistance to TORKi treatment, suggesting that decreased MCL1 is the major mechanism for TORKiinduced apoptosis, and a slight decrease of BCL-XL under a high BCL-2 background has little impact on apoptosis.

Optimizing TORKi treatment in aggressive B-cell lymphoma BH3 profiling not only predicted sensitivity to TORKi treatment, but also provided insights for promoting apoptosis in cancer cells. Based on this analysis, we combined TORKi with the BCL-XL selective inhibitor WEHI-539 to treat wild-type Ramos cells, which exhibited a remarkable synergy in inducing apoptosis (Figure 6B). Notably, inhibition of BCL-2 by ABT-199 or ABT-737 together with TORKi treatment resulted in substantial apoptosis in both Mino and Z138 cells (Figure 6B; Online Supplementary

Figure S3). This is consistent with the above data demonstrating that BCL-2 overexpression is the major determinant for TORKi resistance. To further explore the necessity of inhibiting BCL-2 in TORKi treatment, we analyzed the mRNA expression of BCL-2 family genes in aggressive B-cell lymphoma cohorts from the Lymphoma/Leukemia Molecular Profiling Project (LLMPP) database. Gene expression profiling data revealed that BCL-2 expression was increased by at least 2-fold in nearly all MCL cases, compared to normal naïve B cells. BCl-2 expression was also increased by at least 2-fold in more than half of DLBCL cases, especially in the activated B-cell (ABC) subtype, and also in a small proportion of BL cases, compared to normal germinal center B cells (Figure 6C). The finding implies that BCL-2 deregulation is a common defect in aggressive B-cell lymphomas, which may induce resistance to TORKi treatment. It is worth noting that only a few cases showed a slight increase in MCL1 levels, but the expression of NOXA (PMAIP1) was greatly decreased in almost all of the lymphoma cases, leaving MCL1 free to bind other apoptosis activators or sensitizers. These data suggest that for a large number of aggressive B-cell lymphomas, BCL-2 and MCL1 may collaborate to sustain cell survival. The inhibition of both mTORC1/MCL1 and BCL-2 is therefore likely to provide an effective treatment for these lymphomas. Intriguingly, this strategy was also shown to be very effective in DHL, as the combination of Torin1 and ABT-199 synergistically induced significant apoptosis in Ros-50 and Dohh2 cells (Online Supplementary Figure S3).

A

B

Figure 4. Overexpression of eIF4E rescues lymphoma cells, at least partially, from TORKi-induced apoptosis. (A) Ramos, Mino and Dohh2 cells were transduced with a retroviral vector expressing eIF4E or GFP and immunoblotted by the indicated antibodies. Relative MCL1 expressions were calculated by normalizing individual levels with that of control cells (vector only). (B) Transduced cells were treated with various doses of AZD8055 (AZD) or Torin1 (Tor), and apoptosis was evaluated by Annexin V and PI double staining 48 h after the treatment. Data shown are the average of two experiments and are presented as mean ± SEM. **P≤0.01; ***P<0.001. Ctr: control.

760

haematologica | 2017; 102(4)


mTOR kinase inhibitor treatment for B-cell lymphomas

Lastly, we intended to test the efficacy of TORKi treatment in aggressive B-cell lymphomas in vivo. First, xenografts of the sensitive cell line Ramos were established and treated with Torin1 daily for 11 days. We found that Torin1 alone is sufficient to suppress tumor growth. The tumor weight at sacrifice was decreased by 60% compared to the placebo group (Figure 7A,B), and an increased number of apoptotic cells were observed in the drug treatment group (Figure 7C,D). Subsequently, we established Mino xenografts and then treated engrafted mice with daily doses of Torin1 and ABT-199, individually or in combination for 14 days. Although each drug alone inhibited tumor growth, the effect was quite limited. In contrast, the combined treatment completely blocked tumor growth, as determined by both tumor volume and weight (Figure 7E,F). Immunostaining for cleaved caspase3 showed markedly increased apoptosis upon combined treatment, (Ratio of apoptosis: Control: 0.15 ± 0.21%; Torin1 treatment: 0.67 ± 0.26%; ABT-199 treatment: 17.39 ± 2.82%; Torin1 + ABT-199: 69.04 ± 8.59%), indicating that the combination of TORKi and the BCL-2 inhibitor may be of great value in treating aggressive B-cell lymphoma (Figure 7G-J).

Discussion mTOR complexes are essential for cellular homeostasis as well as controlling cell growth and proliferation. To date, the function and regulation of mTORC1 have been

well studied, whereas little is known about mTORC2. One important function of mTORC2 is phosphorylating AGC kinase family members, especially AKT, implicating the regulatory roles of mTORC2 in multiple physiological processes.35,36 Notably, a recent study using transgenic mice has demonstrated that mTORC2 is required for the homeostasis and function of normal B cells through the regulation of the AKT and NF-κB pathways.37 However, we found that knockout of Rictor had little impact on lymphoma cell survival, suggesting that mTORC2 becomes dispensable in B-cell lymphomas, most likely due to the constitutive activation of AKT and other oncogenic pathways.38,39 Furthermore, we showed that AKTSer473 is a direct target of mTORC2, and its phosphorylation was diminished upon TORKi treatment alone. Although this phosphorylation is thought to precede the phosphorylation of AKTThr308 and is required for full AKT activation,35 our results demonstrated that AKT activity is primarily determined by the phosphorylation of AKTThr308. This is consistent with previous studies, which demonstrated that the phosphorylation of AKTThr308, rather than AKTSer473, is correlated with AKT activity and predicts survival in human non-small cell lung cancer and acute myeloid leukemia.40,41 It is noteworthy that the highest phosphorylation of AKTThr308 and GSK3β are both present at the low concentration of TORKi treatment. In this condition, phosphorylation of AKTS473, although decreased, was not totally abrogated, supporting the notion that the phosphorylation of AKTS473 does enhance the phosphorylation of AKTThr308. Besides, despite the irrelevance of mTORC2 inhibition in

A C

D

E

B

F

G

Figure 5. Overexpression of MCL1 and BCL-XL protects lymphoma cells from TORKi-induced apoptosis. (A-D) Exogenous GFP, BCL-2, MCL1 or BCL-XL was overexpressed in lymphoma cells by retrovirus transduction, and the cells were then treated with various doses of AZD8055 for 24 hr. Corresponding protein levels as well as Caspase-3 and PARP were evaluated by immunoblotting. (E-G) Transduced cells were treated with various doses of AZD8055 for 48 h, and apoptosis was evaluated using flow cytometry with Annexin V and PI double staining. Data shown are the average of two experiments and are presented as mean ± SEM. *P<0.05; **P<0.01; ***P<0.001. AZD: AZD8055; Tor: Torin1; Ctr: control.

haematologica | 2017; 102(4)

761


C. Bi et al.

TORKi-induced apoptosis, targeting mTORC2 may be still of great value in lymphoma treatment, provided that mTORC2 signaling plays an important part in cell migration and tumor metastasis.42,43 The two major substrates of mTORC1 have different roles in cancer development. We found that TORKi-induced apoptosis in lymphoma cells is primarily through the inhibition of 4EBP phosphorylation. However, it does not imply that the S6K pathway is dispensable. In fact, in 4EBP1/2 double knockout cells treated with TORKi, apoptosis is markedly abrogated, yet cell proliferation is still remarkably suppressed (data not shown). These data corroborate a previous study which demonstrated that rapalogs, which inhibited S6K alone, have a merely cytostatic effect in B-cell lymphoma cells,44 and partially explained their limited effect in lymphoma clinical trials, even when combined with rituximab.9-12 Among the mTORC1-regulated 4EBP proteins, 4EBP1 is the most recognized eIF4E binding protein in mammalian cells. Nevertheless, 4EBP1 knockout only partially rescued the most sensitive lymphoma cells from mTOR inhibition, which implied the importance of 4EBP2 in lymphoma cells. We analyzed the expression of 4EBPs in primary aggressive B-cell lymphomas and found that the abun-

dance of 4EBP2 is higher than, or at least similar to, that of 4EBP1, whereas 4EBP3 is hardly detectable (data not shown). As 4EBP1 and 4EBP2 share a conserved eIF4E binding motif and are both regulated by mTORC1, one could compensate for loss of the other.45,46 Interestingly, a recent study has shown that in Val, a 4EBP1-deficient cell, TORKi treatment increased the amount of 4EBP2 bound to eIF4E, but it seemed to be ineffective in blocking the formation of the initiation complex.47 One likely explanation is that eIF4E expression largely exceeds that of eIF4G and 4EBPs in Val cells, making 4EBP2 unable to sequester eIF4E from other eIF4F components. This further indicates that the ratio of eIF4E to 4EBPs, rather than the abundance of individual proteins, determines the efficacy of mTOR inhibition.8,31 In the study herein, we found that reduced MCL1 expression is a crucial determinant of TORKi sensitivity, which makes using BH3 profiling to predict the sensitivity to mTOR inhibition possible. One of the advantages of BH3 profiling is that it can directly ascertain whether lymphoma cells will undergo apoptosis upon MCL1 blockade by NOXA. Notably, a large proportion of aggressive B-cell lymphomas express relatively high levels of BCL-2, which is the major determinant for TORKi resistance.

A

B

C

Figure 6. BH3 profiling predicts sensitivity to TORKi treatment. (A) Lymphoma cells were treated with various peptides as indicated in Ramos, Ramos with BCL-2 overexpression, Mino and Z138. Mitochondrial depolarization was determined by JC-1 staining. Results are shown as mean (n =3), compared to solvent control DMSO values; error bar represents standard deviation. (B) Cells were treated with the combination of TORKi with BCL-XL selective inhibitor WEHI-539 or BCL-2/BCL-XL inhibitor ABT-737 or BCL-2 inhibitor ABT-199 for 48 h, and then apoptosis was quantified by flow cytometry. Coefficients of drug interaction (CDI) were calculated based on the inhibitory effect of the individual drugs and combined treatment. Data shown are the average of two experiments and are presented as mean Âą SEM. (C) The heatmap illustrates the relative mRNA level (log2 fold change) of BCL-2 family genes with DLBCL and BL compared to that of normal centrocytes and MCL compared to that of normal naĂŻve B cells. GCB-DLBCL: diffuse large B-cell lymphoma germinal center B-cell subtype; ABC DLBCL: diffuse large B-cell lymphoma activated B-cell subtype: BL: burkitt lymphoma; MCL: mantle cell lymphoma; AZD: AZD8055; Tor: Torin1.

762

haematologica | 2017; 102(4)


mTOR kinase inhibitor treatment for B-cell lymphomas

A

E

C

D

G

H

I

J

B

F

Figure 7. Animals with xenograft lymphomas were treated with Torin1 and/or BCL-2 inhibitor, ABT-199. Ramos and Mino xenografts were established in NOD/SCID mice, and the mice were treated with the indicated drugs. (A and E) Tumor growth was evaluated by the ratio of tumor volume at the indicated time to that at the initiation of drug treatment, results are shown as mean ± SEM (n =6 in each group). (B and F) Tumor weights were obtained at autopsy and compared across different groups of treatment at 11 and 14 days after the initiation of treatment for Ramos and Mino models, respectively. (C and D) H&E staining of the Ramos xenograft tumors. Markedly increased apoptosis was observed in mice treated with Torin1. (G-J) Immunohistochemical staining using antibody against cleaved caspase-3 in Mino xenograft tumors treated with Torin1 (Tor) and ABT-199 (ABT), individually or in combination. ****P<0.0001. Ctr: control.

Combining TORKi with a BCL-2 antagonist, such as ABT199, will thus likely provide a synergistic anti-tumor effect in aggressive B-cell lymphomas. A previous study has demonstrated that dual PI3K/mTOR inhibitors can overcome resistance to ABT-199 by decreasing MCL1 and BCL-XL in lymphoid malignancies. Our study, from the other aspect, suggested that blockade of BCL-2 is of great value in counteracting resistance to TORKi treatment. Interestingly, the effectiveness of combining TORKi with ABT-199 in DHL cells makes it a promising strategy for this intractable entity. Nevertheless, there are some B-cell lymphomas which do not rely on BCL-2 for survival, including those “unprimed” cases.34 Therefore, predicting sensitivity to BCL-2 inhibition is also valuable. Unfortunately, due to the lack of a unique BCL-2 binding peptide, it is hard to predict the precise effect of BCL-2 in cancer cells using a peptide-based BH3 profiling method.

Reference 1. Tarella C, Passera R, Magni M, et al. Risk factors for the development of secondary malignancy after high-dose chemotherapy and autograft, with or without rituximab: a 20-year retrospective follow-up study in patients with lymphoma. J Clin Oncol. 2011;29(7):814-824. 2. Ng AK, LaCasce A, Travis LB. Long-term complications of lymphoma and its treatment. J Clin Oncol. 2011;29(14):1885-1892. 3. Laplante M, Sabatini DM. mTOR signaling in growth control and disease. Cell. 2012; 149(2):274-293. 4. Hay N, Sonenberg N. Upstream and down-

haematologica | 2017; 102(4)

However, a similar method which measures mitochondrial depolarization following a direct challenge by the BCL2 inhibitor has been applied in recent studies.48,49 In conclusion, we demonstrated that TORKi-induced apoptosis is predominantly dependent on the loss of mTORC1-mediated 4EBP phosphorylation, and subsequent MCL1 downregulation. Overexpression of BCL-2 confers resistance to TORKi treatment, and BH3 profiling is useful in predicting TORKi sensitivity in lymphoma cells. Combined treatment with TORKi and a BCL-2 antagonist exerts significant synergistic anti-tumor effects both in vitro and in vivo. Funding This study was supported by the Fred & Pamela Buffett Cancer Center Grant (P30CA036727) and the National Natural Science Foundation of China (81372539).

stream of mTOR. Genes Dev. 2004;18(16): 1926-1945. 5. Larsson O, Morita M, Topisirovic I, et al. Distinct perturbation of the translatome by the antidiabetic drug metformin. Proc Natl Acad Sci USA. 2012;109(23):8977-8982. 6. Green AS, Chapuis N, Maciel TT, et al. The LKB1/AMPK signaling pathway has tumor suppressor activity in acute myeloid leukemia through the repression of mTORdependent oncogenic mRNA translation. Blood. 2010;116(20):4262-4273. 7. Zang C, Eucker J, Liu H, Muller A, Possinger K, Scholz CW. Concurrent inhibition of PI3-kinase and mTOR induces cell death in diffuse large B cell lymphomas, a

mechanism involving down regulation of Mcl-1. Cancer Lett. 2013;339(2):288-297. 8. Hsieh AC, Costa M, Zollo O, et al. Genetic dissection of the oncogenic mTOR pathway reveals druggable addiction to translational control via 4EBP-eIF4E. Cancer Cell. 2010;17(3):249-261. 9. Witzig TE, Geyer SM, Ghobrial I, et al. Phase II trial of single-agent temsirolimus (CCI-779) for relapsed mantle cell lymphoma. J Clin Oncol. 2005;23(23):53475356. 10. Ansell SM, Tang H, Kurtin PJ, et al. Temsirolimus and rituximab in patients with relapsed or refractory mantle cell lymphoma: a phase 2 study. Lancet Oncol.

763


C. Bi et al. 2011;12(4):361-368. 11. Smith SM, van Besien K, Karrison T, et al. Temsirolimus has activity in non-mantle cell non-Hodgkin's lymphoma subtypes: The University of Chicago phase II consortium. J Clin Oncol. 2010;28(31):4740-4746. 12. Witzig TE, Reeder CB, LaPlant BR, et al. A phase II trial of the oral mTOR inhibitor everolimus in relapsed aggressive lymphoma. Leukemia. 2011;25(2):341-347. 13. Thoreen CC, Kang SA, Chang JW, et al. An ATP-competitive mammalian target of rapamycin inhibitor reveals rapamycinresistant functions of mTORC1. J Biol Chem. 2009;284(12):8023-8032. 14. Kang SA, Pacold ME, Cervantes CL, et al. mTORC1 phosphorylation sites encode their sensitivity to starvation and rapamycin. Science. 2013;341(6144): 1236566. 15. Kim DH, Sarbassov DD, Ali SM, et al. mTOR interacts with raptor to form a nutrient-sensitive complex that signals to the cell growth machinery. Cell. 2002; 110(2):163-175. 16. Yip CK, Murata K, Walz T, Sabatini DM, Kang SA. Structure of the human mTOR complex I and its implications for rapamycin inhibition. Mol Cell. 2010; 38(5):768-774. 17. Dal Col J, Zancai P, Terrin L, et al. Distinct functional significance of Akt and mTOR constitutive activation in mantle cell lymphoma. Blood. 2008;111(10):5142-5151. 18. Wanner K, Hipp S, Oelsner M, et al. Mammalian target of rapamycin inhibition induces cell cycle arrest in diffuse large B cell lymphoma (DLBCL) cells and sensitises DLBCL cells to rituximab. Br J Haematol. 2006;134(5):475-484. 19. Chresta CM, Davies BR, Hickson I, et al. AZD8055 is a potent, selective, and orally bioavailable ATP-competitive mammalian target of rapamycin kinase inhibitor with in vitro and in vivo antitumor activity. Cancer Res. 2010;70(1):288-298. 20. Liu Q, Xu C, Kirubakaran S, et al. Characterization of Torin2, an ATP-competitive inhibitor of mTOR, ATM, and ATR. Cancer Res. 2013;73(8):2574-2586. 21. Gupta M, Hendrickson AE, Yun SS, et al. Dual mTORC1/mTORC2 inhibition diminishes Akt activation and induces Puma-dependent apoptosis in lymphoid malignancies. Blood. 2012;119(2):476-487. 22. Naing A, Aghajanian C, Raymond E, et al. Safety, tolerability, pharmacokinetics and pharmacodynamics of AZD8055 in advanced solid tumours and lymphoma. Br J Cancer. 2012;107(7):1093-1099. 23. Zeng Z, Shi YX, Tsao T, et al. Targeting of mTORC1/2 by the mTOR kinase inhibitor PP242 induces apoptosis in AML cells

764

24. 25.

26.

27.

28.

29.

30. 31.

32.

33.

34.

35.

36.

37.

under conditions mimicking the bone marrow microenvironment. Blood. 2012; 120(13):2679-2689. Guertin DA, Sabatini DM. Defining the role of mTOR in cancer. Cancer Cell. 2007; 12(1):9-22. Hsu PP, Kang SA, Rameseder J, et al. The mTOR-regulated phosphoproteome reveals a mechanism of mTORC1-mediated inhibition of growth factor signaling. Science. 2011;332(6035):1317-1322. Harrington LS, Findlay GM, Gray A, et al. The TSC1-2 tumor suppressor controls insulin-PI3K signaling via regulation of IRS proteins. J Cell Biol. 2004;166(2):213-223. Thoreen CC, Chantranupong L, Keys HR, Wang T, Gray NS, Sabatini DM. A unifying model for mTORC1-mediated regulation of mRNA translation. Nature. 2012; 485(7396):109-113. Ngo VN, Davis RE, Lamy L, et al. A loss-offunction RNA interference screen for molecular targets in cancer. Nature. 2006;441(7089):106-110. Sanjana NE, Shalem O, Zhang F. Improved vectors and genome-wide libraries for CRISPR screening. Nat Methods. 2014; 11(8):783-784. Ryan J, Letai A. BH3 profiling in whole cells by fluorimeter or FACS. Methods. 2013; 61(2):156-164. Alain T, Morita M, Fonseca BD, et al. eIF4E/4E-BP ratio predicts the efficacy of mTOR targeted therapies. Cancer Res. 2012;72(24):6468-6476. Pause A, Belsham GJ, Gingras AC, et al. Insulin-dependent stimulation of protein synthesis by phosphorylation of a regulator of 5'-cap function. Nature. 1994; 371(6500):762-767. Poulin F, Gingras AC, Olsen H, Chevalier S, Sonenberg N. 4E-BP3, a new member of the eukaryotic initiation factor 4E-binding protein family. J Biol Chem. 1998; 273(22):14002-14007. Deng J, Carlson N, Takeyama K, Dal Cin P, Shipp M, Letai A. BH3 profiling identifies three distinct classes of apoptotic blocks to predict response to ABT-737 and conventional chemotherapeutic agents. Cancer Cell. 2007;12(2):171-185. Sarbassov DD, Guertin DA, Ali SM, Sabatini DM. Phosphorylation and regulation of Akt/PKB by the rictor-mTOR complex. Science. 2005;307(5712):1098-1101. Zoncu R, Efeyan A, Sabatini DM. mTOR: from growth signal integration to cancer, diabetes and ageing. Nat Rev Mol Cell Biol. 2011;12(1):21-35. Lee K, Heffington L, Jellusova J, et al. Requirement for Rictor in homeostasis and function of mature B lymphoid cells. Blood.

2013;122(14):2369-2379. 38. Rudelius M, Pittaluga S, Nishizuka S, et al. Constitutive activation of Akt contributes to the pathogenesis and survival of mantle cell lymphoma. Blood. 2006;108(5):1668-1676. 39. Abubaker J, Bavi PP, Al-Harbi S, et al. PIK3CA mutations are mutually exclusive with PTEN loss in diffuse large B-cell lymphoma. Leukemia. 2007;21(11):2368-2370. 40. Vincent EE, Elder DJ, Thomas EC, et al. Akt phosphorylation on Thr308 but not on Ser473 correlates with Akt protein kinase activity in human non-small cell lung cancer. Br J Cancer. 2011;104(11):1755-1761. 41. Gallay N, Dos Santos C, Cuzin L, et al. The level of AKT phosphorylation on threonine 308 but not on serine 473 is associated with high-risk cytogenetics and predicts poor overall survival in acute myeloid leukaemia. Leukemia. 2009;23(6):10291038. 42. Gulhati P, Bowen KA, Liu J, et al. mTORC1 and mTORC2 regulate EMT, motility, and metastasis of colorectal cancer via RhoA and Rac1 signaling pathways. Cancer Res. 2011;71(9):3246-3256. 43. Charest PG, Shen Z, Lakoduk A, Sasaki AT, Briggs SP, Firtel RA. A Ras signaling complex controls the RasC-TORC2 pathway and directed cell migration. Dev Cell. 2010; 18(5):737-749. 44. Haritunians T, Mori A, O'Kelly J, Luong QT, Giles FJ, Koeffler HP. Antiproliferative activity of RAD001 (everolimus) as a single agent and combined with other agents in mantle cell lymphoma. Leukemia. 2007;21(2):333-339. 45. Lin TA, Lawrence JC, Jr. Control of the translational regulators PHAS-I and PHASII by insulin and cAMP in 3T3-L1 adipocytes. J Biol Chem. 1996;271(47): 30199-30204. 46. Mader S, Lee H, Pause A, Sonenberg N. The translation initiation factor eIF-4E binds to a common motif shared by the translation factor eIF-4 gamma and the translational repressors 4E-binding proteins. Mol Cell Biol. 1995;15(9):4990-4997. 47. Mallya S, Fitch BA, Lee JS, So L, Janes MR, Fruman DA. Resistance to mTOR kinase inhibitors in lymphoma cells lacking 4EBP1. PloS one. 2014;9(2):e88865. 48. Colak S, Zimberlin CD, Fessler E, et al. Decreased mitochondrial priming determines chemoresistance of colon cancer stem cells. Cell Death Differ. 2014; 21(7):1170-1177. 49. Zeuner A, Francescangeli F, Contavalli P, et al. Elimination of quiescent/slow-proliferating cancer stem cells by Bcl-XL inhibition in non-small cell lung cancer. Cell Death Differ. 2014;21(12):1877-1888.

haematologica | 2017; 102(4)


ARTICLE

Non-Hodgkin Lymphoma

Safety and efficacy of obinutuzumab with CHOP or bendamustine in previously untreated follicular lymphoma

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Andrew Grigg,1 Martin J.S. Dyer,2 Marcos González Díaz,3 Martin Dreyling,4 Simon Rule,5 Guiyuan Lei,6 Andrea Knapp,7 Elisabeth Wassner-Fritsch7 and Paula Marlton8

Department of Clinical Haematology, Austin Hospital, Heidelberg, Australia; 2Ernest and Helen Scott Haematological Research Institute, University of Leicester, UK; 3Department of Hematology, University Hospital of Salamanca, Spain; 4Department of Medicine III, LMU, Munich, Germany; 5Department of Haematology, Derriford Hospital, Plymouth, UK; 6 Roche Products Ltd, Welwyn Garden City, UK; 7F. Hoffmann-La Roche Ltd, Basel, Switzerland and 8Department of Haematology, Princess Alexandra Hospital and University of Queensland School of Medicine, Brisbane, Australia 1

Haematologica 2017 Volume 102(4):765-772

ABSTRACT

T

he GAUDI study assessed safety and preliminary efficacy of induction therapy with obinutuzumab plus chemotherapy, followed by maintenance therapy with obinutuzumab alone, in previously untreated patients with follicular lymphoma. Assignment to chemotherapy was decided on a per-center basis before the patients’ enrollment. Patients (n=81) received four to six cycles of obinutuzumab plus bendamustine every 4 weeks or six to eight cycles of obinutuzumab plus CHOP every 3 weeks. Patients with an end-of-treatment response were eligible for obinutuzumab maintenance therapy every 3 months for 2 years or until disease progression. Induction treatment was completed by 90% of patients in the obinutuzumab plus bendamustine group and 95% in the obinutuzumab plus CHOP group, while maintenance was completed by 81% and 72% of patients, respectively. All patients experienced at least one adverse event during induction, most commonly infusion-related reactions (58%), the majority of which were grade 1/2. The most common hematologic adverse event was grade 3/4 neutropenia (36% during induction and 7% during maintenance). One treatmentrelated death occurred during the maintenance phase. At the end of induction, 94% of patients had achieved an overall response, with complete response based on computed tomography in 36%. The progression-free survival rate at 36 months was 90% in the obinutuzumab plus bendamustine group and 84% in the obinutuzumab plus CHOP group. These results demonstrate that induction therapy with obinutuzumab plus bendamustine or obinutuzumab plus CHOP, followed by obinutuzumab maintenance, is associated with tolerable safety and promising efficacy. This study is registered at ClinicalTrials.gov as NCT00825149.

Correspondence: andrew.grigg@austin.org.au

Received: July 11, 2016. Accepted: December 22, 2016. Pre-published: December 23, 2016. doi:10.3324/haematol.2016.152272 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/765 ©2017 Ferrata Storti Foundation

Introduction Chemoimmunotherapy utilizing the type I anti-CD20 monoclonal antibody rituximab is the standard-of-care treatment for advanced follicular lymphoma (FL),1 with the chemotherapy component generally consisting of bendamustine or CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) in the first-line setting.2 However, as some patients do not respond to treatment, and most will relapse after an initial response,3 new treatments with improved anti-tumor efficacy are needed. Obinutuzumab (GA101; G) is a glyco-engineered type II, humanized, anti-CD20 monoclonal antibody that has reduced core fucosylation compared with rituximab. haematologica | 2017; 102(4)

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

765


A. Grigg et al.

In preclinical studies, obinutuzumab showed increased direct cell death and antibody-dependent cellular cytotoxicity, but reduced complement activation, when compared with rituximab in vitro and improved survival in human lymphoma xenograft models in vivo.4-8 In a clinical setting, obinutuzumab monotherapy was well tolerated with promising activity in relapsed or refractory patients with indolent non-Hodgkin lymphoma (NHL), including FL.9-12 Obinutuzumab has been investigated in combination with chemotherapy. In the open-label, phase 1b GAUDI study, induction therapy with obinutuzumab plus either CHOP (G-CHOP) or fludarabine and cyclophosphamide (G-FC), followed by obinutuzumab maintenance, was associated with encouraging efficacy and safety outcomes for patients with relapsed or refractory FL.13,14 The GAUDI study also investigated the safety and efficacy of G-CHOP or obinutuzumab plus bendamustine (G-B) followed by obinutuzumab maintenance in patients with previously untreated FL. Data from the induction phase, and preliminary data from the maintenance phase, for patients receiving first-line treatment showed a safety profile consistent with that reported for the relapsed or refractory subset.15-16 This paper presents the final analysis of the subset of previously untreated patients from the open-label, multicenter, phase 1b GAUDI study, the primary aim of which was to establish the safety and efficacy of G-CHOP and G-B induction therapy in patients with previously untreated CD20+ FL.

Methods

monotherapy (1000 mg iv) starting 12 weeks after the last chemoimmunotherapy dose and administered every 3 months for 2 years or until disease progression. Details of prophylactic medication and chemotherapy dose modifications are provided in the Online Supplementary Appendix.

Assessments The primary endpoint was safety of induction treatment for GB and G-CHOP. Secondary endpoints included overall response rate, complete response rate, progression-free survival, obinutuzumab pharmacokinetics, B-cell depletion and recovery, and safety of obinutuzumab maintenance therapy. Adverse events and serious adverse events were monitored according to National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0. Response rates were assessed by computed tomography scan according to the revised response criteria for non-Hodgkin lymphoma.17 Further details of assessments are provided in the Online Supplementary Appendix.

Statistical analysis Eighty patients were planned for the safety evaluation. All patients who received ≥1 dose of G-chemotherapy were eligible for the safety and efficacy analyses. With approximately 40 patients per regimen, there was at least an 80% chance of observing an adverse event with a true incidence ≥4%. The two-sided 95% confidence interval for overall response rate for each regimen was expected to be ± 0.11 from the observed rate, assuming an expected overall response rate of 0.85 based on data from previous studies of rituximab plus CHOP or bendamustine.18 For the efficacy evaluation, response rates and 95% PearsonClopper confidence intervals were estimated. Progression-free survival was assessed using Kaplan-Meier methodology.

Patients

Eligible patients were aged ≥18 years, had documented CD20+ FL with no prior systemic therapy, were deemed in need of treatment by the investigator, had ≥1 bi-dimensionally measurable lesion (>1.5 cm at its largest dimension by computed tomography scan), had a life expectancy >12 weeks, an Eastern Cooperative Oncology Group performance status of 0-2, and no disease transformation based on lymph node biopsy or re-biopsy within 5 months of the start of treatment. Key exclusion criteria included central nervous system lymphoma, a history of malignancy within 2 years of study entry, and evidence of significant, uncontrolled comorbidities. All patients provided written informed consent. The study was conducted in accordance with the Declaration of Helsinki and the International Conference on Harmonization Good Clinical Practice guidelines, and was approved by the appropriate local ethics committees.

Treatment Assignment to chemotherapy regimen was decided on a percenter basis before enrollment. Patients received obinutuzumab [1000 mg intravenously (iv), days 1 and 8 of cycle 1, and day 1 of subsequent cycles] plus bendamustine (4-6 cycles at 4-week intervals: 90 mg/m2 iv on days 2 and 3 of cycle 1, and days 1 and 2 of subsequent cycles) or CHOP (6-8 cycles at 3-week intervals: cyclophosphamide, 750 mg/m2 iv day 1; doxorubicin, 50 mg/m2 iv day 1; vincristine, 1.4 mg/m2 capped at 2 mg iv day 1; prednisone, 100 mg orally days 1-5). The first obinutuzumab infusion was administered at an initial rate of 50 mg/h escalating to a maximum of 400 mg/h while subsequent infusions were administered at an initial rate of 100 mg/h escalating to the same maximum rate. Patients with a complete response or partial response at the end of induction were eligible for maintenance with obinutuzumab 766

Results Patients’ disposition and baseline characteristics Eighty-one patients were enrolled between August 2010 and September 2011; 41 were allocated to the G-B group and 40 to the G-CHOP group. Results of the final analysis are presented. Baseline characteristics, including prognostic indicators as measured by the FL International Prognostic Index (FLIPI) score,19 are summarized in Table 1. Most patients (91%) had Ann Arbor stage III-IV disease, had an intermediate/high FLIPI score (82%), and had extra-nodal involvement (67%); 43% had bulky disease. The treatment allocation and study flow are summarized in Figure 1. All patients entered in the study received at least one dose of treatment. During the induction period, 90% of patients in the G-B group completed four cycles of treatment, and 85% received the maximum six cycles. In the G-CHOP group, 95% of patients completed six cycles and 33% received the maximum eight cycles. Due to the protocol design, whereby patients received obinutuzumab in combination with four to six cycles of bendamustine or six to eight cycles of CHOP, the mean cumulative dose of obinutuzumab during induction was lower in the G-B group (6534 mg) than in the G-CHOP group (7271 mg). Fifteen patients in the G-B group received steroid prophylaxis after cycle 1, six of whom had not experienced an infusion-related reaction. Nine patients (G-B, 5; G-CHOP, 4) discontinued the study during induction or completed induction but did not proceed to maintenance for reasons summarized in the legend to Figure 1. haematologica | 2017; 102(4)


Obinutuzumab plus chemotherapy in first-line FL

Seventy-two patients started the maintenance phase; 81% of G-B patients and 72% of G-CHOP patients completed the maximum eight cycles of obinutuzumab, with mean cumulative doses for patients who started maintenance of 7222 mg and 6833 mg, respectively. Seventeen patients discontinued maintenance for reasons summarized in the legend to Figure 1; seven patients in the G-B group (19% of patients entering maintenance) and ten patients in the G-CHOP group (28%). Eleven discontinuations occurred during the first to fourth administrations and six during the fifth to eighth administrations. Overall, 73 patients entered follow-up [G-B, 38 (93%); G-CHOP, 35 (88%)], comprising eight patients who entered post-induction without proceeding to maintenance and 65 patients who entered post-maintenance (55 who completed maintenance and 10 who discontinued maintenance prematurely). Eight patients did not enter follow-up for various reasons, most commonly insufficient therapeutic response (4 patients with progressive disease and 1 with stable disease). At the time of analysis, 14 of the 73 patients had discontinued follow-up, most commonly due to disease progression. The median observation time (time from first drug administration to last date alive) on study was 51 months (range, 0.3-60).

Safety Hematologic adverse events and the most common non-hematologic adverse events by treatment group and phase are shown in Tables 2 and 3, respectively. Induction. All patients experienced at least one adverse event during the induction phase, with 64% (G-B, 51%; GCHOP, 78%) experiencing grade 3/4 adverse events. Infusion-related reactions were the most common adverse events (occurring in 58% of patients); the majority occurred during cycle 1 and were grade ≤2 in intensity. Dose delays or modifications (including infusion interruptions due to

Table 1. Patients’ baseline demographics and disease characteristics.

Characteristic

G-B (N=41)

Age, years, median (range) 57.0 (34-84) Male, n. (%) 17 (41) Ann Arbor stage, n. (%) I 2 (5) II 3 (7) III 15 (37) IV 21 (51) Bulky disease (≥7 cm), n. (%) Yes 17 (41) No 24 (59) Bone marrow involvement, n. (%) Negative 20 (49) Positive 19 (46) Indeterminate 2 (5) FLIPI score, n. (%) Low risk (0-1) 8 (20) Intermediate risk (2) 14 (34) High risk (3-5) 19 (46) FLIPI 2* score, n. (%) Low risk (0-1) 16 (39) Intermediate risk (2) 6 (15) High risk (3-5) 17 (41) Unknown 2 (5) Hemoglobin, n. (%) <120 g/L 13 (32) ≥120 g/L 28 (68) Extra-nodal involvement, n. (%) Yes 24 (59) No 17 (41)

G-CHOP (N=40)

Total (N=81)

53.5 (27-74) 19 (48)

55.0 (27-84) 36 (44)

— 2 (5) 10 (25) 28 (70)

2 (2) 5 (6) 25 (31) 49 (60)

18 (45) 22 (55)

35 (43) 46 (57)

19 (48) 21 (53) —

39 (48) 40 (49) 2 (2)

7 (18) 15 (38) 18 (45)

15 (19) 29 (36) 37 (46)

12 (30) 13 (33) 13 (33) 2 (5)

28 (35) 19 (23) 30 (37) 4 (5)

8 (20) 32 (80)

21 (26) 60 (74)

30 (75) 10 (25)

54 (67) 27 (33)

*Analysis of FLIPI 2 scores28 was not specified in the study protocol and data were calculated retrospectively. FLIPI: Follicular Lymphoma International Prognostic Index; G-B: obinutuzumab plus bendamustine; G-CHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

Figure 1. Disposition of patients in the study. *Reasons for discontinuation from G-B induction therapy: insufficient therapeutic response (n=2), administrative/other (n=1), and withdrawal of consent (n=1; this patient did not enter post-induction follow-up). †Reasons for discontinuation from G-CHOP induction therapy: adverse event (AE)/intercurrent illness (n=1) and administrative/other (n=1). ‡Reasons patients did not start Gmaintenance treatment (G-B group): AE/intercurrent illness (n=1). §Reasons patients did not start G-maintenance treatment (G-CHOP group): administrative/other (n=2). IReasons for withdrawal from maintenance treatment (G-B group): AE/intercurrent illness (n=5) and insufficient therapeutic response (n=2). ¶ Reasons for withdrawal from maintenance treatment (G-CHOP group): AE/intercurrent illness (n=4), insufficient therapeutic response (n=3), administrative/other (n=2), and death (n=1).

haematologica | 2017; 102(4)

767


A. Grigg et al. Table 2. Hematologic adverse events* occurring in >5% of patients in either group during any treatment phase, n. (%).

G-B Total patients with ≥1 AE Neutropenia Anemia Thrombocytopenia Febrile neutropenia

Induction (N=41) All grades

Grade 3/4

All grades

16 (39) 12 (29) 3 (7) 4 (10) —

15 (37) 12 (29) 3 (7) 2 (5) —

8 (22) 5 (14) 3 (8) 2 (6) 1 (3)

All grades

Grade 3/4

All grades

23 (58) 18 (45) 2 (5) 1 (3) 4 (10)

22 (55) 17 (43) 1 (3) 1 (3) 4 (10)

— — — — —

All grades

Grade 3/4

All grades

39 (48) 30 (37) 5 (6) 5 (6) 4 (5)

37 (46) 29 (36) 4 (5) 3 (4) 4 (5)

8 (11) 5 (7) 3 (4) 2 (3) 1 (1)

G-CHOP Total patients with ≥1 AE Neutropenia Anemia Thrombocytopenia Febrile neutropenia

Induction (N=40)

Total Total patients with ≥1 AE Neutropenia Anemia Thrombocytopenia Febrile neutropenia

Induction (N=81)

Maintenance (N=36) Grade 3/4 6 (17) 5 (14) 1 (3) 2 (6) 1 (3)

Maintenance (N=36) Grade 3/4 — — — — —

Maintenance (N=72) Grade 3/4 6 (8) 5 (7) 1 (1) 2 (3) 1 (1)

*Hematologic adverse events were defined as adverse events in the Medical Dictionary for Regulatory Activities (MedDRA) System Organ Class ‘Blood and Lymphatic System Disorders’. AE: adverse event; G-B: obinutuzumab plus bendamustine; G-CHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

infusion-related reactions) were required in 65% of patients (G-B, 59%; G-CHOP, 73%). Fifty patients (G-B, 23; GCHOP, 27) had 74 (G-B, 33; G-CHOP, 41) dose delays or interruptions of obinutuzumab because of adverse events (no dose reductions were allowed per protocol), most commonly because of infusion-related reactions: 17 patients in the G-B group, and 19 patients in the G-CHOP group. Twenty-nine patients (36% overall: 29% of G-B patients and 43% of G-CHOP patients) had 43 (G-B, 16; G-CHOP, 27) dose delays or modifications to chemotherapy because of adverse events, most commonly neutropenia. Only one patient in the G-CHOP group discontinued induction due to an infusion-related reaction. The most common grade 3/4 hematologic adverse event was neutropenia, which occurred in 36% of patients (G-B, 29%; G-CHOP, 43%) during induction, although febrile neutropenia was rare (Table 2). Of note, primary granulocyte colony-stimulating factor (G-CSF) prophylaxis was not used in this trial. Neutropenia resolved within 3-33 days from onset in the G-B group and 4-32 days of onset in the G-CHOP group. Three patients had neutropenia unresolved by day 28; in two patients in the G-B group it resolved on days 30 and 33 and in one patient in the GCHOP group it resolved on day 32. Eight patients in the GB group and five patients in the G-CHOP group experienced dose delays due to neutropenia. In the G-B group, all 12 patients who experienced grade 3/4 neutropenia subsequently received G-CSF, as did 14 of 17 patients experiencing neutropenia in the G-CHOP group. One patient in the G-CHOP group developed grade 3 thrombocytopenia during cycle 1, which led to a delay in starting cycle 2. Grade 3/4 non-hematologic adverse events overall were uncommon (Table 3). Grade 3/4 infections occurred in 13 patients (16%), predominantly in the context of neutropenia (9 patients). P. jirovecii pneumonia was reported in one patient. 768

Table 3. Non-hematologic adverse events occurring in >15% of patients in either group, % total (% G-B, % G-CHOP). Induction [N=81 (G-B, 41; G-CHOP, 40)] Infusion-related reaction Infections and infestations* Nausea Fatigue Headache Diarrhea Constipation Cough Vomiting Pyrexia Dyspepsia Alopecia Mucosal inflammation Maintenance [N=72 (G-B, 36; G-CHOP, 36)] Infections and infestations* Cough

All grades

Grade 3/4

58 (59, 58) 58 (54, 63) 51 (59, 43) 37 (39, 35) 32 (34, 30) 26 (34, 18) 26 (20, 33) 23 (22, 25) 20 (17, 23) 19 (22, 15) 15 (12, 18) 11 (2, 20) 10 (2, 18)

7 (10, 5) 16 (10, 23) 4 (5, 3) 4 (2, 5) 1 (0, 3) 1 (2, 0) — 1 (0, 3) 1 (0, 3) 2 (5, 0) 1 (0, 3) — —

65 (72, 58) 15 (19, 11)

15 (17, 14) —

*Reported as Medical Dictionary for Regulatory Activities (MedDRA) System Organ Class ‘Infections and Infestations’ as adverse events of special interest. G-B: obinutuzumab plus bendamustine; G-CHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

Maintenance. Overall, 27 of 72 eligible patients experienced grade 3-5 adverse events during maintenance. Nine patients withdrew from obinutuzumab treatment due to an adverse event, five in the G-B group (due to giardiasis with anemia, neutropenic infection, flare-up of Crohn disease, nasopharyngitis, and neutropenia in one patient each) and four in the G-CHOP group (3 due to infection and 1 due to peripheral sensory neuropathy). Eight patients (G-B, 6; G-CHOP, 2) had obinutuzumab dose delays or interruptions. The only treatment-related death haematologica | 2017; 102(4)


Obinutuzumab plus chemotherapy in first-line FL

Table 4. Summary of efficacy parameters.

Variable Overall response rate, % (95% CI) CR at end of induction, % (95% CI) CR at 30 months, % (95% CI) PFS at 36 months, % (95% CI) Progression/death (n) Deaths due to progressive disease (n)

G-B (N=41)

G-CHOP (N=40)

Total (N=81)

93 (80.1, 98.5) 37 (22.1, 53.1) 63 (46.0, 78.2) 90 (0.80, 0.99) 6 1

95 (83.1, 99.4) 35 (20.6, 51.7) 58 (40.8, 74.5) 84 (0.72, 0.96) 11 2

94 (86.2, 98.0) 36 (25.4, 47.2) 61 (NA, NA) 87 (0.79, 0.94) 17 3

CI: confidence interval; CR: complete response; G-B: obinutuzumab plus bendamustine; G-CHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone; NA: not analyzed; PFS: progression-free survival.

occurred in a patient in the G-CHOP group, 59 days after the only dose of maintenance treatment, due to lactic acidosis in the context of an underlying respiratory infection (pathogen not identified) in the absence of neutropenia. The most common class of non-hematologic adverse events was infections, with 11 patients (G-B, 6; G-CHOP, 5) experiencing a variety of grade 3 infections and one patient in the G-B group having a grade 4 neutropenic infection. No further cases of P. jirovecii pneumonia were reported during maintenance. Eight patients experienced hematologic adverse events during maintenance, all in the G-B group (Table 2); it should be noted that blood tests were only mandatory prior to each 3-monthly cycle. Six patients (8%) developed grade 3/4 neutropenia (n=5) or febrile neutropenia (n=1), noted 81-91 days after the last dose of obinutuzumab. The duration of neutropenia was highly variable, ranging from 4 days to more than 265 days, remaining unresolved at the last follow-up at 265 days in one patient. Only two patients experienced febrile/infective complications. Three patients with prompt resolution of neutropenia (of 4, 8, and 22 days duration after 5, 1, and 2 doses of maintenance, respectively) were re-challenged following response to G-CSF (n=2) or spontaneous resolution (n=1) and went on to complete eight cycles of maintenance without further neutropenia or G-CSF. The other three patients had prolonged neutropenia of 85-265 daysâ&#x20AC;&#x2122; duration; of these, one completed maintenance and two discontinued therapy. G-CSF was used variably in these patients with neutropenia. Two patients did not receive G-CSF. One with grade 4 neutropenia after the first maintenance dose recovered (white cell count â&#x2030;Ľ1x109 cells/L) within 8 days and resumed maintenance. The other patient with grade 3 neutropenia after the seventh dose of maintenance received the eighth and final dose of maintenance after improvement to grade 1 but subsequently had ongoing grade 3 neutropenia that had not resolved at 265 days; a marrow biopsy was not performed. The other four patients in this group received G-CSF. As mentioned, two responded quickly and resumed maintenance. A third developed grade 4 neutropenia after the fifth dose of maintenance, received G-CSF for 3 days without effect and remained persistently neutropenic (white cell count ranging from 0.24-0.8x109 cells/L) until recovery (3.9x109 cells/L) 8 months after onset; a marrow biopsy was not performed. The fourth patient withdrew from the study 3 months after the first maintenance dose due to a late-onset neutropenic infection. Grade 4 neuhaematologica | 2017; 102(4)

tropenia and febrile neutropenia were recurrent in this patient over an 8-month period with seven discrete episodes (1 of febrile neutropenia) lasting from 9-21 days despite ongoing titrated G-CSF therapy. These were associated with grade 2-4 thrombocytopenia on several occasions. Bone marrow examinations during four of the episodes (this was the only patient in the group who had a bone marrow examination for investigation of neutropenia) all showed similar findings with maturation arrest of the myeloid series, adequate megakaryopoiesis and no evidence of lymphoma. Prednisolone therapy was instituted at 8 months and neutropenia resolved; maintenance GCSF was slowly weaned and stopped 15 months after initial onset of the neutropenia with no further recurrence after 33 months of subsequent follow up. Follow-up. No serious adverse events were observed in the eight patients who entered follow-up directly after induction. Three patients experienced serious adverse events during post-maintenance follow-up. In the G-B group, one patient had lower abdominal pain (grade 3); in the G-CHOP group, one patient had an abnormal liver function test (grade 4) and one had dyspnea (grade 3). Blood tests were conducted every 3 months during follow-up. One of 38 evaluable G-B patients experienced grade 3 neutropenia during follow-up 6 months after completing maintenance. Spontaneous recovery was recorded after 85 days and no further neutropenic events were recorded at the 12-month follow-up.

Efficacy The overall response rate was 94% at the end of induction and the complete response rate was 37% at the end of induction and 61% at 30 months (Table 4). The estimated progression-free survival rate at 36 months was 87% (Figure 2). At the final analysis, 17 events defining progression/death had occurred in 81 patients: one event (progression) occurred during induction, six during maintenance (5 progression and 1 death, including one patient in the G-CHOP group with transformation to diffuse large B-cell lymphoma), and ten after maintenance. A summary of efficacy parameters by study group is presented in Table 4.

Pharmacokinetics Patients in both treatment groups had similar mean serum concentrations of obinutuzumab during induction and maintenance, with a similar half-life observed for both treatments (G-B, 37.3 days; G-CHOP, 35.5 days; Figure 3). Cmax and Cmin values increased over the induction period. 769


A. Grigg et al.

Pharmacodynamics B-cell levels decreased rapidly in both treatment groups following the first obinutuzumab infusion, with all patients showing B-cell depletion throughout the treatment period (Figure 4). The majority of patients remained depleted throughout follow-up. The median time from end of treatment to B-cell recovery was 24 months for the G-B group and 29 months for the G-CHOP group. Median levels of T cells and natural killer cells were reduced after the first cycle and were low throughout induction treatment, with full recovery to baseline levels during maintenance or follow-up. By day 28 of follow-up, levels of CD3+, CD4+, CD8+, and CD16+/56+ cells had recovered to 59%, 42%, 78%, and 44% of baseline levels, respectively. Median IgG levels fell from a baseline of 8.3 g/L (normal range, 5-12 g/L) to 7.54 g/L after the end of induction, but

A

B

remained within the normal range and steady thereafter to the end of maintenance [6.95 g/L (based on 66 patients)] and during follow-up [6.76 g/L at 60 weeks of follow-up (49 patients) and 6.5 g/L at 122 weeks (31 patients)]. One patient had hypogammaglobulinemia of IgG (lowest value, 2.74 g/L) during follow-up without clinical consequences and was treated monthly with gammaglobulins until recovery at 28 months after onset. Median levels of IgA and IgM were decreased compared with baseline during the induction phase and remained so throughout maintenance; median levels generally remained within the normal range for IgA (0.5-3.5 g/L) and IgM (0.3-2.3 g/L).

Discussion The results of this study in patients with previously untreated FL demonstrate that induction therapy with GB or G-CHOP followed by obinutuzumab maintenance is associated with a tolerable safety profile and promising efficacy. The decision to start therapy was at the discretion of the investigator, although as 60% of patients with an assessable FLIPI 2 score had intermediate-high risk disease, 43% of patients had bulky disease (>7.5cm), 91% had Ann Arbor stage III-IV, and 67% had extranodal disease, the study population is likely to represent a cohort in which therapy was justified. The majority of patients completed the minimum number of prescribed cycles of G-chemotherapy. There were no new safety signals compared with previous studies of G-monotherapy9,10,12 or G-chemotherapy (G-CHOP or GFC) followed by G-maintenance14 in relapsed or refractory FL. Seventy-six percent of patients (55 of 72 who started maintenance) completed the treatment course. Nonhematologic adverse events were mostly grade 1/2. One death, due to non-neutropenic sepsis, occurred during the study. The most common toxicity was infusion-related reactions in cycle 1, which were generally low grade and at a rate consistent with or lower than in previous studies of

C

Figure 2. Progression-free survival. Progression-free survival (PFS) for (A) the total study population, the (B) G-B group, and (C) the G-CHOP group. G-B: obinutuzumab plus bendamustine; G-CHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

770

Figure 3. Mean serum concentration of obinutuzumab throughout the induction and maintenance periods. G-B: obinutuzumab plus bendamustine; GCHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

haematologica | 2017; 102(4)


Obinutuzumab plus chemotherapy in first-line FL

Figure 4. B-cell levels throughout the induction and maintenance periods. BSL: baseline; EoT: end of treatment; GB: obinutuzumab plus bendamustine; GCHOP: obinutuzumab plus cyclophosphamide, doxorubicin, vincristine, and prednisone.

obinutuzumab in relapsed or refractory indolent NHL.9,10,12,14 In particular, the rate of grade 3/4 infusionrelated reactions (7%) compares favorably with the rate of 21% in the currently indicated use of obinutuzumab in combination with chlorambucil for chronic lymphocytic leukemia.20 Grade 3/4 neutropenia was experienced by 36% of patients (29% of G-B patients and 43% of G-CHOP patients) during induction. In previous studies of patients with indolent NHL receiving induction treatment of rituximab with bendamustine or CHOP, grade 3/4 neutropenia occurred in 29% to 49% of patients given rituximab plus bendamustine and in 69% to 87% of those given rituximab plus CHOP.2,21 The incidence of febrile neutropenia and grade 3/4 infections during induction was modest (5% and 16%, respectively). Grade 3/4 neutropenia was also documented in seven G-B patients, but no G-CHOP patients, during the maintenance phase (n=6) or follow-up (n=1) indicating a potential increase in susceptibility to neutropenia with this combination, although febrile neutropenia only occurred in one of these patients. As neutropenia was monitored only by routine blood tests every 3 months, the incidence and timing of onset is unclear. Delayed onset neutropenia, occurring more than 1 month after the last antibody administration, is a well-recognized complication of rituximab therapy with an incidence of 8% to 25%, and is often associated with myeloid maturation arrest on bone marrow examination as observed in one patient who underwent a marrow investigation in this context.22-24 Neutropenia resolved quickly in three patients; however, in the other four, the neutropenia lasted from 85 days to more than 8 months with variable responses to G-CSF. Additional data from ongoing trials are awaited to examine further the incidence, clinical significance and natural history of neutropenia in this context. In this study complete response was defined by com-

References 1. Dreyling M, Ghielmini M, Marcus R, et al. Newly diagnosed and relapsed follicular lymphoma: ESMO clinical practice guidelines for diagnosis, treatment and follow-up.

haematologica | 2017; 102(4)

puted tomography findings, which in FL can be difficult to interpret in the short term and may give lower complete response rates than with assessment based on positron emission tomography.25 With this caveat, the complete response rate at 30 months was 61% and was substantially higher after obinutuzumab maintenance than after induction; whether this represents the natural history of nodal regression in FL after therapy or a true anti-lymphoma effect of maintenance is impossible to determine in the absence of prospective randomized data. The progression-free rate at 36 months was 87%. Comparison of these results with those from similar studies of rituximab-chemotherapy induction followed by rituximab maintenance (the progression-free survival at 36 months in the PRIMA study was 75%) are not strictly valid as study populations may differ;26,27 however, the impression from this study is that obinutuzumabchemotherapy is unlikely to be inferior. Based on the clinical potential seen for obinutuzumab in this and other phase 1 and 2 studies, the confirmatory phase 3 GALLIUM study is ongoing, comparing obinutuzumab plus chemotherapy with rituximab plus chemotherapy, followed by immunotherapy maintenance with obinutuzumab or rituximab, respectively, in previously untreated patients with indolent NHL (NCT01332968). Acknowledgments Research support for this study was provided by F. HoffmannLa Roche Ltd. The authors would like to thank the patients and their families, and the study investigators, study coordinators, and nurses who assisted with the obinutuzumab clinical program. This study was sponsored by F. Hoffmann-La Roche Ltd. Medical writing support, under the direction of the lead author, was provided by Susan Browne, PhD, Gardiner-Caldwell Communications (Macclesfield, UK). Support for third-party writing assistance was provided by F. Hoffmann-La Roche Ltd.

Ann Oncol. 2014;25(Suppl 3):iii76-iii82. 2. Rummel MJ, Niederle N, Maschmeyer G, et al. Bendamustine plus rituximab versus CHOP plus rituximab as first-line treatment for patients with indolent and mantle-cell lymphomas: an open-label, multicentre, randomised, phase 3 non-inferiority trial.

Lancet. 2013;381(9873):1203-1210. 3. Keating GM. Rituximab: a review of its use in chronic lymphocytic leukaemia, lowgrade or follicular lymphoma and diffuse large B-cell lymphoma. Drugs. 2010;70(11):1445-1476. 4. Glennie MJ, French RR, Cragg MS, Taylor

771


A. Grigg et al.

5.

6.

7.

8.

9.

10.

11.

12.

772

RP. Mechanisms of killing by anti-CD20 monoclonal antibodies. Mol Immunol. 2007;44(16):3823-3837. Heinrich DA, Weinkauf M, Hutter G, et al. Differential regulation patterns of the antiCD20 antibodies obinutuzumab and rituximab in mantle cell lymphoma. Br J Haematol. 2015;168(4):606-610. Herter S, Herting F, Mundigl O, et al. Preclinical activity of the type II CD20 antibody GA101 (obinutuzumab) compared with rituximab and ofatumumab in vitro and in xenograft models. Mol Cancer Ther. 2013;12(10):2031-2042. Herting F, Friess T, Bader S, et al. Enhanced anti-tumor activity of the glycoengineered type II CD20 antibody obinutuzumab (GA101) in combination with chemotherapy in xenograft models of human lymphoma. Leuk Lymphoma. 2014;55(9):2151-2160. Mössner E, Brünker P, Moser S, et al. Increasing the efficacy of CD20 antibody therapy through the engineering of a new type II anti-CD20 antibody with enhanced direct and immune effector cell-mediated Bcell cytotoxicity. Blood. 2010;115(22):43934402. Salles G, Morschhauser F, Lamy T, et al. Phase 1 study results of the type II glycoengineered humanized anti-CD20 monoclonal antibody obinutuzumab (GA101) in B-cell lymphoma patients. Blood. 2012;119(22):5126-5132. Salles GA, Morschhauser F, Solal-Céligny P, et al. Obinutuzumab (GA101) in patients with relapsed/refractory indolent nonHodgkin lymphoma: results from the phase II GAUGUIN study. J Clin Oncol. 2013;31(23):2920-2926. Sehn LH, Assouline SE, Stewart DA, et al. A phase 1 study of obinutuzumab induction followed by 2 years of maintenance in patients with relapsed CD20-positive B-cell malignancies. Blood. 2012;119(22):51185125. Sehn LH, Goy A, Offner FC, et al. Randomized phase II trial comparing obinutuzumab (GA101) with rituximab in

13.

14.

15.

16.

17. 18.

19.

patients with relapsed CD20+ indolent Bcell non-Hodgkin lymphoma: final analysis of the GAUSS study. J Clin Oncol. 2015;33(30):3467-3474. Davies A, Radford J, Cartron G, et al. Obinutuzumab (GA101) plus CHOP or FC in relapsed/refractory follicular lymphoma: final data from the maintenance phase of the phase 1b GAUDI study (BO21000). Blood. 2013;122:(21):1814. Radford J, Davies A, Cartron G, et al. Obinutuzumab (GA101) plus CHOP or FC in relapsed/refractory follicular lymphoma: results of the GAUDI study (BO21000). Blood. 2013;122(7):1137-1143. Dyer MJ, Grigg A, González Díaz M, et al. Obinutuzumab (GA101) in combination with cyclophosphamide, doxorubicin, vincristine and prednisone (CHOP) or bendamustine in patients with previously untreated follicular lymphoma (FL): results of the phase Ib GAUDI study (BO21000). Blood. 2012;120:(21):3686. Dyer MJ, Grigg A, González Díaz M, et al. Obinutuzumab (GA101) in combination with CHOP (cyclophosphamide, doxorubicin, vincristine and prednisone) or bendamustine for the first-line treatment of follicular non-Hodgkin lymphoma: final results from the maintenance phase of the phase Ib GAUDI study. Blood. 2014;124:(21):1743. Cheson BD, Pfistner B, Juweid ME, et al. Revised response criteria for malignant lymphoma. J Clin Oncol. 2007;25(5):579-586. Hiddemann W, Kneba M, Dreyling M, et al. Frontline therapy with rituximab added to the combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) significantly improves the outcome for patients with advanced-stage follicular lymphoma compared with therapy with CHOP alone: results of a prospective randomized study of the German Low-Grade Lymphoma Study Group. Blood. 2005;106 (12):3725-3732. Solal-Céligny P, Roy P, Colombat P, et al. Follicular lymphoma international prognostic index. Blood. 2004;104(5):1258-1265.

20. Goede V, Fischer K, Busch R, et al. Obinutuzumab plus chlorambucil in patients with CLL and coexisting conditions. N Engl J Med. 2014;370(12):1101-1110. 21. Flinn IW, van der Jagt R, Kahl BS, et al. Randomized trial of bendamustine-rituximab or R-CHOP/R-CVP in first-line treatment of indolent NHL or MCL: the BRIGHT study. Blood. 2014;123(19):2944-2952. 22. Voog E, Morschhauser F, Solal-Celigny P. Neutropenia in patients treated with rituximab. N Engl J Med. 2003;348(26):26912694. 23. Nitta E, Izutsu K, Sato T, et al. A high incidence of late-onset neutropenia following rituximab-containing chemotherapy as a primary treatment of CD20-positive B-cell lymphoma: a single-institution study. Ann Oncol. 2007;18(2):364-369. 24. Lemieux B, Tartas S, Traulle C, et al. Rituximab-related late-onset meutropenia after autologous stem cell transplantation for aggressive non-Hodgkin’s lymphoma. Bone Marrow Transplant. 2004;33(9):921-923. 25. Trotman J, Luminari S, Boussetta S, et al. Prognostic value of PET-CT after first-line therapy in patients with follicular lymphoma: a pooled analysis of central scan review in three multicentre studies. Lancet Haematol. 2014;1(1):e17-e27. 26. Salles G, Seymour JF, Offner F, et al. Rituximab maintenance for 2 years in patients with high tumour burden follicular lymphoma responding to rituximab plus chemotherapy (PRIMA): a phase 3, randomised controlled trial. Lancet. 2011;377 (9759):42-51. 27. Morschhauser F, Seymour J, Feugier P, et al. Impact of induction chemotherapy regimen on response, safety and outcome in the PRIMA study. Ann Oncol. 2011;22(Suppl 4):89. 28. Federico M, Bellei M, Marcheselli L, et al. Follicular lymphoma international prognostic index 2: a new prognostic index for follicular lymphoma developed by the international follicular lymphoma prognostic factor project. J Clin Oncol. 2009;27(27):4555-4562.

haematologica | 2017; 102(4)


ARTICLE

Plasma Cell Disorders

IL21R expressing CD14+CD16+ monocytes expand in multiple myeloma patients leading to increased osteoclasts

Marina Bolzoni,1 Domenica Ronchetti,2,3 Paola Storti,1,4 Gaetano Donofrio,5 Valentina Marchica,1,4 Federica Costa,1 Luca Agnelli,2,3 Denise Toscani,1 Rosanna Vescovini,1 Katia Todoerti,6 Sabrina Bonomini,7 Gabriella Sammarelli,1,7 Andrea Vecchi,8 Daniela Guasco,1 Fabrizio Accardi,1,7 Benedetta Dalla Palma,1,7 Barbara Gamberi,9 Carlo Ferrari,8 Antonino Neri,2,3 Franco Aversa1,4,7 and Nicola Giuliani1,4,7

1 Myeloma Unit, Dept. of Medicine and Surgery, University of Parma; 2Dept. of Oncology and Hemato-Oncology, University of Milan; 3Hematology Unit, “Fondazione IRCCS Ca’ Granda”, Ospedale Maggiore Policlinico, Milan; 4CoreLab, University Hospital of Parma; 5 Dept. of Medical-Veterinary Science, University of Parma; 6Laboratory of Pre-clinical and Translational Research, IRCCS-CROB, Referral Cancer Center of Basilicata, Rionero in Vulture; 7Hematology and BMT Center, University Hospital of Parma; 8Infectious Disease Unit, University Hospital of Parma and 9“Dip. Oncologico e Tecnologie Avanzate”, IRCCS Arcispedale Santa Maria Nuova, Reggio Emilia, Italy

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Haematologica 2017 Volume 102(4):773-784

ABSTRACT

B

one marrow monocytes are primarily committed to osteoclast formation. It is, however, unknown whether potential primary alterations are specifically present in bone marrow monocytes from patients with multiple myeloma, smoldering myeloma or monoclonal gammopathy of undetermined significance. We analyzed the immunophenotypic and transcriptional profiles of bone marrow CD14+ monocytes in a cohort of patients with different types of monoclonal gammopathies to identify alterations involved in myeloma-enhanced osteoclastogenesis. The number of bone marrow CD14+CD16+ cells was higher in patients with active myeloma than in those with smoldering myeloma or monoclonal gammopathy of undetermined significance. Interestingly, sorted bone marrow CD14+CD16+ cells from myeloma patients were more pro-osteoclastogenic than CD14+CD16cells in cultures ex vivo. Moreover, transcriptional analysis demonstrated that bone marrow CD14+ cells from patients with multiple myeloma (but neither monoclonal gammopathy of undetermined significance nor smoldering myeloma) significantly upregulated genes involved in osteoclast formation, including IL21R. IL21R mRNA over-expression by bone marrow CD14+ cells was independent of the presence of interleukin-21. Consistently, interleukin-21 production by T cells as well as levels of interleukin-21 in the bone marrow were not significantly different among monoclonal gammopathies. Thereafter, we showed that IL21R over-expression in CD14+ cells increased osteoclast formation. Consistently, interleukin-21 receptor signaling inhibition by Janex 1 suppressed osteoclast differentiation from bone marrow CD14+ cells of myeloma patients. Our results indicate that bone marrow monocytes from multiple myeloma patients show distinct features compared to those from patients with indolent monoclonal gammopathies, supporting the role of IL21R over-expression by bone marrow CD14+ cells in enhanced osteoclast formation. haematologica | 2017; 102(4)

Correspondence: nicola.giuliani@unipr.it

Received: August 5, 2016. Accepted: December 23, 2016. Pre-published: January 5, 2017. doi:10.3324/haematol.2016.153841 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/773 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

773


M. Bolzoni et al.

Introduction Multiple myeloma (MM) is characterized by bone destruction, osteolytic lesions and consequently a higher risk of fractures1 due to an increase in bone marrow (BM) osteoclast formation and osteoblast suppression.2-4 Conversely, patients with indolent gammopathies such as smoldering MM (SMM) and monoclonal gammopathy of undetermined significance (MGUS) are characterized by the absence of lytic lesions, although they may have an increase in osteoclastic bone resorption.5-9 Since the close relationship between plasma cells and BM microenvironment plays a pivotal role in the pathogenesis of MM,10 ongoing studies are focusing on the presence of potential molecular alterations in the microenvironment.11-13 Transcriptional profile alterations have been reported in mesenchymal stromal cells and osteoblasts of MM patients correlated to osteolytic lesions and when compared to healthy donors but not MGUS patients.13 BM monocytes play a pivotal role in bone disease,2,4,14 angiogenesis15 and immune system dysfunction,16 which are hallmarks of active MM. Enhanced bone resorption in MM patients occurs in the BM in close contact with plasma cell infiltration.5 Contact between MM cells and BM stromal cells stimulates the production of the receptor activator of nuclear factor kappa-B ligand (RANKL), the main pro-osteoclastogenic cytokine involved in osteoclast differentiation, through its receptor RANK on the monocyte surface.2,17 Moreover, different factors produced by MM cells can stimulate osteoclastogenesis, including interleukin (IL)-618 and macrophage inflammatory protein (MIP)-1α.19,20 Recently, it has also been reported that IL-3 is increased in BM plasma from MM patients and that it induces Activin A production by BM monocytes which, in turn, stimulates osteoclastogenesis in a RANKL-independent mechanism.21 This suggests that BM monocytes may be directly involved in enhanced osteoclastogenesis in MM. Immunophenotypic analysis of peripheral monocytes has shown that CD14+CD16+ cells account for 5–10% of monocytes in healthy individuals but this sub-population is significantly expanded in cancer22 and inflammatory conditions.23,24 In psoriatic arthritis, CD14+CD16+ cells have been associated with bone erosion and identified as the main source of osteoclast progenitors.25 More recently, it has been reported that the proportion of CD14+CD16+ cells increased in MM patients with the tumor load26,27 and that these cells are potential markers of osteoclast progenitors.27 However, it is not known whether there are alterations in BM monocytes in MM patients. Thus, the aim of this study was to characterize the immunophenotypic and transcriptional profiles of BM CD14+ cells across a cohort of patients with different types of monoclonal gammopathies in order to identify genes that are potentially involved in enhanced osteoclastogenesis and possibly druggable as new therapeutic targets.

the study gave their written informed consent, according to the Declaration of Helsinki. The Institutional Review Board of the University of Parma (Italy) approved all the study protocols. To define the presence of osteolytic lesions in MM patients we used X-ray skeletal survey as the first imaging procedure and alternatively a low-dose computed tomography scan or computed tomography/positron emission tomography evaluation, as recently updated by the International Myeloma Working Group.28 The skeleton was evaluated in MM patients in the same period as the BM aspirates were taken. The presence of at least one lytic lesion on X-ray or computed tomography scan or computed tomography/positron emission tomography scan was used to define osteolytic patients. The presence of three or more lytic lesions was used to define patients with “high bone disease”. MM patients with fewer than three lytic lesions or without bone lesions were considered as having “low bone disease”.29 Not all patients’ samples were suitable for all the analyses. Table 1 reports the number of patients analyzed by the different techniques used in the study.

Immunophenotyping Details of the immunophenotypic analyses performed are reported in the Online Supplementary Data.

Isolation of primary CD14+ cells and CD14+ cell sorting from bone marrow samples CD14+ monocytes were isolated from BM and peripheral blood mononuclear cells by an immunomagnetic method with antiCD14 monoclonal antibody conjugated with microbeads (Miltenyi Biotech; Bergisch-Gladbach, Germany). With the same protocol, CD138+, CD3+, CD4+ and CD8+ cells were isolated from BM samples. The presence of potential contaminating cells in each fraction was evaluated by flow cytometry analysis, using the fluorescence-activated flow cytometer BD FACS Canto II with Diva software [Becton, Dickinson and Company (BD); Franklin Lakes, NJ, USA]. The purity of monocyte samples was >92% and an example of purity analysis is shown in Online Supplementary Figure S1. All the antibodies (anti-human CD14-PerCP-Cy5.5, clone MφP9; anti-human CD138 PE, clone MI15; anti-human CD3 FITC, clone SK7; anti-human CD4 FITC, clone L120; anti-human CD8 PE, clone SK1) were obtained from BD. Primary BM mesenchymal stromal cells were obtained from the CD14-CD138- fraction of BM mononuclear cells. Cells were incubated until confluence for 2 weeks in alpha minimum essential medium (αMEM) supplemented with glutamine, at 15% fetal bovine serum (FBS) (all these reagents purchased from Invitrogen Life Technologies; Carlsbad, CA, USA). For CD14+ cell sorting, purified BM CD14+ cells were stained with PE-Cy7-conjugated anti-CD16 antibody and sorted according to the gating strategy shown in Online Supplementary Figure S2. The cells and gates were analyzed by FACSDiva 7 software (BD) and the cells were sorted on a FACS Aria III instrument.

Table 1. Number of patients with the three types of monoclonal gammopathy analyzed for monocyte features by the different techniques.

Methods Patients A total cohort of 50 patients with newly diagnosed active MM, 32 patients with SMM, and 20 patients with MGUS admitted to our hematology institute from 2010 until 2016 were included in the analysis of monocyte features. All of the subjects involved in 774

MGUS SMM MM

Immunophenotype

GEP

Real-Time PCR

9 15 28

9 15 32*

6 11 13

GEP: gene expression profiling; PCR: polymerase chain reaction. *9 out of 32 samples, displaying high CD138 expression, were excluded from the statistical analysis.

haematologica | 2017; 102(4)


Molecular features of monocytes in myeloma

Fluorescence in situ hybridization, microarray analysis and real-time polymerase chain reaction

Bone marrow interleukin-21 levels in patients with monoclonal gammopathies

These methodologies are detailed in the Online Supplementary Methods section.

The levels of IL-21 in BM plasma were measured by enzymelinked immunosorbent assay, as reported in the Online Supplementary Data.

Monocyte treatment with cytokines Primary CD14+ cells were cultured in the presence or absence of recombinant human (rh) IL-21 (30 pg/mL) (Biovision Inc.; Milpitas, CA, USA) for 24 h and then collected for mRNA analysis. The human monocytic cell line THP-1 was purchased from the American Type Culture Collection (Rockville, MD, USA) and was recently authenticated and tested for mycoplasma contamination. THP-1 cells were treated with rhIL-6 (20 ng/mL) (Thermo Scientific; Rockford, IL, USA) and/or rh tumor necrosis factor (TNF)-α (10 ng/mL) (OriGene; Rockville, MD, USA) for 48 h and then collected for mRNA analysis.

Lentiviral infections The amplified IL21R complementary DNA sequence was cloned into the pLenti-GIII-CMV-GFP-2A-Puro lentiviral vector (Applied Biological Materials Inc.; Richmond, BC, Canada). Recombinant lentivirus was produced by transient transfection of 293T cells.30 Primary CD14+ cells were transduced following published protocols.31 Briefly, 8x106 CD14+ cells purified from peripheral blood buffy coats of healthy donors were placed in wells in 2 mL of αMEM with 10% FBS and rh monocyte colony-stimulating factor (M-CSF; 25 ng/mL) (Peprotech, Rocky Hill, NJ, USA) in the presence of either empty or IL21R vector. As a control, CD14+ cells were also seeded in the same conditions without adding the lentiviral vector. After 18 h, 2 mL of fresh αMEM with 10% FBS and rhM-CSF (25 ng/mL) were added. After 3 days, cells were collected and seeded for osteoclastogenesis assays and for IL21R mRNA analysis.

Osteoclastogenesis assays After cell sorting, both CD14+CD16- and CD14+CD16+ populations were seeded at 2x105 cells/well in 96-well plates in αMEM with 10% FBS, rhM-CSF at 25 ng/mL and rhRANKL (Peprotech) at 60 ng/mL and then cultured for 28 days, replacing half the medium every 2-3 days. In another osteoclastogenesis assay setting, 2x105 lentiviral transduced CD14+ cells/well were seeded in 96-well plates in αMEM with 10% FBS, rhM-CSF 10 ng/mL, rhRANKL 50 ng/mL and the IL-21R signaling inhibitor Janex 1 (10 μM) (Cayman Chemical Company; Ann Arbor, MI, USA) and then cultured for 28 days, replacing half the medium every 2-3 days. Finally, in the third osteoclastogenesis assay setting, MM BM mononuclear cells or purified BM CD14+ cells were used (n=18). Mononuclear cells (4x105/well) or CD14+ cells (2x105/well) were seeded in 96-well plates in αMEM with 10% FBS, rhM-CSF 25 ng/mL and rhRANKL 20 or 60 ng/mL in the presence or absence of rhIL-21 (30 pg/mL) and Janex 1 (10 μM), and then cultured for 28 days, replacing half the medium every 2-3 days. In all in vitro osteoclastogenesis assays, each condition was performed at least in triplicate. The osteoclasts were identified at the end of the culture period as multinucleated (>3 nuclei) cells positive for tartrate-resistant acid phosphatase (TRAP) (Sigma Aldrich; Saint Louis, MO, USA) and counted by light microscopy. Osteoclast areas were quantified using ImageJ software (National Institutes of Health, Bethesda, MD, USA).

STAT3 activity assay The specific procedures adopted for the STAT3 activity analysis are detailed in the Online Supplementary Methods section. haematologica | 2017; 102(4)

Statistical analysis Quantitative variables were compared using the non-parametric Kruskal-Wallis and Mann-Whitney tests, or the parametric two-tailed Student t-test. Results are considered statistically significant at P<0.05. GraphPad Prism 5.0TM was used for all the statistical analyses.

Results Patients with multiple myeloma have higher numbers of bone marrow CD14+CD16+ cells compared to patients with monoclonal gammopathy of undetermined significance The immunophenotype of BM CD14+ cells was evaluated in BM aspirates from 28 patients with active MM (median age 73 years, range 48-89; 46% female, 54% male; International Staging System stage: I=3 , II=8, III=17); 15 with SMM (median age 57 years, range 38-82; 33% female, 67% male) and nine with MGUS (median age 57 years, range 37-78; 44% female, 56% male). Forty-eight percent of the patients with active MM had evidence of osteolytic lesions. A representative example of flow cytometry analysis is reported in Figure 1A. We found that the median percentage of CD14+CD16+ cells in BM samples increased significantly across the different types of monoclonal gammopathies from MGUS to active MM: MGUS: 1.9% (range, 0-3%); SMM: 3.5% (range, 1.0-7.5%); active MM: 5.25% (range, 0-20.0%) (P=0.0071). In particular a statistically significant difference was observed comparing MM and MGUS patients (P=0.0144) (Figure 1B). There was no statistical difference in the median percentages of the BM CD14+CD16+ population between osteolytic MM patients (n=13) versus not-osteolytic ones (n=15): 4% (range, 0-20%) versus 7.8% (range, 0-20%), respectively (Figure 1C). Similarly, comparing MM patients with high bone disease (n=9) versus those with low bone disease (n=19), we did not find a statistically significant difference in the median percentages of the BM CD14+CD16+ population: 4.7% (range, 0-20%) versus 5.5% (range, 020%), respectively (Figure 1D). In the same way, analyzing all myeloma patients (SMM and MM), there was no significant difference between the osteolytic patients compared to those without osteolytic lesions [osteolytic MM (n=13) versus not-osteolytic MM plus SMM (n=30): 4% (range, 0-20%) versus 4.3% (range, 0-20%), respectively].

Bone marrow CD14+CD16+ cells in multiple myeloma patients are pro-osteoclastogenic in ex vivo cultures To investigate whether the increased number of CD14+CD16+ cells observed in BM samples from MM patients could be associated with enhanced pro-osteoclastogenic activity, we sorted the BM CD14+CD16+ monocyte population by FACS and then tested its ex vivo pro-osteoclastogenic differentiation properties in comparison with the CD14+CD16- cell fraction. A representative example of 775


M. Bolzoni et al.

flow cytometry analysis of monocyte sub-populations before and after cell sorting is reported in Online Supplementary Figure S2. A median of 3.5x105 CD14+CD16+ cells was obtained after cell sorting from MM BM samples. Due to the limited numbers of cells, we were only able to perform osteoclastogenesis assays. Interestingly, we found that, compared to the CD14+CD16- population, CD14+CD16+ cells generated more TRAP-positive cells with a higher number of osteoclasts showing five or more nuclei (Figure 1 E).

A

C

A different transcriptional fingerprint characterizes bone marrow CD14+ cells from patients with multiple myeloma as compared to smoldering multiple myeloma and monoclonal gammopathy of undetermined significance We performed gene expression profiling of purified primary BM monocytes. We checked the intensity value of specific CD14 and CD138 probe sets (Online Supplementary Table S1) and discarded nine samples displaying high CD138 expression from the analysis to further ensure that

B

D

E

Figure 1. Immunophenotype of bone marrow monocytes in patients with monoclonal gammopathies and their pro-osteoclastogenic ex vivo properties. (A) CD14 and CD16 expression by BM monocytes in patients with monoclonal gammopathies: example of plots of flow cytometry data. (B) Box plots representing the median percentage values of CD14+CD16+ cells evaluated in BM samples obtained from patients with MGUS, SMM or MM (P calculated using the Mann-Whitney test). (C, D) Box plots representing the median percentage values of CD14+CD16+ cells evaluated in BM samples obtained (C) from patients with (w) or without (w/o) osteolysis and (D) from patients with high bone disease (HBD) or low bone disease (LBD). (E) Osteoclast assay. CD14+ cells were purified from BM samples of patients with monoclonal gammopathies by an immunomagnetic method and then sorted into the two sub-populations CD14+CD16- and CD14+CD16+ by a flow cell sorter as described in the Methods section. CD14+CD16- or CD14+CD16+ cells (200,000 cells/well) were seeded in 96-well plates in ÎąMEM with 10% FBS, rhM-CSF 25 ng/mL and rhRANKL 60 ng/mL. After 28 days of culture, osteoclasts were identified as multinucleated TRAP-positive cells and counted by light microscopy. (E) Bar graph represents the median number of osteoclasts/well of each condition, divided into osteoclasts with â&#x2030;Ľ5 nuclei or <5 nuclei (P calculated using the Mann-Whitney test, CD14+CD16- versus CD14+CD16+ cells). On the right there is a representative image of the osteoclastogenesis assay stained by TRAP from BM sorted CD14+CD16and CD14+CD16+ cells (original magnification, 4x).

776

haematologica | 2017; 102(4)


Molecular features of monocytes in myeloma

Table 2. Main clinical characteristics of the cohort of patients eligible for gene expression analysis.

Patient

Sex

Age

MGUS1 MGUS2 MGUS3 MGUS4 MGUS5 MGUS6 MGUS7 MGUS8 MGUS9 SMM1 SMM2 SMM3 SMM4 SMM5 SMM6 SMM7 SMM8 SMM9 SMM10 SMM11 SMM12 SMM13 SMM14 SMM15 MM1 MM2 MM3 MM4 MM5 MM6 MM7 MM8 MM9 MM10 MM11 MM12 MM13 MM14 MM15 MM16 MM17 MM18 MM19 MM20 MM21 MM22 MM23

F F M M F M M M F F F F M F F F M M F F M M M M M M F F M F M M M M M F M M F M F F M M F F F

43 52 42 78 54 77 43 91 50 67 72 58 58 64 61 64 56 70 67 76 83 47 41 65 83 75 59 80 76 79 69 79 57 81 60 78 73 74 73 70 84 86 73 56 50 75 77

ISS

Osteolysis

HBD

Type

del(13q)

H

MM cell genetic alterations del(17p) t(11;14) t(4;14)

t(14;16)

l l k l k k l l l l k k l k k k

III III II II II II I II I II II II III II II I III III II III II I III

+

-

+ + + + +

+ + +

+ + -

+ + -

+ + + -

+ + -

k k k k k l k k l l l k l l k k l k l l k k l l k k k k

-

-

-

-

-

-

+

-

-

+

-

-

+

-

-

-

-

-

-

-

+

-

-

+

-

-

+

-

-

-

+

-

+ + + +

+ -

-

+ -

-

+ -

+ -

+

+ -

-

-

-

-

-

-

-

+

-

+

-

-

-

-

-

-

-

-

-

-

-

-

+

-

-

+

-

-

-

-

-

-

-

MGUS: monoclonal gammopathy of undetermined significance; SMM: smoldering multiple myeloma; MM: multiple myeloma; F: female; M: male; ISS: International Staging System; HBD: high bone disease; H: hyperdploid; l: lambda k: kappa.

777


M. Bolzoni et al. A

B

C

D

Figure 2. Transcriptional fingerprints evaluated by gene expression profiling of purified bone marrow CD14+ cells from patients with different monoclonal gammopathies. (A) Heatmap of the trancriptional profiles resulting from the unsupervised analysis of all the MM, SMM and MGUS monocyte samples. (B) Scatterplots showing the correlation between IL21R expression and that of CD40, SLAMF7 and CCR5 by BM monocytes as determined from gene expression analysis. The lines represent the linear regression between each couple of genes. (C) Quantitative real-time polymerase chain reaction of IL21R, CCR5, CD40 and SLAMF7 genes performed on BM monocytes purified from patients with monoclonal gammopathies. Values represent the median of the -Î&#x201D;Ct values of the reactions (*: fold change >1.5). (D) Quantitative real-time polymerase chain reaction of IL21R, CCR5, CD40 and SLAMF7 genes performed on BM monocytes purified from MM patients with (w) or without (w/o) osteolysis. Values represent the mean of the -â&#x2C6;&#x2020;Ct values of the reactions.

malignant plasma cells were not included in the analysis. Hence, BM monocyte samples included in the gene expression analysis were obtained from 23 MM, 15 SMM and nine MGUS patients. The main characteristics of patients eligible for gene expression analysis are reported in Table 2. Unsupervised analysis significantly clustered together MM samples (P=0.0024, Fisher exact test) whereas SMM and MGUS were scattered along the dendrogram. (Figure 2A). A multiclass analysis identified 99 differentially expressed genes in CD14+ cells between the three classes of patients (Online Supplementary Figure S3A; Online Supplementary Table S2), whereas 78 genes (18 up- and 60 down-regulated) were differentially expressed in monocytes of MM patients as compared to SMM patients (Online Supplementary Figure S3B; Online Supplementary Table S3). The comparison of MM with samples from asymptomatic patients (SMM and MGUS) identified 254 genes differentially expressed in CD14+ cells; specifically, there were 62 up-regulated and 192 down-regulated genes (Online Supplementary Figure S3C, Online Supplementary Table S4). Functional annotation analysis of genes differentially expressed in symptomatic patients was performed using standard procedures with the Database for 778

Annotation, Visualization and Integrated Discovery (DAVID) and Gene Set Enrichment Analysis (GSEA) tools (Table 3). Among the identified gene sets, it is worth mentioning those associated with the cytokine-cytokine receptor interaction pathway, Jak-STAT signaling pathway, and the interferon alpha and gamma responses. Among the differentially expressed genes, chemokines and chemokine and cytokine receptors with pro-osteoclastogenic properties such as CCR5, IL21R and CD40, were specifically upregulated in CD14+ cells from MM patients. Importantly, monocytes in MM samples up-regulate SLAMF7, which is selectively expressed in plasma cells and natural killer cells in MM leading to antibody-dependent cellular cytotoxicity and direct natural killer cell activation.32 Particularly, IL21R over-expression by BM CD14+ cells in MM was demonstrated (q-value=0), both in the multiclass analysis and when comparing MM versus SMM plus MGUS (Online Supplementary Tables S2 and S4). Interestingly, IL21R gene expression in the complete database significantly correlated with the expression of CCR5 (P=0.0197), CD40 (P<0.0001), and SLAMF7 (P=0.0002) genes (Figure 2B). Moreover a further analysis between osteolytic versus notosteolytic patients with active MM identified 12 genes haematologica | 2017; 102(4)


Molecular features of monocytes in myeloma

Table 3. Functional annotations* of the representative genes distinguishing BM monocytes as emerging from supervised analysis of MM versus MGUS plus SMM patients.

Database for Annotation, Visualization and Integrated Discovery (DAVID ) Genes§

Pathway Database

Term

KEGG

Cytokine-cytokine receptor interaction Jak-STAT signaling pathway Chemokine signaling pathway Metabolism of carbohydrates

REACTOME

Functional group category

CCR5, IL21R, CSF3R, CD40 IL21R, CSF3R, GNGT2, CCR5 GOT2, GPI, PGM1, PGD, GYS1

Gene Set Enrichment Analysis (GSEA) Genes§

INTERFERON ALPHA RESPONSE INTERFERON GAMMA RESPONSE

EPSTI1, IFI27, IFITM1, ISG20, LAP3 CD40, EPSTI1, IFI27, ISG20, LAP3, PIM1, SLAMF7, STAT4, VAMP5

*The DAVID Functional Annotation Tool v6.8 (https://david.ncifcrf.gov) and GSEA tool (http://software.broadinstitute.org/gsea) were used to classify genes into functional categories. §Genes down-regulated in MM monocytes are underlined.

(SERPINB10; CDCA5; MYBL2; SELENBP1; TK1; GYPA; KIF18A; SPC25; HJURP; TAL1; SKA1; E2F8) that were down-regulated in not-osteolytic MM patients. On the other hand, we did not find a significantly different gene expression signature between patients with active MM with high bone disease versus low bone disease. Thereafter, in a subgroup of patients, we confirmed a significant up-regulation of IL21R, CCR5, CD40, and SLAMF7 genes in CD14+ cells from MM patients compared to SMM and/or MGUS patients, by real-time polymerase chain reaction (Figure 2C). Consistently with the gene expression profiling data, we did not find a significantly different expression of IL21R, CCR5, CD40, and SLAMF7 genes between osteolytic versus not-osteolytic patients with active MM (Figure 2D).

Bone marrow CD14+ cells over-express IL21R mRNA in multiple myeloma patients irrespective of interleukin-21 Based on the gene expression data and previous evidence that IL-21 is a growth factor for MM cells33,34 and that the IL-21/IL21R axis promotes osteoclastogenesis and bone destruction in pathological conditions35,36 we further investigated the possible role of IL21R over-expression by CD14+ cells in MM-induced osteoclastogenesis. Firstly, by means of real-time polymerase chain reaction, we confirmed that IL21R mRNA levels significantly increased in BM CD14+ cells across different plasma cell dyscrasias in a cohort of patients with MM, SMM and MGUS (P=0.036) (Figure 3A). We showed that BM CD14+ cells expressed significantly higher levels of IL21R mRNA in MM patients than in MGUS patients (P=0.023, Figure 3A), and in MM patients compared to SMM plus MGUS patients (P=0.005, Figure 3B). The up-regulation of IL21R mRNA was also observed in MM CD14+ cells obtained from peripheral blood (n=3, data not shown). The mean difference between IL21R expression (expressed as -ΔCt) by BM and peripheral blood monocytes purified from the same patient was 0.46±0.49 (P=0.64). The expression of IL21R was also investigated at the protein level by flow cytometry: in line with the evidence that the CD14+CD16+ population was increased in MM patients, we found that BM CD14+ cells in MM patients expressed IL-21R/CD360 and that the CD14+CD16+ population showed higher median fluorescence intensity (MFI) compared to CD14+CD16- cells in each tested patient (mean ΔMFICD14+CD16+MFICD14+CD16- ±SD= 6.1±2.4) (Figure 3C). haematologica | 2017; 102(4)

Furthermore, we checked the levels of active STAT3 in BM CD14+ cells, as it is well known that the signaling pathway down-stream of IL-21R leads to the activation of Jak3 and STAT3.37,38 We found that MM CD14+ cells had significantly higher levels of active STAT3 compared to MGUS cells (P=0.0029) and cells from asymptomatic patients (MGUS plus SMM) (P=0.0093, Figure 3D). To investigate the possible mechanisms involved in IL21R mRNA over-expression, we treated purified BM CD14+ cells obtained from MM, SMM or MGUS patients with the rhIL-21 concentration reached in the BM plasma of our cohort of patients. The addition of rhIL-21 (30 pg/mL) slightly increased IL21R mRNA in BM CD14+ cells from MGUS and SMM patients but not from MM patients, without the difference reaching statistical significance (Figure 3E), suggesting a constitutive IL21R mRNA expression in MM patients irrespective of the presence of IL-21. On the other hand, using the well-established39 monocytic cell line THP-1, we found that combined treatment with the pro-inflammatory cytokines rhIL-6 (20 ng/mL) and rhTNF-α (10 ng/mL) significantly increased IL21R mRNA expression compared to that of untreated controls (P=0.005, Figure 3F).

Bone marrow IL21 expression and protein levels did not significantly differ across patients with monoclonal gammopathies Next, we evaluated IL21 mRNA expression levels in our cohort of patients. BM mesenchymal stromal cells, CD14+ and primary MM cells did not express the IL21 gene, which was otherwise expressed by CD3+ cells, in the CD4+ fraction, as checked by real-time polymerase chain reaction (Online Supplementary Table S5). We failed to find a significant difference in IL21 mRNA expression by CD3+ cells among MM, SMM and MGUS patients, as shown in Figure 4A. Consistently, no significant difference was found in BM levels of IL-21 across the different monoclonal gammopathies, as detected by enzyme-linked immunosorbent assay (Figure 4B). The median BM IL-21 level was 32.4 pg/mL for MGUS, 24.4 pg/mL for SMM and 34.1 pg/mL for newly diagnosed MM patients.

IL21R over-expression by CD14+ cells is involved in osteoclastogenesis Since IL21R was identified among the genes overexpressed by BM CD14+ cells in MM patients, we investi779


M. Bolzoni et al.

gated its role in osteoclast differentiation. We induced IL21R over-expression in CD14+ cells obtained from three different healthy donors. The over-expression of IL21R gene was evaluated by real-time polymerase chain reaction in CD14+ cells infected with IL21R vector (CD14+ IL21R vector) and compared to that of cells infected with the empty vector (CD14+ empty vector, Figure 5A). Subsequently, we performed in vitro osteoclastogenesis assays. The number of osteoclasts in this set of experiments was low because the infection with lentiviral vectors strongly affects primary monocyte viability (Figure 5B,C). Interestingly, in vitro osteoclastogenesis assays showed that the over-expression of IL21R increased the

number and median area of osteoclasts in the presence of RANKL and M-CSF compared to controls; consistently, the presence of Janex 1, a JAK3 inhibitor known to block IL21R signaling, significantly reduced osteoclast formation and size in CD14+ IL21R vector cells (Figure 5B,C).

Blocking interleukin-21 receptor signaling inhibits osteoclastogenesis To further confirm the role of IL21R over-expression in osteoclastogenesis, we performed in vitro osteoclastogenesis assays with or without rhIL-21 in the presence or absence of Janex 1. The presence of rhIL-21 did not affect the number or area of TRAP-positive osteoclasts obtained

B

A

C

D

E

F

Figure 3. IL-21R over-expression by bone marrow CD14+ cells from patients with multiple myeloma compared to those from patients with smoldering multiple myeloma or monoclonal gammopathy of undetermined significance. (A) IL21R mRNA expression was evaluated by real-time polymerase chain reaction (PCR) in purified BM CD14+ obtained from patients with monoclonal gammopathies. Box plots show the median –ΔCt levels (P calculated by the Mann-Whitney test). (B) IL21R mRNA expression was evaluated by real-time PCR in purified BM CD14+ obtained from MM patients versus SMM plus MGUS patients. Box plots show the median – ΔCt levels (P calculated by the Mann-Whitney test). (C) CD360/IL-21R expression was evaluated by flow cytometry in BM CD14+CD16- and CD14+CD16+ cells. Comparison between CD360 expression by CD14+CD16- and CD14+CD16+ monocyte populations, stained with anti-CD360 or control IgG1, is shown in three representative MM patients (i, ii, and iii) (MFI: median fluorescence intensity). (D) Active STAT3 levels were determined by the STAT family assay kit in nuclear extracts of purified BM CD14+ cells obtained from MGUS (n=3), SMM (n=3) and MM (n=3) patients. The bar chart represents the mean±SD level of active STAT3 checked as optical density (OD) at 450 nm with a reference wavelength of 620 nm, after subtracting the blank. (E) BM CD14+ cells purified from patients with MM, SMM or MGUS were treated with or without rhIL-21 (30 pg/mL) for 24 h. IL21R mRNA level was evaluated by real-time PCR. The bar chart represents the median –ΔCt of IL21R mRNA of three replicates (Con: untreated control). (F) The monocytic cell line THP-1 was treated for 48 h with or without rhIL-6 (20 ng/mL) or TNF-α (10 ng/mL) or both cytokines. IL21R mRNA levels were evaluated by real-time PCR in three independent experiments (P calculated using the t test). The bar chart represents the mean±SD fold change of mRNA IL21R (Con: untreated control).

780

haematologica | 2017; 102(4)


Molecular features of monocytes in myeloma

A

B

Figure 4. IL21 mRNA expression by the bone marrow microenvironment and levels of interleukin-21 in the bone marrow in patients with multiple myeloma, smoldering multiple myeloma or monoclonal gammopathy of undetermined significance. (A) IL21 mRNA expression by purified CD3+ cells from MM (n=5), SMM (n=7) or MGUS (n=7) patients evaluated by real-time polymerase chain reaction. The bar chart shows the mean±SD –ΔCt of IL21 mRNA. (B) BM IL-21 levels were evaluated by enzyme linked immunosorbent assay in a cohort of 76 newly diagnosed MM, 42 SMM and 41 MGUS patients. The scatter dot plot represents BM IL-21 levels in the cohort of patients with the lines representing median levels.

B

A

C

Figure 5. IL21R over-expression by a lentiviral vector in monocytes increases the differentiation of osteoclasts. (A) IL21R was over-expressed in peripheral blood CD14+ cells obtained from three different healthy donors transduced with a specific lentiviral vector for IL21R (CD14+ IL21R vector) as compared to those infected with the empty control vector (CD14+ empty vector) or not transduced (CD14+). IL21R mRNA levels were checked by real-time polymerase chain reaction. Bar graph represents the median –ΔCt levels of three independent experiments. CD14+ transduced cells with IL21R or empty lentiviral vectors (200,000 cells/well) were seeded in 96-well plates in αMEM with 10% FBS, rhM-CSF 10 ng/mL and rhRANKL 50 ng/mL in the presence or absence of the IL-21R signaling inhibitor Janex 1 (10 μM) or vehicle (DMSO). After 28 days of culture, osteoclasts were identified as multinucleated (>3 nuclei) TRAP-positive cells and counted by light microscopy. (B) The bar graph shows the mean±SD number of osteoclasts for each well (P calculated using the t test) in three independent experiments with CD14+ from three different healthy donors (left panel). The box plot represents the osteoclast area (P calculated by the Mann-Whitney test) in a representative experiment performed at least in triplicate (right panel). (C) Images of one representative experiment of the osteoclastogenesis assay stained by TRAP of CD14+ IL21R vector and CD14+ empty vector cells performed in the presence or absence of Janex 1 (original magnification, 4x).

haematologica | 2017; 102(4)

781


M. Bolzoni et al.

from total mononuclear cells or CD14+ cells purified from MM BM aspirates (Figure 6A,B). No significant differences were found in osteoclast number or area between tumor and not-tumor samples treated with IL-21 (data not shown). On the other hand, Janex 1 significantly suppressed osteoclastogenesis from either total BM mononuclear cells or BM CD14+ cells obtained from MM patients, both in the presence (P<0.001) and in the absence (P<0.001) of the rhIL-21, as shown in Figure 6A,B. Moreover, the presence of Janex 1 significantly reduced the median osteoclast area both in the presence (P=0.012) and in the absence (P<0.001) of rhIL-21, compared to untreated controls (Figure 6A).

Discussion Monoclonal gammopathies are characterized by the activation of bone resorption with a progressive increase in the number of osteoclasts from MGUS to MM.5 Several studies have evaluated the gene expression profiles of plasma cells obtained both from patients with newly diagnosed MM or MGUS and from healthy donors to identify genes potentially related to the progression of MM.40,41 However, while MGUS and MM can be distinguished from normal plasma cells, these two conditions cannot be easily differentiated from each other.40,41 Transcriptional data were used to strat-

A

ify MM patients with lytic lesions, identifying DKK1 as the main over-expressed gene when focal bone lesions occur.42 Studies analyzing BM microenvironment cells indicate that mesenchymal stromal cells and osteoblasts in MM patients have different transcriptional profiles compared to those of healthy donors and based on the occurrence of osteolytic lesions.13 However, not all the BM biological alterations from MGUS to SMM and, finally, to active MM are clear yet. As regards the immunophenotypic profile, we found that the median percentage of CD14+CD16+ cells in BM samples increased among the different types of monoclonal gammopathies, being significantly higher in MM than in MGUS. In our study, for the first time we sorted BM CD14+CD16+ cells and showed that they represent the osteoclastogenic fraction of CD14+ cells in MM patients, supporting the notion that inflammatory monocytes are involved in MM-induced osteoclastogenesis. In addition, CD14+CD16+ cells might contribute to the high production of inflammatory cytokines, such as TNF-α,23 which are increased in the BM of MM patients and involved in osteoclast formation.43,44 Consistent with the immunophenotypic profile, the transcriptome of CD14+ cells obtained from MM patients showed up-regulation, as compared to the expression in SMM and MGUS, of genes involved in immune response, chemotaxis and osteoclastogenesis. We focused on genes

B

Figure 6. Interleukin-21 receptor signaling inhibition blocks Interleukin-21 driven osteoclastogenesis. BM mononuclear cells, obtained from MM patients, were seeded at the concentration of 4x105 cells/well in 96-well plates in αMEM with 10% FBS, rhM-CSF 25 ng/mL and rhRANKL 20 ng/mL in the presence or absence of rIL-21 (30 pg/mL) and the IL-21R signaling inhibitor Janex 1 (10 μM) or vehicle (DMSO) for 28 days, replacing the medium every 3 days. At the end of the culture period osteoclasts were identified as multinucleated (>3 nuclei) TRAP-positive cells and counted by light microscopy (Con: untreated control). (A) The bar graph represents the mean±SD osteoclast number for each well (P calculated using the t test) (upper panel). The box plot shows the osteoclast (OC) area (P calculated by the Mann-Whitney test) in one representative experiment performed at least in triplicate (lower panel). (B) Representative images of osteoclasts stained with TRAP after 28 days of culture (original magnification, 4x).

782

haematologica | 2017; 102(4)


Molecular features of monocytes in myeloma

potentially involved in osteoclastogenesis. The up-regulated genes included CCR5, whose role in bone destruction in MM has already been extensively investigated.19,20 IL21R mRNA was also over-expressed by BM CD14+ cells in MM but not in SMM or MGUS. The analysis of patients with active MM according to the presence of osteolysis identified only 12 genes that were down-regulated in not-osteolytic patients. Nevertheless, we did not find a significantly different gene expression signature between the MM patients with high bone disease versus those with low bone disease. The lack of major differences in the immunophenotypic and transcriptional profiles of monocytes from osteolytic and not-osteolytic patients are not surprising because the main pathophysiological difference between osteolytic and non-osteolytic MM patients is the suppression of osteoblast formation rather than increased osteoclast formation and activity. Our data are supported by previous reports that all MM patients have a significant increase of bone resorption rate with unbalanced bone remodeling.5,9 In addition, MGUS patients have a significant increase of bone resorption rate.5,7 Consistently, in this study we did not find a large number of differentially expressed genes by monocytes across patients with the different monoclonal gammopathies. In this study, we demonstrated the potential involvement of the IL-21/IL-21R axis in the increased osteoclastogenesis that occurs in MM patients. The ligand of IL-21R, the cytokine IL-21, is a growth factor for MM cells33,34 and it is mainly produced by T cells.37,45 The binding of IL-21 to its receptor leads to the activation of the Jakâ&#x20AC;&#x201C;STAT pathway, in particular Jak1, Jak3, STAT1, and STAT3.37,46,47 Interestingly, a previous study showed that IL-21 up-regulation in the synovium and the serum of patients with rheumatoid arthritis is involved in osteoclastogenesis and bone destruction.35 Nevertheless, the role of IL-21 in MMinduced osteoclast formation is largely unknown. In this study we found significant IL21R mRNA over-expression by CD14+ cells correlated with the other osteoclastogenic genes identified such as CCR5, but also with CD40 and SLAMF7. Interestingly, in line with the immunophenotyp-

References 1. Melton LJ 3rd, Kyle RA, Achenbach SJ, Oberg AL, Rajkumar SV. Fracture risk with multiple myeloma: a population-based study. J Bone Miner Res. 2005;20(3):487-493. 2. Giuliani N, Colla S, Rizzoli V. New insight in the mechanism of osteoclast activation and formation in multiple myeloma: focus on the receptor activator of NF-kappaB ligand (RANKL). Exp Hematol. 2004;32(8):685-691. 3. Giuliani N, Rizzoli V, Roodman GD. Multiple myeloma bone disease: pathophysiology of osteoblast inhibition. Blood. 2006;108(13):3992-3996. 4. Roodman GD. Pathogenesis of myeloma bone disease. Leukemia. 2009;23(3):435441. 5. Bataille R, Chappard D, Basle MF. Quantifiable excess of bone resorption in monoclonal gammopathy is an early symptom of malignancy: a prospective study of 87 bone biopsies. Blood. 1996;87(11):47624769.

haematologica | 2017; 102(4)

ic profile of BM CD14+ in MM patients, IL21R was expressed at high intensity in the CD14+CD16+ fraction. The up-regulation of IL21R in MM patients was associated with an increase of STAT3 signaling and was independent of the presence of IL-21. On the other hand, as the combination of the pro-inflammatory cytokines IL-6 and TNF-Îą increases IL21R mRNA expression in monocytes, we might suppose that these cytokines are involved in IL21R over-expression by BM CD14+ cells. The pathophysiological role of IL21R over-expression by CD14+ cells in the enhanced osteoclastogenesis that occurs in MM patients was further demonstrated by a lentiviral approach. These data also support the role of IL-21R signaling as a potential therapeutic target. Accordingly, it is worth remembering that the clinically approved JAK3 inhibitor tofacitinib suppresses osteoclast-mediated structural damage to arthritic joints and decreases RANKL production.48 This study was not designed to evaluate the role of IL-21R overexpression as a potential biomarker of MM progression; however, our data suggest that IL-21R expression level could be a potential biomarker of myeloma progression. Clearly, only an appropriate prospective study evaluating IL-21R expression would be able to address this point. In conclusion, this study supports the notion that a proinflammatory profile of BM CD14+ cells is involved in osteoclastogenesis in MM patients, in line with a considerable amount of evidence from the literature.19,49,50 For the first time, we highlighted IL21R over-expression in BM monocytes from MM patients and demonstrated its role in increased osteoclastogenesis, suggesting that IL-21R signaling could be a potential new therapeutic target for MM bone disease. Acknowledgments This work was supported in part by a grant from the Associazione Italiana per la Ricerca sul Cancro (AIRC) IG2014 n.15531 (NG) Fondazione Italiana per la Ricerca sul Cancro fellowship [id. 18152 (MB),and id. 16462 (DT)] and AIRC IG16722 (AN). This work was also supported by a fellowship from ParmAIL (Associazione Italiana Contro le LeucemieLinfomi e Myeloma, Parma) (VM).

6. Kyle RA, Durie BG, Rajkumar SV, et al. Monoclonal gammopathy of undetermined significance (MGUS) and smoldering (asymptomatic) multiple myeloma: IMWG consensus perspectives risk factors for progression and guidelines for monitoring and management. Leukemia. 2010;24(6):11211127. 7. Ng AC, Khosla S, Charatcharoenwitthaya N, et al. Bone microstructural changes revealed by high-resolution peripheral quantitative computed tomography imaging and elevated DKK1 and MIP-1alpha levels in patients with MGUS. Blood. 2011;118(25): 6529-6534. 8. Agarwal A, Ghobrial IM. Monoclonal gammopathy of undetermined significance and smoldering multiple myeloma: a review of the current understanding of epidemiology, biology, risk stratification, and management of myeloma precursor disease. Clin Cancer Res. 2013;19(5):985-994. 9. Bataille R, Chappard D, Marcelli C, et al. Mechanisms of bone destruction in multiple myeloma: the importance of an unbalanced

10. 11.

12.

13.

14.

process in determining the severity of lytic bone disease. J Clin Oncol. 1989;7(12):19091914. Roodman GD. Role of the bone marrow microenvironment in multiple myeloma. J Bone Miner Res. 2002;17(11):1921-1925. Corre J, Mahtouk K, Attal M, et al. Bone marrow mesenchymal stem cells are abnormal in multiple myeloma. Leukemia. 2007;21(5):1079-1088. Garayoa M, Garcia JL, Santamaria C, et al. Mesenchymal stem cells from multiple myeloma patients display distinct genomic profile as compared with those from normal donors. Leukemia. 2009;23(8):1515-1527. Todoerti K, Lisignoli G, Storti P, et al. Distinct transcriptional profiles characterize bone microenvironment mesenchymal cells rather than osteoblasts in relationship with multiple myeloma bone disease. Exp Hematol. 2010;38(2):141-153. Yaccoby S, Wezeman MJ, Henderson A, et al. Cancer and the microenvironment: myeloma-osteoclast interactions as a model. Cancer Res. 2004;64(6):2016-2023.

783


M. Bolzoni et al.

15. Ribatti D, Vacca A. The role of monocytesmacrophages in vasculogenesis in multiple myeloma. Leukemia. 2009;23(9):1535-1536. 16. Kawano Y, Moschetta M, Manier S, et al. Targeting the bone marrow microenvironment in multiple myeloma. Immunol Rev. 2015;263(1):160-172. 17. Giuliani N, Bataille R, Mancini C, Lazzaretti M, Barille S. Myeloma cells induce imbalance in the osteoprotegerin/osteoprotegerin ligand system in the human bone marrow environment. Blood. 2001;98(13):3527-3533. 18. Bataille R, Chappard D, Klein B. The critical role of interleukin-6, interleukin-1B and macrophage colony-stimulating factor in the pathogenesis of bone lesions in multiple myeloma. Int J Clin Lab Res. 1992;21(4):283287. 19. Choi SJ, Cruz JC, Craig F, et al. Macrophage inflammatory protein 1-alpha is a potential osteoclast stimulatory factor in multiple myeloma. Blood. 2000;96(2):671-675. 20. Han JH, Choi SJ, Kurihara N, Koide M, Oba Y, Roodman GD. Macrophage inflammatory protein-1alpha is an osteoclastogenic factor in myeloma that is independent of receptor activator of nuclear factor kappaB ligand. Blood. 2001;97(11):3349-3353. 21. Silbermann R, Bolzoni M, Storti P, et al. Bone marrow monocyte-/macrophage-derived activin A mediates the osteoclastogenic effect of IL-3 in multiple myeloma. Leukemia. 2014;28(4):951-954. 22. Lee HW, Choi HJ, Ha SJ, Lee KT, Kwon YG. Recruitment of monocytes/macrophages in different tumor microenvironments. Biochim Biophys Acta. 2013;1835(2):170179. 23. Belge KU, Dayyani F, Horelt A, et al. The proinflammatory CD14+CD16+DR++ monocytes are a major source of TNF. J Immunol. 2002;168(7):3536-3542. 24. Ziegler-Heitbrock L. The CD14+ CD16+ blood monocytes: their role in infection and inflammation. J Leukoc Biol. 2007;81(3):584592. 25. Chiu YG, Shao T, Feng C, et al. CD16 (FcRgammaIII) as a potential marker of osteoclast precursors in psoriatic arthritis. Arthritis Res Ther. 2010;12(1):R14. 26. Sponaas AM, Moen SH, Liabakk NB, et al. The proportion of CD16(+)CD14(dim) monocytes increases with tumor cell load in bone marrow of patients with multiple myeloma. Immun Inflamm Dis. 2015;3(2): 94-102.

784

27. Petitprez V, Royer B, Desoutter J, et al. CD14+ CD16+ monocytes rather than CD14+ CD51/61+ monocytes are a potential cytological marker of circulating osteoclast precursors in multiple myeloma. A preliminary study. Int L Lab Hematol. 2015;37(1):29-35. 28. Rajkumar SV, Dimopoulos MA, Palumbo A, et al. International Myeloma Working Group updated criteria for the diagnosis of multiple myeloma. Lancet Oncol. 2014;15 (12):e538-548. 29. Palma BD, Guasco D, Pedrazzoni M, et al. Osteolytic lesions, cytogenetic features and bone marrow levels of cytokines and chemokines in multiple myeloma patients: role of chemokine (C-C motif) ligand 20. Leukemia. 2016;30(2):409-416. 30. Storti P, Donofrio G, Colla S, et al. HOXB7 expression by myeloma cells regulates their pro-angiogenic properties in multiple myeloma patients. Leukemia. 2011;25(3):527-537. 31. Ramnaraine ML, Mathews WE, Clohisy DR. Lentivirus transduction of human osteoclast precursor cells and differentiation into functional osteoclasts. Bone. 2012;50(1):97-103. 32. Markham A. Elotuzumab: first global approval. Drugs. 2016;76(3):397-403. 33. Brenne AT, Ro TB, Waage A, Sundan A, Borset M, Hjorth-Hansen H. Interleukin-21 is a growth and survival factor for human myeloma cells. Blood. 2002;99(10):37563762. 34. Menoret E, Maiga S, Descamps G, et al. IL21 stimulates human myeloma cell growth through an autocrine IGF-1 loop. J Immunol. 2008;181(10):6837-6842. 35. Kwok SK, Cho ML, Park MK, et al. Interleukin-21 promotes osteoclastogenesis in humans with rheumatoid arthritis and in mice with collagen-induced arthritis. Arthritis Rheum. 2012;64(3):740-751. 36. Kim KW, Kim HR, Kim BM, Cho ML, Lee SH. Th17 Cytokines regulate osteoclastogenesis in rheumatoid arthritis. Am J Pathol. 2015;185(11):3011-3024. 37. Mehta DS, Wurster AL, Grusby MJ. Biology of IL-21 and the IL-21 receptor. Immunol Rev. 2004;202:84-95. 38. Ma J, Ma D, Ji C. The role of IL-21 in hematological malignancies. Cytokine. 2011;56 (2):133-139. 39. Millet P, Vachharajani V, McPhail L, Yoza B, McCall CE. GAPDH binding to TNF-alpha mRNA contributes to posttranscriptional

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

repression in monocytes: a novel mechanism of communication between inflammation and metabolism. J Immunol. 2016;196 (6):2541-2551. Zhan F, Hardin J, Kordsmeier B, et al. Global gene expression profiling of multiple myeloma, monoclonal gammopathy of undetermined significance, and normal bone marrow plasma cells. Blood. 2002;99(5):17451757. Davies FE, Dring AM, Li C, et al. Insights into the multistep transformation of MGUS to myeloma using microarray expression analysis. Blood. 2003;102(13):4504-4511. Tian E, Zhan F, Walker R, et al. The role of the Wnt-signaling antagonist DKK1 in the development of osteolytic lesions in multiple myeloma. N Engl J Med. 2003;349 (26):2483-2494. Lichtenstein A, Berenson J, Norman D, Chang MP, Carlile A. Production of cytokines by bone marrow cells obtained from patients with multiple myeloma. Blood. 1989;74(4):1266-1273. Portier M, Zhang XG, Ursule E, et al. Cytokine gene expression in human multiple myeloma. Br J Haematol. 1993;85(3): 514-520. Caprioli F, Sarra M, Caruso R, et al. Autocrine regulation of IL-21 production in human T lymphocytes. J Immunol. 2008;180(3):1800-1807. Asao H, Okuyama C, Kumaki S, et al. Cutting edge: the common gamma-chain is an indispensable subunit of the IL-21 receptor complex. J Immunol. 2001;167(1):1-5. Zeng R, Spolski R, Casas E, Zhu W, Levy DE, Leonard WJ. The molecular basis of IL21-mediated proliferation. Blood. 2007;109 (10):4135-4142. LaBranche TP, Jesson MI, Radi ZA, et al. JAK inhibition with tofacitinib suppresses arthritic joint structural damage through decreased RANKL production. Arthritis Rheum. 2012;64(11):3531-3542. Giuliani N, Lisignoli G, Colla S, et al. CCchemokine ligand 20/macrophage inflammatory protein-3alpha and CC-chemokine receptor 6 are overexpressed in myeloma microenvironment related to osteolytic bone lesions. Cancer Res. 2008;68(16):6840-6850. Tucci M, Stucci S, Savonarola A, et al. Immature dendritic cells in multiple myeloma are prone to osteoclast-like differentiation through interleukin-17A stimulation. Br J Haematol. 2013;161(6):821-831.

haematologica | 2017; 102(4)


ARTICLE

Plasma Cell Disorders

Serum B-cell maturation antigen: a novel biomarker to predict outcomes for multiple myeloma patients

EUROPEAN HEMATOLOGY ASSOCIATION

Ferrata Storti Foundation

Michael Ghermezi,1 Mingjie Li,1 Suzie Vardanyan,1 Nika Manik Harutyunyan,1 Jillian Gottlieb,1 Ariana Berenson,1 Tanya M. Spektor,2 Claudia Andreu-Vieyra,2 Sophia Petraki,2 Eric Sanchez,1 Kyle Udd,1 Cathy S. Wang,1 Regina A. Swift,3 Haiming Chen1 and James R. Berenson1,2,3

Institute for Myeloma & Bone Cancer Research, West Hollywood, CA; Oncotherapeutics, West Hollywood, CA and 3James R. Berenson, MD, Inc., West Hollywood, CA, USA 1 2

Haematologica 2017 Volume 102(4):785-795

ABSTRACT

B

-cell maturation antigen is expressed on plasma cells. In this study, we have identified serum B-cell maturation antigen as a novel biomarker that can monitor and predict outcomes for multiple myeloma patients. Compared to healthy donors, patients with multiple myeloma showed elevated serum B-cell maturation antigen levels (P<0.0001). Serum B-cell maturation antigen levels correlated with the proportion of plasma cells in bone marrow biopsies (Spearman’s rho = 0.710; P<0.001), clinical status (complete response vs. partial response, P=0.0374; complete response vs. progressive disease, P<0.0001), and tracked with changes in M-protein levels. Among patients with non-secretory disease, serum B-cell maturation antigen levels correlated with bone marrow plasma cell levels and findings from positron emission tomography scans. Kaplan-Meier analysis demonstrated that serum B-cell maturation antigen levels above the median levels were predictive of a shorter progression-free survival (P=0.0006) and overall survival (P=0.0108) among multiple myeloma patients (n=243). Specifically, patients with serum Bcell maturation antigen levels above the median level at the time of starting front-line (P=0.0043) or a new salvage therapy (P=0.0044) were found to have shorter progression-free survival. Importantly, serum B-cell maturation antigen levels did not show any dependence on renal function and maintained independent significance when tested against other known prognostic markers for multiple myeloma such as age, serum β2 microglobulin, hemoglobin, and bone disease. These data identify serum B-cell maturation antigen as a new biomarker to manage multiple myeloma patients.

Introduction Multiple myeloma (MM) is a bone marrow (BM)-based B-cell malignancy of terminally differentiated plasma cells.1-3 These clonal cells produce excessive amounts of monoclonal immunoglobulins (Ig).3 The clinical course of MM patients is quite variable and, with the currently available tools, predicting individual patient outcomes is a difficult task. The introduction of several new therapeutic agents for MM patients has resulted in a dramatic increase in the number of effective combination therapies and led to a marked improvement in their median overall survival (OS).4 More recently, therapeutic options for MM patients have expanded to include immune-based approaches.5 Unfortunately, the methods for evaluating MM patients’ disease status have not kept pace with this expanding profile. Thus, developing more effective methods to haematologica | 2017; 102(4)

Correspondence: jberenson@imbcr.org

Received: June 17, 2016. Accepted: December 22, 2016. Pre-published: December 29, 2016. doi:10.3324/haematol.2016.150896 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/785 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

785


M. Ghermezi et al.

characterize and follow these patients is becoming increasingly important. The Durie-Salmon classification system is commonly used to determine MM patients’ outcomes.6 Though a correlation between disease stage and length of survival was demonstrated, a number of the parameters used in this staging system were shown to have considerable shortcomings. The number of lytic lesions identified is subject to the radiologist’s interpretation, creating inconsistencies in the determination of stage.7 Elevated serum creatinine and reduced hemoglobin levels have many different etiologies, and may not necessarily be related to the patient’s MM.8,9 Serum beta 2 microglobulin (β2M) has also been used to help stratify myeloma patients and predict OS.10,11 Its utility is limited, however, in the presence of renal failure from any cause, since β2M cannot be effectively cleared by the kidneys. With compromised renal function, serum β2M levels remain elevated and confound the interpretation of its test results. Predicting patients’ outcomes is no longer feasible under these conditions.12,13 Although serum free light chain (SFLC) levels also have a rapid turnover, their reliability as an early determinant of response has been less than optimal.14,15 Thus, establishing more effective prognostic indicators remains an area of interest. The International Staging System (ISS) has largely replaced the Durie-Salmon staging system to predict OS for MM patients, relying on serum β2M and albumin levels at the time of diagnosis.16 However, Bataille et al. have indicated that ISS staging may, in fact, be a rather potent staging system for “pathological aging” rather than a specific MM staging system.17 Consistent with this, recent studies have shown that ISS is not consistent in predicting outcomes in the era of novel targets.18,19 Our recent findings have also demonstrated that there were no significant differences in OS between patients with different ISS stages.20 Besides the problems with using serum β2M levels mentioned above, the other component of the ISS staging system, serum albumin, is similarly subject to numerous limitations in that its levels are influenced by many factors unrelated to MM, including poor nutrition, acute or chronic inflammation, and loss of albumin via gastrointestinal or renal disorders.21 A variety of other prognostic markers and indicators have also been evaluated, including plasma cell labeling index, C-reactive protein, plasmablast morphology, cytogenetics, and BM angiogenesis.10,22,23 Combining these factors with genetic markers based on cytogenetics, fluorescent in situ hybridization and gene expression profiling from BM specimens have been used to assign different risk categories for MM patients.24-27 However, in addition to the hardships associated with undergoing invasive procedures to obtain suitable material, these tests can be both quite costly and lead to inconsistent results depending upon the quality and the amount of malignant tumor cells obtained. Thus, there remains much confusion regarding which factors and methods can be effectively used to predict outcomes for MM patients. The levels of several other serum proteins, such as soluble interleukin (sIL)-6, syndecan-1 and sclerostin have also been explored for MM patients and their usefulness as prognostic markers evaluated. Studies have shown a correlation between high levels of sIL-6 and progressive disease (PD).28,29 sIL-6 levels have also been used to distinguish monoclonal gammopathy of undetermined significance (MGUS) from MM.27 Although sIL-6 is not directly 786

produced by MM cells, its levels correlate with OS for MM patients. However, sIL-6 is only present in a small minority of MM patients, limiting its usefulness as a prognostic marker.28,29 Syndecan-1 and sclerostin levels have been determined in MM patients. Both proteins accumulate in the BM and their levels are primarily reflective of myeloma-related bone disease. Serum syndecan-1 levels are elevated in MM patients and correlated with poor prognosis, whereas serum sclerostin levels have not been found to be significantly different from that of a control group.30-33 There are also significant limitations with respect to the tests that are currently used to monitor these patients’ course of disease. Changes in M-protein levels from results of protein electrophoresis, measured in serum, urine or both are widely accepted as markers of disease status.4,14,23 However, measurement of M-protein can be unreliable, especially given that its levels are determined subjectively. Furthermore, Katzmann et al. have suggested that the 24-hour urine collection be eliminated as a test to monitor MM patients due to the advent of the similarly effective SFLC assay.34 The measurements of specific paraprotein types, such as IgA can also be problematic when assessed using an electrophoretic assay making their levels unreliable to monitor disease status.23,34 Other traditional methods of following MM include the quantitative evaluation of Igs, but this method also has its limitations. Differences in reagents over time lead to variations in these results, and the levels do not specifically measure only the monoclonal antibody that can be problematic especially among patients with low levels of paraprotein. The HevyLite assay has been used to help identify the involved Ig, but it still does not specifically identify the paraprotein within all of the matching isotype pairs.35 Notably, the serum half-life of Ig is also long,36,37 so that real-time measurements do not necessarily reflect current disease activity. More recently, the measurement of SFLC levels has been used; its levels and ratios and differences between the involved and uninvolved light chains are useful for monitoring and predicting outcomes for MM patients.34,38 The rapid turnover of these levels in the blood also allows for a more current indication of tumor load than measurement of conventional M-protein levels. However, SFLCs are not always elevated among patients with active disease and previous studies have reported a high degree of variability from results of the test especially during the first few weeks following initiation of a new therapy and among patients with renal impairment.15,34,38 A major drawback of all of the currently used tests to follow the course of disease is their inability to determine changes in clinical status among MM patients with nonsecretory disease (NSD).39,40 Currently, these individuals are treated by following an expensive and invasive paradigm involving frequent BM biopsies with variable results depending upon the site where the sample is obtained and positron emission tomography (PET) scans.40 B-cell maturation antigen (BCMA), a member of the tumor necrosis factor receptor family, is another protein whose function has been implicated in B-cell malignancies. It is expressed on the cell surface of mature and malignant B lymphocytes41,42 and is known to bind B-lymphocyte stimulator (BLyS), also known as B cell-activating factor (BAFF), a protein which plays a significant role in the growth and survival of MM cells.43 We have recently identified BCMA in serum and shown its levels are higher haematologica | 2017; 102(4)


sBCMA is a novel multiple myeloma biomarker

Table 1. Patients’ demographics.

All patient samples IgA patients Number of subjects Age (years)* Sex (male/female) Race (W/AA^/Asian/Other) ISS (1/2/3/n/a) β2 microglobulin (mg/L)* IgA (mg/dL)* IgG (mg/dL)* IgM (mg/dL)* Follow up from Dx (months)**

243 66 (35-89) 141 / 102 209 / 7 / 11 / 16 85 / 42 / 54 / 62 3.5 (1.0-26.4) 41 (3-12400) 1074.5 (20-9450) 24 (2-1190) 51 (3-367)

48 67 (45-83) 28 / 20 38 / 0 / 4 / 6 20 / 7 / 13 / 8 3.8 (1.3-26.4) 1180 (7-12400) 446 (34-1300) 20 (2-191) 59 (3-198)

IgG patients

IgM patients

Light chain patients

Non-secretory patients

156 65 (35-89) 92 / 64 139 / 4 / 6 / 7 57 / 29 / 29 / 41 3.4 (1.0-22.2) 26 (3-457) 1820 (219-9450) 41 (4-374) 51 (3-367)

3 77 (71-83) 2/1 3/0/0/0 0/0/1/2 4.1 (2.6-6.7) 210 (7-248) 80 (20-1010) 921 (21-1190) 60 (58-61)

33 65 (50-86) 18 / 15 26 / 3 / 1 / 3 7 / 5 / 11 / 10 4.3 (1.0-19.2) 52 (4-408) 486 (60-1330) 24 (4-113) 32 (4-215)

3 54 (41-67) 1/2 3/0/0/0 1/1/0/1 2.4 (2.4-2.4) 43 (25-61) 655.5 (579-732) 86 (21-51) 11 (11-53)

*Mean; **Median; ^AA African American; Dx: diagnosis.

among patients with monoclonal gammopathies.44 In mice, studies have shown that soluble BCMA has a halflife of approximately 24-36 hours45 which is much faster than the turnover rates for IgG (21 days) and IgA (7 days),36,37 suggesting that soluble BCMA levels can be used to evaluate the effect of a given treatment much more rapidly than through measurement of monoclonal antibody levels. In the current preliminary study, we determined the diagnostic, prognostic and monitoring values of serum (s)BCMA levels in a population of MM patients including patients with non-secretory disease (NSD) and also compared it to other known independent prognostic and monitoring markers.

Methods Patients Serum was collected from 243 MM patients followed in a single clinic that specialized in the treatment of MM (James R. Berenson, MD, Inc.). All patients provided informed consent in accordance with local institutional review board requirements and the Declaration of Helsinki. The International Myeloma Working Group Uniform Response Criteria (IMWG-URC) was used to determine the patient’s clinical status.7

Assessment of serum BCMA Frozen serum samples were thawed and diluted 1:500. An enzyme-linked immunosorbent assay (ELISA) was used to determine sBCMA levels according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN, USA; catalogue # DY193E) and as previously published.44 sBCMA levels were represented as the mean of triplicate samples for each specimen. Its levels were compared with the proportion of plasma cells in BM biopsy specimens. sBCMA was compared to the patients’ current clinical status and changes in M-protein levels during their course of disease. In addition, for patients with NSD, sBCMA was correlated with changes in their PET scan findings and plasma cell involvement in their BM.

Statistical analysis The percentage of plasma cell infiltration as measured by bone marrow biopsy was scatter plotted against serum BCMA and a haematologica | 2017; 102(4)

line of best fit was regressed through the data using an exponential function. To correlate the two, sBCMA levels were log10 transformed and the results were compared using Spearman’s correlation coefficient. Mann-Whitney U-test was used to compare sBCMA levels of smoldering MM or active MM patients to healthy controls and to compare sBCMA levels among patients with differences in clinical status. An ROC curve threshold analysis was used to determine the sensitivity and specificity of sBCMA when used as a tool to determine presence or absence of active disease requiring therapy. Kaplan-Meier survival analysis and log-rank comparison tests were used to determine the associations between sBCMA levels and OS and PFS. OS was evaluated from the time of initial sBCMA measurement to date of last follow up or death from any cause. PFS was evaluated from the time of initial sBCMA measurement to the first day of disease progression or death from any cause. Proportional Hazard Regression Analysis was utilized to determine the predictive ability of sBCMA on OS and other prognostic factors, including age, serum creatinine, serum hemoglobin, and ISS staging. One-way univariate analysis of variance was performed to study the relationship of sBCMA with bone disease status, based on presence of osteopenia, osteolytic lesions and/or fractures40 as well as the patient’s clinical status. A multivariate analysis of sBCMA with the other covariates including age, serum β2M, hemoglobin, creatinine, and bone disease was also performed to analyze the relationship of sBCMA levels with other covariates. All of the statistical analysis was performed using JMP Pro for Windows by SAS, GraphPad Prism 4 (San Diego, CA, USA) and R. P<0.05 was considered statistically significant

Results Patients Multiple myeloma patients’ characteristics (n=243) at baseline are shown in Table 1. The Ig isotype was found to be IgA, IgG, and IgM in 48 (19.8%), 156 (64.4%), and 3 (1.2%) patients, respectively, whereas 33 (16.6%) showed the absence of a serum monoclonal heavy chain-containing Ig component and 3 were non-secretory patients. The median serum β2M level was 3.5 mg/L (range 1.0-26.4 mg/L). According to ISS criteria, 85 (35.1%), 42 (17.3%) 787


M. Ghermezi et al.

and 54 (22.3%) patients had stage I, II and III disease at diagnosis, respectively, and 61 patients (25.3%) could not be classified. The median follow-up time from the time of their diagnosis for all MM patients, those without a detectable serum M-protein, and NSD were 51 months (range 3-367 months), 32 months (range 4-215 months) and 11 months (range 11-53 months), respectively.

A sBCMA

Serum BCMA levels correlate with bone marrow plasma cell involvement and are elevated in MM patients Among 57 patients with MM, the percentage of plasma cells in BM biopsy specimens was compared to their sBCMA levels, and found to correlate with one another (Spearman’s r=0.710; P<0.001) (Figure 1A and B). sBCMA levels were evaluated for 43 age-matched healthy donors, 46 patients with smoldering MM and 44 patients with untreated active MM (Figure 2), and compared using the Mann-Whitney U-test. Results demonstrated that sBCMA levels were higher among smoldering and untreated active MM patients than healthy donors (P<0.0001). Specifically, healthy donors had a median sBCMA of 36.8 ng/mL whereas SMM patients had higher levels (median 88.9 ng/mL; P<0.0001) and patients with active untreated MM had the highest levels (median 505.9 ng/mL, P<0.0001 compared to healthy subjects and P<0.0001 compared to SMM patients) (Figure 2A). In addition, a threshold analysis was performed using a Receiver Operator Characteristic (ROC) curve. We compared the sBCMA levels of patients diagnosed as age-matched healthy donors or patients with smoldering MM (n=89) and the sBCMA levels of patients diagnosed with active, untreated MM (n=44). Patients who were part of the healthy donor and smoldering MM group were assigned a value of 0 and were therefore treated as a negative result. Patients who were part of the active, untreated MM group were assigned a 1 and were therefore treated as a positive result. The ROC curve analyzed the optimal level of sBCMA at which the number of false positives and false negatives were minimized. The analysis determined that the optimal threshold level of sBCMA for indication of active, untreated MM was 107.6 ng/mL [sensitivity (i.e. detection of active untreated MM cases) 93.2% (95%CI: 84%100%); specificity (i.e. identifying only cases of active untreated MM above the threshold) 76.4% (95%CI: 67.4%-85.3%); area under curve 0.8805] (Figure 2B).

Serum BCMA levels correlate with M-protein levels sBCMA levels were correlated with the levels of conventional serum protein markers and involved SFLCs during the course of disease (Figure 3 and Online Supplementary Table S1). Figure 3A demonstrates changes in conventional M-protein levels closely matched the changes in sBCMA levels in 2 representative MM patients out of 44 MM patients (Online Supplementary Table S1) whose sBCMA was evaluated during their course of disease. Consistent with the M-protein findings, sBCMA levels also correlated with the levels of the involved SFLC levels. Two examples, Patients 1547 and 2116, are shown in Figure 3B.

Serum BCMA levels among patients with non-secretory disease To determine whether sBCMA can be used to monitor MM patients with NSD, sBCMA levels were compared 788

B sBCMA

Figure 1. Correlation of the proportion of plasma cells in bone marrow (BM) biopsy specimens with serum B-cell maturation antigen (sBCMA) levels in multiple myeloma (MM) patients. (A) Scatter plot showing correlation between the percentage of plasma cells in BM biopsy specimens and sBCMA levels in 57 MM patients was generated using GraphPad Prism 4. A regression line was generated using exponential growth model (Y=Start*exp [K*X]) with best-fit values (START=73.88; K=0.03404). (B) Scatter plot of Log10 transformation of sBCMA levels showing linearized correlation: Spearman correlation assessment (P<0.0001; Spearman’s rho=0.710). A regression line was generated using first order polynomial (Y=intercept +slope[X]; equation Y=0.01533[X] + 1.674; R2=0.5125.

with changes in patients’ PET scans and BM findings. NSD patients showed changes in sBCMA levels that correlated with changes in their PET scan and the percentage of BM plasma cells during their course of disease (Figure 4).

Analysis of serum BCMA levels and MM patients’ response to treatment and clinical status One hundred and sixty-four MM patients were grouped either based on the quality of response (Figure 5A) as complete response (CR) [n=23 (14%)]; partial response (PR) and minor response (MR) [n=46 (28%)]; or ND [n=95 (58%)], or by the IMWG-URC (Figure 5B) as CR [n=23 (14%)]; PR [n=33 (20%)]; less than PR [n=42 (26%)]; or PD [n=66 (40%)], and their sBCMA levels were compared. sBCMA levels showed a high correlation with these patients’ current clinical status. Using the quality of response (Figure 5A), patients with CR had markedly lower sBCMA levels (median 38.9 ng/mL) than those with a response less than CR (PR or MR; median 99.7 ng/mL; P=0.0045) or ND (median 195.3 ng/mL; P<0.0001) (Figure 5A). Notably, the median sBCMA level among those patients in CR was similar to healthy subjects (median 36.8 ng/mL) (Figure 2). When patients were grouped based on IMWG-URC haematologica | 2017; 102(4)


sBCMA is a novel multiple myeloma biomarker

A

B

Age Matched Healthy Donors Smoldering MM

Active Untreated MM

Figure 2. Serum B-cell maturation antigen (sBCMA) levels are elevated in multiple myeloma (MM) patients. (A) Specifically, 43 age-matched healthy donors (●) had significantly lower sBCMA levels (median 36.8 ng/mL) than 46 patients with smoldering MM (■) (median BCMA 88.9 ng/mL; P<0.0001) and 44 patients with active MM (▲) prior to any treatment (median BCMA 505.9 ng/mL; P<0.0001). (B) A threshold determination performed by ROC curve analysis indicated that the optimal threshold of sBCMA level to compare patients diagnosed as age-matched healthy donors or patients with smoldering MM (n=89) and patients diagnosed with active, untreated MM (n=44) was 107.6 ng/mL [sensitivity 93.2% (95%CI: 84%-100%); specificity 76.4% (95%CI: 67.4%-85.3%); area under curve 0.8805]. For our analysis “sensitivity” refers to the ability of sBCMA levels above a chosen threshold to be able to detect active untreated MM cases and “specificity” to sBCMA levels above a chosen threshold to identify only cases of active untreated MM above the threshold.

A

B

Figure 3. Serum B-cell maturation antigen (sBCMA) levels correlate with changes in M-protein and serum free light chain (SFLC) levels in individual multiple myeloma (MM) patients. (A) Analysis of sBCMA (▲) versus M-protein (■) levels during the course of disease in 4 MM patients among 44 patients analyzed. (B) Analysis of sBCMA (▲) versus SFLC (■) in 2 representative MM patients during the course of their disease.

haematologica | 2017; 102(4)

789


M. Ghermezi et al.

(Figure 5B), sBCMA levels were significantly lower among patients in CR as compared to all other groups (CR vs. PR, P=0.0374; CR vs. less than PR, P=0.0002; CR vs. PD, P<0.0001). In addition, levels of sBCMA were significantly higher among PD patients as compared to all other groups (PD vs. PR, P<0.0001; PD vs. less than PR, P=0.0005).

A

Serum BCMA levels predict progression-free survival and overall survival Figure 6A shows PFS for 184 MM patients whose sBCMA was determined just prior to the start of a new treatment regimen. PFS was longer (P=0.0006) for patients with sBCMA levels below the median (326.4 ng/mL)

Non-secretory patient 1977

Non-secretory patient 2460

B

Non-secretory patient 2482

C

Figure 4. Serum B-cell maturation antigen (sBCMA) levels during the course of disease among 3 patients with non-secretory disease (NSD). (A) non-secretory Patient #1977, (B) nonsecretory Patient #2460 and (C) non-secretory Patient #2482. Notably, the 3 patients with NSD showed a correlation between changes in sBCMA levels and their clinical status as reflected by positron emission tomography (PET) scan and bone marrow findings during their disease course.

790

haematologica | 2017; 102(4)


sBCMA is a novel multiple myeloma biomarker

when compared with those with levels above the median (median 9.0 vs. 3.6 months) (Figure 6A, left). When comparing patients with sBCMA levels in the highest quartile (>971.0 ng/mL) to those with levels in the lower three quartiles, a marked difference in PFS was observed (P=0.0002; median 3.1 vs. 7.2 months) (Figure 6A, right). Prior to initiation of their first treatment (n=38) (Figure 6B) or a salvage therapy (n=146) (Figure 6C), the median PFS was longer among patients with sBCMA levels below the median (Figure 6B and C, left) or in the lowest three quartiles (Figure 6B and C, right). In the subgroup analysis of 99 MM patients who achieved MR or less, patients with baseline sBCMA levels above the median (261.7 ng/mL) or in the highest quartile (> 852.2 ng/mL) both achieved shorter PFS (Figure 6D). Overall survival was longer among those with sBCMA levels below the median compared with those above (median 155 vs. 96 months; P=0.0108) (Figure 7A). Similarly, those in the lower three quartiles had a prolonged OS compared with those in the highest quartile

A

B

(P=0.0075) (Figure 7B). To estimate the relative risk of an event and its 95% Confidence Interval, a proportionalhazard regression model was used. It showed that sBCMA levels significantly correlated with OS, whereas there was a lack of correlation with age, creatinine, hemoglobin, or ISS and/or stage (Online Supplementary Figure S3).

Serum BCMA levels are independent of other prognostic markers Statistical comparison of sBCMA levels from MM patients with different clinical outcomes was performed to show that the serum levels of BCMA were specific to the clinical status of the response. The entire data set was grouped into different sub-cohorts based on the current clinical status. sBCMA level of patients in each group was compared using Dunnett’s test, where patients belonging to the CR group were used as the control group. As shown in Online Supplementary Figure S1, sBCMA levels of cohorts grouped under PD and untreated patients showed a significant difference (P<0.05) from the control group whereas other sub-cohorts such as MR or PR did not. This analysis demonstrated that sBCMA levels among patients with PD and untreated patients were significantly different and could be used as an indicator of the disease status. The independence of sBCMA levels from other prognostic markers such as age, serum β2M, hemoglobin, and creatinine was investigated using multivariate analysis and a proportional hazards regression model. As shown in Online Supplementary Table S2 and Online Supplementary Figure S2, sBCMA levels were independent of other prognostic variables and did not display any substantial correlation with the other prognostic markers. Lastly, the entire data set was grouped into two cohorts based on the presence of bone disease. The sBCMA level of patients with bone disease (group 1) was compared to those without bone disease (group 0), using Dunnett’s test. As shown in Online Supplementary Figure S3, sBCMA levels of group 1 did not show a difference from group 0 (P=0.45). This analysis demonstrated that sBCMA levels among patients with and without bone disease did not differ from one another; indicating that sBCMA levels are independent of an MM patient’s bone disease status.

Discussion

Figure 5. Analysis of serum B-cell maturation antigen (sBCMA) levels and multiple myeloma (MM) patients’ response to treatment and clinical status. sBCMA levels correlated with patients’ clinical status at the time of its determination. Specifically, (A) patients with complete response (CR) (●) had significantly lower sBCMA levels (median 38.6 ng/mL) than those with partial response (PR) or minor response (MR) (■) (median 99.7 ng/mL; P=0.0045) and non-responsive disease (ND) (▲), including those with either stable or progressive disease (PD) (median 195.3 ng/mL; P<0.0001). (B) Using International Myeloma Working Group (IMWG) criteria, patients with CR (●) had significantly lower sBCMA levels (median 38.6 ng/mL) than those with PR (▲; P=0.0374; median 81.7 mg/mL), less than PR (MR+SD; o; P=0.0002; median 100.6 mg/mL) and PD ( ; P<0.0001; median 301.4 mg/mL). Total number of patients=164; *P versus CR.

haematologica | 2017; 102(4)

The results from this study demonstrate that the measurement of sBCMA levels provides important diagnostic, prognostic and monitoring information for patients with MM. sBCMA levels were highly correlated with the percentage of plasma cell infiltration in BM biopsies from MM patients. Next, we demonstrated that sBCMA levels were elevated among patients with MM compared with healthy individuals, and MM patients with smoldering disease showed lower sBCMA levels than those with active disease. A strong correlation between patients’ current clinical status and sBCMA level was also identified. Specifically, patients who were in CR displayed lower sBCMA levels than those who were in MR or PR; those with ND had the highest levels. The ability of sBCMA to predict clinical outcomes, PFS and OS, was also examined. Patients with lower BCMA levels at the time of starting initial or a new salvage therapy were both found to have a much longer 791


M. Ghermezi et al.

A

B

C

D

Figure 6. Correlation of serum B-cell maturation antigen (sBCMA) levels with progression-free survival (PFS) in multiple myeloma (MM) patients. MM patients were assessed according to their sBCMA levels being above or below the median level (left) or in the top quartile or bottom three quartiles (right). Kaplan-Meier analysis was performed to analyze PFS of 184 MM patients before start of a new treatment (A) with a baseline sBCMA level above or below the median level of 326.4 ng/mL (left) and for those with levels in the lowest three quartiles (range 14.3-970.9 ng/mL) or the highest quartile (≥ 971.0 ng/mL) (right). (B) Progression-free survival (PFS) of 38 MM patients before starting their front-line treatment according to whether their baseline sBCMA was above or below the median level of 430.5 ng/mL (left) and for those with levels in the lowest three quartiles (range 17.8-861.7 ng/mL) or highest quartile (≥ 861.8 ng/mL) (right). (C) PFS of 146 MM patients from time of starting a new salvage therapy according to whether their baseline sBCMA was above or below the median level of 281.9 ng/mL (left) and for those in the lowest three quartiles (range 14.3-962.3 ng/mL) or highest quartile (≥ 962.4 ng/mL) (right). (D) Among 99 MM patients who had a sBCMA determination just prior to starting a new treatment and achieved at least a minor response (MR), PFS was determined from the time of determination of their baseline sBCMA according to whether it was above or below a median of 261.7 ng/mL (left) and for those in the lowest three quartiles (range 17.8-852.1 ng/mL) or highest quartile (>852.2 ng/mL) (right).

792

haematologica | 2017; 102(4)


sBCMA is a novel multiple myeloma biomarker

PFS. OS was also longer for patients with sBCMA levels below the median or in the lowest three quartiles. Our results indicate that sBCMA represents a novel prognostic marker capable of predicting current disease status, PFS and OS for MM patients. To further evaluate sBCMA as an independent determinant of OS, we performed a multivariate analysis comparing sBCMA to other known prognostic factors linked to MM: age, β2M, hemoglobin, and creatinine. We also used a proportional-hazards regression model to examine the relative risk of events for each of these covariates. Notably, prognostic effect of sBCMA was found to be significant and independent of all tested covariates, including serum creatinine. Renal impairment commonly occurs in MM patients and remains a major disadvantage of utilizing renal dependent biomarkers such as serum β2M (as a prognostic indicator) and SFLC (as a laboratory test) to monitor disease.7,34 Similarly, the manifestations from bone involvement in MM patients can have negative clinical effects as well as shorten OS.46 In this study, we have demonstrated that sBCMA levels in MM patients were not influenced by patients’ bone disease status and, therefore, further validates its value as an MM-specific biomarker. To demonstrate that sBCMA is predictive of treatment benefit and can be used to monitor patients’ disease status, its levels were monitored through the disease course of 44 MM patients. Using M-protein and involved SFLC levels as standard reference monitoring tests, we demonstrated that changes in these patients’ levels correlated with those observed in their M-protein and involved SFLC levels. In addition, sBCMA was also shown to effectively monitor MM patients with NSD. The monitoring of these patients has been problematic, relying on invasive and expensive procedures including frequent BM aspirates and biopsies and radiological evaluation with tests such as PET CT scans. The availability of a reliable serum test should help considerably in their care and importantly provide the opportunity for NSD patients to be able to enroll in clinical trials for which they have been previously excluded. Although many other markers have been previously used to monitor these patients, sBCMA offers several potential advantages. Its rapid turnover in the blood (24-36 hours half life)45 compared to conventional Ig levels36,37 would potentially allow for early determination of changes in clinical status. Preliminary results from monitoring our patients suggest that this is the case (JR Berenson, 2017, unpublished observations). This should identify when a patient is not responding to their current therapy much quicker, allowing them to change to another therapy more rapidly. Another fortuitous byproduct of this would be to reduce a patient’s exposure to ineffective treatments and, as a result, prevent them from experiencing unnecessary side effects from their therapy. Moreover, although SFLC levels turn over rapidly, early data do not demonstrate consistent results with early evaluation for responses to new therapy. The variability of the results of this test, especially among patients with renal failure, makes this assay problematic for determining response status accurately.15,34 In addition, M-protein and SFLC levels cannot be used to follow disease progression for all MM patients, as some have, and others eventually develop NSD. Our results show that sBCMA can now be used to monitor these types of patients. We have recently shown that circulating BCMA has haematologica | 2017; 102(4)

A

B

Figure 7. Correlation of serum B-cell maturation antigen (sBCMA) levels with overall survival (OS) in 243 multiple myeloma (MM) patients. MM patients were assessed according to their sBCMA levels being above or below the median (A), or in the highest quartile compared with the bottom three quartiles (B). KaplanMeier analysis showed that OS was longer among patients with BCMA below the median (136.2 ng/mL), and was shorter in the highest quartile (> 470.1 ng/mL) compared with the other three quartiles (range 14.4-470.0 ng/mL).

other important roles. Specifically, the circulating protein has been implicated in the pathway that leads to immune deficiency in MM patients, a hallmark of the disease.45 The shed TNF receptor acts as a decoy receptor and binds Bcell ligands including BAFF preventing them from performing their normal function to stimulate late B-cell development and their production of normal antibodies. Many groups are developing antibodies to BCMA as a therapeutic approach to combat myeloma and other plasma cell malignancies.47,48 It will be interesting to see whether sBCMA levels will predict responsiveness to these targeted approaches. On the one hand, it may be possible that patients with high sBCMA will have more BCMA on their plasma cells making them better candidates for this targeted approach. Conversely, it is also possible that patients with high levels of circulating BCMA may derive less clinical benefit because these soluble proteins may bind to the targeted antibodies preventing them from attaching to their intended target, the malignant plasma cell. Patients with other late B-cell malignancies also show elevated levels of circulating BCMA, including those with chronic lymphocytic leukemia (CLL), lymphoma and Waldenstrom macroglobulinemia.49,50 These levels were 793


M. Ghermezi et al.

found to correlate with patients’ clinical status. In CLL patients, sBCMA correlates with standard prognostic factors including IgVH mutational status, ZAP-70 expression, chromosome 13 deletion, and white blood cell counts.50 sBCMA levels also can be used to monitor changes in CLL patients’ disease status and predict both their time to first treatment and OS. We have identified a specific serum protein, BCMA, as a novel independent marker for both monitoring and predicting outcomes for MM patients. We have shown that sBCMA is elevated in MM patients, and can be used to follow their disease status, PFS and OS. It also provides, for the first time, a non-invasive blood test to monitor the disease of MM patients with NSD. Additional studies are warranted to further explore the role of sBCMA to predict

References 1. Smith ML, Newland AC. Treatment of myeloma. QJM. 1999;92(1):11-16. 2. Morgan GJ. Advances in the biology and treatment of myeloma. Br J Haematol. 1999;105(Suppl 1):4-6. 3. Kastrinakis NG, Gorgoulis VG, Foukas PG, et al. Molecular aspects of multiple myeloma. Ann Oncol. 2000;11(10):1217-1228. 4. Kumar SK, Rajkumar SV, Dispenzieri A, et al. Improved survival in multiple myeloma and the impact of novel therapies. Blood. 2008;111(5):2516-2520. 5. van de Donk NW, Moreau P, Plesner T, et al. Clinical efficacy and management of monoclonal antibodies targeting CD38 and SLAMF7 in multiple myeloma. Blood. 2016;127(6):681-695. 6. Durie BG. Staging and kinetics of multiple myeloma. Semin Oncol. 1986;13(3):300309. 7. Durie BG, Harousseau JL, Miguel JS, et al. International uniform response criteria for multiple myeloma. Leukemia. 2006;20(9): 1467-1473. 8. Kleber M, Ihorst G, Deschler B, et al. Detection of renal impairment as one specific comorbidity factor in multiple myeloma: multicenter study in 198 consecutive patients. Eur J Haematol. 2009;83(6):519527. 9. Mittelman M. The implications of anemia in multiple myeloma. Clin Lymphoma. 2003;4(Suppl 1):S23-29. 10. Bataille R, Boccadoro M, Klein B, et al. Creactive protein and beta-2 microglobulin produce a simple and powerful myeloma staging system. Blood. 1992;80(3):733-737. 11. Turesson I, Abildgaard N, Ahlgren T, et al. Prognostic evaluation in multiple myeloma: an analysis of the impact of new prognostic factors. Br J Haematol. 1999; 106(4):10051012. 12. Cuzick J, Cooper EH, MacLennan IC. The prognostic value of serum beta 2 microglobulin compared with other presentation features in myelomatosis. Br J Cancer. 1985;52(1):1-6. 13. Merlini G, Waldenström JG, Jayakar SD. A new improved clinical staging system for multiple myeloma based on analysis of 123 treated patients. Blood. 1980;55(6):10111019. 14. Kyle RA, Gertz MA, Witzig TE, et al. Review of 1027 patients with newly diag-

794

15.

16.

17.

18.

19.

20.

21. 22.

23. 24.

25.

26.

outcomes for the plethora of new agents that are now being used to treat MM patients. Given that sBCMA helps orchestrate the immune deficient state in these patients,46 it will also be of interest to determine whether its levels predict responsiveness. It will also be of interest to see if it can be used to monitor patients undergoing the immunebased treatments that have recently become available and in clinical development, including research into BCMA itself.47,48 Acknowledgments Dr. Berenson is a director, employee and shareholder in OncoTracker, Inc. a company that is involved in the development of BCMA as a marker. Haiming Chen, Eric Sanchez, Mingjie Li, and Cathy Wang are shareholders in OncoTracker, Inc.

nosed multiple myeloma. Mayo Clin Proc. 2003;78:21-33. Jacobs JF, Tate JR, Merlini G. Is accuracy of serum free light chain measurement acheiveable? Clin Chem Lab Med. 2016; 54(6):1021-1030. Greipp PR, San Miguel J, Durie BG, et al. International staging system for multiple myeloma. J Clin Oncol. 2005;23(15):34123420. Bataille R, Annweiler C, Beauchet O. Multiple Myeloma International Staging System “Staging” or Simply “Aging” system? Clin Lymphoma Myeloma Leuk. 2013;13(6):635-637. Maltezas D, Dimopoulos MA, Katodritou I, et al. Re-evaluation of prognostic markers including staging, serum free light chains or their ratio and serum lactate dehydrogenase in multiple myeloma patients receiving novel agents. Hematol Oncol. 2013; 31(2):96-102. Majithia N, Rajkumar VS, Lacy MQ, et al. Outcomes of primary refractory multiple myeloma and the impact of novel therapies. Am J Hematol. 2015;90(11):981-985. Berenson A, Vardanyan S, David M, et al. Improved Clinical Outcomes for Multiple Myeloma Patients Treated at a Single Specialty Clinic. Ann Hematol. 2016 Dec 2. [Epub ahead of print] Arques S, Ambrosi P. Human serum albumin in the clinical syndrome of heart failure. J Card Fail. 2011;17(6):451-458. Greipp PR, Leong T, Bennett JM, et al. Plasmablastic morphology−an independent prognostic factor with clinical and laboratory correlates: Eastern Cooperative Oncology Group (ECOG) myeloma trial E9486 report by the ECOG Myeloma Laboratory Group. Blood. 1998; 91(7):2501-2507. Rajkumar SV, Greipp PR. Prognostic factors in multiple myeloma. Hematol Oncol Clin North Am. 1999;13(3):1295-1314. Avet-Loiseau H, Andree-Ashley LE, Moore D, et al. Molecular cytogenetic abnormalities in multiple myeloma and plasma cell leukemia measured using comparative genomic hybridization. Genes Chromosomes Cancer. 1997;19(2):124-133. Dewald GW, Kyle RA, Hicks GA, et al. The clinical significance of cytogenetic studies in 100 patients with multiple myeloma, plasma cell leukemia, or amyloidosis. Blood. 1985;66(2):380-390. Broyl A, Hose D, Lokhorst H, et al. Gene

27. 28.

29.

30.

31.

32.

33.

34.

35.

36. 37.

38.

expression profiling for molecular classification of multiple myeloma in newly diagnosed patients. Blood. 2010;116(14):25432553. Rajan AM, Rajkumar SV. Interpretation of cytogenetic results in multipe myeloma for clinical practice. Blood Cancer J. 2015;5:e365 Wierzbowska A, Urbanska-Rys H, Robak T. Circulating IL-6-type cytokines and sIL-6R in patients with multiple myeloma. Br J Haematol. 1999;105(2):412-419. Lauta VM. Interleukin-6 and the network of several cytokines in multiple myeloma: an overview of clinical and experimental data. Cytokine. 2001;16(3):79-86. Lovell R, Dunn JA, Begum G, et al. Soluble syndecan-1 level at diagnosis is an independent prognostic factor in multiple myeloma and the extent of fall from diagnosis to plateau predicts for overall survival. Br J Haematol. 2005;130(4):542-548. Brunetti G, Oranger A, Mori G, et al. Sclerostin is overexpressed by plasma cells from multiple myeloma patients. Ann NY Acad Sci. 2011;1237:19-23. Delgado-Calle J, Bellido T, Roodman GD. Role of osteocytes in multiple myeloma bone disease. Curr Opin Support Palliat Care. 2014;8(4):407-413. Kim JM, Lee JA, Cho IS, et al. Soluble syndecan-1 at diagnosis and during follow up of multiple myeloma: a single institution study. Korean J Hematol. 201;45(2):115-119. Katzmann JA, Dispenzieri A, Kyle RA, et al. Elimination of the need for urine studies in the screening algorithm for monoclonal gammopathies by using serum immunofixation and free light chain assays. Mayo Clin Proc. 2006;81:1575-1578. Harutyunyan NM, Vardanyan S, Ghermezi M, et al. Levels of uninvolved immunoglobulins predict clinical status and progression free survival for multiple myeloma patients. Br J Haematol. 2016; 174(1):81-87. Waldmann TA, Strober W. Metabolism of immunoglobulins. Prog Allergy. 1969; 13(1):1-110. Mariani G, Strober W. Immunoglobulin metabolism. In: Metzger H (ed.). Fc receptors and the action of antibodies. Am Soc Microbiol. 1990;94-177. Hansen CT, Pedersen PT, Nielsen LC, et al. Evaluation of the serum free light chain (sFLC) analysis in prediction of response in symptomatic multiple myeloma patients: rapid profound reduction in involved FLC

haematologica | 2017; 102(4)


sBCMA is a novel multiple myeloma biomarker

39. 40. 41.

42.

43.

predicts achievement of VGPR. Eur J Haematol. 2014;93(5):407-413. Dreicer R, Alexanian R. Non-secretory multiple myeloma. Am J Hematol. 1982;13(4): 313-318. Lonial S, Kaufman JL. Non-secretory myeloma: a clinicianâ&#x20AC;&#x2122;s guide. Oncology (Williston Park). 2013;27:924-928. Laabi Y, Gras MP, Brouet JC, et al. The BCMA gene, preferentially expressed during lymphoid maturation, is bidirectionally transcribed. Nucleic Acids Res. 1994; 22(7):11471154. Novak AJ, Darce JR, Arendt BK, et al. Expression of BCMA, TACI, and BAFF-R in multiple myeloma: a mechanism for growth and survival. Blood. 2004; 103(2):689-694. Tai YT, Li XF, Breitkreutz I, et al. Role of Bcell-activating factor in adhesion and growth

haematologica | 2017; 102(4)

44.

45.

46.

47.

of human multiple myeloma cells in the bone marrow microenvironment. Cancer Res. 2006;66(13):6675-6682. Sanchez E, Li M, Kitto A, et al. Serum B-cell maturation antigen is elevated in multiple myeloma and correlates with disease status and survival. Br J Haematol. 2012; 158 (6):727-738. Sanchez E, Gillespie A, Tang G, et al. Soluble B-cell maturation antigen mediates tumor induced immune deficiency in multiple myeloma. Clin Cancer Res. 2016; 22(13): 3383-3397. Melton LJ, Kyle RA, Achenbach SJ, et al. Fracture risk with multiple myeloma: a population-based study. J Bone Miner Res. 2005;20(3):487-493. Ryan MC, Hering M, Peckham D, et al. Antibody targeting of B-cell maturation anti-

gen on malignant plasma cells. Mol Cancer Ther. 2007;6(11):3009 3018. 48. Tai YT, Mayes PA, Acharya C, et al. Novel anti-B-cell maturation antigen antibodydrug conjugate (GSK2857916) selectively induces killing of multiple myeloma. Blood. 2014;123(20):3128-3138. 49. Vardanyan S, Meid K, Udd KA, et al. Serum levels of B-cell maturation antigen are elevated in Waldenstromâ&#x20AC;&#x2122;s macroglobulinemia patients and correlate with disease status and conventional M-protein and IgM levels. Blood. 2015;126:(Abstract 1778). 50. Udd KA, Rassenti LZ, David ME, et al. Plasma B-cell maturation antigen levels are elevated and correlate with disease activity in patients with chronic lymphocytic leukemia. Blood. 2015;126: (Abstract 2931).

795


ARTICLE EUROPEAN HEMATOLOGY ASSOCIATION

Cell Therapy & Immunotherapy

Ferrata Storti Foundation

Increased age-associated mortality risk in HLA-mismatched hematopoietic stem cell transplantation

Daniel Fürst,1,2 Dietger Niederwieser,3 Donald Bunjes,4 Eva M. Wagner,5 Martin Gramatzki,6 Gerald Wulf,7 Carlheinz R. Müller,8,9 Christine Neuchel,1,2 Chrysanthi Tsamadou,1,2 Hubert Schrezenmeier,1,2 and Joannis Mytilineos1,2,9

Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Transfusion Service, Baden Wuerttemberg – Hessen, Ulm; 2Institute of Transfusion Medicine, University of Ulm; 3Department of Hematology/Oncology, University of Leipzig; 4Department of Internal Medicine III, University of Ulm; 5 Department of Medicine III, Johannes Gutenberg-University Mainz; 6Division of Stem Cell Transplantation and Immunotherapy, 2nd Department of Medicine, University of Kiel; 7 Department of Hematology/Oncology, Georg-August-University Göttingen; 8ZKRD Zentrales Knochenmarkspender-Register für Deutschland (German National Bone Marrow Donor Registry), Ulm and 9DRST – German Registry for Stem Cell Transplantation, Ulm, Germany 1

Haematologica 2017 Volume 102(4):796-803

ABSTRACT

W

Correspondence: j.mytilineos@blutspende.de

Received: June 20, 2016. Accepted: December 28, 2016. Pre-published: January 5, 2017. doi:10.3324/haematol.2016.151340 Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/102/4/796 ©2017 Ferrata Storti Foundation Material published in Haematologica is covered by copyright. All rights are reserved to the Ferrata Storti Foundation. Use of published material is allowed under the following terms and conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode. Copies of published material are allowed for personal or internal use. Sharing published material for non-commercial purposes is subject to the following conditions: https://creativecommons.org/licenses/by-nc/4.0/legalcode, sect. 3. Reproducing and sharing published material for commercial purposes is not allowed without permission in writing from the publisher.

796

e investigated a possible interaction between age-associated risk and HLA-mismatch associated risk on prognosis in different age categories of recipients of unrelated hematopoietic stem cell transplants (HSCT) (n=3019). Patients over 55 years of age transplanted with 8/10 donors showed a mortality risk of 2.27 (CI 1.703.03, P<0.001) and 3.48 (CI 2.49-4.86, P<0.001) when compared to 10/10 matched patients in the same age group and to 10/10 matched patients aged 18-35 years, respectively. Compared to 10/10 matched transplantations within each age category, the Hazards Ratio for 8/10 matched transplantation was 1.14, 1.40 and 2.27 in patients aged 18-35 years, 3655 and above 55 years. Modeling age as continuous variable showed different levels of risk attributed to age at the time of transplantation [OS: 10/10: Hazards Ratio 1.015 (per life year); 9/10: Hazards Ratio: 1.019; 8/10: Hazards Ratio 1.026]. The interaction term was significant for 8/10 transplantations (P=0.009). Findings for disease-free survival and transplant-related mortality were similar. Statistical models were stratified for diagnosis and included clinically relevant predictors except cytomegalovirus status and Karnofsky performance status. The risk conferred by age at the time of transplantation varies according to the number of HLA-mismatches and leads to a disproportional increase in risk for elderly patients, particularly with double mismatched donors. Our findings highlight the importance of HLA-matching, especially in patients over 55 years of age, as HLA-mismatches are less well tolerated in these patients. The interaction between age-associated risk and HLA-mismatches should be considered in donor selection and in the risk assessment of elderly HSCT recipients.

Introduction Unrelated hematopoietic stem cell transplantation is a rapidly evolving field offering a curative therapy for various hematologic diseases. In particular, the proportions of older patients and patients transplanted with unrelated donors have increased over the last decade.1,2 One prerequisite was the introduction of reduced intensity conditioning regimens (RIC) as an alternative to myeloablative conditioning (MAC) in elderly patients as well as in patients with co-morbidities.3,4 There is haematologica | 2017; 102(4)


Age risk and HLA-matching Table 1. Patients' characteristics. Median age Number of patients Number of centers Diagnosis AML ALL AL CML CLL MDS NHL MM HLA-matching status 10/10 9/10 8/10 Ethnicity Caucasian Asian African Disease stage Early Intermediate Advanced Karnofsky performance score 80-100 <80 Data missing Conditioning regimen Myeloablative Reduced intensity GvHD prophylaxis CsA ± MTX ± other Tacrolimus ± other MMF ± other MTX ± other T-cell depletion Other Data missing ATG treatment Yes No Data missing Stem cell source BM PBSC Recipient-donor sex match male-male male-female female-male female-female Patient HLA-C KIR-ligand status C1C1 C1C2 C2C2 CMV status (patient-donor) neg-neg neg-pos pos-neg pos-pos Data missing Year of transplantation 1997-2003 2004-2007 2008-2011 Distribution of 8/10 mismatches Only HLA-class I MM HLA-Class I + class II MM Only HLA-class II MM

P

Age 18-35 years

Age 36-55 years

Age >55 years

Total n

27 529 28

48 1295 26

62 1195 25

n.a. 3019 29

187 (35.3) 178 (33.6) 13 (2.5) 59 (11.2) 2 (0.4) 37 (7.0) 49 (9.3) 4 (0.8)

393 (30.3) 142 (11.0) 66 (5.1) 111 (8.6) 60 (4.6) 201 (15.5) 184 (14.2) 138 (10.7)

344 (28.8) 55 (4.6) 95 (7.9) 31 (2.6) 74 (6.2) 342 (28.6) 145 (12.1) 109 (9.1)

924 375 174 201 136 580 378 251

<0.001

295 (55.8) 172 (32.5) 62 (11.7)

774 (59.8) 397 (30.7) 124 (9.6)

778 (65.1) 342 (28.6) 75 (6.3)

1847 911 261

<0.001

527 (99.6) 1 (0.2) 1 (0.2)

1289 (99.5) 4 (0.3) 2 (0.2)

1193 (99.8) 2 (0.2) 0 (0.0)

3009 7 3

n.s.

250 (47.3) 170 (32.1) 109 (20.6)

516 (39.8) 468 (36.1) 311 (24.0)

455 (38.1) 403 (33.7) 337 (28.2)

1221 1041 757

<0.001

235 (44.4) 16 (3.0) 278 (52.6)

732 (56.5) 31 (2.4) 532 (41.1)

856 (71.6) 67 (5.6) 272 (22.8)

1823 114 1082

<0.001

460 (87.0) 69 (13.0)

928 (71.7) 367 (28.3)

487 (40.8) 708 (59.2)

1875 1144

<0.001

231 (43.7) 22 (4.2) 7 (1.3) 2 (0.4) 16 (3.0) 10 (1.9) 241 (45.6)

670 (51.7) 72 (5.6) 10 (0.8) 6 (0.5) 22 (1.7) 11 (0.8) 504 (38.9)

720 (60.3) 79 (6.6) 6 (0.5) 7 (0.6) 21 (1.8) 17 (1.4) 345 (28.9)

1621 173 23 15 59 38 1090

<0.001

329 (62.2) 125 (23.6) 75 (14.2)

810 (62.5) 319 (24.6) 166 (12.8)

723 (60.5) 416 (34.8) 56 (4.7)

1862 860 297

<0.001

109 (20.6) 420 (79.4)

166 (12.8) 1129 (87.2)

50 (4.2) 1145 (95.8)

325 2694

<0.001

244 (46.1) 71 (13.4) 127 (24.0) 87 (16.4)

575 (44.4) 174 (13.4) 359 (27.7) 187 (14.4)

586 (49.0) 159 (13.3) 297 (24.9) 153 (12.8)

1405 404 783 427

n.s.

212 (40.1) 240 (45.4) 77 (14.6)

499 (38.5) 580 (44.8) 216 (16.7)

480 (40.2) 542 (45.4) 173 (14.5)

1191 1362 466

n.s.

112 (21.2) 44 (8.3) 87 (16.4) 57 (10.8) 229 (43.3)

312 (24.1) 91 (7.0) 213 (16.4) 242 (18.7) 437 (33.7)

241 (20.2) 82 (6.9) 301 (25.2) 373 (31.2) 198 (16.6)

665 217 601 672 864

<0.001

149 (28.2) 182 (34.4) 198 (37.4)

222 (17.1) 437 (33.7) 636 (49.1)

38 (3.2) 320 (26.8) 837 (70.0)

409 939 1671

<0.001

39 19 (12, 63.2) 4 (4, 100)

81 32 (18, 56.3) 11 (11, 100)

47 22 (16, 72.7) 6 (6, 100)

167 73 21

n.s

AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia; AL: unclassified acute leukemia; CML: chronic myeloid leukemia; CLL: chronic lymphocytic leukemia; MDS: myelodysplastic syndrome; NHL: non-Hodgkin lymphoma; MM: multiple myeloma; BM: bone marrow; PBSC: peripheral blood stem cells. Distribution of 8/10 mismatches section - MM: mismatch, for groups including HLA-Class II mismatches, the number and percentage of cases involving HLA-DQB1 mismatches are given in parentheses; n.s.: not significant.

haematologica | 2017; 102(4)

797


D. FĂźrst et al.

already a wealth of data showing that RIC is a safe and effective treatment form for patients previously not eligible for hematopoietic stem cell transplantation (HSCT).5,6 As a consequence, therapeutic schemes for elderly patients have been established which now include HSCT as treatment option in some clinical instances.7 Nevertheless, classical risk factors still apply, and while increasing age did not influence the incidence of acute or chronic graft-versus-host disease (GvHD),8 transplant-associated morbidity and mortality as well as disease relapse still pose challenges in elderly patients.9,10 One study investigating a significant number of transplanted ALL patients aged over 45 years showed a substantially higher rate for transplant-related mortality (TRM) in MAC-treated patients with HLA-mismatches when compared to the RIC-treated cohort, prompting the authors to discourage MAC conditioning in this patient group altogether.11 This observation suggests an interaction between transplantation-associated mortality caused by age-associated risk and HLA-mismatching. Age and HLAmatching status are important clinical predictors for the outcome of HSCT and are used among others for risk assessment in HSCT.12 We analyzed the relationship between age-risk and HLA-risk in a large cohort of patients transplanted with unrelated donors and tested the hypothesis that age-risk varies according to HLA-matching status. Such a differentiation might have an impact on donor search and selection recommendations.

Methods

Statistical analysis For univariate analysis of overall survival (OS), the KaplanMeier method and logrank testing was applied. Multivariate analysis for OS and disease-free survival (DFS) was performed using extended Cox-proportional hazards models.15 For TRM and RI, univariate competing risks analysis and multivariate competing risks regression for stratified data was used.16 Backward stepwise exclusion was used for multivariate model selection. Evaluated covariates were: patient age, HLA-matching status, disease stage, conditioning regimen intensity, treatment with anti-thymocyte globulin (ATG), year of transplantation, time

A

B

Patients A total of 3019 adult patients transplanted for malignant hematologic disorders were included in this analysis. Transplantations were performed at German transplant centers between 1997 and 2011. All patients received a first allogeneic unrelated transplant from bone marrow (BM) or peripheral blood stem cells (PBSC) with no more than 2 HLA-mismatches on 5-loci (HLA-A, -B, -C, -DRB1 and -DQB1). Disease stage definitions were adopted from a previous study defining the European Group for Blood and Marrow Transplantation (EBMT) risk score.12 MAC was defined according to the recommendations of the EBMT Central Registry Office (MedAB manual forms).13 Treatments with busulfan 16 mg/kg + cyclophosphamide 120-200 mg/kg, cyclophosphamide 120 mg/kg fractionated total body irradiation (TBI) 12Gy, etoposide VP-16 30-60 mg/kg + TBI 12Gy fractionated/10Gy single dose, BEAM polychemotherapy, CBV polychemotherapy or TBI 10-14Gy; busulfan 16 mg/kg are considered as myeloablative. Less intense regimens were considered as RIC. Patient and donor consent for HLA typing and for the analysis of clinical data were obtained. The study was approved by the ethical review board of the University of Ulm (project number 263/09).

C

HLA-typing All patients and donors were high resolution typed for HLA-A, -B, -C, -DRB1 and -DQB1. Ambiguities within exons 2+3 for HLAclass I and exon 2 for HLA-class II alleles were resolved. Ambiguities involving non-expressed (null) alleles were resolved according to NMDP confirmatory typing requirements. Differences in exon 2 and 3 for HLA-class I alleles and exon 2 for HLA-class II alleles were considered as HLA-mismatch irrespective of the vector of mismatches.14 Patient HLA-C KIR ligand status was inferred from high resolution HLA-C typing (C1=Asn80; C2=Lys80). Resulting phenotypes were C1C1, C1C2 and C2C2. 798

Figure 1. Kaplan-Meier estimates for overall survival. Kaplan-Meier estimates for overall survival according to HLA-matching status (10/10 black lines, 9/10 blue lines, 8/10 red lines) in different age categories. (A) Age 18-35 years, P=not significant. (B) age 36-55 years, P<0.001. (C) age >55 years, P<0.001.

haematologica | 2017; 102(4)


Age risk and HLA-matching

to transplantation, graft source, donor-recipient sex combination, KIR ligand status, and donor origin (national vs. international). For antithymocyte globulin (ATG) treatment, some data were missing (Table 1). Models were validated by inclusion of missing values as a separate group and by omission of cases with missing values, and no bias was found.17 Stratification was used to account for heterogeneity of diagnosis. Violations of the proportional hazards assumption (PHA) by disease stage, conditioning regimen intensity and transplantation before 2004 were adjusted using time-dependent modeling of these covariates.15 A significant center effect was adjusted using a frailty term with gamma distribution.18 To assess the relationship between age and HLA-compatibility, subgroups were formed and analyzed as factors: age group 18-35 years (HLA-match: 10/10, 9/10 and 8/10), 36-55 years (HLA-match: 10/10, 9/10 and 8/10), and over 55 years (HLA-match: 10/10, 9/10 and 8/10). The cut-off value of 55 years for elderly patients has been used in previous studies and the cut-off value of 35 years is close to the arithmetic mean between the age boundaries in the remaining patients.19 In addition, an interaction model between age and number of HLA-mismatches was investigated. The relative risk conferred by age was visualized as age-dependent risk in different HLA-match categories relative to an 18-year old patient transplanted with a 10/10 matched donor as baseline. In this model, the covariate age was included as a continuous variable and no violation of the PHA was found. P=0.05 was considered statistically significant.

Results Patients' characteristics are given in Table 1. Patients over 55 years of age formed the second largest age group (n=1195, 39.6%). The distribution of diagnoses reflects the current spectrum of indications, with acute myeloid leukemia (AML) being the most frequent diagnosis (n=924, 30.6%). Single HLA-mismatches were present in 30.2% (n=911) and double mismatches occurred in 8.7% (n=261) of all patients. Although the proportion of HLADQ mismatches among double mismatched transplantations was slightly higher in older patients, there was no statistically significant difference in the distribution of 8/10 mismatches. Ethnicity was almost exclusively Caucasian. MAC was used in 62.1% (n=1875) of the patients, with peripheral blood stem cells (PBSC) being the leading graft source (n=2694, 89.2%). More than half of the transplantations were performed in the years between 2008 and 2011 (n=1671, 55.4%). Median follow-up time was 29 months. Table 2 and Figure 1 show the results of the univariate OS analysis in patients according to their HLA-matching status and age group. Logrank-testing showed no significant difference between 10/10, 9/10 and 8/10 matched transplantations in the youngest age group (aged 18-35 years). In the intermediate age group (36-55 years) a highly significant difference (P<0.001) was found with higher mortality for patients transplanted with single or double mismatches. In patients over 55 years of age, the differences were even more pronounced, showing high mortality, especially in the 8/10 matching group (P<0.001). In multivariate modeling, these results could be confirmed for OS showing no significant differences between single and double mismatched transplantations in the younger age group (Table 3). Risk sharply increased with age in the respective mismatch groups, reaching the highest relative risk in the age group over 55 years (HR: 3.48, CI 2.49-4.86, P<0.001). Similar patterns were seen for DFS haematologica | 2017; 102(4)

and TRM with hazard ratios spreading with increasing numbers of HLA-mismatches and increasing age, thus conferring highest risk for patients aged over 55 years with double HLA-mismatches [DFS: Hazard Ratio (HR) 2.74, CI 2.00-3.76, P<0.001 and TRM: HR 3.79, CI 2.296.30, P<0.001]. No significant differences were observed for relapse incidences. Modeling an interaction term between age and number of HLA-mismatches allowed estimation of age risk within matched, single-mismatched and double mismatched patient groups. Age risk showed increasing risk estimates with increasing number of HLA-mismatches. In 10/10 matched transplantations, this additional risk per life year at time of transplantation was lowest (HR: 1.015, CI 1.010-1.020; P<0.001). It increased, however, with the decreasing degree of HLA-compatibility between donor and patient (9/10, HR: 1.019, CI 1.014-1.024, P<0.001 and 8/10 HR: 1.026, CI 1.020-1.031, P<0.001). The interaction term for age and 2 HLA-mismatches was significant (P=0.009). The Cox regression model is a multiplicative hazard model. In order to visualize the component of agerisk within the respective HLA-match groups, the change of risk contributed to the prognosis by age at the time of transplantation was plotted relatively to an 18-year-old 'baseline' patient with a 10/10 matched donor. This visualization is based on the different age-associated risk estimates within each HLA-match category as observed in the multivariate model for OS, and it illustrates the change in risk with increasing age (Figure 2).

Discussion We found a statistically significant interaction between HLA-matching status and age-associated risk. This interaction can be interpreted as different levels of age-associated risk according to the number of HLA-mismatches. Our findings substantiate that transplantation for patients aged over 55 years with two HLA-mismatches are particularly risky with a highly significant hazard ratio of 3.48 (CI 2.49-4.86; P<0.001) when compared to 10/10 matched patients younger than 35 years. If compared to 10/10

Figure 2. Age risk by HLA-matching status. Relative risk contributed by the continuous covariate age at the time of transplantation according to different levels of HLA-mismatches (completely matched 10/10: black, single mismatched 9/10: blue, double mismatched 8/10: red).

799


D. FĂźrst et al.

transplantations within each age category, double mismatches increased mortality risk for OS by a factor of 1.14 in the lowest age group, by a factor of 1.40 in the middle age group, and 2.27 in patients aged over 55 years. This disproportional increase and the poor one-year survival rate of only 19% in double mismatched transplantations for elderly patients highlights the importance of HLAmatching especially in this group of patients. Luckily, donors with 2 HLA-mismatches had to be accepted only in a small fraction of patients aged over 55 years (6.3%). The age cohorts showed expected structural differences in composition with regard to diagnosis and conditioning regimen, as well as graft source. Multivariate analysis adjusted for differences in conditioning treatment, while graft source showed no differential impact on survival end points. It is known that older patients tolerate conditioning related toxicity less well than younger patients, which is

the reason for the development and the use of conditioning regimes with reduced intensity.6,20,21 Treatment-associated toxicity correlates strongly with transplant-related mortality and therefore it greatly influences OS. HLA-mismatches also associate strongly with treatment-related morbidity and -mortality. This relationship explains our findings from the perspective of transplant biology, suggesting that older patients tolerate HLA-mismatches less well than younger patients as it is also the case for treatment-related toxicity. On the other hand, it cannot be deduced from this data whether younger patients benefit less from bettermatched donors, as life expectancy is higher and HLAassociated risk cumulates over time. This finding was only made possible because of the relatively high proportion of older patients in our dataset. As most of the transplantations were performed in the years between 2008 and 2011, our dataset reflects the substan-

Table 2. Univariate analysis in different age categories.

Overall survival

Age group

HLA-compatibility

N

1-year

3-year

P

18-35 (N=529)

10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10

295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75

0.67 (0.61-0.73) 0.61 (0.54-0.69) 0.64 (0.53-0.78) 0.63 (0.59-0.67) 0.50 (0.45-0.56) 0.45 (0.37-0.55) 0.59 (0.55-0.63) 0.47 (0.42-0.54) 0.27 (0.17-0.40) 0.59 (0.53-0.65) 0.53 (0.46-0.62) 0.49 (0.38-0.64) 0.53 (0.49-0.57) 0.42 (0.37-0.47) 0.40 (0.32-0.50) 0.49 (0.45-0.53) 0.40 (0.34-0.46) 0.20 (0.12-0.33) 0.22 (0.17-0.27) 0.24 (0.18-0.31) 0.26 (0.15-0.38) 0.22 (0.19-0.25) 0.24 (0.20-0.28) 0.23 (0.16-0.31) 0.22 (0.19-0.25) 0.24 (0.19-0.29) 0.30 (0.19-0.41) 0.15 (0.11-0.20) 0.19 (0.13-0.26) 0.23 (0.13-0.34) 0.19 (0.16-0.22) 0.29 (0.25-0.34) 0.39 (0.30-0.48) 0.22 (0.18-0.25) 0.30 (0.25-0.36) 0.40 (0.28-0.52)

0.53 (0.47-0.60) 0.54 (0.46-0.63) 0.45 (0.33-0.61) 0.49 (0.45-0.53) 0.39 (0.34-0.45) 0.35 (0.27-0.45) 0.41 (0.37-0.46) 0.34 (0.29-0.41) 0.19 (0.11-0.32) 0.45 (0.39-0.52) 0.50 (0.42-0.59) 0.37 (0.26-0.52) 0.38 (0.34-0.42) 0.32 (0.27-0.37) 0.32 (0.24-0.42) 0.30 (0.26-0.35) 0.24 (0.19-0.30) 0.12 (0.06-0.25) 0.28 (0.22-0.34) 0.26 (0.19-0.33) 0.36 (0.24-0.49) 0.30 (0.26-0.33) 0.29 (0.24-0.34) 0.29 (0.21-0.38) 0.32 (0.28-0.36) 0.30 (0.25-0.36) 0.35 (0.23-0.47) -

n.s.

36-55 (N=1295)

>55 (N=1195)

18-35 (N=529) Disease-free survival

36-55 (N=1295) >55 (N=1195)

18-35 (N=529)

Relapse incidence

36-55 (N=1295)

>55 (N=1195)

18-35 (N=529) Transplant-related mortality

36-55 (N=1295) >55 (N=1195)

<0.001

<0.001

n.s. 0.016 <0.001

n.s.

n.s.

n.s.

n.s. <0.001 <0.001

N: number within the respective group; 95% confidence interval in parentheses; n.s.: not significant.

800

haematologica | 2017; 102(4)


Age risk and HLA-matching

tial increase in elderly patients transplanted in Germany in recent years. Other large studies investigating the impact of risk factors in HSCT contained significantly fewer older patients, which is why this interaction may have remained unnoticed in these studies.22-24 Interestingly, in the youngest age group, no significant difference was found between completely 10/10 matched transplantations and single or double mismatched transplantations. However, this age category was the smallest, consisting of only 17.5% of the cases, which limits interpretation of this particular result. Testing for proportional hazards assumption in our models showed no significant violation for the covariate age, which was treated as a continuous variable in the interaction model and in the prediction plot (Figure 2). Thus, the way we chose to visualize the disproportional increase in hazard ratios for age-risk at the time of transplantation is justified.

Our results were obtained from a cohort transplanted with allogeneic unrelated PBSC or bone marrow as a graft source. In our analysis, graft source did not differentially impact outcome, which is why no separate analysis for each graft source was made. Similar findings were reported in other studies.25,26 Data on the impact of haploidentical transplantation or cord blood transplantations on the outcome of HSCT in elderly patients are very limited, so that a sensible risk-benefit comparison of our data with alternative graft or transplant sources is difficult. However, cord blood transplantation has been reported to result in similar outcomes in a small cohort of single mismatched transplantations in elderly patients treated with RIC.27 In multivariate analysis (Table 4), some predictors showed violation of the proportional hazards assumption (PHA). These violations can be explained by a higher early mortality for patients transplanted in advanced disease stage, transplanted before 2004 and treated with MAC. To

Table 3. Risk estimates for HLA mismatches according to age categories.

End point

Overall survival

Age group

HLA

N

HR (95% CI)

Within group

18-35 (N=529)

10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10 10/10 9/10 8/10

295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75 295 172 62 774 397 124 778 342 75

1.00 0.94 (0.70-1.26) 1.14 (0.77-1.71) 1.26 (1.01-1.57) 1.57 (1.25-1.98) 1.76 (1.32-2.36) 1.54 (1.22-1.92) 1.93 (1.50-2.47) 3.48 (2.49-4.86) 1.00 0.87 (0.66-1.14) 1.11 (0.77-1.60) 1.23 (1.01-1.49) 1.42 (1.15-1.75) 1.58 (1.21-2.08) 1.45 (1.18-1.77) 1.67 (1.34-2.09) 2.74 (2.00-3.76) 1.00 1.04 (0.71-1.51) 1.24 (0.77-1.99) 1.02 (0.77-1.34) 1.00 (0.74-1.36) 1.10 (0.74-1.63) 0.97 (0.73-1.29) 1.00 (0.72-1.39) 1.23 (0.76-2.00) 1.00 1.15 (0.71-1.85) 1.51 (0.81-2.82) 1.39 (0.98-1.99) 2.11 (1.47-3.04) 3.01 (1.96-4.62) 1.70 (1.18-2.45) 2.32 (1.58-3.41) 3.79 (2.29-6.30)

1.00 0.94 (0.70-1.26) 1.14 (0.77-1.71) 1.00 1.25 (1.05-1.48) 1.40 (1.09-1.80) 1.00 1.25 (1.05-1.50) 2.27 (1.70-3.03) 1.00 0.87 (0.66-1.14) 1.11 (0.77-1.60) 1.00 1.16 (0.99-1.35) 1.29 (1.02-1.64) 1.00 1.15 (0.98-1-36) 1.89 (1.44-2.50) 1.00 1.04 (0.71-1.51) 1.24 (0.77-1.99) 1.00 0.99 (0.78-1.25) 1.08 (0.76-1.54) 1.00 1.03 (0.79-1.34) 1.27 (0.81-1.98) 1.00 1.15 (0.71-1.85) 1.51 (0.81-2.82) 1.00 1.51 (1.18-1.95) 2.16 (1.53-3.05) 1.00 1.36 (1.05-1.77) 2.23 (1.47-3.37)

36-55 (N=1295)

>55 (N=1195)

18-35 (N=529) Disease-free survival

36-55 (N=1295) >55 (N=1195)

18-35 (N=529)

Relapse incidence

36-55 (N=1295)

>55 (N=1195)

18-35 (N=529) Transplant-related mortality

36-55 (N=1295) >55 (N=1195)

P n.s. n.s. 0.012 0.008 0.014 <0.001 n.s. n.s. n.s. 0.033 n.s. <0.001 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. 0.001 <0.001 0.022 <0.001

N: number within the respective group; HR: Hazard ratio; CI: Confidence Interval; n.s.: not significant. P-values are computed for the comparison of 9/10 and 8/10 groups with the 10/10 matched transplantations within each age category. Relative risk computation is also performed for each age group (within group). The P-values refer to this comparison. Other covariates included: disease stage, KIR-ligand status, national donor versus international donor, conditioning treatment, year of transplantation.

haematologica | 2017; 102(4)

801


D. Fürst et al. Table 4. Multivariate analysis.

End point

Predictor

Overall survival

Age-risk (10/10 HLA) Age-risk (9/10 HLA) Age-risk (8/10 HLA) Intermediate disease stage Advanced disease stage until day 314 post Tx Advanced disease stage after day 314 post Tx Patient C2C2 KIR ligand status National donor RIC vs. MAC until day 96 RIC vs. MAC after day 96 Tx before 2004 until day 198 post Tx Tx before 2004 after day 198 post Tx Age-risk (10/10 HLA) Age-risk (9/10 HLA) Age-risk (8/10 HLA) Intermediate disease stage Advanced disease stage until day 253 post Tx Advanced disease stage after day 253 post Tx Patient C2C2 KIR ligand status National donor RIC vs. MAC until day 81 RIC vs. MAC after day 81 Tx before 2004 until day 205 post Tx Tx before 2004 after day 205 post Tx

Disease-free survival

HR (95% CI)

P

1.015 (1.010-1.020) 1.019 (1.014-1.024) 1.026 (1.020-1.031) 1.37 (1.19-1.57) 2.37 (2.04-2.74) 1.03 (0.78-1.36) 1.25 (1.08-1.43) 0.83 (0.73-0.95) 0.57 (0.46-0.70) 1.13 (0.98-1.31) 1.43 (1.18-1.72) 1.05 (0.82-1.34) 1.014 (1.010-1.018) 1.016 (1.012-1.021) 1.023 (1.017-1.028) 1.51 (1.33-1.71) 2.35 (2.05-2.70) 1.36 (1.08-1.71) 1.17 (1.03-1.33) 0.84 (0.74-0.94) 0.83 (0.70-0.97) 1.01 (0.88-1.16) 1.27 (1.07-1.51) 0.80 (0.62-1.03)

<0.001 <0.001 <0.001 <0.001 <0.001 n.s. 0.002 0.005 <0.001 n.s. <0.001 n.s. <0.001 <0.001 <0.001 <0.001 <0.001 0.009 0.019 0.003 0.021 n.s. 0.006 n.s.

HR: Hazard Ratio; HLA: Human Leukocyte Antigen; RIC: reduced intensity conditioning; MAC: myeloablative conditioning; Tx: transplantation; n.s.: not significant.

reflect this relationship, an extended Cox regression model was fitted to obtain regression estimates for the respective predictors according to time periods where PHA is satisfied, as we have shown before.28 In analysis of OS, advanced disease stage showed a substantially higher mortality risk until day 314 but not thereafter. Patients treated with RIC showed a significantly lower early mortality until day 96 and a non-significantly different risk afterwards. In addition, patients transplanted before 2004 showed a higher mortality risk until day 198 after transplantation but not thereafter. Similar findings were present in an analysis of DFS. In our models, also a patient C2C2 KIR-ligand status as well as an international donor status was associated with adverse outcome, which we have reported before.29 ATG treatment was not included in the final models because it did not reach statistical significance. Our analysis encompassed some simplifications, namely that any HLA-mismatch was considered equally. HLADPB1 mismatches were not included and the vector of mismatches was also not regarded. We included HLA-DQB1 mismatches in this study, because a previous analysis on the same dataset has shown that these mismatches are associated with higher mortality risk.29 HLA-DPB1 mismatches have been shown to influence outcome of HSCT, but due to lower linkage disequilibrium, HLA-DPB1-mismatches in HLA-A, -B, -C, -DRB1 and

802

-DQB1 matched and mismatched transplantations are almost equally distributed.30 Therefore, we may assume that our results are not biased by not including HLADPB1. The vector of mismatches was not considered, because no significant differences in survival outcome have been seen for unidirectional mismatches when compared to bidirectional mismatches for the end points analyzed in our study.14 We refrained from including Karnofsky performance status and donor-recipient cytomegalovirus status due to the high proportion of missing data for these variables, which is a limitation of our analysis. When selecting donors for elderly patients, the additional risk associated with HLA-mismatches in this age group should be considered. Especially when only donors with double HLA-mismatches are available for such a patient, the substantial risk conferred in this situation must be carefully weighed against the benefit of transplantation. Cord blood transplantation might be an alternative in such cases, although data regarding the impact of alternative graft sources for transplantation of elderly patients are still limited. Funding This work was supported by the Deutsche José Carreras Leukämie-Stiftung e.V. (Grant No. DJCLS 11/10), and the German Red Cross Blood Transfusion Service, BadenWuerttemberg/Hessen, Germany.

haematologica | 2017; 102(4)


Age risk and HLA-matching

References 1. Gratwohl A, Baldomero H. Trends of hematopoietic stem cell transplantation in the third millennium. Curr Opin Hematol. 2009;16(6):420-426. 2. Passweg JR, Baldomero H, Bregni M, et al. Hematopoietic SCT in Europe: data and trends in 2011. Bone Marrow Transplant. 2013;48(9):1161-1167. 3. Mohty M, Labopin M, Volin L, et al. Reduced-intensity versus conventional myeloablative conditioning allogeneic stem cell transplantation for patients with acute lymphoblastic leukemia: a retrospective study from the European Group for Blood and Marrow Transplantation. Blood. 2010; 116(22):4439-4443. 4. Niederwieser D, Lange T, Cross M, Basara N, Al-Ali H. Reduced intensity conditioning (RIC) haematopoietic cell transplants in elderly patients with AML. Best Pract Res Clin Haematol. 2006;19(4):825-838. 5. Huisman C, Meijer E, Petersen EJ, Lokhorst HM, Verdonck LF. Hematopoietic stem cell transplantation after reduced intensity conditioning in acute myelogenous leukemia patients older than 40 years. Biol Blood Marrow Transplant. 2008;14(2):181-186. 6. McSweeney PA, Niederwieser D, Shizuru JA, et al. Hematopoietic cell transplantation in older patients with hematologic malignancies: replacing high-dose cytotoxic therapy with graft-versus-tumor effects. Blood. 2001;97(11):3390-3400. 7. Brandwein JM, Geddes M, Kassis J, et al. Treatment of older patients with acute myeloid leukemia (AML): a Canadian consensus. Am J Blood Res. 2013;3(2):141-164. 8. Sorror ML, Sandmaier BM, Storer BE, et al. Long-term outcomes among older patients following nonmyeloablative conditioning and allogeneic hematopoietic cell transplantation for advanced hematologic malignancies. JAMA. 2011;306(17):1874-1883. 9. Brunner AM, Kim HT, Coughlin E, et al. Outcomes in patients age 70 or older undergoing allogeneic hematopoietic stem cell transplantation for hematologic malignancies. Biol Blood Marrow Transplant. 2013;19(9):1374-1380. 10. Jindra P, Muzik J, Indrak K, et al. The outcome of allogeneic HSCT in older AML patients is determined by disease biology and not by the donor type: An analysis of 96 allografted AML patients >/=50 years from the Czech acute leukaemia clinical register

haematologica | 2017; 102(4)

(alert). Neoplasma. 2013;60(5):576-583. 11. Tanaka J, Kanamori H, Nishiwaki S, et al. Reduced-intensity vs myeloablative conditioning allogeneic hematopoietic SCT for patients aged over 45 years with ALL in remission: a study from the Adult ALL Working Group of the Japan Society for Hematopoietic Cell Transplantation (JSHCT). Bone Marrow Transplant. 2013; 48(11):1389-1394. 12. Gratwohl A, Stern M, Brand R, et al. Risk score for outcome after allogeneic hematopoietic stem cell transplantation: a retrospective analysis. Cancer. 2009; 115(20):4715-4726. 13. EBMT.org [Internet]. London: EBMT Central Registry Office. MED-AB FORMS MANUAL; c2010-2016 [updated 2016 December 12; cited 2016 December 19]. Available from: http://www.ebmt.org/. 14. Hurley CK, Woolfrey A, Wang T, et al. The impact of HLA unidirectional mismatches on the outcome of myeloablative hematopoietic stem cell transplantation with unrelated donors. Blood. 2013;121(23):4800-4806. 15. Therneau T, Grambsch P. Modeling Survival Data: Extending the Cox Model. 1st ed. New York: Springer; 2000. 16. Zhou B, Latouche A, Rocha V, Fine J. Competing risks regression for stratified data. Biometrics. 2011;67(2):661-670. 17. Iacobelli S. Suggestions on the use of statistical methodologies in studies of the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant. 2013;48(Suppl 1):S1-37. 18. Glidden DV, Vittinghoff E. Modelling clustered survival data from multicentre clinical trials. Stat Med. 2004;23(3):369-388 19. Shimoni A, Kroger N, Zabelina T, et al. Hematopoietic stem-cell transplantation from unrelated donors in elderly patients (age >55 years) with hematologic malignancies: older age is no longer a contraindication when using reduced intensity conditioning. Leukemia. 2005;19(1):7-12. 20. Slavin S, Nagler A, Naparstek E, et al. Nonmyeloablative stem cell transplantation and cell therapy as an alternative to conventional bone marrow transplantation with lethal cytoreduction for the treatment of malignant and nonmalignant hematologic diseases. Blood. 1998;91(3):756-763. 21. Estey E, de Lima M, Tibes R, et al. Prospective feasibility analysis of reducedintensity conditioning (RIC) regimens for hematopoietic stem cell transplantation

22.

23.

24.

25.

26.

27.

28.

29.

30.

(HSCT) in elderly patients with acute myeloid leukemia (AML) and high-risk myelodysplastic syndrome (MDS). Blood. 2007;109(4):1395-1400. Flomenberg N, Baxter-Lowe LA, Confer D, et al. Impact of HLA class I and class II highresolution matching on outcomes of unrelated donor bone marrow transplantation: HLA-C mismatching is associated with a strong adverse effect on transplantation outcome. Blood. 2004;104(7):1923-1930. 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. 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. Nagler A, Labopin M, Shimoni A, et al. Mobilized peripheral blood stem cells compared with bone marrow as the stem cell source for unrelated donor allogeneic transplantation with reduced-intensity conditioning in patients with acute myeloid leukemia in complete remission: an analysis from the Acute Leukemia Working Party of the European Group for Blood and Marrow Transplantation. Biol Blood Marrow Transplant. 2012;18(9):1422-1429. Eapen M, Rocha V, Sanz G, et al. Effect of graft source on unrelated donor haemopoietic stem-cell transplantation in adults with acute leukaemia: a retrospective analysis. Lancet Oncol. 2010;11(7):653-660. Malard F, Furst S, Loirat M, et al. Effect of graft source on mismatched unrelated donor hemopoietic stem cell transplantation after reduced intensity conditioning. Leukemia. 2013;27(11):2113-2117. Fuerst D, Mueller C, Beelen DW, et al. Timedependent effects of clinical predictors in unrelated hematopoietic stem cell transplantation. Haematologica. 2016; 101(2):241247. 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. 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.

803


Haematologica, volume 102, issue 4  
Read more
Read more
Similar to
Popular now
Just for you